Article pubs.acs.org/JPCA
Distance Dependence of Electron Spin Polarization during Photophysical Quenching of Excited Naphthalene by TEMPO Radical Vinayak Rane and Ranjan Das* Tata Institute of Fundamental Research, Homi Bhabha Road, Mumbai 400005, India S Supporting Information *
ABSTRACT: Quenching of excited states by a free radical is generally studied in systems where these two are separate entities freely moving in a liquid solution. Random diffusive encounters bring them together to cause the quenching and leave the spins of the radical polarized. In the dynamics of the radical−triplet pair mechanism of the generation of electron spin polarization (ESP), the distancedependent exchange interaction plays a crucial role. To investigate how the distance between the excited molecule and the radical influences the ESP, we have covalently linked a naphthalene moiety to a TEMPO free radical through a spacer group of three different lengths. We compared the ESP process of these linked compounds with that of the usual “unlinked system” of naphthalene and TEMPO through timeresolved EPR experiments at low temperature in n-hexane solution. The time evolution of both the linked and the “unlinked system” was treated on a similar footing. The time-dependent EPR signal was analyzed by combining photophysical kinetics and time-dependent Bloch equations incorporating spin dynamics. Sequential quenching of the singlet state and the triplet state of naphthalene was seen in all the systems, as revealed through the spin-polarized TREPR spectra of opposite phase. The magnitudes of the ESP in the linked molecules were higher than those of the “unlinked system,” showing that when the two moieties are held together greater mixing of quartet−doublet states takes place. The magnitudes of ESP steadily decrease with increasing the length of the spacer group. The polarization magnitudes due to triplet quenching and singlet quenching are very similar, differing by a factor of only ∼2. These characteristics show that for all the linked molecules the quenching takes place in the “weak exchange” regime and at almost the same distance of separation between the two moieties. Our results also showed that observation of small absorptive TREPR signals does not necessarily imply that its magnitude of polarization is small.
■
INTRODUCTION Quenching of excited molecules by free radicals has been studied for several decades. In photophysical quenching, where no chemical reaction takes place, quenching of excited singlet and triplet states generally takes place through energy-transfer mechanism or enhanced intersystem crossing process. That such quenching could also be electron spin selective was demonstrated in the EPR spectra of radicals showing nonBoltzmann spin distribution.1−3 From the early observation of such electron spin polarized EPR spectra, a host of studies on a variety of molecular systems and under different experimental conditions have been reported.4 From these studies, several general features of the dynamics of the quenching process of the generation of electron spin polarization (ESP) have emerged. The mechanism, known as the radical-triplet pair mechanism (RTPM), invokes an important feature of the pair, which is its overall spin state. Because the radical (R) has a doublet state and the excited molecule can be a singlet (S) or a triplet (T), the overall spin state of the pair can be a doublet or a quartet. The energy difference between them is the doublet− quartet exchange energy, J(r), whose magnitude depends exponentially on r, the distance between R and the excited molecule. Generation of ESP in the radical after quenching © 2015 American Chemical Society
involves mixing of states between the quartet (Q) and the doublet (D) states, caused primarily by the zero-field splitting of the triplet and, to a lesser extent, by the hyperfine interaction of the radical. The mixing of states is most favorable when J(r) is comparable to the Zeeman energy of the radical in the magnetic field of the EPR spectrometer. A direct consequence is that in solutions of higher viscosity ESP of a larger magnitude is expected because the radical-excited molecule pair spends more time in the crossing region than in solutions of lower viscosity. Experimental observations have been in accord with this. Most experimental studies of ESP, however, have been conducted on systems where the radical and the excited molecule are separate entities freely moving in a homogeneous solution. Random diffusive encounters between them lead to the formation of a radical-excited molecule pair and quenching and generation of ESP on the separated radical. The observed spin-polarized EPR spectra therefore reveal only an average effect of the otherwise distance-dependent interaction, J(r), which plays a crucial role in the whole process. Received: February 28, 2015 Revised: April 29, 2015 Published: May 22, 2015 5515
DOI: 10.1021/acs.jpca.5b01989 J. Phys. Chem. A 2015, 119, 5515−5523
Article
The Journal of Physical Chemistry A On photoexcitation, the molecule first goes to an excited singlet state and then to a triplet state. The free radical can quench both the states. Because of rather short lifetime of the singlet excited state of most organic molecules, however, quenching by the radical through a bimolecular quenching process is not very probable within its lifetime. Thus, one usually observes the quenching of only the relatively long-lived triplet states by the radical. As a result, in general, the characteristic EPR spectra of emissive phase with a negative spin polarization PT are seen. Quenching of singlet excited states by free radicals has been observed only in molecules such as coronene and pyrene,5,6 which have long singlet-state lifetimes. These systems give EPR spectra in an enhanced absorptive phase with a positive spin polarization PS. In either case, polarization is defined as P ≡ (nβ − nα)/(nβ + nα), where nα and nβ are the population of radicals with α and β spin states. Whether the singlet or the triplet state is quenched by the radical, the essential features of the ESP generation remain the same. Observation of very different magnitudes of PS and PT in the same system, in experiments conducted under identical conditions, thus becomes an important issue. Such observations have been rationalized either by invoking that the singlet and the triplet quenching take place at very different values of r and not in the neighborhood of the level crossing region6,7 or by postulating the presence of another electronic state, often called a charge-transfer state, facilitating the quenching.6 The distance dependence of J(r) is usually written in the following exponential manner J(r ) = J0 e−α(r − d)
sequently, these intermediates remained undetected for almost half a decade after the RTPM was proposed. The first successful recording of these intermediates was done by the Corvaja’s group 9 who achieved this by covalently linking a C60 chromophore to a pyrrolidene radical, thereby increasing the lifetime of the intermediate states. The time-resolved EPR (TREPR) spectra of this system gave two sets of EPR lines, one set corresponding to the ground doublet state D0 and the other to the quartet state Q1. The excited doublet state D1 was not observed because its lifetime was shorter than the instrument response time due to the allowed D1 → D0 transition. Shortly afterward, Yamauchi’s group10 detected both the D1 and Q1 states of a linked system of tetraphenylporphinato zinc and ppyridyl nitronyl nitroxide by recording the TREPR spectrum at 20 K. The low temperature helped to slow down the dynamics, facilitating the detection of the D1 state. The phase of the spin-polarized EPR spectra depends not only on the electronic state of the excited molecule but also on the sign of the exchange interaction J. Its sign decides the relative ordering of the D1 and Q1 states. As with a radical pair, J depends on the particular chromophore and radical involved and also on their mutual distance11 and relative orientation.12 Corvaja’s group12 reported that the phase of the TREPR spectra of the radical changes when the number of bonds between the two moieties increases from five to six. They explained this observation by assuming that the sign of J changes in the two cases and that the sign is determined by the relative orientations of the two moieties. Yamauchi’s group13 also demonstrated the phase of the TREPR spectra in the linked system of metallo-porphyrin and pyridyl nitronyl nitroxide radical (nit-py) to depend on the specific isomer of nit-py. They rationalized this by proposing that the sign of J depended on whether the nitroxide orbital interacted directly with prophyrin triplet (ortho and meta isomers of nit-py) or via the intermediate pyridine ring (para isomer of nit-py). The group of Wasielewski14,15 covalently attached t-butylphenyl niroxyl free radical to perylene-3,4:9,10-bis(dicarboximide) at specific distances and orientations. On photoexcitation, the observed ESP of the radical was found to depend on the molecular structure14 and rationalized by invoking that the electronic coupling depends on the through-bond coupling between the two moieties. They also proposed that the radical and the triplet experience weak exchange coupling.15 All these examples highlight the importance of studying the ESP of linked molecular systems of the radical and chromophore in a systematic manner, with special attention to their separation and orientation. In the work reported here, we have chosen naphthalene (Nap) and 2,2,6,6-tetramethylpiperidine N-oxyl free radical (TEMPO) pair and covalently linked them through a spacer group of variable length. The general symbol for these molecules is Nap-Spacer-TEMPO. Naphthalene was selected because its photochemistry and photophysics are well known. On photoexcitation, it forms a long-lived triplet state with a quantum efficiency of 0.75 to 0.80.16 Quenching of this state has been extensively studied by both optical and EPR techniques. We chose TEMPO because of its stability and the availability of its various substituted derivatives, which can be used as starting materials in the synthesis. We compared the ESP behavior of these linked molecules with the “unlinked system” consisting of naphthalene and TEMPO present as separate entities in the solution. In our preliminary investigations on the spin-polarized EPR spectra, we observed unusual shapes of the TREPR spectra of the linked
(1)
where J0 and α are parameters and d is the distance of closest approach of the radical and the excited molecule. Depending on the relative magnitudes of J0 and the Zeeman energy, ω0, systems have been classified under strong exchange (|J0| ≫ ω0) and weak exchange (|J0| ≪ ω0) regimes. The observed ESP has been rationalized by postulating that the quenching takes place at very different values of r in the two regimes.7 Theoretical work of Shushin8 shows P to be independent of J0 in the strong exchange regime but proportional to J0 in the weak exchange regime. The above discussion shows that studying the ESP on a specific radical-excited molecule pair, where the only variable is their mutual distance r, one should be able to gain important insight. Such studies cannot be carried out if the radical and the excited molecules are separate entities. By linking the two, however, the distance between them can be kept largely fixed. There are several reports on the ESP studies on such linked systems, demonstrating several finer details of the quenching process. As previously mentioned, one of the key postulates of the RTPM theory is the existence of doublet and quartet spin intermediates, which are formed during the collisional encounters of the excited molecule and the radical. These intermediates are difficult to characterize by optical techniques, as their electronic properties are very similar to those of separate chromophore and the radical, but because their spin states are different, they can be characterized by EPR technique. In the case of freely diffusing excited molecule and the radical, however, these intermediates are difficult to detect. This is because their lifetime in solution is determined by the lifetime of the collisional encounter, which is usually much shorter than the typical response time of an EPR spectrometer. Con5516
DOI: 10.1021/acs.jpca.5b01989 J. Phys. Chem. A 2015, 119, 5515−5523
Article
The Journal of Physical Chemistry A Chart 1. Structures and Symbolic Names of the Molecular Systems Studied in This Work
a
“Unlinked system” denotes an equimolar homogeneous mixture of naphthalene and TEMPO.
molecule.17 These shapes evolved with time, appearing as extra splitting of each of the nitrogen hyperfine lines. Careful analysis of the time evolution of the TREPR spectra revealed that they are the result of a detailed interplay of photophysical and magnetic resonance dynamics of the sequential quenching of the singlet and the triplet naphthalene moiety by TEMPO linked to it. Here we found the quenching efficiency to decrease with increasing the distance between the two moieties. This decrease correlated with the magnitude of the ESP generation.
