Excited-State Dynamics in Nitro-Naphthalene Derivatives: Intersystem

Article pubs.acs.org/JPCA

Excited-State Dynamics in Nitro-Naphthalene Derivatives: Intersystem Crossing to the Triplet Manifold in Hundreds of Femtoseconds R. Aaron Vogt, Christian Reichardt, and Carlos E. Crespo-Hernández* Department of Chemistry, Center for Chemical Dynamics, Case Western Reserve University, 10900 Euclid Avenue, Cleveland, Ohio 44106, United States S Supporting Information *

ABSTRACT: Femtosecond transient absorption experiments and density functional calculations are presented for 2-methyl-1-nitronaphthalene, 2-nitronaphthalene, and 1nitronaphthalene in cyclohexane and acetonitrile solutions. Excitation of 2-methyl-1nitronaphthalene at 340 nm populates the Franck−Condon singlet state, which bifurcates into two barrierless decay channels with sub-200-fs lifetimes. The primary decay channel connects the Franck−Condon singlet excited state with a receiver triplet state, whereas the second, minor channel involves conformational relaxation to populate an intramolecular charge-transfer state, as previously reported for 1-nitronaphthalene (J. Chem. Phys. 2009, 113, 224518). Conversely, the experimental and computational data for 2-nitronaphthalene shows that almost the entire Franck−Condon singlet excited-state population intersystem crosses to the triplet state in less than 200 fs due to a sizable energy barrier of ca. 5 kcal/mol that must be surmounted to access the intramolecular charge-transfer state. Our results lend support to the idea that the probability of population transfer to the triplet manifold in these nitronaphthalene derivatives is controlled not only by the small energy gap between the Franck−Condon singlet excited state and the receiver triplet state but also by the region of configuration space sampled in the singlet excited-state potential energy surface at the time of excitation. It is proposed that the ultrafast intersystem crossing dynamics in these nitronaphthalene molecules most likely occurs between nonequilibrated excited states in the strongly nonadiabatic regime. including solvent effects,6,10 and also by minimum energy path calculations in the gas phase at the CASPT2//CASSCF level of theory.22 A question arises as to whether subpicosecond ISC dynamics is a general phenomenon exhibited by other nitronaphthalene derivatives or just an isolated photochemical event in 1NN. In addition, it is unknown whether the proximity of a Tn(nπ*) state to the Franck−Condon region of the S1-state potential energy surface is the only factor that controls the efficiency of population transfer to the triplet manifold in 1NN. Recent investigations provide evidence that a distribution of nitro-aromatic torsion angles is available in the ground state at room temperature,10,16 which may be expected to play a role in the ultrafast population of the triplet state in 1NN and in other nitronaphthalene derivatives (NNDs). In this contribution, femtosecond broad-band transient absorption experiments and quantum-chemical calculations for 2-nitronaphthalene (2NN) and 2-methyl-1-nitronaphthalene (2M1NN) are reported. In addition, we compare the transient absorption and computational results gathered in these two NNDs with new results obtained in this work for 1NN, as well as with those previously presented.10 Taken together, the experimental and computational results provide further evidence that ultrafast intersystem crossing to the triplet manifold in these NNDs is controlled not only by the small

1. INTRODUCTION The time-resolved photochemistry of nitro aromatic compounds1−3 and of nitro-polycyclic aromatic hydrocarbons (NPAHs)4−16 has received significant attention recently primarily because of their distinctive ultrafast nonadiabatic dynamics. Additionally, NPAHs have been identified as persistent organic pollutants in the environment17 and can absorb sunlight with the potential of increasing the formation of products showing equal or higher toxicity than the parent compounds.18,19 Early work has provided evidence that NPAHs containing two to four aromatic rings have vanishingly small fluorescence quantum yields in solution at room temperature.20,21 More recently, femtosecond fluorescence upconversion and transient absorption experiments in combination with quantum-chemical calculations have shown that the small fluorescence yield in these compounds is a consequence of ultrafast population of one or more nonradiative decay pathways in the excited states.4−6,9,10,14,22 Of particular relevance to this work, recent studies on 1nitronaphthalene (1NN) have provided compelling evidence that the proximity of the bright, Franck−Condon excited singlet state (SFC 1 ) to a receiver, high-energy triplet state with nπ* character, Tn(nπ*), controls the ultrafast population of the triplet state and, hence, the fluorescence yield.6,9,10,16,22 Ultrafast intersystem crossing (ISC) dynamics in 1NN is supported by the calculation of singlet and triplet excitation energies performed at the density functional level of theory, © 2013 American Chemical Society

Received: June 7, 2013 Published: June 26, 2013 6580

dx.doi.org/10.1021/jp405656n | J. Phys. Chem. A 2013, 117, 6580−6588

The Journal of Physical Chemistry A

Article

Table 1. Lifetimes Obtained from a Target and Global Fit Analysis of the Transient Absorption Data Using a Sequential Kinetic Modela cyclohexane

a

acetonitrile

lifetime/ps

1NN10

2M1NN

2NN

1NN10

2M1NN

2NN

τ1 τ2 τ3

0.11 ± 0.05 2.3 ± 0.2 10.3 ± 0.3

0.37 ± 0.07 1.4 ± 0.3 7.1 ± 0.9

0.11 ± 0.05 2.1 ± 0.1 10 ± 1

0.14 ± 0.05 2.8 ± 0.2 11.2 ± 0.4

0.21 ± 0.05 0.6 ± 0.1 5.9 ± 0.3

0.17 ± 0.05 1.9 ± 0.1 9.1 ± 0.1

Errors are reported as twice the standard deviation of three or more individual measurements.

