52
J . Phys. Chem. 1992,96, 52-56 NH(a,u”=O)
+ Xe
++
NH(X,u>O)
+ Xe
can be excluded. The observation of NH(X,u>O) in the other experiment (see below) indicates that the absence of NH(X,u) cannot be explained by a rapid vibrational quenching in the electronic ground state. Thus, at least for u” = 0 the proof can be given that the vibrational state is not changed by the collision-induced ISC with Xe, Le., the electronic excitation energy is not converted into vibrational energy. For NH(a,u”=l) the following process NH(a,u”=l) + Xe NH(X,u=l) + Xe
-
can be assumed, but a change in the vibrational state cannot be excluded. The reaction rate of NH(a,u”=O) with Xe (k4 = 7.2 x 10I2cm3/(mol s ) ) ~was nearly identical with the reaction rate of NH(a,u”=l) with Xe ( k , = 7.7 X lo1*cm3/(mol s)). This indicates that vibrational quenching is of minor importance in reaction 5. The appearance of NH(X,u=0,1,2) with time was followed and compared to the NH(a,u”=O,l) depletiona6 To determine from these measurements the quenching mechanism, however, calibrations of [NH(a,u”)] and [NH(X,u)] are needed. For N, as the quenching gas, no NH(X) with u > 0 was observed in both experiments with NH(a,u”=O) and NH(a,u”20). As in the case for Xe a quenching process NH(a,u”=O) + N2 ++ NH(X,u) N2
+
can be excluded. The fact that no NH(X,u>O) is observed can be explained by a fast vibrational quenching
NH(a,u’+O)
+ N2
kdb
NH(a,u”=O)
+ N,
with hb= 1.2 X lo1’cm3/(mol s).I7 This quenching process can occur very efficiently via the crossing of the lA’(u’’=l) and IA”(u”=O) surfaces28as illustrated in Figure 4. Vibrational quenching contributes mainly to the depletion of NH(a,u”Z 1). The direct ISC pathway NH(a,u”=l) + N2 NH(X,u=O,l) + N2
-
is of minor importance. A calibration of the [NH(X,u=l)] detection limit will give an upper limit for the rate constant. For the N2 quenching of NH(ap’9, produced in the photolysis of HN, at X = 266 nm, A d a m and Pasternack6detected the ratio [NH(X,u= l)] / [NH(X,u=O)] I 0.02. Since the initial vibrational population (NH(a,u”)) obtained at X = 266 nm is very similar to the population at X = 248 nm these results can be compared directly. The results in reference6 and in this work are in good agreement. It can be concluded that all experiments can be described assuming that the vibrational quantum number is unchanged during the ISC.
Acknowledgment. We are greatly indebted to Prof. H. Gg. Wagner for his generous support and stimulating interest. Financial support by the DFG, SFB 93 “Photochemie mit Lasem” is acknowledged. Registry No. NH, 13774-92-0; N1, 7727-37-9; Xe,7440-63-3; HN,, 7782-79-8; ND, 15123-00-9.
First Observation of a Radical-Triplet Pair Mechanism (RTPM) with Doublet Precursor Akio Kawai and Kinichi Obi* Department of Chemistry, Tokyo Institute of Technology, Ohokayama, Meguroku, Tokyo 152, Japan (Received: July 26, 1991; In Final Form: September 1 1 , 1991)
AbsorptiveCIDEP signals of nitroxide radical were first observed in fluorantheneTEMP0,coroneneTEMP0, pyreneTEMPO, and naphthaleneTEMP0 systems by time-resolved ESR. The absorptive spin polarization was interpreted by RTPM with doublet precursor of the radical triplet pair. From the CIDEP decay profiles and triplet quenching experiments, it was concluded that the generated CIDEP in these systems is explainable by the simultaneous operation of both doublet and quartet precursor RTPMs.
