J. Phys. Chem. A 2010, 114, 6897–6903
6897
Is There Symmetry Breaking in the First Excited Singlet State of 2-Pyridone Dimer? Espen Sagvolden*,† and Filipp Furche* UniVersity of California, IrVine, Department of Chemistry, 1102 Natural Sciences II, IrVine California 92697-2025 ReceiVed: March 24, 2010; ReVised Manuscript ReceiVed: May 12, 2010
We investigate the S1 state potential energy surface of 2-pyridone dimer (2PY)2 using time-dependent density functional and coupled cluster theory. Although the ground and S2 excited states of (2PY)2 have C2h symmetry, the S1 state shows symmetry breaking and localization of the excitation on one of the two monomers upon relaxation of the geometry. This localization is rationalized using a simple diabatic curve crossing model. As a consequence of the symmetry breaking, S1 to S0 transitions become optically allowed. We hypothesize that the band at 30 776 cm-1 observed in the excitation spectrum of (2PY)2 might be attributed to the S1 state rather than the S2 state; the S2 state origin is predicted 3000-4000 cm-1 above the S1 state by hybrid density functional and coupled cluster methods. Asymmetric transfer of one hydrogen atom leads to a second S1 state minimum that can rapidly decay to the ground state. This suggests that photoinduced tautomerization of (2PY)2 occurs in a stepwise fashion, with only one hydrogen transfer taking place on the S1 surface. Introduction 2-Pyridone (2PY, Figure 1) has attracted theoretical and experimental interest for decades because of its solventdependent tautomerism with 2-hydroxypyridine (2HP, Figure 1),1-7 its amenity to gas-phase spectroscopy,6,8-10 and its role as a prototype cis-amide possessing the same molecular recognition motif as thymine and uracil.11 2PY and 2HP form three different hydrogen-bonded dimers: (2PY)2, (2HP)2, and the mixed dimer 2-pyridone · 2-hydroxypyridine (2PY · 2HP) (Figure 1). (2PY)2 has been studied extensively as a uracil dimer model.12-19 As opposed to uracil dimer, which is spectroscopically elusive, (2PY)2 fluoresces and has narrow absorption bands.14 Besides nucleotide pairs, hydrogen bonded cis-amide structures also occur (albeit atypically) in peptides,20,21 and the 2-pyridone dimer offers the possibility to study a hydrogenbonded cis-amide structure in the gas phase. The 2-pyridone dimer has also been used to study whether the environment influences hydrogen-bond lengths.12,22 Mu¨ller et al.16 recently used (2PY)2 to study the rate of interstrand excitation energy transfer (EET) in DNA. Using twocolor resonant two-photon ionization spectroscopy on monodeuterated (2PY)2, they found that the S1/S2 splitting of (2PY)2 is ∼50 cm-1. The S1/S2 splitting is related to Fo¨rster’s rate expression for EET23 through first-order degenerate perturbation theory. The moderate size of (2PY)2 permits accurate calculations of spectroscopic properties. Recently, Sagvolden, Furche, and Ko¨hn studied the vertical splitting between the S1 and S2 states of (2PY)2 at the ground-state equilibrium structure.24 Coupledcluster and hybrid density functional theory predicted a much larger splitting of ∼1100 cm-1, which is more than an order of magnitude larger than the experimental result. This discrepancy is puzzling because 2PY is a prominent model system with a wealth of experimental data available. It is imperative to * Corresponding author. E-mail:
[email protected]; filipp.furche@ uci.edu. † Current address: SINTEF Materials and Chemistry, Forskningsvn 1, NO-0314 Oslo, Norway.