(S1−Tn) energy gap but also by a distribution of the nitroaromatic torsion angles that are thermodynamically accessible in the ground state at the time of excitation and by the topology of the S1 state in the Franck−Condon region potential energy surface of each molecule. Furthermore, the results presented in this work lend strong support to the idea that subpicosecond intersystem crossing dynamics, most likely occurring in a strongly nonadiabatic regime, is a general phenomenon of these NNDs.

recrystallization, the weak emission was reduced to nearly undetectable levels, showing that it was in fact due to an impurity. Importantly, after careful analysis of the data we find no indication that the transient absorption experiments reported below for 2NN may be compromised by the presence of this minor impurity. 2.4. Femtosecond Transient Absorption Measurements. The experimental setup used for this research has been described in detail elsewhere.10 Briefly, the output of a Quantronix Integra-i/e 3.5 Laser (100 fs fwhm centered at 800 nm) was fed into an optical parametric amplifier (TOPAS, Quantronix/Light Conversion) that generates the femtosecond pulses used for excitation of the samples. The excitation wavelength used was 340 nm. Contributions from other wavelengths to the excitation pulses were removed by a reflective wavelength filter and a Glan-Taylor prism. Data were acquired using a broad-band transient absorption spectrometer (Helios, Ultrafast Systems, LLC) with a homemade data acquisition routine developed in the LabView visual programming environment. The white-light probe pulses were corrected for group velocity dispersion,28 as described elsewhere.10 The absorbance for all solutions used in the transient absorption experiments was 1.0 ± 0.2 at 340 nm in a 2 mm optical path length cell, which corresponds to a concentration of ca. 10−3 M for all three compounds. The probed volume of the samples was continuously renewed using a Teflon-coated stir bar and a magnetic stirrer. The samples were replaced with fresh solutions if a decrease of 5% in the steady-state absorbance was observed. Importantly, no changes in the transient absorption spectra or decay signals were evident in samples that showed 5% or less decrease in ground-state absorption during the time-resolved experiments. The excitation intensity at the sample was 0.5 μJ in all cases, and the ratio of the pump to probe beam size was 3:1. As shown in Figure S1, Supporting Information, transient signals for 2M1NN deviate from linearity at pump intensities greater than 0.6 μJ under the experimental conditions used in this work. All transient absorption measurements were performed under air-saturated conditions. Data analysis was performed using Igor Pro 6.12A software (Wavemetrics, Inc.). From each data set, nineteen evenly spaced traces were selected between 350 and 650 nm. The global and target analysis method29,30 based on a sequential kinetic model was used to obtain the excited-state lifetimes and decay associated spectra (DAS) of the three NNDs studied in this work. The sequential model rate law was composed of three exponential components plus a time-independent constant.31,32 This constant offset was needed because the long-lived transient absorption signal in the three NNDs does not decay within the time window of ∼3 ns of our setup. The fitting function was convoluted with a Gaussian-shaped instrument response function. The instrument response function was estimated to be 200 ± 50 fs from the multiphoton

2. EXPERIMENTAL AND COMPUTATIONAL METHODS 2.1. Chemicals. Cyclohexane (99.9%) was obtained from Fisher Scientific. Acetonitrile (99.6%) was obtained from Acros. Both solvents were used as received. 2NN, 1NN, and 2M1NN were obtained from Sigma-Aldrich (99.7%, 99%, and 99%, respectively). 2NN, 1NN, and 2M1NN are moderately toxic compounds. Proper safety precautions must be taken at all times to limit health risks. 2.2. Density Functional Calculations. All quantum chemical calculations were performed using the Gaussian 09 suite of programs.23 Bulk solvent effects were modeled by using self-consistent reaction field calculations (SCRF) with the polarizable continuum model (PCM)24 within the integral equation formalism (IEFPCM).25 Ground-state optimizations and vertical excitation energy calculations were performed earlier using density functional theory (DFT) at the B3LYP/IEFPCM/6-311++G(d,p) and the TD-PBE0/IEFPCM/6-311++G(d,p) levels of theory,26,27 respectively.10,16 In this work, the S1 states of 2NN, 1NN, and 2M1NN were optimized without any geometry restrictions at the TD-PBE0/IEFPCM/6-311++G(d,p) level of theory. Vibrational frequencies were calculated at the TD-PBE0/ IEFPCM/6-31G level of theory to confirm that the optimized geometries were local minima on the S1-state potential energy surface. Adiabatic potential energy curves (PECs) were calculated using the nitro-aromatic torsion angle as the variable coordinate by fixing this angle at selected values and allowing all other geometrical parameters to optimize in the S1 state. The energy values for these PECs are reported relative to the energy of the S0-state global minimum at the same level of theory. 2.3. Steady-State Measurements. Steady-state absorbance measurements were performed using a Cary 100 UV/vis spectrometer (Varian, Inc.).16 The purity of the nitronaphthalene compounds was verified by high-performance liquid chromatography16 and fluorescence spectroscopy. None of the NNDs showed intrinsic fluorescence emission detectable within the sensitivity of the instrument used (Cary Eclipse, Varian, Inc.). A weak fluorescence emission band was observed for 2NN in acetonitrile after excitation at 340 nm. However, this emission band was assigned to an impurity, as its excitation spectrum did not match the absorption spectrum of 2NN. We proceeded to recrystallized 2NN using methanol. After 6581



Article

CIS and CASPT2//CASSCF calculations. Similar conclusions have been reached by Quenneville and co-workers recently,2 in the quantum-chemical investigations of the excited-state potential energy surfaces in nitrobenzene, 2,4,6-trinitroaniline, and 2,4,6-trinitrotoluene. As shown in Figure 2, adiabatic PECs were also generated by optimizing all the nuclear coordinates in the S1 state except for

absorption signal of the solvent in the probe spectral region used. If needed, a delta function was included to model the solvent coherent or Raman signals at some of the UV probe wavelengths. The reported uncertainties for the lifetimes shown in Table 1 are twice the standard deviation (2σ) obtained from the global analysis of three or more independent set of experiments for each compound.