Introduction Singlet and triplet excited molecules and intermediate radicals coexist in the early photochemical stage. Interactions between these species have extensively been investigated and many interesting phenomena have been well-known, such as SI and T I quenching by radicals, triplet-triplet annihilation, and so on.] Photochemistry and photophysics of these species have been discussed based on the electron spin multiplicities because the spin angular momentum of these intermediate molecules is a good quantum number. For example, quenching of excited states by free radicals and radical recombination reactions selectively occur according to the conservation rule of the spin angular momentum, which are confirmed by magnetic field effects.,~~ Chemically induced dynamic electron polarization (CIDEP) of radicals is generated through the radical pair mechanism (RPM),4 the triplet mechanism (TM),5 and the radical-triplet
pair mechanism (RTPM).637 These mechanism are also discussed based on the spin angular momentum. RPM and RTPM are explained by magnetic interaction acting on the potential surface of spin states of radical pair and radical-triplet pair, respectively. In RTPM, randomly encountered radical and triplet pairs show quartet and doublet spin states. The quenching of triplet molecule occurs through the doublet spin states of radical-triplet pair. During the course of triplet-doublet interaction, the quartet and doublet spin states mix with each other by the zero-field splitting and hyperfine interactions. This mixing generates CIDEP of net emission (E) and an E / A (emission/absorption)multiplet pattern, which polarization is abbreviated as E * / A . RPM and RTPM are so far the only CIDEP mechanisms due to the interaction among the species of initial photochemical processes, but it is still unknown whether CIDEP is generated or not via radical-excited singlet and triplet-triplet pairs. As for
(1) Birks, J. B. Photophysics of Aromafic Molecules; Wiley: New York,
(4) Adrian, F. J. Rev. Chem. Intermed. 1979, 3, 3. Muus, L. T.; Atkins, P. W.; McLauchlan, K. A.; Pedersen, J. B. Chemically Induced Magnetic Polarization; Reidel: Dordrecht, 1977. (5) Wong, S. K.; Wan, J. K. S. J. Am. Chem. Soc. 1972, 94, 7197. (6) Blittler, C.; Jent, F.; Paul, H. Chem. Phys. Lett. 1990, 166, 375. (7) Kawai, A.; Okutsu, T.; Obi, K. J . Phys. Chem. 1991, 95, 9130.
1970.
(2) Razi Naqvi, K.; Gillbro, T. Chem. Phys. Lett. 1977, 49, 160. (3) Sakaguchi, Y.; Nagakura, S.;Hayashi, H. Chem. Phys. Lett. 1980,72, 420. Molin, Yu. N., Ed. Spin polarization and magnetic effect in radical reactions; Elsevier: Amsterdam, 1984.
0022-365419212096-52$03.00/0 0 1992 American Chemical Society
Doublet Precursor Radical-Triplet Pair Mechanism
The Journal of Physical Chemistry, Vol. 96, No. 1, 1992 53
the interaction between the lowest excited singlet molecule and radical, quenching of fluorescence is well-known and is interpreted by the enhanced intersystem crossing (enhanced ISC) of singlet molecule due to radical pert~rbation.**~ Just after this phenomenon, pairs of radical and generated triplet molecule are formed. Magnetic interactions and operation of the RTPM are expected in these pairs of paramagnetic species. In this paper we demonstrate for the coronene-, fluoranthene-, pyrene-, and naphthaleneTEMP0 (2,2,6,6-tetramethyl-1-piperidinyloxyl) systems by time-resolved ESR that those radical-triplet pairs generate an A * / E (absorption*/emission) polarization. The obtained CIDEP signal pattern is explained by an extended RTPM in which we assume the initial spin state of the radicalsinglet pair as the doublet spin states of the radical-triplet pair. In this study, we conclude that there are two types of CIDEP generation in RTPM: one is the doublet precursor RTPM which shows A * / E pattern and the other is the quartet precursor RTPM that shows E * / A pattern. These two types of CIDEP generation are expected to be simultaneously operative in the early stage of photochemical processes, and we demonstrate the CIDEP signal of naphthalene-TEMPO system as the typical example of simultaneous operation of these RTPMs.
