Rotamer-Specific Photoisomerization of Difluorostilbenes from

The Raman spectra of S0 and S1 are in qualitative agreement with calculations. ..... For F2, the fit yields τ1 = 357 ps (81%) and τ2 = 62 ps (19%), ...
0 downloads 0 Views 3MB Size
Article pubs.acs.org/JPCB

Cite This: J. Phys. Chem. B 2018, 122, 1049−1059

Rotamer-Specific Photoisomerization of Difluorostilbenes from Transient Absorption and Transient Raman Spectroscopy M. Quick,*,† A. L. Dobryakov,† I. N. Ioffe,‡ F. Berndt,† R. Mahrwald,† N. P. Ernsting,† and S. A. Kovalenko† †

Department of Chemistry, Humboldt-Universität zu Berlin, Brook-Taylor-Str. 2, D-12489 Berlin, Germany Department of Chemistry, Lomonosov Moscow State University, Moscow, Russia



S Supporting Information *

ABSTRACT: Photoisomerization of 2,2′-, 3,3′-, and 4,4′-difluorostilbene (F2, F3, F4, respectively) in n-hexane, perfluoro-n-hexane, and acetonitrile is studied with broadband transient absorption (TA) and femtosecond stimulated Raman (FSR) spectroscopy and by DFT/TDDFT calculations. F2 and F3 possess three rotamers (rotational isomers) each, while F4 has one single conformation only. These differences are reflected in TA and FSR spectra. Thus F4 reveals a monoexponential decay of TA with τ1 = 172 ps in n-hexane, as expected for a single species. For F2 and F3, the decays are biexponential in all solvents, corresponding to two distinctly discerned rotamers or rotamer fractions. Specifically, for F2 in n-hexane, τ1 = 357 ps (83%) and τ2 = 62 ps (17%), and for F3 in the same solvent, τ1 = 222 ps (57%), and τ2 = 81 ps (43%). The weights in brackets agree with theoretically estimated ground-state abundances of the rotamers. Furthermore, a global fit of the TA and FSR data allows us to extract the spectra of the pure rotamers. The Raman spectra of S0 and S1 are in qualitative agreement with calculations.

I. INTRODUCTION The photophysics and photochemistry of fluoro-substituted stilbenes were first reported by Fischer and coworkers and Luettke and Rauch.1 The authors studied the absorption and fluorescence spectra in a wide temperature range concerning fluorescence and isomerization yields of five selected fluorostilbenes. One of them, 1,1′-difluorostilbene, has recently been investigated by us2 with broadband transient absorption (TA) and femtosecond stimulated Raman (FSR) spectroscopy.3−6 The substitution at the ethylenic position eliminates the photoisomerization barrier, resulting in an ultrafast trans-to-cis torsion. Thus in n-hexane the isomerization proceeds with 0.4 ps, compared with 84 ps for unsubstituted stilbene.2 We extend our study to ring-fluorinated trans-stilbenes, 2,2′-, 3,3′-, and 4,4′-difluorostilbene (F2, F3, F4, respectively) as depicted in Scheme 1. A new interesting feature here (in addition to the photoisomerization) is the rotational isomerism (rotamerism) about the 1-7 or 1′-7′ single bonds. As can be seen, F2 and F3 possess three rotamers, whereas F4 has one single conformer only, quite like in parent stilbene. In general, the rotamers should have different photoisomerization barriers, as can be checked by time-resolved spectroscopy. The beginning of rotamer photochemistry goes back to the early 1960s, when Cherkasov7 explained the spectral anomalies of n-vinylanthracene by a mixture of its rotamers. Later on, Havinga and coworkers detected rotamer-specific photoproducts also from hexa-1,3,5-triene.8 Also transient methods were applied, here by Park and Waldeck,9 who observed a nonexponential decay in 3,3′-dimethylstilbene, which they ascribed to the coexistence of different rotamers in the excited © 2017 American Chemical Society

Scheme 1. F4 Has a Single Rotamer (Rotational Isomer), whereas F2 and F3 Possess Three: Symmetric s and s′ and Asymmetric aa

a Steric hindrance depends on the F···H distance, which is especially small in F2s′. Therefore the latter should be less populated than F2s or F2a. For F3, however, the steric hindrance is negligible; hence all three rotamers may be populated.

state. A large review article was drafted by Mazzucato and Momicchioli that constituted the state of the art until 1991.10 Received: September 18, 2017 Revised: November 27, 2017 Published: November 27, 2017 1049

DOI: 10.1021/acs.jpcb.7b09283 J. Phys. Chem. B 2018, 122, 1049−1059

Article

The Journal of Physical Chemistry B Table 1. Ground-State Energies and Abundances of F2 and F3 PBE0/Def2-TZVPP

mPW2PLYP/Def2-TZVP

rotamer

Erel + ZPE (kJ/mol)

stat. factor

abundance (%)

rotamer

Erel + ZPE (kJ/mol)

stat. factor

abundance (%)