3. RESULTS 3.1. Quantum Chemical Calculations. In our previous works,10,16 the ground-state nitro-aromatic torsion angles, as defined in Figure 1a, were calculated to be 0.1° for 2NN, 33.1°

Figure 2. Potential energy scans of the optimized S1 state as a function of the nitro-aromatic torsion angle for 2NN, 1NN, and 2M1NN in acetonitrile at the TD-PBE0/IEFPCM/6-311++G(d,p) level of theory. Solid symbols represent the torsion angles distribution accessible in the ground state at room temperature for each molecule.

the nitro-aromatic torsion angle, which was held fixed and used as the reaction coordinate. In the case of 2NN, the S1−PEC has a minimum energy at 0°, whereas the S1−PECs of 1NN and 2M1NN have a minimum energy at ∼90°. These results are in excellent agreement with the fully optimized S1-state geometries presented in Figure 1b at the same level of theory. 3.2. Femtosecond Transient Absorption. Transient absorption experiments were performed to measure the evolution of the excited states of 2NN and 2M1NN in cyclohexane and acetonitrile solutions. The transient spectra of 2NN in cyclohexane are shown in Figure 3 (upper left panel). The corresponding DAS are shown in Figure 4. An instrumentresponse limited rise appears with peak amplitude at ∼390 nm after excitation at 340 nm. Following this fast rise is a broad increase of transient absorption centered at ∼450 nm. As this band appears, it becomes more structured, showing vibrational peaks at 392, 432, and 458 nm, as shown in Figure 5 (upper panel). Decay traces and best global fit curves for representative decay signals are presented in Figure 6A,B. The excited-state dynamics of 2NN in acetonitrile are similar to those in cyclohexane, with a few exceptions (Figure 3 upper right panel). At early times, two bands with absorption maxima around 390 and 630 nm rise simultaneously. After a few hundred femtoseconds, the increase of these absorption bands seems to level off while a wide absorption band appears with a maximum around 450 nm. At longer delay times, the band at 450 nm becomes more structured, showing vibronic bands at ∼400 and 465 nm, and a shoulder at ∼500 nm. Figure 3 also shows transient absorption spectra of 2M1NN in cyclohexane (lower left panel) and acetonitrile (lower right panel) solutions. The corresponding DAS are shown in Figure 4. The excited-state dynamics of 2M1NN begin with an instrument-response limited rise with peak absorbance at ∼375 and ∼385 nm in cyclohexane and acetonitrile, respectively, as well as low-amplitude absorption above 500 nm. The UV band decays as a visible band rises in with maxima at 413 and 568 nm in cyclohexane and 420 and 590 nm in acetonitrile. At longer delay times, the visible bands become more structured and undergo blue shifts, as shown in Figure 5 (lower panel). Decay

Figure 1. (a) Optimized S0-state geometries for 2NN (left), 1NN (center), and 2M1NN (right) at the B3LYP/6-311++G(d,p) level of theory (taken from ref 16). Nitro-aromatic torsion angles are defined with asterisks. (b) Optimized S1-state geometries for 2NN (left), 1NN (center), and 2M1NN (right) in acetonitrile at the TD-PBE0/ IEFPCM/6-311++G(d,p) level of theory.

for 1NN, and 56.8° for 2M1NN in cyclohexane. Changing the solvent to acetonitrile resulted in torsion angles of 0.3° for 2NN, 33.9° for 1NN, and 55.4° 2M1NN. Herein, we have further optimized the S1 state of these molecules in cyclohexane and acetonitrile (Figure 1b). The predicted excitation energies and oscillator strengths (in parentheses) of the fully optimized S1 state for 1NN, 2M1NN, and 2NN in acetonitrile are 1.63 (0.0001), 1.60 (0.0000), and 2.67 (0.1241) eV, respectively, at the PBE0/IEFPCM/6-311++G(d,p) level of theory. The corresponding values in cyclohexane are 1.73 (0.0001), 1.67 (0.0000), and 2.93 (0.0724) eV for 1NN, 2M1NN, and 2NN, respectively. The magnitude of the energy values may be underestimated due to the significant charge-transfer character of the S1 state. This is perhaps more likely for the adiabatic S1 energies of 1NN and 2M1NN, for which the frontier Kohn− Sham orbitals show significant charge-transfer character (see below). It is known that TD-DFT methods with low nonlocal exchange functional and without asymptotic corrections can underestimate the excitation energies of charge-transfer states.33−35 Thus, the adiabatic S1 energies of 1NN and 2M1NN are not necessarily accurate. However, we note that optimization of the S1-state minimum of 1NN in acetonitrile using configuration interaction singles (CIS) results in a geometry similar to that predicted by the TD-DFT calculations (Figure 2S, Supporting Information), where the nitro group is out-of-plane and has a pyramidal conformation. The CIS result is also in good agreement with results from CASPT2// CASSCF calculations done for the S1-optimized geometry of 1NN in the gas phase.22 Hence, we argue that the TD-DFT calculations are in qualitative agreement with the results from 6582



Article

Figure 3. Transient absorption spectra of 2NN and 2M1NN in cyclohexane (left) and acetonitrile (right). Solvent stimulated Raman emission bands are observed at short time delays. Time delays in the legends are reported in picoseconds.

Figure 4. Decay associated spectra for 2M1NN, 1NN, and 2NN in cyclohexane (left) and acetonitrile (right) obtained from a target and global fit analysis of the transient absorption data using a sequential kinetic model.

traces and best global fit curves for representative decay signals are shown in Figure 6C,D. The excited-state dynamics of 1NN in several nonpolar, aprotic, and protic solvents were discussed in detail in our previous work,10 whereas the DAS presented for the first time are shown in Figure 4. The DAS for 1NN were obtained by using the sequential kinetic model and the target analysis method described in the Experimental and Computational Methods section. The evolution of 1NN is similar to that of 2M1NN and 2NN following excitation at 340 nm. Briefly, an instrument-response limited rise occurs in both cyclohexane and acetonitrile. The UV band then decays as a visible band rises in with maxima at 378, 402, and 554 nm in cyclohexane and 401 and 578 nm in acetonitrile. As these bands rise in, they narrow, shift toward the blue side of the spectrum, and become more structured.10

Figure 5. Normalized transient absorption spectra of 2NN and 2M1NN at the given delay times. Arrows highlight the band narrowing and blue shifting, characteristic of vibrational cooling dynamics.