Experimental Section A conventional X-band ESR spectrometer (Varian E-1 12) was used to measure TR-ESR spectra. Transient ESR signals obtained without field modulation were transferred to a boxcar integrator (Stanford SR-250) for spectrum measurements or a transient memory (Iwatsu DM-901) for CIDEP decay profiles. The width of the gate time of the boxcar integrator was usually 0.5 ps after 1.0 ps of the laser pulse. The excitation light source was a XeCl excimer laser (Lambda Physik LPX 100) or a nitrogen laser (Molectron UV-24). Details of the equipment and method were described previously.I0 Fluoranthene and pyrene (Tokyo Kasei) were recrystallized from n-hexane. Coronene (Tokyo Kasei) and naphthalene (Kanto Chemicals) were recrystallized from benzene and ethanol, respectively. The other reagents are used as received. GR grade benzene and 2-propanol (Kanto Chemicals) were used as solvents without further purification in room-temperature experiments. Spectrograde 2-methylbutane (Tokyo Kasei) was used as a solvent for the measurement of fluoranthene triplet state. In the room-temperature experiments, the solution was deaerated by bubbling of nitrogen gas and flowed through a quartz flat cell (0.5" interior space) in ESR cavity. In the low-temperature experiments, dissolved oxygen in the sample solution was degassed by freeze-pumpthaw cycles. Afterward the glasses of sample solution were prepared in 5-mm-diameter quartz ESR tube by quick cooling to 77 K. Results and Discussion A */EType CIDEP Signal Figure 1 shows the CIDEP spectra obtained in the fluoranthene-, pyrene-, and coronene-TEMPO systems by 308-nm laser excitation, together with continuous-wave ESR spectrum of TEMPO. Hyperfine peaks appear at the same positions with those of TEMPO, which shows triplet hyperfine structure due to the nitrogen atom. Hence, the signals of these CIDEP spectra are assigned to spin-polarized TEMPO radicals. CIDEP patterns of these spectra are almost same and show two characteristic points: One is the total absorptive spin polarization, and the other is the hypcrfine dependence. The signal of MI = -1 peak is the most intense and the magnitude of absorptive signal diminishes with increasing in the quantum number M I . The CIDEP spectra in Figure 1 are represented as sum of the total absorption and the hyperfine dependent A / E pattern, which results in A * / E pattern. The hyperfine dependence is weak in the coronene-TEMPO system and enhanced in pyrene-TEMPO. In our previous study,' we reported the total emissive and hyperfine dependent ( E / A type) CIDEP signals of nitroxide (8) Watkins, A. J. Chem. Phys. Lett. 1974, 29, 526. (9)Green, J. A.; Singer, L. A.; Parks, J. H. J. Chem. Phys. 1973,58,2690. (10) Murai, H.; Imamura, T.; Obi, K. Chem. Phys. Lett. 1982,87, 295.
J.
Em.
2mT1
Figure 1. Continuous-wave ESR spectrum of TEMPO (0.60 mM) in benzene (a) and TR-ESR spectra of TEMPO in the systems of (b) TEMPO-fluoranthene (12 mM), (c) TEMPO-pyrene (8.2 mM), and (d) TEMPO-coronene (1.1 mM) at room temperature.
radicals in solutions of many organic compounds-radical systems, which were interpreted by triplet-doublet interaction. But the results obtained here show completely opposite phase of CIDEP signal to these systems. RTPM predicts the generation of E * / A type CIDEP signals in the triplet-doublet interaction for J < 0 system. The sign of spin polarization due to RTPM depends on the sign of J value, and thus J > 0 is simply considered for an answer of A * / E type CIDEP signal in these systems. Generally, neutral radical pairs show negative J values though some ionic radical pairs exceptionally show positive v a l ~ e . ~ According ~" to RTPM proposed by us,' all neutral radical-triplet pairs studied previously should have negative J values. It is, therefore, difficult to consider the exceptional opposite sign of J value for the neutral radical-triplet pairs studied here. Electron spin polarization transfer (ESPT) is another plausible mechanism to explain total absorptive CIDEP.I2 If the triplet states of fluoranthene, coronene, and pyrene have j3 spin enhanced population, radicals would show absorptive spin polarization due to ESPT. Figure 2 shows TR-ESR spectrum of the triplet state of fluoranthene obtained in 2-methylbutane at 77 K. The sharp Hmin signal appeared around 150 mT with emissive polarization. If ESPT is effective in this system, the radical should have emissive polarization which is opposite to our result. The CY spin enhanced population in the triplet state of coronene is also demonstrated with microwave-induced delayed phosphore~cence.~~ Hence, ESPT cannot explain our result of CIDEP. Next, we try to interpret the absorptive signals due to the state mixing of quartet and doublet spin states beginning from the doublet state spin pair, which is different from that of RTPM discussed above. Hoytink and Birks reported that ISC from the lowest excited singlet state was enhanced in the presence of free (11) Murai, H.; Kuwata, K. Chem. Phys. Lett. 1989, 164, 567. (12) Imamura, T.; Onitsuka, 0.;Obi, K. J . Phys. Chem. 1986, 90, 6741. (13) Ohno, K.;Nishi, N.; Kinoshita, M.; Inokuchi, H. Chem. Phys. Lett. 1975, 33, 293.