other molecules resulted in either very poor yield (n = 5) or no reaction at all (n = 2, 3). Ground-state absorption, steady-state fluorescence, and phosphorescence measurements were done in a UV/vis spectrophotometer (Lambda 25, PerkinElmer) and JobinYvon spectrofluorimeter, respectively. Fluorescence lifetimes were recorded by TCSPC technique in a laboratory-built setup. No degassing was done for these measurements because no effect of oxygen was observed on the lifetimes of the linked molecules. All these experiments were done in solutions of spectroscopic grade n-hexane (Merck). Lifetimes and quenching constants of triplet states were measured in a nanosecond laser flash photolysis apparatus assembled in the laboratory. Triplet-state EPR measurements at low temperatures were done on a Bruker X-band spectrometer (EMX Micro). Other steady-state EPR experiments were done on a laboratory-built X-band spectrometer modified for steady-state experiments. TREPR experiments were done on the same spectrometer modified for time-resolved experiments. Electron spin−lattice relaxation times (T1) of TEMPO and the linked molecules were obtained by the pulse saturation recovery technique. The response time of the TREPR spectrometer was found to be ∼70 ns. All TREPR experiments were done at 5 mW of microwave power. This choice of the microwave power was based on the criteria of obtaining the most intense EPR spectra with a good S/N and, at the same time, avoid saturation and Torrey oscillations. An excimer laser (Coherent) operating at 248 nm was used as the excitation light for TREPR and nanosecond flash photolysis experiments. To avoid accumulation of contaminants due to photodegradation, all TREPR experiments were done in a flow setup. The solutions were degassed with nitrogen gas (IOLAR grade) for ∼30 min before carrying out TREPR measurements. Steady-state EPR spectra of the sample were recorded before and after the TREPR experiment to check for any photodecomposition. In general, about 5−10% decrease was observed for the linked molecules during the course of the experiments. The size of the laser output was 2 × 1 cm2, and the laser entry hole into the EPR cavity had a diameter of 6 mm. Thus, the entire energy of laser could not reach the sample flowing through a flat cell kept in the cavity. To measure the actual energy entering the cavity, we placed a small slit of the same dimensions as the laser entry hole in a collinear fashion in
■
EXPERIMENTAL SECTION The structures and symbolic names of different molecular systems studied in this work are given in Chart 1. For these linked molecules, we wanted to ensure that the electronic properties of naphthalene did not change on linking it to the free radical. Hence naphthalene was attached to TEMPO via a methylene group at its β position. All the linked molecules therefore had a “Nap-CH2-” moiety. Thus, by preventing electron delocation away from the naphthalene ring, the electronic properties of the naphthalene moiety of the linked molecules are expected to be very similar to those of an isolated naphthalene molecule. Williamson’s method of ether synthesis was used for synthesizing the linked molecules. Detailed procedures for synthesizing and characterizing them will be described elsewhere. The linked molecule Nap-CH2-O-TEMPO was synthesized in one step by the reaction of 2-bromomethyl naphthalene and 4-hydroxy-TEMPO. Linked molecules NapCH2-O-(CH2)n-O-TEMPO were synthesized in two steps. The first step involved reaction of 2-naphthalene methanol with the linker 1,n-dibromoalkane. In this reaction, we observed a unimolecular decomposition of the starting material 2naphthalene methanol under basic conditions in dimethylformamide solvent. It was necessary to circumvent this reaction so as to improve the yield of the required molecule. This was achieved by using the 1,n-dibromoalkane itself as the solvent. This method gave reasonable yields of the required molecule. The product of this first step was then reacted with 4-hydroxyTEMPO, which gave the required linked molecules. Using the above scheme, our objective was to synthesize linked molecules of the type Nap-CH2-O-(CH2)n-O-TEMPO with n varying from 2 through 6; however, we could synthesize only the molecules with n = 4 and 6. Attempts to synthesize the 5517
DOI: 10.1021/acs.jpca.5b01989 J. Phys. Chem. A 2015, 119, 5515−5523
Article
The Journal of Physical Chemistry A
Figure 1. TREPR of (A) “unlinked system” naphthalene and TEMPO, 3 mM each, (B) Nap-CH2-O-TEMPO, (C) Nap-CH2-O-(CH2)4-O-TEMPO, and (D) Nap-CH2-O-(CH2)6-O-TEMPO in degassed n-hexane (concentration 1.5 mM) at various times after a 248 nm laser pulse. Boxcar gate width: 0.1 μs, temperature:−45 °C, microwave power: 5 mW. The extra splittings are clearly visible in the “unlinked system” at 0.1 μs and in NapCH2-O-TEMPO at 0.2 μs. During some scans, the magnetic field was made to skip the regions where no EPR signal was expected, so as to reduce the scan time and exposure of laser light on the sample.
molecules were generally higher than when the two moieties are not linked, and among the linked molecules, magnitude of ESP decreased with increasing the distance between naphthalene and TEMPO moieties. Optical studies of fluorescence quenching also showed that the quenching decreased with increasing the distance. Because the solvent nhexane and the spacer groups of the linked molecules are made of linear hydrocarbon chains, the linked molecules are expected to be present in an extended conformation in the solution. Qualitative results from the spin-polarized EPR spectra and the fluorescence quenching efficiency are consistent with this. Observation of emissive spectra also indicated that J, the energy of separation between the doublet and quartet levels, is negative in all these systems. On lowering the temperature, the S/N ratio of all systems improved considerably. At about −45 °C, TREPR spectra with a moderate S/N could be recorded. Figure 1 shows such spectra at different times after the laser pulse. Molecules NapCH 2 -O-(CH 2 ) 4 -O-TEMPO and Nap-CH 2 -O-(CH 2 ) 6 -OTEMPO showed emissive signals from ∼0.1 μs onward. But Nap-CH2-O-TEMPO and the “unlinked system” showed absorptive signals at early times, which rapidly changed to emissive signals. Thus, the ESP mechanism changed from doublet precursor RTPM to quartet precursor RTPM. This demonstrates that singlet quenching precedes triplet quenching. When the length of the spacer group increased, the contribution of the absorptive signal due to doublet precursor RTPM decreased substantially, and the absorptive component could not be detected in the spectra shown here. Also, we could not detect any signal from doublet or quartet intermediates. Kawai and Obi5 reported TREPR spectra of the naphthaleneTEMPO “unlinked system” in 2-propanol that showed phase inversion from absorptive to emissive, but did not give any detailed analysis of the evolution of the line shape. The spectra
front of, and very close to, the cavity. A power meter was kept between the two to determine the amount of laser energy reaching the sample. Usually, energy of about 8−10 mJ per pulse of 248 nm was entering the cavity for TREPR experiments. The stability of the molecules was the main criterion for the choice of solvent. To check the stability of the linked molecules, we made equimolar solutions of them in acetonitrile, n-hexane, cyclohexane, and 2-propanol. The solutions were then irradiated with a 266 nm filtered output of a UV lamp in a continuous mode, and the fluorescence signal of naphthalene was monitored. The least amount of change in the fluorescence was observed in n-hexane and cyclohexane on prolonged exposure to the UV light, implying maximum stability of the linked molecule in these two solvents. We then chose n-hexane, as it has a much lower freezing point than cyclohexane, as all our experiments would be conducted at temperatures well below −40 °C.
■
RESULTS AND DISCUSSION Qualitative Observations and Inference from TREPR Spectra. Initial TREPR experiments of the unlinked and linked molecules in n-hexane solution at room temperature of 21 °C showed EPR spectra of poor S/N in all systems except NapCH2-O-TEMPO at 0.5 μs after the 248 nm laser excitation. The best S/N ratio of ∼8 was obtained with Nap-CH2-O-TEMPO. Emissive spectra were seen for all the linked molecules as well as for the “unlinked system” in solution. The emissive polarization of TEMPO is in line with the results of Kawai et al.5 on the naphthalene-TEMPO system in 2-propanol; however, in our case, the S/N for the “unlinked system” was very poor. We attributed this small signal to the low viscosity of n-hexane compared with 2-propanol. Qualitative conclusions from these spectra are that the ESP magnitudes of the linked 5518
DOI: 10.1021/acs.jpca.5b01989 J. Phys. Chem. A 2015, 119, 5515−5523
Article
The Journal of Physical Chemistry A of the “unlinked system” seen here are qualitatively similar to those of Kawai and Obi,5 but we noticed that at a very early time of 100 ns, all three hyperfine lines showed a shape resembling the letter M, as if they were split into two. Such a line shape has not been previously reported. With the passage of time, the line became purely emissive and then relaxed to thermally equilibrated signals. In the ESP of phthalocyaninatosilicon covalently linked to TEMPO, Ishii et al.18 interpreted the emissive spectra of the radical to arise from radical−quartet pair mechanism involving intermolecular interaction between the quartet excited state of one molecule and the doublet ground state of another. To investigate if the similar process could also operate in our linked systems, we recorded the time profiles of the EPR signals for all of these systems at various concentrations. Figure 2 shows the
interaction, making the hyperfine interaction negligible in comparison. Modeling the Dynamics of the Quenching Process. On the basis of the above qualitative observation, we now model the detailed scheme of the quenching process giving rise to the evolution of the spin system. We have treated the dynamics of the “unlinked system” and the linked molecule essentially on a similar footing, except that the intermolecular second-order quenching of excited naphthalene by TEMPO of the “unlinked system” was replaced by the respective intramolecular processes in the linked molecules (Scheme 1). The various processes with their corresponding rate constants are shown below. 1. S1 → S0
Radiative (fluorescence) decay,
2. S1 → T1
k fl Intersystem crossing of S1, kISC
3. S1 + R → T1 + RS* Quenching of S1 by radical via EISC, kqs 4. T1 + R → S0 + RT * Quenching of T1 by radical via EISC, kqt 5. T1 → S0
Intersystem crossing of T1, k T
S0, S1, T1, and R refer to the ground state, excited singlet and triplet states of naphthalene, and the doublet TEMPO radical with Boltzmann spin distribution, respectively. RS* (RT*) denotes spin-polarized radicals from doublet (quartet) precursor with absorptive (emissive) polarization. Decay of the S1 state through internal conversion according to the following pathways has been neglected. 6. S1 → S0 Internal conversion, kIC 7. S1 + R → S0 + R Enhanced internal conversion, kEIC
Figure 2. Time profiles of the TREPR signals due to electronic quenching of naphthalene by TEMPO in (A) “unlinked system” ([Nap] = 1.5 mM and [TEMPO] = 0.36 mM (red), 0.75 mM (blue), 1.5 mM (green)) and (B) Nap-CH2-O-TEMPO (conc. = 0.2 mM (blue), 0.5 mM (red), 1 mM (green)) in n-hexane at −45 °C. Microwave power: 5 mW. The downward sharp spike on the left was due to the jitter in the laser emission.