6583



Article

picoseconds and remains practically constant within the 3 ns time window probed in our experiments. This assignment is supported by the high triplet yields recently reported for 1NN (0.64), 2M1NN (0.33), and 2NN (0.93) in acetonitrile16 and in other solvents.36,37 Furthermore, the similarity of these longlived transient spectra to the triplet−triplet absorption spectra previously reported for 1NN9,10,38 and 2NN38 on the nanosecond to microsecond time scale in similar solvents lends further support to this assignment. To our knowledge, the triplet−triplet transient absorption spectrum of 2M1NN has not been previously reported. Comparison of the DAS corresponding to the triplet−triplet absorption spectra for 1NN and 2M1NN in Figure 4 show that they are very similar but red-shifted by ∼20 nm in the case of 2M1NN. The red shift is due in part to the electron-donor properties of the methyl group. The resemblance of the transient absorption spectra in 1NN and 2M1NN suggests that methylation of 1NN in the ortho position does not perturb significantly the transient absorption species populated in 2M1NN versus those in 1NN. The band narrowing and blue shift shown in Figure 5 indicate that the T1 state in 2NN and 2M1NN is populated with excess vibrational energy. It cools down with an average lifetime of tens of picoseconds, represented by τ3 in Table 1. Population of the hot T1 state has been previously shown for 1NN in a wide variety of solvents,10 supporting this assignment. The ultrafast population of the triplet manifold and the ΔE(T3−T1) energy gap of close to 1 eV reported in refs 10 and 16 lend further support for the population of the T1 state with excess vibrational energy. In fact, a correlation seems to exist between the average vibrational cooling lifetimes reported in Table 1 and the magnitude of the ΔE(T3−T1) energy gap reported in refs 10 and 16 in both solventsthe larger the energy gap, the longer it takes for the T1 state to dissipate excess vibrational energy to the environment. In addition, 2M1NN displays a shorter vibrational cooling lifetime than 1NN and 2NN. The shorter lifetime is likely due to the additional vibrational degrees of freedom from the methyl group. We assign the intermediate decay lifetime of a few picoseconds in 2M1NN and 2NN (τ2 in Table 1) to a slow internal conversion (IC) process from the receiver T3(nπ*) state to the T1(ππ*) state. The population in the T3(nπ*) state decays with a lifetime between 0.6 and 2.8 ps in these NNDs, which correlates with the ΔE(T3−T1) energy gap reported in refs 10 and 16. This is also in accordance with the energy gap law for nonradiative decay,39,40 as reported previously for 1NN.10 We also note that slow IC pathways, on the picosecond time scales, from upper triplet states to the T1 state have been documented previously in other naphthalene derivatives.10,32,41,42 This suggests that slow IC in the triplet manifold is a general phenomenon observed in naphthalene derivatives. 4.2. Kinetic Model and Computational Evidence Supporting the Dissociation of the Intramolecular Sdiss(CT) State. In this section we propose a unified kinetic mechanism that can explain the time-resolved and steady-state photochemistry of these NNDs. This mechanism is based on that previously postulated for 1NN in the same solvents,10 but it is further elaborated to take into consideration the new experimental and computational data presented in this work (Scheme 1). In this model, two competitive nonradiative relaxation channels are populated from the SFC state. The 1 state to the primary relaxation channel connects the SFC 1 receiver triplet state with a rate constant k1a whereas the

Figure 6. Representative decay traces at selected probe wavelengths for 2NN and 2M1NN in cyclohexane (A and C, respectively) and acetonitrile (B and D, respectively). Solid lines represent the best global-fit curves to the transient absorption data. The time zero has been offset in the representation of this figure.

4. DISCUSSION 4.1. Assignment of the Transient Absorption Bands. The steady-state experiments show that fluorescence emission in 1NN, 2NN, and 2M1NN is insignificant within the sensitivity of the spectrophotometer used. The lack of fluorescence emission in these NNDs is in accord with the reported fluorescence lifetimes for 1NN in cyclohexane (∼100 fs) and acetonitrile (∼50 fs).6 Fluorescence decay lifetimes for 2NN and 2M1NN have not been reported but are expected to be on the subpicosecond time scale on the basis of the negligible fluorescence emission. This leaves three possible channels for population decay from the S1FC state: (1) nonradiative decay to the ground state, (2) intersystem crossing to the triplet manifold, and (3) product formation. We assign the initial rise in the transient absorption spectra in Figure 3 to the population of the receiver T3(nπ*) state on the basis of the observations that (1) the S1(ππ*) and T3(nπ*) states in 1NN, 2NN, and 2M1NN are almost isoenergetic in the Franck−Condon region;16 (2) the triplet quantum yield of these NNDs is high;16,36,37 and (3) the fluorescence lifetime of 1NN is ≤100 fs.6 Our assignment of the initial transient absorption spectrum with a maximum below 400 nm to the receiver triplet state is in accord with that previously suggested for 1NN,9,10 and with recent CASPT2//CASSCF calculations for the same molecule.22 The inability to detect the excitedstate absorption spectrum of the S1 state can be rationalized as follows: the S1 absorption spectrum may have a small absorption cross section in the spectral probe region monitored and/or it decays on a faster time scale than the time resolution of our experimental setup. The lack of stimulated emission in the transient absorption spectra suggests that the time resolution of our setup is inadequate to detect the S1 absorption spectra in these NNDs. The long-lived transient absorption is assigned to the T1 state (i.e., Tn ← T1 transitions). This state is populated in a few 6584