understand the origin of this discrepancy for (2PY)2 before reliable predictions of energy transfer rates in larger, and less well-characterized systems can be made. Here we suggest a new interpretation of the two-color resonant two-photon ionization (2C-R2PI) spectra, showing that the splitting found in the monodeuterated samples may not be the S1/S2 splitting. A second motivation for our investigation of the (2PY)2 S1 and S2 surfaces is related to a long-standing controversy about the mechanism of dual-proton phototautomerism of DNA base pairs following photoexcitation, which has been discussed as a possible pathway to mutations.25 The mechanism of this photoinduced tautomerization has been much debated; Kasha and coworkers26 argue that both protons are transferred simultaneously, whereas Zewail and others27 favor a stepwise transfer. Most existing work focuses on the 7-azaindole dimer, despite the spectroscopic amenity and analogy of (2PY)2 to DNA base pairs. Our article is organized as follows: After summarizing the computational methods, we present ground- and excited-state optimized structures and energies. We start with the monomer structures because they form a basis for the understanding of the dimers. We discuss and explain the discrepancy between the splitting we found and the splitting found by Mu¨ller et al. and present an alternative interpretation of the spectroscopic data. Finally, we address the question whether photoinduced double proton transfer is concerted or stepwise in (2PY)2.
Figure 1. 2-Pyridone, 2-hydroxypyridine, and their dimers.
10.1021/jp102637e 2010 American Chemical Society Published on Web 06/08/2010
6898
J. Phys. Chem. A, Vol. 114, No. 25, 2010
Sagvolden and Furche
TABLE 1: S1 States in 2PY and 2HPa symmetry assignment transition ∆E (BHandH-LYP) ∆E (CC2) ∆E (exptl)b f (BHandH-LYP)
2PY(S1)
2HP(S1)
Cs 21A′ π f π* 34 899 30 886 29 832 0.12
Cs 21A′ π f π* 41 937 37 166 36 136 0.10
a All energies are in inverse centimeters. ∆E is the adiabatic S0 to S1 excitation energy including zero-point vibrational energy(ZPVE). f is the oscillator strength in the length representation. b Ref 2.
Figure 2. 2PY monomer HOMO (4a′′) and LUMO (5a′′).
Computational Details Ground- and excited-state structures were optimized using the BHandH-LYP density functional28-31 containing 50% Hartree-Fock exchange. All structures were confirmed to be local minima by force constant calculations using analytical32 (ground states) or numerical (excited states) differentiation. The same computations supplied vibrational eigenmodes and frequencies. The coordinates of the structures are given as Supporting Information. Due to considerable charge-transfer character of parts of the S1 surface, hybrid functionals with large fractions of Hartree-Fock exchange are necessary for (2PY)2. We have no indication that BHandH-LYP suffers from spurious admixture of charge-transfer excitations close to the (2PY)2 and (2HP)2 S1 structure. BHandH-LYP is generally considered to be reliable for the prediction of excited-state structures.33 At the BHandH-LYP structures, we computed single-point energies with the approximate coupled-cluster singles-doubles method CC234,35 within the resolution-of-the-identity (RI) approximation.36-38 In our recent study,24 we found that BHandH-LYP and CC2 were consistent for vertical splittings in (2PY)2. All density-functional computations used polarized split-valence basis sets (def2-SVP39) and m4 grids40 and all CC2 computations used polarized triple-ζ valence basis sets (def2TZVPP41). No CC2 excited-state structures were used because unconstrained optimization of the S1 state of (2PY)2 using CC2 lead to highly distorted spurious minima in disagreement with experiment. An 8 × 8 -point 2D potential energy surface for the dualproton transfer tautomerism between (2PY)2 and (2HP)2 was computed by constraining the length of both N-H bonds and the symmetry to Cs and optimizing using BHandH-LYP. The N-H bond distances varied from 99.74 to 204.63 pm (the bond lengths of a biradical structure (which we will call structure B) found in the S1 surface) in steps of 17.48 pm. The N-H bond distances in the (2PY)2 S0 and S1 and (2HP)2 S0 and S1 structures are contained in this interval. At the structures found, S0 and S1 energies were computed using CC2. Transition states between (2PY)2 and (2HP)2 on the S0 potential energy surface and between (2PY)2 and the biradical minimum on the S1 surface were fully optimized at the BHandH-LYP level. The contour value in MO plots is 0.065. All computations were performed using the Turbomole package.35,37,40,42-47
Figure 3. 2HP monomer HOMO (4a′′) and LUMO (5a′′).