F2s (C2h + C2) F2a (C1) F2s′ (C2h) F3s (C2h) F3a (Cs) F3s′ (C2h)

0 3.3 4.1 1.8 1.0 0.0

3 4 1 1 2 1

71 25 4 17 47 36

F2s (C2 + Ci) F2a (C1) F2s′ (C2h) F3s (C2) F3a (C1) F3s′ (C2)

0.0 2.8 2.7 1.3 0.7 0.0

4 4 1 1 2 1

71 23 6 19 49 32

And more recently, Saltiel et al.11 and Karatsu et al.12 reported rotamer-specific adiabatic cis−trans photoisomerization of 1-(2anthryl)-2-phenylethene. Coming back to Scheme 1, it is instructive to clarify two points. The first one concerns the interconversion between the rotamers. Previous calculations10 showed that the relevant barriers in the S0 state range ∼20 kJ/mol, corresponding to an interconversion time of ∼1 ns. This is much shorter than the diffusional time scale of chemical reactions in solution, thus explaining why the common (ground-state) chemistry does not differentiate the rotamers. However, the picture changes in the excited state. Upon S0 → S1 excitation, single bonds 1−7 and 1′−7′ strengthen, thus freezing the ground-state structure of the rotamer, while the ethylenic bond weakens, allowing for a fast (∼100 ps) photisomerization. This is the so-called principle of nonequilibration of excited rotamers (NEER)10 that seems to be valid for many conjugated systems, including those considered here. The second point deals with the choice of a substituent group. The fluorine atom has a relatively small van der Waals volume, 10 Å3, compared with 28 Å3 of the rather large methyl group. Accordingly, the steric hindrance is minimized by fluorines so that a stable planar rotamer structure is preferred. Nonetheless, in F2s′ the fluorines are tightly spaced to the ethylenic hydrogens that destabilize this rotamer in favor of F2s. In the case of F3 the influence on steric hindrance should be negligible, and both F3s and F3s′ may be populated in similar amounts. The rest of the paper is organized as follows. Section II reports our quantum-chemical calculations in the S0 and S1 state of the rotamers. Experimental results are presented in Section III, where we first discuss the symmetric F4 and then the rotamers of F2 and F3. A comparison of experimental results with quantum-chemical calculations is discussed in detail in Section IV.

the thermodynamic functions can be comparable in magnitude to the energy difference between the rotamers, so possible inaccuracies may considerably distort the picture. Obviously, the phenyl rotational modes cannot be treated harmonically, but ambiguities regarding the exact shape of the S0 energy surface and possible anharmonic coupling between vibrational modes16−18 suggest that even sophisticated approaches may still lack reliability. In view of that we restrict ourselves to comparison of the electronic energies with conventional harmonic ZPE correction. In the case of F2, PBE0 and mPW2PLYP predict slightly different exact stationary point geometries for the F2s rotamer. At the PBE0/Def2-TZVPP level, F2s has isoenergetic C2h and C2 minima. (In the C2 case, phenyl rings are slightly rotated out of plane.) At the mPW2PLYP/Def2-TZVP level, a nonplanar Ci minimum is found instead of the C2h one, and the C2 and Ci minima are again isoenergetic. The F2s′ rotamer is found to be planar with both exchange-correlation functionals. Asymmetric F2a shows no nontrivial symmetry; the geometry of its respective halves resembles those of F2s′ and of nonplanar F2s. Considering symmetry numbers and chirality, the statistical factors to the rotameric abundance of F2s:F2a:F2s′ are thus 3:4:1 at the PBE0 level and 4:4:1 at the mPW2PLYP level. The calculated energies and abundances are reported in Table 1. We also checked the ground-state barriers between the rotamers and found them in agreement with previous findings10 to be ∼20 kJ/mol, thus suggesting that the femtosecond experiment indeed deals with an equilibrated ground-state rotamer mixture. In the case of F3, the lack of direct F···H interactions with ethylenic hydrogens results in higher geometric similarity and lower energy differences between the rotamers. F3s and F3s′ are both found to be either C2h-(PBE0) or C2-symmetric (mPW2PLYP), with F3a being C s - or C 1 -symmetric, respectively. The abundances are comparable, yet with somewhat higher content of F3a and F3s′. The photoisomerization barriers S1 → P, from trans to perpendicular conformation P are calculated at TD-PBE0/ Def2-TZVPP level and collected in Table 2. Although the barriers are systematically underestimated, they still reveal the correct relative trends. In all rotamers, the optimized transition-state geometry was highly similar, with twisting angle between 131 and 127° (except for nearly 137° in F2s′).