The fact that the internal conversion is negligible in naphthalene is based on the two key observations: (1) the sum of the fluorescence quantum yield (Process 1) and the intersystem crossing (Process 2) is unity19 and (2) there is no difference in the fluorescence quantum yield between naphthalene-h8 and naphthalene-d8,20 showing that the C−H vibration, the dominant factor contributing to the internal conversion process in rigid aromatic molecules, is ineffective in naphthalene. For the same reason, Process 7, being intermolecular, is also neglected. For the linked molecules, again for the same reason, we do not expect that replacing the hydrogen of the C−H group of naphthalene by the spacer moiety could significantly change vibrational activities to induce internal conversion. Under the pseudo-first-order condition satisfied by the concentrations of S1, T1, and R in our experiments, the effective rate constants of Processes 3 and 4 are given by kqs[R] and kqt[R], respectively. The spin−lattice relaxation process, kSLR, brings the polarized radicals, RS* and RT*, to Boltzmann distribution. The linked molecules are denoted by S−R or T−R, where S and T denote the singlet and triplet naphthalene, respectively. In particular, S1−R (T1−R) denotes the electronic state of the linked molecule, where the naphthalene moiety in its singlet (triplet) excited state is linked to the doublet radical, giving rise to the so-called “sing-doublet” (“trip-doublet”) state. Kinetic steps similar to those previously shown have been used
results. The time profile of the “unlinked system” depended strongly on the concentration of the radical. This is expected, because the intermolecular quenching in this system is bimolecular. In contrast, the EPR time profile of Nap-CH2O-TEMPO did not show any observable dependence on its concentration in the similar range. Similar behavior was observed for the other linked molecules Nap-CH2-O-(CH2)4O-TEMPO and Nap-CH2-O-(CH2)6-O-TEMPO also (data not shown). This proved that the quenching in linked molecules is intramolecular. The intensities of the three lines in the TREPR spectra of the “unlinked system”, displayed in Figure 1, showed considerable dependence on the spin of the nitrogen nucleus, giving an E + E/A pattern. This is again in accord with the observation of Kawai and Obi.5 In contrast, in all linked systems, such dependence was much weaker. The variation of intensities seen in Figure 1 for these linked molecules is no more than the general instability or reproducibility in different scans of the spectra. This indicated that in the linked molecules, ESP due to the zero-field splitting of triplet naphthalene was the dominant 5519
DOI: 10.1021/acs.jpca.5b01989 J. Phys. Chem. A 2015, 119, 5515−5523
Article
The Journal of Physical Chemistry A Scheme 1. Photophysical Pathways and their Rate Constants for the “Unlinked” (A) and the Linked (B) Systems
Table 1. Values of Various Kinetic Parameters Used in the Simulations of the Time-Dependent EPR Signal at −56 °Ca molecular system Nap+TEMPO (“unlinked system”) Nap-CH2-O-TEMPO Nap-CH2-O-(CH2)4-OTEMPO Nap-CH2-O-(CH2)6-OTEMPO
T1/μs, fittedb
T1/μs, measuredc
T2/μs
0.34
0.38
0.050
1.0 1.3
0.64 0.65
1.4
0.67
kqs/s−1
kqt/s−1
0.062 0.062
kqs[R] = (12.3 ± 0.2) × 106d (3.0 ± 0.9) × 109 (1.1 ± 1.0) × 109
kqt[R] = (1.8 ± 0.1) × 106d (30 ± 7) × 106 (12 ± 1) × 106
0.062
(0.35 ± 0.25) × 109
(15 ± 4) × 106
PS/Peq
PT/Peq
4.1 ± 0.2
−7.7 ± 0.3
21.0 ± 1.0 6.5 ± 1.5
−29.0 ± 1.0 −15.0 ± 2.0
4.5 ± 1.0
−9.0 ± 1.0
a Errors in T1: ± 0.02 μs, T2: ± 0.002 μs. bT1 obtained from simulation. cT1 measured by the pulse saturation recovery technique on the same solution that was used for TREPR experiments. dFor the “unlinked system”, the pseudo-first-order values of the rate constants are given. These values were calculated at [TEMPO] = 1.5 mM, using the bimolecular singlet quenching constant kqs = of 8.2 × 109 M−1 s−1 and triplet quenching constant kqt = 1.8 × 109 M−1 s−1. For linked molecules, kqs was taken to be the inverse of their fluorescence lifetimes, and kqt values were obtained by treating them as an adjustable parameter while fitting the simulations to the observed EPR signal.
Information for details. Table 1 shows the values of various kinetic parameters used in these calculations. The time evolution of the EPR line shape of Nap-CH2-OTEMPO and its curious appearance resembling the letter W has been already described.17 A consequence of the timedependent Bloch equations is that the width of an EPR line arising from a magnetization evolving with time is also time dependent. During the quenching of an excited state, the spins of the radical suddenly undergo a change of magnetization. As the quenching continues with time, the radicals that quench the excited state at early times acquire narrower EPR line widths, and those that quench the excited state at later times acquire broader line widths. The superposition of these shapes gives the overall appearance of the EPR spectra at a given time. This way the complete evolution of the EPR lines of the Nap-CH2-OTEMPO could be well reproduced.17 To understand the origin of the unusual EPR spectra for the unlinked system, we
in describing the evolution of line shapes of Nap-CH2-OTEMPO.17 The intrinsic triplet decay of naphthalene (Process 5), associated with the rate constant kT, has been omitted for the linked systems. This was done because of it being too slow to be of any consequence, in comparison with the intramolecular quenching process given by kqt. In Scheme 1, no detailed dynamics of RTPM has been considered, for example, the time evolution of state mixing. Thus, the scheme should be considered as phenomenological. The rate equations for the various states after the laser pulse were set up and solved using Laplace transform technique to obtain their concentrations as a function of time. Then, the time-dependent Bloch equations were solved to obtain the magnetization of the nitroxyl radical as a function of time and magnetic field. The calculated timedependent EPR signal intensity was convolved with the instrument response function of the EPR spectrometer for comparing with the observed intensity. See the Supporting 5520
DOI: 10.1021/acs.jpca.5b01989 J. Phys. Chem. A 2015, 119, 5515−5523
Article
The Journal of Physical Chemistry A
Figure 3. (A) Experimental TREPR spectra, corresponding to the low-field hyperfine line of the “unlinked system” (naphthalene and TEMPO, 3 mM each) in n-hexane at −45 °C, at various delay times after the 248 nm laser pulse. Boxcar gate width: 50 ns, microwave power: 5 mW. (B) Calculated spectra at the denoted times with 50 ns integration time.
linked to the radicals, some molecules of naphthalene and the radical can come very close, much closer than the linked molecules studied here. TEMPO is known to quench excited states by electron transfer, too.21 Thus, one could envisage the following scheme to be another channel for the photochemical quenching of naphthalene, when it is very close to the radical.
recorded its TREPR spectra with improved S/N and at small increments of the delay time after the laser pulse. Simulations were carried out using the relevant values of various kinetic processes, and the results are shown in Figure 3. As previously stated, at the shortest possible time this system gave a line shape similar to the letter M. Another kind of shape was seen around the transition time from the absorptive to the emissive phase at ∼250 ns. Its shape was like the letter W and very similar to what has been seen in Nap-CH2-O-TEMPO.17 The fact that the two kinds of “splitting” were not same was evident from their temporal behavior. The M-shaped splitting was present right from the early time and then disappeared at ∼150 ns. The W-shaped splitting appeared late, at ∼250 ns, when the absorptive RTPM changed to emissive RTPM. The quenching here being bimolecular, the time of appearance of these features depended on the concentration of the radical. It is apparent from the simulated spectra that the M-shaped feature at early time was not reproduced in our calculations. What was needed to simulate the M shape was a small component of a narrow emissive signal riding on the broad absorptive signal. This emissive signal should be present at a time earlier than the absorptive signal, so that its narrow width could be justified. Because the absorptive signal is due to the quenching of the S1 state of naphthalene, the narrow emissive signal requires quenching, possibly of a triplet state earlier than this S1 state. It is not possible to imagine a photophysical process of this type. A possible cause of the M-shaped lines could be depletion of the radicals due to the laser pulse. Our TREPR spectra were recorded in a dual boxcar setup. Here the gate of one boxcar (gate B) opened ∼0.3 μs before the laser pulse, and that of the other boxcar (gate A) opened at the desired time after the laser pulse. The boxcar operated in the “A−B” mode, giving the difference in the microwave signals at these two gate times. Thus, depletion of radical by the laser pulse will appear as a negative signal. We propose that that was the origin behind the appearance of the M-shaped spectra at early times in the “unlinked system”. Because naphthalene molecules are not
Nap*(S1) + R· → Nap− + R+ → 1or3 Nap(S0) + R·
1
If a small fraction undergoes such a fast electron-transfer reaction, the EPR signal should show depletion, giving an impression of emissive signal. Subsequent back electron transfer would show the disappearance of the emissive-looking signal. The same calculations that reproduced the EPR line shapes at different times should also be able to describe the time evolution of the EPR signal at a fixed magnetic field. In this, the range of possible values of various kinetic parameters can be tightly controlled. For that purpose, we recorded the EPR signal at the peak of the low-field hyperfine line. By improving the cooling system, we could lower the temperature of the samples in the EPR cavity to about −56 °C. This improved the S/N further, and after signal averaging, high-quality EPR signals were obtained. These are shown in Figure 4 along with the calculated signals. The calculated signals reproduced the observed ones satisfactorily well for all systems. Magnitudes of Singlet and Triplet Polarization, PS and PT. The values of PS and PT determined from the simulations show that the magnitudes of the polarization of all linked molecules are higher than those of the “unlinked system”. This shows that holding the chromophore and radical at a given distance helps increase the probability of mixing of D and Q states, which is akin to increasing the viscosity of the solution for the “unlinked system”. Because the distance remained fixed for the linked molecules, the rapid decrease in the magnitude of polarization with increasing the length of the spacer showed that all of them were in the “weak exchange” regime, proposed by Shushin.8 We estimated the approximate distance between the center of the naphthalene moiety and the nitrogen atom of 5521
DOI: 10.1021/acs.jpca.5b01989 J. Phys. Chem. A 2015, 119, 5515−5523
Article
The Journal of Physical Chemistry A
magnitudes of PT and PS to be the same, or the ratio PT/PS should be very nearly equal to 1. The reason for PS being smaller than PT may lie in our assuming that the triplet naphthalene is generated, after laser excitation, through Processes 2 and 3 only, producing RS*. It is likely that a fraction of the molecules populate the trip-doublet states without the radicals being spin-polarized. In other words, the T1−R states can be populated by the EISC route causing PS polarization and also by a parallel route without leading to absorptive spin polarization. This parallel path could be through the charge-transfer interaction.11,22 Quenching from T1−R brings the entire population to S0−R, giving a larger value of PT polarization. Recovery of Spin-Polarized Signal and Spin−Lattice Relaxation. Finally, we noted that the spin−lattice relaxation times that we measured (T1,measured) by the pulse saturation recovery technique did not fit well the time dependence of the growth of the emissive EPR signal to the thermally equilibrated signal. A small, but systematic, discrepancy between this relaxation time and the T1 needed to fit (T1,fitted) the observed signals was seen; T1,fitted was a little longer than T1,measured in all systems. This curious trend needs to be investigated in detail. We tentatively believe that Heisenberg spin exchange could be playing a role here. In pulse saturation recovery experiments, the hyperfine lines other than the one whose spin−lattice relaxation is being measured can provide alternative paths for relaxation through Heisenberg spin exchange, but in the ESP through the photophysical quenching, all hyperfine lines are almost equally populated, as the contribution of the hyperfine interaction to ESP is small. Thus, Heisenberg spin exchange, even if it takes place, cannot offer extra paths for spins to relax. Thus, the time dependence of the emissive signal to return to thermal equilibrium will be governed by the true, intrinsic T1 of the radical.