Article

significant charge-transfer character and negligible oscillator strength; Sdiss(CT) state in Scheme 1. This is not the case for 2NN. Though the S1-state energy minimum for 2NN has planar geometry with a nitro-aromatic torsion angle of 0° in both solvents, the calculations predict the S1 state in 1NN and 2M1NN has the nitro group closer to perpendicular to the aromatic moiety (torsion angle of 106° in 1NN and 104° in 2M1NN) and it acquires a pyramidal conformation (Figures 1b and 2S, Supporting Information). The S1-state optimizations further support the idea that the nitro-aromatic torsion angle is the reaction coordinate that plays a key role in the population of the intramolecular Sdiss(CT) state, competing with ultrafast intersystem crossing to the triplet manifold in the case of 1NN and 2M1NN. The fully optimized S1 state of 2NN is due to a 100% LUMO ← HOMO transition with electron density distribution primarily localized in the π-Kohn−Sham orbitals of the aromatic moiety (Figure 7). Only a small fraction of the electron density accumulates in the π* orbitals of the oxygen atoms in 2NN. This explains the smaller oscillator strength of the adiabatic S1 state in 1NN and 2M1NN relative to that in 2NN. In contrast, the fully optimized S1 state in 1NN and 2M1NN is composed of 16−17% LUMO ← HOMO and 84− 83% LUMO ← HOMO−2 Kohn−Sham orbital transitions (Figure 7). In both transitions, a major fraction of the electron density is transferred from the π-Kohn−Sham orbitals of the naphthalene moiety to the antibonding orbitals of the nitro group. This transfer of electron density is expected to increase the probability of the C−N bond dissociation and can also explain the pyramidal conformation of the nitro group in the S1-state minimum. Previous works have shown that nitroaromatic radical anions with and without ortho substituents prefer geometrical structures where the nitro group is perpendicular to the aromatic moiety and has a pyramidal conformation in solution.45−47 The accumulation of electron density in the antibonding orbitals together with the strong charge-transfer character of the Sdiss(CT) state supports the idea that dissociation of the nitro group can occur directly from this intramolecular charge-transfer state5,10,11,43 and can be expected to be partially responsible for the reported photodegradation yields in these NNDs.16,48

Scheme 1. Proposed Kinetic Mechanism To Explain the Excited-State Dynamics and Photochemistry of 1NN, 2NN, and 2M1NN in Nonpolar and Polar Aprotic Solventsa

a

Excitation of the distribution of conformations available in the ground state in each molecule results in bifurcation of the SFC 1 state to two primary relaxation pathways: (1) ultrafast intersystem crossing to the triplet manifold and (2) conformational relaxation (CR) to the Sdiss(CT) state. The Sdiss(CT)-state population can dissociate to form the aryloxy (ArO) and the nitrogen(II) oxide (NO) radicals, resulting in product formation. The ArO and NO radicals are not formed in 2NN due to a sizable energy barrier in the S1 state to populate the Sdiss(CT) state.

second, minor channel connects the SFC 1 state to a relaxed S1 state with a rate constant k1b. As discussed in more detail in the next paragraphs, the relaxed S1 state has significant chargetransfer and dissociative character and it is thus labeled Sdiss(CT) in Scheme 1. This intramolecular Sdiss(CT) state is thought to be responsible for the dissociation of the NNDs to form the aryloxy (ArO) and nitrogen(II) oxide (NO) radicals through a dissociation-recombination mechanism.10,11,16,43 Ultrafast branching of the SFC 1 -state population to an intramolecular charge-transfer state and to an ISC decay channel, mediated by the rotation of the nitro group, has been reported previously in other nitro-aromatic compounds.1−3,44 Figure 2 shows that the S1 PECs of 1NN and 2M1NN have dissociative character along the nitro-aromatic torsion angle reaction coordinate, in agreement with previous vertical excitation-energy calculations.10,16 This decay pathway is proposed to be populated on an ultrafast time scale (τ1 = (k1a + k1b)−1 in Table 1), reaching a region in the PEC that has

Figure 7. Kohn−Sham orbitals participating in the configuration-interaction transitions of the optimized S1 state in 2NN (top), 1NN (middle), and 2M1NN (bottom) in acetonitrile at the PBE0/IEFPCM/6-311++G(d,p) level of theory. Similar configuration-interaction transitions were obtained in cyclohexane (not shown). 6585



Article

dependence on the energy parameter ΔEad − EM, where EM is the molecular reorganization energy for the two electronic states under consideration.40 When the spin−orbit interaction of the two electronic states is small compared to ΔEad, and the density of final vibrational states at the energy of the initial state is high, the golden rule approximation can be used to predict the ISC rates.39,50 Herein, we invoke this approximation to rationalized the ultrafast nature of the ISC dynamics in these NNDs. The results presented in this work provide convincing evidence that ISC dynamics take place on the subpicosecond time scale, thus effectively competing with internal conversion. If we assume that the Franck−Condon overlap factors between the S1 state and the receiver Tn state are similar in 1NN, 2NN, and 2M1NN, the femtosecond lifetime of the ISC relaxation pathway should be qualitatively explained by the small energy gap between the S1 and Tn states, as well as by a strong spin− orbit coupling interaction in the region of the potential energy surface where the population transfer occurs. In agreement with this idea, large spin−orbit coupling values of ca. 65 cm−1 and small ΔEad of ca. 0.1−0.2 eV have been recently reported for 1NN in vacuum from minimum energy path calculations using the CASPT2//CASSCF level of theory.22 These energy values have been used by Orozco-Gonzalez and co-workers22 to estimate the ISC rate in 1NN in the range of 1.8 × 1011 to 5.2 × 1011 s−1. The ISC rate estimated by these authors is in reasonable agreement with our experimental results, taking into consideration that, on the basis of our proposed kinetic model, τ1 in Table 1 is not the ISC lifetime between the S1 state and receiver Tn state but rather the lifetime for the branching of population between both Sdiss(CT) and ISC channels. The energy gap and spin−orbit coupling for the adiabatic transition between the S1 and Tn states in 2NN and in 2M1NN are unknown. However, assuming that the reported vertical energy gap between the S1 and Tn states can be used as an indication of the magnitude of the energy gap in the region of the potential energy surface where the ISC takes place, the small energy gap differences of ca. 0.1−0.2 eV in cyclohexane and of ca. 0.02−0.04 eV in acetonitrile reported for these molecules16 support the ultrafast ISC dynamics in 2NN and 2M1NN. The differences observed in the magnitude of the τ1 lifetimes in Table 1 could simply reflect different bifurcation rates in the population of the Sdiss(CT) and the Tn states in these molecules. Alternatively, the different magnitude of the τ1 lifetimes might reflect a change in the Franck−Condon overlap factors and/or a change in the topology of the potential energy surface in the vicinity where the avoided crossing between the S1 and the Tn adiabatic potential energy surfaces occurs. Clearly, a more in-depth analysis of the nuclear and electronic factors that play a major role in the subpicosecond ISC dynamics in these molecules awaits the results of high-level ab initio and molecular dynamics simulations that can explore nonadiabatic dynamics and spin−orbit coupling interactions simultaneously and on equal grounds.51,52 Regardless, the experimental and computational data gathered thus far on these NNDs support the idea that ISC occurs between nonequilibrated excited states in the strongly nonadiabatic limit, where the ability of active vibrational modes in the S1 state to couple and explore the singlet−triplet crossing region might control the ISC rate.