II. QUANTUM-CHEMICAL CALCULATIONS Electronic ground S0 and excited S1 states of F2, F3, and F4 were studied at the DFT/TDDFT level of theory using the Firefly package13 partly based on the GAMESS(US) source code14 and with Gaussian09.15 The Firefly package was used for (TD-)PBE0/Def2-TZVPP calculations of the twisting barriers of rotamers in the S1 state and of nonresonant Raman activities in S0 and S1. The TDDFT calculations were performed with frozen chemical core. The Gaussian09 was employed to supply mPW2PLYP/Def2-TZVP data for the relative ground-state stability of the rotamers in F2 and F3 to verify the qualitative agreement between the hybrid and double-hybrid exchangecorrelation functionals. The task of accurately calculating relative rotameric abundances is complicated by low-frequency modes due to rotation of the phenyl rings. Contributions of these modes to

III. EXPERIMENTAL RESULTS III.A. Experimental Section. Our TA setup with applications has been described elsewhere.2−4 The probe range is 275−690 nm, with a 0.1 ps pump−probe intensity cross-correlation time (fwhm) and timing precision of 0.02 ps. Solutions of F2, F3, and F4 (each with a 30 μM concentration) 1050

DOI: 10.1021/acs.jpcb.7b09283 J. Phys. Chem. B 2018, 122, 1049−1059

Article

The Journal of Physical Chemistry B Table 2. S1 → P Isomerization Barriers Eb (TD-PBE0/Def2TZVPP) stationary point symmetry in S1

Eb (kJ/mol)

F2s (C2h) F2a (C1) F2s′ (C2) F3s (C2) F3a (C1, nearly Cs) F3s′ (C2h) F4 (C2)

10.1 5.8 2.2 6.5 7.3 8.0 8.0

in n-hexane, acetonitrile, and perfluoro-n-hexane (pfh) flow through a sample cell of a 0.4 mm internal thickness. TA spectra at the magic angle, ΔA(λ,t), are recorded upon S0 → S1 optical excitation, λexc = 325 nm, or 310 nm for pfh. The spectra are also measured for parallel ΔA∥ (λ,t) and perpendicular ΔA⊥(λ,t) pump−probe polarization. Transient anisotropy is calculated as ρ(λ , t ) = (ΔA − ΔA⊥)/(ΔA + 2ΔA⊥) = (ΔA − ΔA⊥)/(3ΔA)

Figure 1. Extinction spectra ε(λ) and normalized SE spectra E(λ) = λ4F(λ) in n-hexane, where F(λ) are fluorescence spectra.

(1)

The FSR setup5,6 is similar to that for TA. The picosecond Raman pump, of ∼10 cm−1 width and ∼0.1 μJ energy, is tuned to λR = 621 nm. The polychromators are adjusted to cover a 1000 cm−1 probe range. Stokes Raman signals are recorded by chopping the Raman-pump beam, with actinic excitation at λac = 313 nm. In this registration scheme, signals at negative delays correspond to S0 Raman contributions from both solute and solvent. For positive delays these contributions are eliminated by subtracting the time-averaged signal at negative delays. The FSR spectra are recorded with parallel actinic/Raman/probe polarization. The magic-angle Raman signal ΔA(λ,t) = (ΔA∥ + 2ΔA⊥)/3 is recalculated with the help of eq 1 ΔA(λ , t ) =

is convenient to start with F4, which exists as a single rotamer only. Its TA spectra and kinetics in n-hexane are shown in Figure 3. Let us consider the transient spectra at left. Early Franck− Condon (FC) spectra about time-zero develop within the pump−probe cross-correlation (top). Negative signals originate from bleach and stimulated emission (SE), while excited-state absorption (ESA) is positive. Arrows indicate the direction of signal evolution. Relaxation away from the FC region occurs with τFC = 0.2 ps (middle). It is seen by vanishing SE structure (∼325 nm) and as a small red shift of ESA. Not directly recognizable in the TA spectra is a small intensity-gain of ESA (by 3%) with τf = 2.4 ps. The corresponding decay-associated spectrum (DAS) exhibits a negative sign in this region, describing a rise of optical density. Such a phenomenon can be attributed to an increase in oscillator strength of the Sn ← S1 transition as part of relaxation in S1. Afterward, the bleach, SE, and ESA decay nearly monoexponentially with τ1 = 172 ps (bottom). The isomerization products, that is, cis- and transisomers, are evident in the bleach region, λ < 330 nm at infinite delay times. In addition, there are traces of triplets and possibly dihydrophenanthrene (DHP) in the range of 400−500 nm. The latter might be formed from the cis-configuration that is populated via an adiabatic trans−cis isomerization. In general, the spectral bands and their evolution are very similar to those of trans-stilbene in n-hexane,22,23 although the decay time in the present case is nearly twice as long. This decay is associated with thermally activated S1 → P intramolecular torsion when the molecule acquires the perpendicular conformation P. The subsequent P → S0 step that completes the isomerization in the ground state should be ultrafast (