Figure 4. Time evolution of the observed EPR signals (noisy blue curves) and along with their simulation (smooth red curves). (A) “Unlinked system” (naphthalene and TEMPO, 3 mM each), (B) NapCH2-O-TEMPO, (C) Nap-CH2-O-(CH2)4-O-TEMPO, and (D) NapCH2-O-(CH2)6-O-TEMPO. Temperature: −56 °C, microwave power: 5 mW. See the Supporting Information for the steps involved in processing the observed signals before they were compared with the simulated signals.
TEMPO moiety, assuming that they are in an extended conformation. As J changes exponentially with distance (eq 1), the logarithm of P should decrease linearly with the distance. With only the three linked molecules studied here, we could see an approximate linearity of this kind. It is important to study more linked molecules with shorter spacer groups, so that distance of separation lies in-between that of Nap-CH2-OTEMPO and Nap-CH 2 -O-(CH 2 ) 4 -O-TEMPO, before a definitive conclusion can be drawn on the applicability of Shushin’s equation8 and also on the form of the distance dependence of J (eq 1) that should apply to the RTPM process. As mentioned in the Introduction, whenever both PS and PT have been reported, their values are usually very different. Such different values have been proposed to arise from different interaction regions for the two processes. Our values of PS and PT are very similar, and the ratio PT/PS was about 2 for all the linked systems. This showed that both singlet and triplet quenching processes are almost equally effective in generating ESP, and they take place extremely rapidly in quick succession. This is also consistent with the time evolution of the signal intensity. The absorptive component was seen to be much smaller than the emissive component, even though the magnitudes of PT and PS were not very different. The absorptively polarized radicals are generated by converting the singlet excited naphthalene moiety to triplet. These naphthalene moieties are also attached to the radical; they are quenched almost immediately, leaving their radical partner emissively polarized. Hence the absorptive EPR signal does not last long, and its EPR signal is unable to reach its full intensity. On the basis of these results, we wish to propose a general rule: Observation of small absorptive TREPR signals does not necessarily imply low polarization. Because the mechanism of ESP from a doublet precursor is very similar to that from a quartet precursor, we believe the
■
SUMMARY AND CONCLUSIONS We attempted to investigate how the ESP would be affected when the quencher radical and the excited molecule are not allowed to move freely in a solution but are held together. We linked a naphthalene moiety to a TEMPO free radical, separated by a spacer group of three different lengths. Sequential quenching of the singlet state and the triplet state of naphthalene was seen in all the systems and revealed through the spin-polarized TREPR spectra of opposite phase. The polarization magnitudes in the linked molecules were higher than that of the “unlinked system”, showing that when the two moieties are held together, the mixing probability of D and Q states is greatly enhanced. The magnitudes of the ESP steadily decreased with increasing the length of the spacer group. The polarization magnitudes due to triplet quenching and singlet quenching are very similar, differing by only a factor of about 2. These characteristics show that for all of the linked molecules, both of the quenching processes take place in the “weak exchange” regime and at almost the same distance between the two moieties. Our results also showed that observation of small absorptive TREPR signals does not necessarily imply that its magnitude of polarization is small. Finally, our measurements showed that the time constant for recovery of the spinpolarized EPR signal to thermal equilibrium is slower than the electron spin−lattice relaxation times measured by the pulse saturation recovery technique. We proposed this discrepancy to arise from the contribution of Heisenberg spin exchange to the electron spin relaxation process. 5522
DOI: 10.1021/acs.jpca.5b01989 J. Phys. Chem. A 2015, 119, 5515−5523
Article
The Journal of Physical Chemistry A
■
(14) Giacobbe, E. M.; Mi, Q.; Colvin, M. T.; Cohen, B.; Ramanan, C.; Scott, A. M.; Yeganeh, S.; Marks, T. J.; Ratner, M. A.; Wasielewski, M. R. Ultrafast Intersystem Crossing and Spin Dynamics of Photoexcited Perylene-3,4:9,10-bis(dicarboximide) Covalently Linked to a Nitroxide Radical at Fixed Distances. J. Am. Chem. Soc. 2009, 131, 3700−3712. (15) Colvin, M. T.; Giacobbe, E. M.; Cohen, B.; Miura, T.; Scott, A. M.; Wasielewski, M. R. Competitive Electron Transfer and Enhanced Intersystem Crossing in Photoexcited Covalent TEMPO-Perylene3,4:9,10-bis(dicarboximide) Dyads: Unusual Spin Polarization Resulting from the Radical-Triplet Interaction. J. Phys. Chem. A 2010, 114, 1741−1748. (16) Murov, S. L.; Carmichael, I.; Hug, G. L. Handbook of Photochemistry, 2nd ed.; Marcel Dekker, Inc.: New York, 1993; pp 30−31. (17) Rane, V.; Das, R. Observation of Splitting of EPR Spectral Lines without Any Concomitant Splitting in Energy Levels. J. Phys. Chem. A 2014, 118, 8689−8694. (18) Ishii, K.; Hirose, Y.; Kobayashi, N. Electron Spin Polarizations of Phthalocyaninatosilicon Covalently Linked to One TEMPO Radical in the Excited Quartet and Doublet Ground States. J. Phys. Chem. A 1999, 103, 1986−1990. (19) Kellogg, R. E.; Bennett, R. G. Radiationless Intermolecular Energy Transfer. III. Determination of Phosphorescence Efficiencies. J. Chem. Phys. 1964, 41, 3042−3045. (20) Lim, E. C.; Laposa, J. D. Radiationless Transitions and Deuterium Effect on Luminescence of Some Aromatics. J. Chem. Phys. 1964, 41, 3257−3259. (21) Green, S.; Fox, M. A. Intramolecular Photoinduced Electron Transfer from Nitroxyl Radicals. J. Phys. Chem. 1995, 99, 14752− 14757. (22) Kawai, A. Dynamic Electron Polarization Created by the Radical-Triplet Pair Mechanism: Application to the Studies on Excited State Deactivation Processes by Free Radicals. Appl. Magn. Reson. 2004, 23, 349−367.
ASSOCIATED CONTENT
S Supporting Information *
Detailed descriptions of the simulation of TREPR spectra, processing of observed EPR signals for determination of the absolute magnitudes of the polarization, and determination of the response time of the EPR spectrometer. The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpca.5b01989.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail: [email protected]. Notes
The authors declare no competing financial interest.
■ ■
ACKNOWLEDGMENTS We sincerely thank Dr. Prasanna S. Ghalsasi of M. S. University for his help and advice in the synthesis of molecules. REFERENCES
(1) Blättler, C.; Jent, F.; Paul, H. A Novel Radical-Triplet Pair Mechanism for Chemically Induced Electron Polarization (CIDEP) of Free Radicals in Solution. Chem. Phys. Lett. 1990, 166, 375−380. (2) Blättler, C.; Paul, H. Cidep after Laser Flash Irradiation of Benzil in 2-Propanol. Electron Spin Polarization by the Radical-Triplet Pair Mechanism. Res. Chem. Intermed. 1991, 16, 201−211. (3) Kawai, A.; Okutsu, T.; Obi, K. Spin Polarization Generated in the Triplet-Doublet Interaction: Hyperfine-Dependent Chemically Induced Dynamic Electron Polarization. J. Phys. Chem. 1991, 95, 9130− 9134. (4) Kawai, A.; Shibuya, K. Electron Spin Dynamics in a Pair Interaction Between Radical and Electronically-Exicted molecule as Studied by a Time-Resolved ESR Method. J. Photochem. Photobiol. C 2006, 7, 89−103. (5) Kawai, A.; Obi, K. First Observation of a Radical-Triplet Pair Mechanism (RTPM) with Doublet Precursor. J. Phys. Chem. 1992, 96, 52−56. (6) Mitsui, M.; Kobori, Y.; Kawai, A.; Obi, K. Quenching Mechanism of Excited Coronene by a Nitroxide Radical Studied by Probing Dynamic Electron Polarization. J. Phys. Chem. A 2004, 108, 524−531. (7) Terazono, H.; Kawai, A.; Tsuji, K.; Shibuya, K. Enhanced Intersystem Crossings of S1-T1 and T1-S0 in Coronene- and Pyrenegalvinoxyl Systems as Studied by a Pulsed ESR Method. J. Photochem. Photobiol., A 2006, 183, 22−30. (8) Shushin, A. I. The Relaxation Mechanism of Net CIDEP generation in Triplet-Radical Quenching. Chem. Phys. Lett. 1993, 208, 173−178. (9) Corvaja, C.; Maggini, M.; Prato, M.; Scorrano, G.; Venzin, M. C60 Derivative Covalently Linked to a Nitroxide Radical: Time-Resolved EPR Evidence of Electron Spin Polarization by Intramolecular RadicalTriplet Pair Interaction. J. Am. Chem. Soc. 1995, 117, 8857−8858. (10) Ishii, K.; Fujisawa, J.; Ohba, Y.; Yamauchi, S. A Time-Resolved Electron Paramagnetic Resonance Study on the Excited States of Tetraphenylporphinatozinc(II) Coordinated by p-Pyridyl Nitronyl Nitroxide. J. Am. Chem. Soc. 1996, 118, 13079−13080. (11) Kawai, A. FT-EPR Study on the Sign of Exchange Interaction in Triplet Naphthalene-Galvinoxyl Encounter Pair. Appl. Magn. Reson. 2004, 26, 213−221. (12) Mazzoni, M.; Conti, F.; Corvaja, C. The Sign of the Exchange Interaction between Triplet Excited Fullerene and Nitroxide Free Radicals. Appl. Magn. Reson. 2000, 18, 351−361. (13) Fujisawa, J.; Ishii, K.; Ohba, Y.; Yamauchi, S.; Fuhs, M.; Möbius, K. X- and W-Band Time-Resolved Electron Paramagnetic Resonance Studies on Radical-Excited Triplet Pairs between Metalloporphyrins and Axial-Ligating Nitroxide Radicals. J. Phys. Chem. A 1997, 101, 5869−5876. 5523
DOI: 10.1021/acs.jpca.5b01989 J. Phys. Chem. A 2015, 119, 5515−5523
Distance Dependence of Electron Spin Polarization during Photophysical Quenching of Excited Naphthalene by TEMPO Radical Vinayak Rane and Ranjan Das* Tata Institute of Fundamental Research, Homi Bhabha Road, Mumbai 400005, India S Supporting Information *
ABSTRACT: Quenching of excited states by a free radical is generally studied in systems where these two are separate entities freely moving in a liquid solution. Random diffusive encounters bring them together to cause the quenching and leave the spins of the radical polarized. In the dynamics of the radical−triplet pair mechanism of the generation of electron spin polarization (ESP), the distancedependent exchange interaction plays a crucial role. To investigate how the distance between the excited molecule and the radical influences the ESP, we have covalently linked a naphthalene moiety to a TEMPO free radical through a spacer group of three different lengths. We compared the ESP process of these linked compounds with that of the usual “unlinked system” of naphthalene and TEMPO through timeresolved EPR experiments at low temperature in n-hexane solution. The time evolution of both the linked and the “unlinked system” was treated on a similar footing. The time-dependent EPR signal was analyzed by combining photophysical kinetics and time-dependent Bloch equations incorporating spin dynamics. Sequential quenching of the singlet state and the triplet state of naphthalene was seen in all the systems, as revealed through the spin-polarized TREPR spectra of opposite phase. The magnitudes of the ESP in the linked molecules were higher than those of the “unlinked system,” showing that when the two moieties are held together greater mixing of quartet−doublet states takes place. The magnitudes of ESP steadily decrease with increasing the length of the spacer group. The polarization magnitudes due to triplet quenching and singlet quenching are very similar, differing by a factor of only ∼2. These characteristics show that for all the linked molecules the quenching takes place in the “weak exchange” regime and at almost the same distance of separation between the two moieties. Our results also showed that observation of small absorptive TREPR signals does not necessarily imply that its magnitude of polarization is small.