Kawai and Obi
54 The Journal of Physical Chemistry, Vol. 96, No. 1, 1992
A
Abs.
a)
Em.
I
1.o
I
I
2.0
I
1
30
I
1-
4.0
5.0 (x10'mT)
Figure 2. TR-ESR spectrum of fluoranthene (0.6 mM) obtained in 2-methylbutane glass at 77 K.
radi~a1s.l~As for the pyrene- and coronene-radical systems, the enhanced SI-Tl ISC occurred with the quenching ratels of about 1O1O M-I s-l, which indicated khat the triplet state was formed with the rise time of 0.1 p s at radical concentration of M. The relative fluorescence quantum yield of fluoranthene was measured for various radical concentrations and reduced to be about 0.65 under the experimental condition of Figure 1b, compared to that in the absence of free radical. Fluoranthene, coronene, and pyrene have long SI lifetimes (250ns),I6 and thus the enhanced ISC of these molecules makes a major contribution to the generation of triplet state. The radical-triplet pair formed through the Sl-T, enhanced ISC from the SI-radical complex should have the same spin states as the complex,I7 that is, ID f where D denotes the doublet spin state and *I/, represents spin magnetic quantum number. Thus, the doublet spin states are selectively populated in the radical-triplet pair by this mechanism, and quartet-doublet mixing in the generated radical-triplet pair will start under the initial condition where only the doublet states are populated but no quartet states. If we apply this initial condition to RTPM? A * / E type of CIDEP is predicted to be generated in radicals. Our experimental results are consistent with this prediction and thus, we attributed this A * / E signal to doublet precursor RTPM. The CIDEP signals obtained here are the first observation of doublet precursor RTPM as far as we know. Switching of Quartet Precursor RTPM to Doublet Precursor RTPM. We reported CIDEP generated through the quartet precursor RTPM in the previous paper' and observed CIDEP brought about by the doublet precursor RTPM in this study. Thus, both RTPMs are expected to be simultaneously operative to generate CIDEP in a system where the rate of &-TI ISC is comparable with that of enhanced ISC. A good example of such a case is naphthalene; the rate of enhanced ISC becomes comparable with the ratel* of SI-TI ISC, lo6 s-l, at the radical concentration of about M. The CIDEP spectra of TEMPO in the naphthalene-TEMPO system were obtained a t different gate times as shown in Figure 3. Phase of CIDEP signal changes from absorption in early gate time (Figure 3a) to emission in late gate time (Figure 3b). Hyperfine dependence of CIDEP is also observed in this system and is almost the same between former and latter spectra. These facts indicate that the absorptive CIDEP signals are generated in the course of enhanced ISC and emissive signals results from the interaction of radicals with the triplet (14) Hoytink, G. J. Acc. Chem. Res. 1969, 2, 114. Birks, J. B. J . Lumin. 1970, 1-2, -154. (15) Suzuki, T.; Obi, K., to be published. (16) Murov, S. L. Handbook of Photochemistry; Marcel Dekker: New Yoik, '1913. (17) Razi Naqvi, K. J . Phys. Chem. 1981, 85, 2303. (1 8) Turro, N. J. Modern Molecular Photochemistry; Benjamin/Cummings: New York, 1978.
Figure 3. TR-ESR spectra of TEMPO (0.60 mM) in the TEMPOnaphthalene (39 mM) system in 2-propanolsolution. The gate is opened from (a) 0.3 to 0.8 and (b) 1.5 to 2.0 ps. The top shows the continuous ESR spectrum of TEMPO. a)
. Abs
v
4 Em.
.1
Em.
Figure 4. Time profiles of CIDEP signal of TEMPO (0.60 mM) at the M I= -1 peak of (a) TEMPO-1-chloronaphthalene (37 mM) in 2propanol, (b) TEMPO-benzophenone (55 mM) in benzene and (c) TEMPO-naphthalene (39 mM) in 2-propanol. Microwave power is 1 mW.
naphthalene molecules produced by ISC. To investigate details of transient CIDEP, we measured the time profiles in various systems, as shown in Figure 4. No significant change of decay profiles was observed for microwave
Doublet Precursor Radical-Triplet Pair Mechanism
The Journal of Physical Chemistry, Vol. 96, No. 1, 1992 55
T In
Abs
.-
VI E
1
C
J.
1
Em.