This idea is further supported by the marked differences in the slope of the SFC 1 region accessed by the distribution of torsion angles available at room temperature in each molecule (Figure 2). A steeper gradient toward the intramolecular charge-transfer channel is predicted for 2M1NN than for 1NN, which should result in a higher photodegradation yield in 2M1NN than in 1NN, as observed experimentally.10,16 In 2NN, however, the distribution of torsion angles available at room temperature results in a 4.6 kcal/mol adiabatic energy barrier that must be surmounted to access the Sdiss(CT) state, which is expected to significantly lower the probability of photochemical reaction from the Sdiss(CT) state. In line with this idea, 2NN is photoinert and its triplet state yield is close to unity (0.93).16 In fact, the steepness of the S1-state potential energy surface in the Franck−Condon region is directly proportional to the magnitude of the photodegradation yield and inversely proportional to the magnitude of the triplet quantum yield in each of these NNDs.16 These observations support the idea that the photochemistry of these NNDs is controlled by the population branching to the Sdiss(CT) state and to the Tn state. The population branching is in turn modulated by (1) the energy gap between the S1 and the receiver Tn states, (2) the distribution of torsion angles available in the ground state at the time of excitation, and (3) the topology of the S1 potential energy surface in the Franck−Condon region, as proposed recently.16 As stated above, recent CASPT2//CASSCF calculations support the involvement of a high-energy triplet state in the ultrafast ISC dynamics in 1NN.22 This conclusion is in good agreement with the experimental and computational results presented in this work and elsewhere for these compounds.6,9,10,16 However, whereas TD-DFT calculations suggest that the receiver triplet state is the T3(n,π*) state in 1NN, CASPT2//CASSCF calculations suggest it is instead the T2(n,π*) state. Regardless of whether the receiver state is the T2 state or the T3 state, both levels of theory show that a strong coupling interaction between the S1 state and an upper triplet state with n,π* character plays an important role in the ultrafast ISC dynamics in 1NN. The calculations presented in this work also show this to be the case in 2M1NN and in 2NN derivatives. In the next section, we attempt to further rationalize the ultrafast ISC dynamics observed in these three molecules by invoking Fermi’s golden rule. 4.3. Nonequilibrium Dynamics and the Nature of the Femtosecond Intersystem Crossing Pathway. Traditionally, ISC has been thought to be much slower than IC because it is formally forbidden within the nonrelativistic quantum theory. However, the forbidden character of the ISC pathway is relaxed if two electronic states of different multiplicity are sufficiently close in energy. Jortner and co-workers distinguished between two extreme cases when deriving qualitative rules for probabilities of radiationless transitions.40,49 In the weak coupling limit, the coordinate displacement of each normal mode is assumed to be small and the transition probability has an exponential dependence on the adiabatic energy gap (ΔEad) between the two electronic states involved in the radiationless transition. In other words, the smaller the energy difference between the two electronic states, the larger the transition probability.40 On the other hand, the strong coupling limit is characterized by large relative displacements of one or more nuclear coordinates such that an intersection of the potential energy surfaces can be expected. In this case, the probability of radiationless transition shows a Gaussian 6586



Article

second Fluorescence Up-Conversion. J. Phys. Chem. A 2007, 111, 552−557. (5) Crespo-Hernández, C. E.; Burdzinski, G.; Arce, R. Environmental Photochemistry of Nitro-PAHs: Direct Observation of Ultrafast Intersystem Crossing in 1-Nitropyrene. J. Phys. Chem. A 2008, 112, 6313−6319. (6) Zugazagoitia, J. S.; Almora-Díaz, C. X.; Peon, J. Ultrafast Intersystem Crossing in 1-Nitronaphthalene. An Experimental and Computational Study. J. Phys. Chem. A 2008, 112, 358−365. (7) Mohammed, O. F.; Vauthey, E. Excited-State Dynamics of Nitroperylene in Solution: Solvent and Excitation Wavelength Dependence. J. Phys. Chem. A 2008, 112, 3823−3830. (8) Arce, R.; Pino, E. F.; Valle, C.; Agreda, J. Photophysics and Photochemistry of 1-Nitropyrene. J. Phys. Chem. A 2008, 121, 10294− 10304. (9) Zugazagoitia, J. S.; Collado-Fregoso, E.; Plaza-Medina, E. F.; Peon, J.; Peon, J. Relaxation in the Triplet Manifold of 1Nitronaphthalene Observed by Transient Absorption Spectroscopy. J. Phys. Chem. A 2009, 113, 805−810. (10) Reichardt, C.; Vogt, R. A.; Crespo-Hernández, C. E. On the Origin of Ultrafast Nonradiative Transitions in Nitro-Polycyclic Aromatic Hydrocarbons: Excited-State Dynamics in 1-Nitronaphthalene. J. Chem. Phys. 2009, 131 (224518), 1−15. (11) Vyas, S.; Onchoke, K. K.; Rajesh, C. S.; Hadad, C. M.; Dutta, P. K. Optical Spectroscopic Studies of Mononitrated Benzo[a]pyrenes. J. Phys. Chem. A 2009, 113, 12558−12565. (12) Arce, R.; Pino, E. F.; Valle, C.; Negrón-Encarnación, I.; Morel, M. A Comparative Photophysical and Photochemical Study of Nitropyrene Isomers Occurring in the Environment. J. Phys. Chem. A 2011, 115, 152−160. (13) Plaza-Medina, E. F.; Rodríguez-Córdoba, W.; Peon, J. Role of Upper Triplet States on the Photophysics of Nitrated Polyaromatic Compounds: S1 Lifetimes of Singly Nitrated Pyrenes. J. Phys. Chem. A 2011, 115, 9782−9789. (14) Plaza-Medina, E. F.; Rodríguez-Córdoba, W.; Morales-Cueto, R.; Peon, J. Primary Photochemistry of Nitrated Aromatic Compounds: Excited-State Dynamics and NȮ Dissociation from 9Nitroanthracene. J. Phys. Chem. A 2011, 115, 577−585. (15) García-Berríos, Z. I.; Arce, R. Photodegradation Mechanism of 1-Nitropyrene, an Environmental Pollutant: The Effect of Organic Solvents, Water, Oxygen, Phenols, and Polycyclic Aromatics on the Destruction and Product Yields. J. Phys. Chem. A 2012, 116, 3652− 3664. (16) Vogt, R. A.; Crespo-Hernández, C. E. Confromational Control in the Population of the Triplet State and Photoreactivity of NitroNaphthalene Derivatives, submitted for publication. (17) IPCS INCHEM (2003) Selected Nitro- and Nitro-Oxy-Polycyclic Aromatic Hydrocarbons. Environmental Health Criteria (EHC) Monographs, No. 229; WHO: Geneva, 2003. (18) Katarina, V. Review of the Genotoxicity of Nitrogen Oxides. Mutat. Res. 1994, 317, 43−55. (19) Yu, H. Environmental Carcinogenic Polycyclic Aromatic Hydrocarbons: Photochemistry and Phototoxicity. J. Environ. Sci. Health C 2002, C20, 149−183. (20) Wehry, E. L., Effects of Molecular Structure on Fluorescence and Phosphorescence. In Practical Fluorescence, 2nd ed.; Guilbault, G. G., Ed.; Marcel Dekker, Inc.: New York, 1990; pp 75−125. (21) Döpp, D., Photochemical Reactivity of the Nitro Group. CRC Handbook of Organic Photochemistry and Photobiology; CRC Press, Inc.: Boca Raton, FL, 1995; pp 1019−1062. (22) Orozco-Gonzalez, Y.; Coutinho, K.; Peon, J.; Canuto, S. Theoretical Study of the Absorption and Nonradiative Deactivation of 1-Nitronaphthalene in the Low-Lying Singlet and Triplet Excited States Including Methanol and Ethanol Solvent Effects. J. Chem. Phys. 2012, 137 (054307), 1−8. (23) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; et al. Gaussian 09, Revision A.01; Gaussian, Inc.: Wallingford, CT, 2009.