■
INTRODUCTION Quenching of excited molecules by free radicals has been studied for several decades. In photophysical quenching, where no chemical reaction takes place, quenching of excited singlet and triplet states generally takes place through energy-transfer mechanism or enhanced intersystem crossing process. That such quenching could also be electron spin selective was demonstrated in the EPR spectra of radicals showing nonBoltzmann spin distribution.1−3 From the early observation of such electron spin polarized EPR spectra, a host of studies on a variety of molecular systems and under different experimental conditions have been reported.4 From these studies, several general features of the dynamics of the quenching process of the generation of electron spin polarization (ESP) have emerged. The mechanism, known as the radical-triplet pair mechanism (RTPM), invokes an important feature of the pair, which is its overall spin state. Because the radical (R) has a doublet state and the excited molecule can be a singlet (S) or a triplet (T), the overall spin state of the pair can be a doublet or a quartet. The energy difference between them is the doublet− quartet exchange energy, J(r), whose magnitude depends exponentially on r, the distance between R and the excited molecule. Generation of ESP in the radical after quenching © 2015 American Chemical Society
involves mixing of states between the quartet (Q) and the doublet (D) states, caused primarily by the zero-field splitting of the triplet and, to a lesser extent, by the hyperfine interaction of the radical. The mixing of states is most favorable when J(r) is comparable to the Zeeman energy of the radical in the magnetic field of the EPR spectrometer. A direct consequence is that in solutions of higher viscosity ESP of a larger magnitude is expected because the radical-excited molecule pair spends more time in the crossing region than in solutions of lower viscosity. Experimental observations have been in accord with this. Most experimental studies of ESP, however, have been conducted on systems where the radical and the excited molecule are separate entities freely moving in a homogeneous solution. Random diffusive encounters between them lead to the formation of a radical-excited molecule pair and quenching and generation of ESP on the separated radical. The observed spin-polarized EPR spectra therefore reveal only an average effect of the otherwise distance-dependent interaction, J(r), which plays a crucial role in the whole process. Received: February 28, 2015 Revised: April 29, 2015 Published: May 22, 2015 5515
DOI: 10.1021/acs.jpca.5b01989 J. Phys. Chem. A 2015, 119, 5515−5523
Article
The Journal of Physical Chemistry A On photoexcitation, the molecule first goes to an excited singlet state and then to a triplet state. The free radical can quench both the states. Because of rather short lifetime of the singlet excited state of most organic molecules, however, quenching by the radical through a bimolecular quenching process is not very probable within its lifetime. Thus, one usually observes the quenching of only the relatively long-lived triplet states by the radical. As a result, in general, the characteristic EPR spectra of emissive phase with a negative spin polarization PT are seen. Quenching of singlet excited states by free radicals has been observed only in molecules such as coronene and pyrene,5,6 which have long singlet-state lifetimes. These systems give EPR spectra in an enhanced absorptive phase with a positive spin polarization PS. In either case, polarization is defined as P ≡ (nβ − nα)/(nβ + nα), where nα and nβ are the population of radicals with α and β spin states. Whether the singlet or the triplet state is quenched by the radical, the essential features of the ESP generation remain the same. Observation of very different magnitudes of PS and PT in the same system, in experiments conducted under identical conditions, thus becomes an important issue. Such observations have been rationalized either by invoking that the singlet and the triplet quenching take place at very different values of r and not in the neighborhood of the level crossing region6,7 or by postulating the presence of another electronic state, often called a charge-transfer state, facilitating the quenching.6 The distance dependence of J(r) is usually written in the following exponential manner J(r ) = J0 e−α(r − d)
sequently, these intermediates remained undetected for almost half a decade after the RTPM was proposed. The first successful recording of these intermediates was done by the Corvaja’s group 9 who achieved this by covalently linking a C60 chromophore to a pyrrolidene radical, thereby increasing the lifetime of the intermediate states. The time-resolved EPR (TREPR) spectra of this system gave two sets of EPR lines, one set corresponding to the ground doublet state D0 and the other to the quartet state Q1. The excited doublet state D1 was not observed because its lifetime was shorter than the instrument response time due to the allowed D1 → D0 transition. Shortly afterward, Yamauchi’s group10 detected both the D1 and Q1 states of a linked system of tetraphenylporphinato zinc and ppyridyl nitronyl nitroxide by recording the TREPR spectrum at 20 K. The low temperature helped to slow down the dynamics, facilitating the detection of the D1 state. The phase of the spin-polarized EPR spectra depends not only on the electronic state of the excited molecule but also on the sign of the exchange interaction J. Its sign decides the relative ordering of the D1 and Q1 states. As with a radical pair, J depends on the particular chromophore and radical involved and also on their mutual distance11 and relative orientation.12 Corvaja’s group12 reported that the phase of the TREPR spectra of the radical changes when the number of bonds between the two moieties increases from five to six. They explained this observation by assuming that the sign of J changes in the two cases and that the sign is determined by the relative orientations of the two moieties. Yamauchi’s group13 also demonstrated the phase of the TREPR spectra in the linked system of metallo-porphyrin and pyridyl nitronyl nitroxide radical (nit-py) to depend on the specific isomer of nit-py. They rationalized this by proposing that the sign of J depended on whether the nitroxide orbital interacted directly with prophyrin triplet (ortho and meta isomers of nit-py) or via the intermediate pyridine ring (para isomer of nit-py). The group of Wasielewski14,15 covalently attached t-butylphenyl niroxyl free radical to perylene-3,4:9,10-bis(dicarboximide) at specific distances and orientations. On photoexcitation, the observed ESP of the radical was found to depend on the molecular structure14 and rationalized by invoking that the electronic coupling depends on the through-bond coupling between the two moieties. They also proposed that the radical and the triplet experience weak exchange coupling.15 All these examples highlight the importance of studying the ESP of linked molecular systems of the radical and chromophore in a systematic manner, with special attention to their separation and orientation. In the work reported here, we have chosen naphthalene (Nap) and 2,2,6,6-tetramethylpiperidine N-oxyl free radical (TEMPO) pair and covalently linked them through a spacer group of variable length. The general symbol for these molecules is Nap-Spacer-TEMPO. Naphthalene was selected because its photochemistry and photophysics are well known. On photoexcitation, it forms a long-lived triplet state with a quantum efficiency of 0.75 to 0.80.16 Quenching of this state has been extensively studied by both optical and EPR techniques. We chose TEMPO because of its stability and the availability of its various substituted derivatives, which can be used as starting materials in the synthesis. We compared the ESP behavior of these linked molecules with the “unlinked system” consisting of naphthalene and TEMPO present as separate entities in the solution. In our preliminary investigations on the spin-polarized EPR spectra, we observed unusual shapes of the TREPR spectra of the linked
(1)
where J0 and α are parameters and d is the distance of closest approach of the radical and the excited molecule. Depending on the relative magnitudes of J0 and the Zeeman energy, ω0, systems have been classified under strong exchange (|J0| ≫ ω0) and weak exchange (|J0| ≪ ω0) regimes. The observed ESP has been rationalized by postulating that the quenching takes place at very different values of r in the two regimes.7 Theoretical work of Shushin8 shows P to be independent of J0 in the strong exchange regime but proportional to J0 in the weak exchange regime. The above discussion shows that studying the ESP on a specific radical-excited molecule pair, where the only variable is their mutual distance r, one should be able to gain important insight. Such studies cannot be carried out if the radical and the excited molecules are separate entities. By linking the two, however, the distance between them can be kept largely fixed. There are several reports on the ESP studies on such linked systems, demonstrating several finer details of the quenching process. As previously mentioned, one of the key postulates of the RTPM theory is the existence of doublet and quartet spin intermediates, which are formed during the collisional encounters of the excited molecule and the radical. These intermediates are difficult to characterize by optical techniques, as their electronic properties are very similar to those of separate chromophore and the radical, but because their spin states are different, they can be characterized by EPR technique. In the case of freely diffusing excited molecule and the radical, however, these intermediates are difficult to detect. This is because their lifetime in solution is determined by the lifetime of the collisional encounter, which is usually much shorter than the typical response time of an EPR spectrometer. Con5516
DOI: 10.1021/acs.jpca.5b01989 J. Phys. Chem. A 2015, 119, 5515−5523
Article
The Journal of Physical Chemistry A Chart 1. Structures and Symbolic Names of the Molecular Systems Studied in This Work
a
“Unlinked system” denotes an equimolar homogeneous mixture of naphthalene and TEMPO.
molecule.17 These shapes evolved with time, appearing as extra splitting of each of the nitrogen hyperfine lines. Careful analysis of the time evolution of the TREPR spectra revealed that they are the result of a detailed interplay of photophysical and magnetic resonance dynamics of the sequential quenching of the singlet and the triplet naphthalene moiety by TEMPO linked to it. Here we found the quenching efficiency to decrease with increasing the distance between the two moieties. This decrease correlated with the magnitude of the ESP generation.