Fwre 5. Time profiles of CIDEP signal of TEMPO (0.60 mM) at the M,= -1 peak in the TEMPO-naphthalene (39 mM) system (a) without triplet quencher and (b) with 1,3-pentadiene (35 mM). Microwave power is 1 mW.
power between 0.1 and 50 mW. Decay profiles of CIDEP in 1-chloronaphthalene-TEMPOand benzophenone-TEMPO systems are the typical CIDEP time profile in the quartet precursor RTPM; the phase of CIDEP is emissive throughout the profiles. The SI-TI ISC rates of benzophenone and 1-chloronaphthalene are reportedI8 to be lo1' and lo8 s-l, respectively, which are too fast to generate the doublet precursor RTPM signal. This is why only emissive CIDEP is observed in these systems. A little difference in decay rates between these profiles would be caused by the difference in magnetic relaxation time due to spin-spin interaction between triplet and radical. This is supported by the broad line width of TEMPO in the benzophenone-TEMPO system which is twice as large as that of the l-chloronaphthalene-TEMPO system. On the other hand, in the time profile of the naphthalene-TEMPO system shown in Figure 4c, phase of CIDEP is absorption for the first 0.3-0.8 ps and is turned to emission after 0.8 ps which is continued to the end of its decay. Such kind of phase change has not been seen for other triplet-TEMPO systems observed before. The maximum intensity of absorptive signal is much weaker than that of emissive signal. If such a waving decay profile is due to the magnetic relaxation process known as Torrey o s ~ i l l a t i o n intensity ,~~ of the first maximum absorptive signal should be stronger than that of the fmt maximum emissive signal. Therefore, the phase turning due to magnetic relaxation processes is ruled out. From the kinetic consideration discussed above, the phase change is interpreted by the switching of the CIDEP generation mechanism from doublet precursor RTPM to quartet precursor RTPM with time. The magnitude of spin polarization, (Sz(t)),at M I= -1 peak is determined by the equation (SZ(0) =
+
~ - ~ q s ~ ~ l kq,[T,(Ol~J[Rl ~ o l ~ s
(1)
Positive sign means emissive polarization. kqs and k,! are quenching rate constants of SIand T I by radicals, respectively. Isand 1, are the enhancement factors of CIDEP generation. As TEMPO is a stable free radical, its concentration, [R], is constant. In early gate time, doublet precursor RTPM is dominant because the SIstate of naphthalene is first quenched by radicals and forms the doublet spin states of radical-triplet pair, which corresponds to the first term of eq 1. In the late stage, radicals encounter the triplet state of naphthalene generated through direct or enhanced ISC. The triplet molecules formed through enhanced ISC loose their spin correlation with the counter radicals due to separation of the pair. The quartet spin states survive in triplet-radical pair during triplet quenching, which results in E * / A type CIDEP observed in the late gate time after 0.8 ps. To confirm this mechanism, we added 1,3-pentadiene to the naphthalene-TEMPO system as a triplet quencher, where selective quenching of the T I state must reduce the quartet precursor RTPM signal. The T I state of 1,3-pentadiene is 20700 cm-', ( 1 9 ) Hore, P. J.;
McLauchlan, K. A. J . Mogn. Reson. 1979, 36, 129.
V Em.
b)
4
4 Em. Figure 6. Time profiles of CIDEP signal of TEMPO (0.60 mM) at the M I= -1 peakof (a) TEMPO-fluoranthene (12 mM) and (b) TEMPOcoronene (1.1 mM) in benzene. Microwave power is 1 mW.