5. CONCLUSIONS We present evidence supporting the idea that the branching ratio between the Sdiss(CT) state and the Tn state is primarily controlled by (1) the topology of the S1-state potential energy surface in the Franck−Condon region, (2) the energy difference between the S1 state and the receiver Tn state, and (3) the distribution of nitro-aromatic torsion angles that is thermodynamically available in the ground state at the time of excitation. These observations lead us to propose that the ultrafast ISC dynamics in 1NN, 2NN, and 2M1NN occur in the strongly nonadiabatic limit. Subpicosecond ISC dynamics have now been documented in several naphthalene derivatives,6,10,32 showing that this electronic relaxation pathways is not unique for 1NN. Furthermore, the results presented in this work lend further support to the idea that an important relationship exists between conformational heterogeneity in the ground state and the observed photochemistry in these NNDs.16 Multiconfigurational ab initio methods, coupled to quantumdynamics simulations, in addition to UV resonance Raman and femtosecond time-resolved vibrational techniques should be able to provide further insights about the excited-state structural dynamics and the relationship between ground-state conformational sampling and the photochemistry in these and other NPAHs.

■

ASSOCIATED CONTENT

S Supporting Information *

Dependence of the transient absorption signals on the pump power and optimized excited singlet state geometry of 1nitronaphthalene in acetonitrile at the CIS/IEFPCM/6-31+G(d,p) level of theory. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This research was partially supported by the donors of the American Chemical Society Petroleum Research Fund. The authors thank Case Western Reserve University, as well as the Mississippi Center for Supercomputer Research, and the Ohio Supercomputer Center for generous allotment of computer time. C.R. thanks the Deutsche Forschungsgemeinschaft (DFG) for support (GZ: RE 2918/1-1).

■

REFERENCES

(1) Rafiq, S.; Yadav, R.; Sen, P. Femtosecond Excited-State Dynamics of 4-Nitrophenyl Pyrrolidinemethanol: Evidence of Twisted Intramolecular Charge Transfer and Intersystem Crossing Involving the Nitro Group. J. Phys. Chem. A 2011, 115, 8335−8343. (2) Quenneville, J.; Greenfield, M.; Moore, D. S.; McGrane, S. D.; Scharff, R. J. Quantum Chemistry Studies of Electronically Excited Nitrobenzene, TNA, and TNT. J. Phys. Chem. A 2011, 115, 12286− 12297. (3) Ghosh, R.; Palit, D. K. Ultrafast Dynamics of the Excited States of 1-(p-Nitrophenyl)-2-(hydroxymethyl)pyrrolidine. J. Phys. Chem. A 2012, 116, 1993−2005. (4) Morales-Cueto, R.; Esquivelzeta-Rabell, M.; SaucedoZugazagoitia, J.; Peon, J. Singlet Excited-State Dynamics of Nitropolycyclic Aromatic Hydrocarbons: Direct Measurements by Femto6587