other molecules resulted in either very poor yield (n = 5) or no reaction at all (n = 2, 3). Ground-state absorption, steady-state fluorescence, and phosphorescence measurements were done in a UV/vis spectrophotometer (Lambda 25, PerkinElmer) and JobinYvon spectrofluorimeter, respectively. Fluorescence lifetimes were recorded by TCSPC technique in a laboratory-built setup. No degassing was done for these measurements because no effect of oxygen was observed on the lifetimes of the linked molecules. All these experiments were done in solutions of spectroscopic grade n-hexane (Merck). Lifetimes and quenching constants of triplet states were measured in a nanosecond laser flash photolysis apparatus assembled in the laboratory. Triplet-state EPR measurements at low temperatures were done on a Bruker X-band spectrometer (EMX Micro). Other steady-state EPR experiments were done on a laboratory-built X-band spectrometer modified for steady-state experiments. TREPR experiments were done on the same spectrometer modified for time-resolved experiments. Electron spin−lattice relaxation times (T1) of TEMPO and the linked molecules were obtained by the pulse saturation recovery technique. The response time of the TREPR spectrometer was found to be ∼70 ns. All TREPR experiments were done at 5 mW of microwave power. This choice of the microwave power was based on the criteria of obtaining the most intense EPR spectra with a good S/N and, at the same time, avoid saturation and Torrey oscillations. An excimer laser (Coherent) operating at 248 nm was used as the excitation light for TREPR and nanosecond flash photolysis experiments. To avoid accumulation of contaminants due to photodegradation, all TREPR experiments were done in a flow setup. The solutions were degassed with nitrogen gas (IOLAR grade) for ∼30 min before carrying out TREPR measurements. Steady-state EPR spectra of the sample were recorded before and after the TREPR experiment to check for any photodecomposition. In general, about 5−10% decrease was observed for the linked molecules during the course of the experiments. The size of the laser output was 2 × 1 cm2, and the laser entry hole into the EPR cavity had a diameter of 6 mm. Thus, the entire energy of laser could not reach the sample flowing through a flat cell kept in the cavity. To measure the actual energy entering the cavity, we placed a small slit of the same dimensions as the laser entry hole in a collinear fashion in
■
EXPERIMENTAL SECTION The structures and symbolic names of different molecular systems studied in this work are given in Chart 1. For these linked molecules, we wanted to ensure that the electronic properties of naphthalene did not change on linking it to the free radical. Hence naphthalene was attached to TEMPO via a methylene group at its β position. All the linked molecules therefore had a “Nap-CH2-” moiety. Thus, by preventing electron delocation away from the naphthalene ring, the electronic properties of the naphthalene moiety of the linked molecules are expected to be very similar to those of an isolated naphthalene molecule. Williamson’s method of ether synthesis was used for synthesizing the linked molecules. Detailed procedures for synthesizing and characterizing them will be described elsewhere. The linked molecule Nap-CH2-O-TEMPO was synthesized in one step by the reaction of 2-bromomethyl naphthalene and 4-hydroxy-TEMPO. Linked molecules NapCH2-O-(CH2)n-O-TEMPO were synthesized in two steps. The first step involved reaction of 2-naphthalene methanol with the linker 1,n-dibromoalkane. In this reaction, we observed a unimolecular decomposition of the starting material 2naphthalene methanol under basic conditions in dimethylformamide solvent. It was necessary to circumvent this reaction so as to improve the yield of the required molecule. This was achieved by using the 1,n-dibromoalkane itself as the solvent. This method gave reasonable yields of the required molecule. The product of this first step was then reacted with 4-hydroxyTEMPO, which gave the required linked molecules. Using the above scheme, our objective was to synthesize linked molecules of the type Nap-CH2-O-(CH2)n-O-TEMPO with n varying from 2 through 6; however, we could synthesize only the molecules with n = 4 and 6. Attempts to synthesize the 5517
DOI: 10.1021/acs.jpca.5b01989 J. Phys. Chem. A 2015, 119, 5515−5523
Article
The Journal of Physical Chemistry A
Figure 1. TREPR of (A) “unlinked system” naphthalene and TEMPO, 3 mM each, (B) Nap-CH2-O-TEMPO, (C) Nap-CH2-O-(CH2)4-O-TEMPO, and (D) Nap-CH2-O-(CH2)6-O-TEMPO in degassed n-hexane (concentration 1.5 mM) at various times after a 248 nm laser pulse. Boxcar gate width: 0.1 μs, temperature:−45 °C, microwave power: 5 mW. The extra splittings are clearly visible in the “unlinked system” at 0.1 μs and in NapCH2-O-TEMPO at 0.2 μs. During some scans, the magnetic field was made to skip the regions where no EPR signal was expected, so as to reduce the scan time and exposure of laser light on the sample.
molecules were generally higher than when the two moieties are not linked, and among the linked molecules, magnitude of ESP decreased with increasing the distance between naphthalene and TEMPO moieties. Optical studies of fluorescence quenching also showed that the quenching decreased with increasing the distance. Because the solvent nhexane and the spacer groups of the linked molecules are made of linear hydrocarbon chains, the linked molecules are expected to be present in an extended conformation in the solution. Qualitative results from the spin-polarized EPR spectra and the fluorescence quenching efficiency are consistent with this. Observation of emissive spectra also indicated that J, the energy of separation between the doublet and quartet levels, is negative in all these systems. On lowering the temperature, the S/N ratio of all systems improved considerably. At about −45 °C, TREPR spectra with a moderate S/N could be recorded. Figure 1 shows such spectra at different times after the laser pulse. Molecules NapCH 2 -O-(CH 2 ) 4 -O-TEMPO and Nap-CH 2 -O-(CH 2 ) 6 -OTEMPO showed emissive signals from ∼0.1 μs onward. But Nap-CH2-O-TEMPO and the “unlinked system” showed absorptive signals at early times, which rapidly changed to emissive signals. Thus, the ESP mechanism changed from doublet precursor RTPM to quartet precursor RTPM. This demonstrates that singlet quenching precedes triplet quenching. When the length of the spacer group increased, the contribution of the absorptive signal due to doublet precursor RTPM decreased substantially, and the absorptive component could not be detected in the spectra shown here. Also, we could not detect any signal from doublet or quartet intermediates. Kawai and Obi5 reported TREPR spectra of the naphthaleneTEMPO “unlinked system” in 2-propanol that showed phase inversion from absorptive to emissive, but did not give any detailed analysis of the evolution of the line shape. The spectra
front of, and very close to, the cavity. A power meter was kept between the two to determine the amount of laser energy reaching the sample. Usually, energy of about 8−10 mJ per pulse of 248 nm was entering the cavity for TREPR experiments. The stability of the molecules was the main criterion for the choice of solvent. To check the stability of the linked molecules, we made equimolar solutions of them in acetonitrile, n-hexane, cyclohexane, and 2-propanol. The solutions were then irradiated with a 266 nm filtered output of a UV lamp in a continuous mode, and the fluorescence signal of naphthalene was monitored. The least amount of change in the fluorescence was observed in n-hexane and cyclohexane on prolonged exposure to the UV light, implying maximum stability of the linked molecule in these two solvents. We then chose n-hexane, as it has a much lower freezing point than cyclohexane, as all our experiments would be conducted at temperatures well below −40 °C.
■
RESULTS AND DISCUSSION Qualitative Observations and Inference from TREPR Spectra. Initial TREPR experiments of the unlinked and linked molecules in n-hexane solution at room temperature of 21 °C showed EPR spectra of poor S/N in all systems except NapCH2-O-TEMPO at 0.5 μs after the 248 nm laser excitation. The best S/N ratio of ∼8 was obtained with Nap-CH2-O-TEMPO. Emissive spectra were seen for all the linked molecules as well as for the “unlinked system” in solution. The emissive polarization of TEMPO is in line with the results of Kawai et al.5 on the naphthalene-TEMPO system in 2-propanol; however, in our case, the S/N for the “unlinked system” was very poor. We attributed this small signal to the low viscosity of n-hexane compared with 2-propanol. Qualitative conclusions from these spectra are that the ESP magnitudes of the linked 5518
DOI: 10.1021/acs.jpca.5b01989 J. Phys. Chem. A 2015, 119, 5515−5523
Article
The Journal of Physical Chemistry A of the “unlinked system” seen here are qualitatively similar to those of Kawai and Obi,5 but we noticed that at a very early time of 100 ns, all three hyperfine lines showed a shape resembling the letter M, as if they were split into two. Such a line shape has not been previously reported. With the passage of time, the line became purely emissive and then relaxed to thermally equilibrated signals. In the ESP of phthalocyaninatosilicon covalently linked to TEMPO, Ishii et al.18 interpreted the emissive spectra of the radical to arise from radical−quartet pair mechanism involving intermolecular interaction between the quartet excited state of one molecule and the doublet ground state of another. To investigate if the similar process could also operate in our linked systems, we recorded the time profiles of the EPR signals for all of these systems at various concentrations. Figure 2 shows the
interaction, making the hyperfine interaction negligible in comparison. Modeling the Dynamics of the Quenching Process. On the basis of the above qualitative observation, we now model the detailed scheme of the quenching process giving rise to the evolution of the spin system. We have treated the dynamics of the “unlinked system” and the linked molecule essentially on a similar footing, except that the intermolecular second-order quenching of excited naphthalene by TEMPO of the “unlinked system” was replaced by the respective intramolecular processes in the linked molecules (Scheme 1). The various processes with their corresponding rate constants are shown below. 1. S1 → S0
Radiative (fluorescence) decay,
2. S1 → T1
k fl Intersystem crossing of S1, kISC
3. S1 + R → T1 + RS* Quenching of S1 by radical via EISC, kqs 4. T1 + R → S0 + RT * Quenching of T1 by radical via EISC, kqt 5. T1 → S0
Intersystem crossing of T1, k T
S0, S1, T1, and R refer to the ground state, excited singlet and triplet states of naphthalene, and the doublet TEMPO radical with Boltzmann spin distribution, respectively. RS* (RT*) denotes spin-polarized radicals from doublet (quartet) precursor with absorptive (emissive) polarization. Decay of the S1 state through internal conversion according to the following pathways has been neglected. 6. S1 → S0 Internal conversion, kIC 7. S1 + R → S0 + R Enhanced internal conversion, kEIC
Figure 2. Time profiles of the TREPR signals due to electronic quenching of naphthalene by TEMPO in (A) “unlinked system” ([Nap] = 1.5 mM and [TEMPO] = 0.36 mM (red), 0.75 mM (blue), 1.5 mM (green)) and (B) Nap-CH2-O-TEMPO (conc. = 0.2 mM (blue), 0.5 mM (red), 1 mM (green)) in n-hexane at −45 °C. Microwave power: 5 mW. The downward sharp spike on the left was due to the jitter in the laser emission.