which is 600 cm-' lower than that of naphthalene. Hence, the effective triplet quenching is expected in this system. Figure 5 shows the time profile of CIDEP signal generated in the naphthalene-TEMPO systems without the triplet quencher (a) and with 1,3-pentadiene(b). The absolute intensity in Figure 5 is not calibrated, but the relative intensity of emissive signals is reduced in decay profile (b) as predicted above. In case of l-chloronaphthaleneTEMP0 system, the same reduction of quartet precursor RTPM signal was observed by addition of 1,3-pentadiene. The TIlifetime of 1,3-pentadiene is not reported as far as we know but could be short due to its free-rotor structure. The triplet 1,3-pentadiene-TEMPOinteraction is, therefore, considered to give a minor contribution to whole CIDEP signal. Abnormal decay profiles like that in the naphthalene-TEMPO system are also observed in the fluoranthene- and coroneneTEMPO systems as shown in Figure 6. In these systems, the maximum intensity of absorptive signal is much stronger than that of emissive signal appearing in later window time, which is different from the naphthalene-TEMPO system. If the emissive decay signal is generated as the result of Torrey oscillation of the initial absorptive signal, the decay profile should oscillate with a specific frequency. This is not the case in our systems. Moreover, no phase change is observed in other systems with short-lived SIand efficient ISC. The phase change of CIDEP is, thus, accounted with switching of the doublet precursor RTPM at early stage to the quartet precursor RTPM at late stage. Triplet quenching experiment was also carried out in the coroneneTEMPO system in order to confirm the switching mechanism. Figure 7 shows decay profiles of CIDEP signal in the coroneneTEMPO system by 337-nm laser excitation. In the absence of triplet quencher, phase change was observed as shown in Figure 7a, which is almost the same with Figure 6b, whereas when tram-stilbene was added in this system as a triplet quencher, only absorptive signals were observed without phase change. This result strongly supported the switching mechanism in the coroneneTEMPO system. Table I summarizes SIlifetime,'* triplet quantum yield, @ 1 ~ ~ , 1 6 and generated CIDEP pattern due to RTPM. As seen in this
Kawai and Obi
56 The Journal of Physical Chemistry, Vol. 96, No. I, 1992
a)
1 ) Quartet Precursor RTPM ( Triplet quenching )
TAbS
E t EIA
2 ) Doublet Precursor RTPM ( Singlet quenching )
Em.
SI D +
J
'(So+
D)
+
So+D
Figure 8. Schematic representations of doublet and quartet precursor
RTPMs.
J,
Em.
Figure 7. Time profiles of CIDEP signal of TEMPO (0.60 mM) at the M,= -1 peak in TEMPO-wronene (1.1 mM) system (a) without triplet quencher and (b) with trans-stilbene (83 mM). Microwave power is 50
mW. TABLE I: SILifetime, Ob and CIDEP Patterns Generated in TEMPO Radicals excited molecule SI lifetime, ns ai= CIDEP pattern 450 0.38 A*/E pyrene 302 0.56 A*/E E*/A coronene 53 fluoranthene A*/E E*/A 96 0.70 A*/E E*/A naphthalene 1-naphthol 10.6 E*/A 2.0 1 .oo E*/A acetone 0.005 1.oo E*/A benzophenone
---
table, if SIhas a lifetime longer than 50 ns, we can detect the A * / E type CIDEP. On the other hand, if its lifetime is shorter than 50 ns, it is almost impossible to detect A * / E and only E * / A type is observed. The SI lifetime and ISC rate under our experimental conditions in such as the naphthaleneTEMPO system are just appropriate to observe both A * / E and E * / A type CIDEP depending on the gate time. This dual-spin polarization observed in these systems is another evidence for the idea that the sign of J value is negative in all radical-triplet systems as discussed before, because the phase change is accounted by switching of RTPM and net emissive
polarization is observed at later gate time due to quartet precursor RTPM. Summary of Doublet and Quartet Precursor RTPMs. We observed CIDEP signals due to both quartet and doublet precursor RTPMs and illustrate them in Figure 8. In quartet precursor RTPM, randomly encountered pairs of radical and triplet molecule become quartet and doublet spin states and the latter states undergo the pair of radical and ground-state singlet molecule through triplet quenching. Through this quenching process, quartet spin states are populated much more than doublet spin states. This enhanced population on quartet states produces the E + E / A type CIDEP on radicals due to the quartet4oublet mixing in quartet precursor through magnetic interaction. This type of CIDEP is often observed in the systems with fast SI-TI ISC rates, such as ketones and azaaromatics. On the other hand, in doublet precursor RTPM, the random encounter of the lowest excited singlet molecule and radical yields the doublet spin states of the pair. In this pair, enhanced ISC by radicals occurs and doublet spin states of radical-triplet pair are formed. Hence, the doublet spin states of the pair are initially populated much more than quartet states, which is a condition opposite to that of the quartet precursor RTPM. Therefore, CIDEP generated on radical in this mechanism is the A + A / E type. Doublet precursor RTPM is operative under the following conditions: (1) long lifetime of lowest excited singlet state (350 ns), (2) slow S1-TI ISC rate, and (3) radical concentrations high enough to enhance ISC. Registry No. TEMPO, 2564-83-2; fluoranthene, 206-44-0 coronene, 191-07-1; pyrene, 129-00-0; naphthalene, 91-20-3.