Article

(24) Barone, V.; Cossi, M.; Tomasi, J. A New Definition of Cavities for the Computation of Solvation Free Energies by Polarizable Continuum Model. J. Chem. Phys. 1997, 107, 3210−3221. (25) Cancès, E.; Mennucci, B.; Tomasi, J. A New Integral Equation Formalism for the Polarizable Continuum Model: Theoretical Background and Applications to Isotropic and Anisotropic Dielectrics. J. Chem. Phys. 1997, 107, 3032−3041. (26) Adamo, C.; Scuseria, G. E.; Barone, V. Accurate Excitation Energies from Time-Dependent Density Functional Theory: Assessing the PBE0 Model. J. Chem. Phys. 1999, 111, 2889−2899. (27) Adamo, C.; Barone, V. Toward Reliable Density Functional Methods Without Adjustable Parameters: The PBE0 Model. J. Chem. Phys. 1999, 110, 6158−6170. (28) Nakayama, T.; Amijima, Y.; Ibuki, K.; Hamanoue, K. Construction of a Subpicosecond Double-Beam Laser Photolysis System Utilizing a Femtosecond Ti:Sapphire Oscillator and Three Ti:Sapphire Amplifiers (A Regenerative Amplifier and Two Double Passed Linear Amplifiers), and Measurements of the Transient Absorption Spectra by a Pump-Probe Method. Rev. Sci. Instrum. 1997, 68, 4364−4371. (29) van Stokkum, I. H. M.; Larsen, D. S.; van Grondelle, R. Global and Target Analysis of Time-Resolved Spectra. Biochim. Biophys. Acta 2004, 1657, 82−104. (30) van Stokkum, I. H. M.; Larsen, D. S.; van Grondelle, R. Erratum to “Global and Target Analysis of Time-Resolved Spectra”. Biochim. Biophys. Acta 2004, 1658, 262. (31) Capellos, C.; Bielski, B. H. J. Kinetic Systems; Wiley Interscience: New York, 1972. (32) Vogt, R. A.; Gray, T. G.; Crespo-Hernández, C. E. Subpicosecond Intersystem Crossing in Mono- and Di-(Organophosphine)Gold(I) Naphthalene Derivatives in Solution. J. Am. Chem. Soc. 2012, 134, 14808−14817. (33) Dreuw, A.; Weisman, J. L.; Head-Gordon, M. Long-Range Charge-Transfer Excited States in Time-Dependent Density Functional Theory Require Non-Local Exchange. J. Chem. Phys. 2003, 119, 2943−2946. (34) Dreuw, A.; Head-Gordon, M. Single-Reference Ab Initio Methods for the Calculation of Excited States of Large Molecules. Chem. Rev. 2005, 105, 4009−4037. (35) Richard, R. M.; Herbert, J. M. Time-Dependent DensityFunctional Description of the 1La State in Polycyclic Aromatic Hydrocarbons: Charge-Transfer Character in Disguise? J. Chem. Theory Compt. 2011, 7, 1296−1306. (36) Rusakowicz, R.; Testa, A. C. Phosphorescence Study of Nitronaphthalenes. Spectrochim. Acta A 1971, 27A, 787−792. (37) Hurley, R.; Testa, A. C. Triplet-State Yield of Aromatic Nitro Compounds. J. Am. Chem. Soc. 1968, 90, 1949−1952. (38) Mikula, J. J.; Anderson, R. W.; Harris, L. E. Electronic States and Energy Transfer in Nitronaphthalenes. Adv. Mol. Relax. Proc. 1973, 5, 193−209. (39) Turro, N. J.; Ramamurthy, V.; Scaiano, J. C. Principles of Molecular Photochemistry: An Introduction; University Science Books: Sausalito, CA, 2009. (40) Englman, R.; Jortner, J. Energy Gap Law for Radiationless Transitions in Large Molecules. Mol. Phys. 1970, 18, 145−64. (41) Sakamoto, M.; Cai, X.; Hara, M.; Fujitsuka, M.; Majima, T. Intermolecular Electron Transfer from Naphthalene Derivatives in the Higher Triplet Excited States. J. Am. Chem. Soc. 2004, 126, 9709− 9714. (42) Sakamoto, M.; Cai, X.; Hara, M.; Fujitsuka, M.; Majima, T. Significant Effects of Substituents on Substituted Naphthalenes in the Higher Triplet Excited State. J. Phys. Chem. A 2005, 109, 4657−4661. (43) Brown, H. W.; Pimentel, G. C. Photolysis of Nitromethane and Methyl Nitrite in an Argon Matrix; Infrared Detection of Nitroxyl, HNO. J. Chem. Phys. 1958, 29, 883−888. (44) Kovalenko, S. A.; Schanz, R.; Farztdinov, V. M.; Hennig, H.; Ernsting, N. P. Femtosecond Relaxation of Photoexcited ParaNitroaniline: Solvation, Charge Transfer, Internal Conversion and Cooling. Chem. Phys. Lett. 2000, 323, 312−322.

(45) Geske, D. H.; Ragle, J. L.; Bambenek, M. A.; Balch, A. L. Study of Steric Effects by Electron Spin Resonance Spectroscopy and Polarography. Substituted Nitrobenzenes and Nitroanilines. J. Am. Chem. Soc. 1964, 86, 987−1002. (46) Dirk, J. A.; Ridder, D.; Schenk, H. The Influence of Ortho Substituents on the Twisting Angle of a Nitro Group with Respect to Aromatic Six-Membered Carbon Rings. Acta Crystallogr. 1995, B51, 221−231. (47) Barrows, S. E.; Cramer, C. J.; Truhlar, D. G. Factors Controlling Regioselectivity in the Reduction of Polynitroaromatics in Aqueous Solution. Environ. Sci. Technol. 1996, 30, 3028−3038. (48) Phousongphouang, P. T.; Arey, J. Rate Constants for the Photolysis of the Nitronaphthalenes and Methylnitronaphthalenes. J. Photochem. Photobiol. A 2003, 157, 301−309. (49) Freed, K. F.; Jortner, J. Multiphonon Processes in the Nonradiative Decay of Large Molecules. J. Chem. Phys. 1970, 52, 6272−6291. (50) Marian, C. M. Spin-Orbit Coupling and Intersystem Crossing in Molecules. WIREs Comput. Mol. Sci. 2012, 2, 187−203. (51) Richter, M.; Marquetand, P.; González-Vázquez, J.; Sola, I.; González, L. SHARC - Ab Initio Molecular Dynamics with Surface Hopping in the Adiabatic Representation Including Arbitrary Couplings. J. Chem. Theory Compt. 2011, 7, 1253−1258. (52) Granucci, G.; Persico, M.; Spighi, G. Surface Hopping Trajectory Simulations with Spin-Orbit and Dynamical Couplings. J. Chem. Phys. 2012, 137 (22A501), 1−9.

6588


Excited-State Dynamics in Nitro-Naphthalene Derivatives: Intersystem

Recommend Documents