The fact that the internal conversion is negligible in naphthalene is based on the two key observations: (1) the sum of the fluorescence quantum yield (Process 1) and the intersystem crossing (Process 2) is unity19 and (2) there is no difference in the fluorescence quantum yield between naphthalene-h8 and naphthalene-d8,20 showing that the C−H vibration, the dominant factor contributing to the internal conversion process in rigid aromatic molecules, is ineffective in naphthalene. For the same reason, Process 7, being intermolecular, is also neglected. For the linked molecules, again for the same reason, we do not expect that replacing the hydrogen of the C−H group of naphthalene by the spacer moiety could significantly change vibrational activities to induce internal conversion. Under the pseudo-first-order condition satisfied by the concentrations of S1, T1, and R in our experiments, the effective rate constants of Processes 3 and 4 are given by kqs[R] and kqt[R], respectively. The spin−lattice relaxation process, kSLR, brings the polarized radicals, RS* and RT*, to Boltzmann distribution. The linked molecules are denoted by S−R or T−R, where S and T denote the singlet and triplet naphthalene, respectively. In particular, S1−R (T1−R) denotes the electronic state of the linked molecule, where the naphthalene moiety in its singlet (triplet) excited state is linked to the doublet radical, giving rise to the so-called “sing-doublet” (“trip-doublet”) state. Kinetic steps similar to those previously shown have been used
results. The time profile of the “unlinked system” depended strongly on the concentration of the radical. This is expected, because the intermolecular quenching in this system is bimolecular. In contrast, the EPR time profile of Nap-CH2O-TEMPO did not show any observable dependence on its concentration in the similar range. Similar behavior was observed for the other linked molecules Nap-CH2-O-(CH2)4O-TEMPO and Nap-CH2-O-(CH2)6-O-TEMPO also (data not shown). This proved that the quenching in linked molecules is intramolecular. The intensities of the three lines in the TREPR spectra of the “unlinked system”, displayed in Figure 1, showed considerable dependence on the spin of the nitrogen nucleus, giving an E + E/A pattern. This is again in accord with the observation of Kawai and Obi.5 In contrast, in all linked systems, such dependence was much weaker. The variation of intensities seen in Figure 1 for these linked molecules is no more than the general instability or reproducibility in different scans of the spectra. This indicated that in the linked molecules, ESP due to the zero-field splitting of triplet naphthalene was the dominant 5519
DOI: 10.1021/acs.jpca.5b01989 J. Phys. Chem. A 2015, 119, 5515−5523
Article
The Journal of Physical Chemistry A Scheme 1. Photophysical Pathways and their Rate Constants for the “Unlinked” (A) and the Linked (B) Systems
Table 1. Values of Various Kinetic Parameters Used in the Simulations of the Time-Dependent EPR Signal at −56 °Ca molecular system Nap+TEMPO (“unlinked system”) Nap-CH2-O-TEMPO Nap-CH2-O-(CH2)4-OTEMPO Nap-CH2-O-(CH2)6-OTEMPO
T1/μs, fittedb
T1/μs, measuredc
T2/μs
0.34
0.38
0.050
1.0 1.3
0.64 0.65
1.4
0.67
kqs/s−1
kqt/s−1
0.062 0.062
kqs[R] = (12.3 ± 0.2) × 106d (3.0 ± 0.9) × 109 (1.1 ± 1.0) × 109
kqt[R] = (1.8 ± 0.1) × 106d (30 ± 7) × 106 (12 ± 1) × 106
0.062
(0.35 ± 0.25) × 109
(15 ± 4) × 106
PS/Peq
PT/Peq
4.1 ± 0.2
−7.7 ± 0.3
21.0 ± 1.0 6.5 ± 1.5
−29.0 ± 1.0 −15.0 ± 2.0
4.5 ± 1.0
−9.0 ± 1.0
a Errors in T1: ± 0.02 μs, T2: ± 0.002 μs. bT1 obtained from simulation. cT1 measured by the pulse saturation recovery technique on the same solution that was used for TREPR experiments. dFor the “unlinked system”, the pseudo-first-order values of the rate constants are given. These values were calculated at [TEMPO] = 1.5 mM, using the bimolecular singlet quenching constant kqs = of 8.2 × 109 M−1 s−1 and triplet quenching constant kqt = 1.8 × 109 M−1 s−1. For linked molecules, kqs was taken to be the inverse of their fluorescence lifetimes, and kqt values were obtained by treating them as an adjustable parameter while fitting the simulations to the observed EPR signal.
Information for details. Table 1 shows the values of various kinetic parameters used in these calculations. The time evolution of the EPR line shape of Nap-CH2-OTEMPO and its curious appearance resembling the letter W has been already described.17 A consequence of the timedependent Bloch equations is that the width of an EPR line arising from a magnetization evolving with time is also time dependent. During the quenching of an excited state, the spins of the radical suddenly undergo a change of magnetization. As the quenching continues with time, the radicals that quench the excited state at early times acquire narrower EPR line widths, and those that quench the excited state at later times acquire broader line widths. The superposition of these shapes gives the overall appearance of the EPR spectra at a given time. This way the complete evolution of the EPR lines of the Nap-CH2-OTEMPO could be well reproduced.17 To understand the origin of the unusual EPR spectra for the unlinked system, we
in describing the evolution of line shapes of Nap-CH2-OTEMPO.17 The intrinsic triplet decay of naphthalene (Process 5), associated with the rate constant kT, has been omitted for the linked systems. This was done because of it being too slow to be of any consequence, in comparison with the intramolecular quenching process given by kqt. In Scheme 1, no detailed dynamics of RTPM has been considered, for example, the time evolution of state mixing. Thus, the scheme should be considered as phenomenological. The rate equations for the various states after the laser pulse were set up and solved using Laplace transform technique to obtain their concentrations as a function of time. Then, the time-dependent Bloch equations were solved to obtain the magnetization of the nitroxyl radical as a function of time and magnetic field. The calculated timedependent EPR signal intensity was convolved with the instrument response function of the EPR spectrometer for comparing with the observed intensity. See the Supporting 5520
DOI: 10.1021/acs.jpca.5b01989 J. Phys. Chem. A 2015, 119, 5515−5523
Article
The Journal of Physical Chemistry A
Figure 3. (A) Experimental TREPR spectra, corresponding to the low-field hyperfine line of the “unlinked system” (naphthalene and TEMPO, 3 mM each) in n-hexane at −45 °C, at various delay times after the 248 nm laser pulse. Boxcar gate width: 50 ns, microwave power: 5 mW. (B) Calculated spectra at the denoted times with 50 ns integration time.
linked to the radicals, some molecules of naphthalene and the radical can come very close, much closer than the linked molecules studied here. TEMPO is known to quench excited states by electron transfer, too.21 Thus, one could envisage the following scheme to be another channel for the photochemical quenching of naphthalene, when it is very close to the radical.
recorded its TREPR spectra with improved S/N and at small increments of the delay time after the laser pulse. Simulations were carried out using the relevant values of various kinetic processes, and the results are shown in Figure 3. As previously stated, at the shortest possible time this system gave a line shape similar to the letter M. Another kind of shape was seen around the transition time from the absorptive to the emissive phase at ∼250 ns. Its shape was like the letter W and very similar to what has been seen in Nap-CH2-O-TEMPO.17 The fact that the two kinds of “splitting” were not same was evident from their temporal behavior. The M-shaped splitting was present right from the early time and then disappeared at ∼150 ns. The W-shaped splitting appeared late, at ∼250 ns, when the absorptive RTPM changed to emissive RTPM. The quenching here being bimolecular, the time of appearance of these features depended on the concentration of the radical. It is apparent from the simulated spectra that the M-shaped feature at early time was not reproduced in our calculations. What was needed to simulate the M shape was a small component of a narrow emissive signal riding on the broad absorptive signal. This emissive signal should be present at a time earlier than the absorptive signal, so that its narrow width could be justified. Because the absorptive signal is due to the quenching of the S1 state of naphthalene, the narrow emissive signal requires quenching, possibly of a triplet state earlier than this S1 state. It is not possible to imagine a photophysical process of this type. A possible cause of the M-shaped lines could be depletion of the radicals due to the laser pulse. Our TREPR spectra were recorded in a dual boxcar setup. Here the gate of one boxcar (gate B) opened ∼0.3 μs before the laser pulse, and that of the other boxcar (gate A) opened at the desired time after the laser pulse. The boxcar operated in the “A−B” mode, giving the difference in the microwave signals at these two gate times. Thus, depletion of radical by the laser pulse will appear as a negative signal. We propose that that was the origin behind the appearance of the M-shaped spectra at early times in the “unlinked system”. Because naphthalene molecules are not
Nap*(S1) + R· → Nap− + R+ → 1or3 Nap(S0) + R·
1
If a small fraction undergoes such a fast electron-transfer reaction, the EPR signal should show depletion, giving an impression of emissive signal. Subsequent back electron transfer would show the disappearance of the emissive-looking signal. The same calculations that reproduced the EPR line shapes at different times should also be able to describe the time evolution of the EPR signal at a fixed magnetic field. In this, the range of possible values of various kinetic parameters can be tightly controlled. For that purpose, we recorded the EPR signal at the peak of the low-field hyperfine line. By improving the cooling system, we could lower the temperature of the samples in the EPR cavity to about −56 °C. This improved the S/N further, and after signal averaging, high-quality EPR signals were obtained. These are shown in Figure 4 along with the calculated signals. The calculated signals reproduced the observed ones satisfactorily well for all systems. Magnitudes of Singlet and Triplet Polarization, PS and PT. The values of PS and PT determined from the simulations show that the magnitudes of the polarization of all linked molecules are higher than those of the “unlinked system”. This shows that holding the chromophore and radical at a given distance helps increase the probability of mixing of D and Q states, which is akin to increasing the viscosity of the solution for the “unlinked system”. Because the distance remained fixed for the linked molecules, the rapid decrease in the magnitude of polarization with increasing the length of the spacer showed that all of them were in the “weak exchange” regime, proposed by Shushin.8 We estimated the approximate distance between the center of the naphthalene moiety and the nitrogen atom of 5521
DOI: 10.1021/acs.jpca.5b01989 J. Phys. Chem. A 2015, 119, 5515−5523
Article
The Journal of Physical Chemistry A
magnitudes of PT and PS to be the same, or the ratio PT/PS should be very nearly equal to 1. The reason for PS being smaller than PT may lie in our assuming that the triplet naphthalene is generated, after laser excitation, through Processes 2 and 3 only, producing RS*. It is likely that a fraction of the molecules populate the trip-doublet states without the radicals being spin-polarized. In other words, the T1−R states can be populated by the EISC route causing PS polarization and also by a parallel route without leading to absorptive spin polarization. This parallel path could be through the charge-transfer interaction.11,22 Quenching from T1−R brings the entire population to S0−R, giving a larger value of PT polarization. Recovery of Spin-Polarized Signal and Spin−Lattice Relaxation. Finally, we noted that the spin−lattice relaxation times that we measured (T1,measured) by the pulse saturation recovery technique did not fit well the time dependence of the growth of the emissive EPR signal to the thermally equilibrated signal. A small, but systematic, discrepancy between this relaxation time and the T1 needed to fit (T1,fitted) the observed signals was seen; T1,fitted was a little longer than T1,measured in all systems. This curious trend needs to be investigated in detail. We tentatively believe that Heisenberg spin exchange could be playing a role here. In pulse saturation recovery experiments, the hyperfine lines other than the one whose spin−lattice relaxation is being measured can provide alternative paths for relaxation through Heisenberg spin exchange, but in the ESP through the photophysical quenching, all hyperfine lines are almost equally populated, as the contribution of the hyperfine interaction to ESP is small. Thus, Heisenberg spin exchange, even if it takes place, cannot offer extra paths for spins to relax. Thus, the time dependence of the emissive signal to return to thermal equilibrium will be governed by the true, intrinsic T1 of the radical.
Figure 4. Time evolution of the observed EPR signals (noisy blue curves) and along with their simulation (smooth red curves). (A) “Unlinked system” (naphthalene and TEMPO, 3 mM each), (B) NapCH2-O-TEMPO, (C) Nap-CH2-O-(CH2)4-O-TEMPO, and (D) NapCH2-O-(CH2)6-O-TEMPO. Temperature: −56 °C, microwave power: 5 mW. See the Supporting Information for the steps involved in processing the observed signals before they were compared with the simulated signals.
TEMPO moiety, assuming that they are in an extended conformation. As J changes exponentially with distance (eq 1), the logarithm of P should decrease linearly with the distance. With only the three linked molecules studied here, we could see an approximate linearity of this kind. It is important to study more linked molecules with shorter spacer groups, so that distance of separation lies in-between that of Nap-CH2-OTEMPO and Nap-CH 2 -O-(CH 2 ) 4 -O-TEMPO, before a definitive conclusion can be drawn on the applicability of Shushin’s equation8 and also on the form of the distance dependence of J (eq 1) that should apply to the RTPM process. As mentioned in the Introduction, whenever both PS and PT have been reported, their values are usually very different. Such different values have been proposed to arise from different interaction regions for the two processes. Our values of PS and PT are very similar, and the ratio PT/PS was about 2 for all the linked systems. This showed that both singlet and triplet quenching processes are almost equally effective in generating ESP, and they take place extremely rapidly in quick succession. This is also consistent with the time evolution of the signal intensity. The absorptive component was seen to be much smaller than the emissive component, even though the magnitudes of PT and PS were not very different. The absorptively polarized radicals are generated by converting the singlet excited naphthalene moiety to triplet. These naphthalene moieties are also attached to the radical; they are quenched almost immediately, leaving their radical partner emissively polarized. Hence the absorptive EPR signal does not last long, and its EPR signal is unable to reach its full intensity. On the basis of these results, we wish to propose a general rule: Observation of small absorptive TREPR signals does not necessarily imply low polarization. Because the mechanism of ESP from a doublet precursor is very similar to that from a quartet precursor, we believe the
■
SUMMARY AND CONCLUSIONS We attempted to investigate how the ESP would be affected when the quencher radical and the excited molecule are not allowed to move freely in a solution but are held together. We linked a naphthalene moiety to a TEMPO free radical, separated by a spacer group of three different lengths. Sequential quenching of the singlet state and the triplet state of naphthalene was seen in all the systems and revealed through the spin-polarized TREPR spectra of opposite phase. The polarization magnitudes in the linked molecules were higher than that of the “unlinked system”, showing that when the two moieties are held together, the mixing probability of D and Q states is greatly enhanced. The magnitudes of the ESP steadily decreased with increasing the length of the spacer group. The polarization magnitudes due to triplet quenching and singlet quenching are very similar, differing by only a factor of about 2. These characteristics show that for all of the linked molecules, both of the quenching processes take place in the “weak exchange” regime and at almost the same distance between the two moieties. Our results also showed that observation of small absorptive TREPR signals does not necessarily imply that its magnitude of polarization is small. Finally, our measurements showed that the time constant for recovery of the spinpolarized EPR signal to thermal equilibrium is slower than the electron spin−lattice relaxation times measured by the pulse saturation recovery technique. We proposed this discrepancy to arise from the contribution of Heisenberg spin exchange to the electron spin relaxation process. 5522
DOI: 10.1021/acs.jpca.5b01989 J. Phys. Chem. A 2015, 119, 5515−5523
Article
The Journal of Physical Chemistry A
■
(14) Giacobbe, E. M.; Mi, Q.; Colvin, M. T.; Cohen, B.; Ramanan, C.; Scott, A. M.; Yeganeh, S.; Marks, T. J.; Ratner, M. A.; Wasielewski, M. R. Ultrafast Intersystem Crossing and Spin Dynamics of Photoexcited Perylene-3,4:9,10-bis(dicarboximide) Covalently Linked to a Nitroxide Radical at Fixed Distances. J. Am. Chem. Soc. 2009, 131, 3700−3712. (15) Colvin, M. T.; Giacobbe, E. M.; Cohen, B.; Miura, T.; Scott, A. M.; Wasielewski, M. R. Competitive Electron Transfer and Enhanced Intersystem Crossing in Photoexcited Covalent TEMPO-Perylene3,4:9,10-bis(dicarboximide) Dyads: Unusual Spin Polarization Resulting from the Radical-Triplet Interaction. J. Phys. Chem. A 2010, 114, 1741−1748. (16) Murov, S. L.; Carmichael, I.; Hug, G. L. Handbook of Photochemistry, 2nd ed.; Marcel Dekker, Inc.: New York, 1993; pp 30−31. (17) Rane, V.; Das, R. Observation of Splitting of EPR Spectral Lines without Any Concomitant Splitting in Energy Levels. J. Phys. Chem. A 2014, 118, 8689−8694. (18) Ishii, K.; Hirose, Y.; Kobayashi, N. Electron Spin Polarizations of Phthalocyaninatosilicon Covalently Linked to One TEMPO Radical in the Excited Quartet and Doublet Ground States. J. Phys. Chem. A 1999, 103, 1986−1990. (19) Kellogg, R. E.; Bennett, R. G. Radiationless Intermolecular Energy Transfer. III. Determination of Phosphorescence Efficiencies. J. Chem. Phys. 1964, 41, 3042−3045. (20) Lim, E. C.; Laposa, J. D. Radiationless Transitions and Deuterium Effect on Luminescence of Some Aromatics. J. Chem. Phys. 1964, 41, 3257−3259. (21) Green, S.; Fox, M. A. Intramolecular Photoinduced Electron Transfer from Nitroxyl Radicals. J. Phys. Chem. 1995, 99, 14752− 14757. (22) Kawai, A. Dynamic Electron Polarization Created by the Radical-Triplet Pair Mechanism: Application to the Studies on Excited State Deactivation Processes by Free Radicals. Appl. Magn. Reson. 2004, 23, 349−367.
ASSOCIATED CONTENT
S Supporting Information *
Detailed descriptions of the simulation of TREPR spectra, processing of observed EPR signals for determination of the absolute magnitudes of the polarization, and determination of the response time of the EPR spectrometer. The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpca.5b01989.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail: [email protected]. Notes
The authors declare no competing financial interest.
■ ■
ACKNOWLEDGMENTS We sincerely thank Dr. Prasanna S. Ghalsasi of M. S. University for his help and advice in the synthesis of molecules. REFERENCES
(1) Blättler, C.; Jent, F.; Paul, H. A Novel Radical-Triplet Pair Mechanism for Chemically Induced Electron Polarization (CIDEP) of Free Radicals in Solution. Chem. Phys. Lett. 1990, 166, 375−380. (2) Blättler, C.; Paul, H. Cidep after Laser Flash Irradiation of Benzil in 2-Propanol. Electron Spin Polarization by the Radical-Triplet Pair Mechanism. Res. Chem. Intermed. 1991, 16, 201−211. (3) Kawai, A.; Okutsu, T.; Obi, K. Spin Polarization Generated in the Triplet-Doublet Interaction: Hyperfine-Dependent Chemically Induced Dynamic Electron Polarization. J. Phys. Chem. 1991, 95, 9130− 9134. (4) Kawai, A.; Shibuya, K. Electron Spin Dynamics in a Pair Interaction Between Radical and Electronically-Exicted molecule as Studied by a Time-Resolved ESR Method. J. Photochem. Photobiol. C 2006, 7, 89−103. (5) Kawai, A.; Obi, K. First Observation of a Radical-Triplet Pair Mechanism (RTPM) with Doublet Precursor. J. Phys. Chem. 1992, 96, 52−56. (6) Mitsui, M.; Kobori, Y.; Kawai, A.; Obi, K. Quenching Mechanism of Excited Coronene by a Nitroxide Radical Studied by Probing Dynamic Electron Polarization. J. Phys. Chem. A 2004, 108, 524−531. (7) Terazono, H.; Kawai, A.; Tsuji, K.; Shibuya, K. Enhanced Intersystem Crossings of S1-T1 and T1-S0 in Coronene- and Pyrenegalvinoxyl Systems as Studied by a Pulsed ESR Method. J. Photochem. Photobiol., A 2006, 183, 22−30. (8) Shushin, A. I. The Relaxation Mechanism of Net CIDEP generation in Triplet-Radical Quenching. Chem. Phys. Lett. 1993, 208, 173−178. (9) Corvaja, C.; Maggini, M.; Prato, M.; Scorrano, G.; Venzin, M. C60 Derivative Covalently Linked to a Nitroxide Radical: Time-Resolved EPR Evidence of Electron Spin Polarization by Intramolecular RadicalTriplet Pair Interaction. J. Am. Chem. Soc. 1995, 117, 8857−8858. (10) Ishii, K.; Fujisawa, J.; Ohba, Y.; Yamauchi, S. A Time-Resolved Electron Paramagnetic Resonance Study on the Excited States of Tetraphenylporphinatozinc(II) Coordinated by p-Pyridyl Nitronyl Nitroxide. J. Am. Chem. Soc. 1996, 118, 13079−13080. (11) Kawai, A. FT-EPR Study on the Sign of Exchange Interaction in Triplet Naphthalene-Galvinoxyl Encounter Pair. Appl. Magn. Reson. 2004, 26, 213−221. (12) Mazzoni, M.; Conti, F.; Corvaja, C. The Sign of the Exchange Interaction between Triplet Excited Fullerene and Nitroxide Free Radicals. Appl. Magn. Reson. 2000, 18, 351−361. (13) Fujisawa, J.; Ishii, K.; Ohba, Y.; Yamauchi, S.; Fuhs, M.; Möbius, K. X- and W-Band Time-Resolved Electron Paramagnetic Resonance Studies on Radical-Excited Triplet Pairs between Metalloporphyrins and Axial-Ligating Nitroxide Radicals. J. Phys. Chem. A 1997, 101, 5869−5876. 5523
DOI: 10.1021/acs.jpca.5b01989 J. Phys. Chem. A 2015, 119, 5515−5523