TABLE 2: Monomer Bond Lengths in Picometers 2PY(S0) d(N1 - H12) d(C2 - O7) d(O7 - H12)
100.7 120.7
2HP(S0) 133.3 95.8
100.7 123.2
2HP(S1) 131.4 96.4
by Frey et al.9 based on a fit of the vibrationally resolved absorption spectrum of 2PY using a symmetric double-minimum potential. The question of puckering is addressed in refs 9 and 10, and there is disagreement about the nature and magnitude of the out-of-plane distortion. Whereas our present results do not support a puckered structure, we cannot exclude a small, energetically unimportant distortion. For the dimers, we are not aware of any results supporting nonplanar S1 structures. The gas-phase ground-state energy of 2HP computed with CC2 is 235 cm-1 lower than that of 2PY, consistent with a tautomeric equilibrium at room temperature.1,48 In both monomers, the S0 f S1 excitation is a π f π* HOMO-LUMO transition at the BHandH-LYP level. The HOMO(4a′′) and LUMO(5a′′) of 2PY and 2HP are illustrated in Figures 2 and 3, respectively. The S1 excitation energy of 2PY is computed to be ∼6300 cm-1 lower than that of 2HP (Table 1), in good agreement with the results of ref 2. The most important ground- and excited-state bond lengths are given in Table 2. Figure 4 shows the atom labels used. Atomic coordinates and more complete tables of bond lengths are available as Supporting Information. The bond length alternation is consistent with expectations from simple valencebond theory.
Results Monomer Structures. The monomer S1 states are displayed in Table 1. Both 2PY and 2HP have Cs symmetry in their S0 and S1 states at the BHandH-LYP/SVP level. This is at variance with Held and Pratt’s8 conclusion that the 2PY S1 structure is slightly puckered, with the nitrogen atom protruding out of the molecular plane. Puckering of the S1 structure was confirmed
2PY(S1)
Figure 4. Atomic labels in 2PY and 2HP.
S1 State Potential Energy Surface of (2PY)2
J. Phys. Chem. A, Vol. 114, No. 25, 2010 6899
TABLE 3: Excited States of Dimer Structures in Inverse Centimetersa (2PY)2
symmetry ∆E (BHandH-LYP) ∆E (CC2) ∆E (exptl)b osc. strength (BHandH-LYP)
2HP · 2PY
(2HP)2
S1
S2
S1
S2
S1
Cs 35 974 31 586 30 776 0.09
C2h 38 913 34 789
Cs 40 130 35 165
C2h 42 503 37 739
0.24
0.08
0.20
Cs 35 900 31 441 30 656 0.11
a ∆E is the adiabatic excitation energy from S0, including the change in zero-point vibration energy. b Ref 16 peak at 30 776 assigned to S2.
Figure 5. S1 structure of (2PY)2, the A minimum of the S1 surface.
Dimer Structures. We found S0, S1, and S2 states minima in (2PY)2 and (2HP)2 and S0 and S1 state minima in 2PY · 2HP. This is in line with observed15,16 fluorescence in (2PY)2 and 2PY · 2HP. Table 3 assigns the S1 and S2 states of (2PY)2 and (2HP)2 and the S1 state of 2PY · 2HP. Structural data are presented in Table 4. In the following, the (2PY)2 and (2HP)2 S0 minima will, respectively, be denoted X and Y and the S1 minima A and C. The S0 and S2 states of (2PY)2 have C2h symmetry, with 11Ag as S0 and 11Bu as S2. However, the S1 state has Cs symmetry with one monomer in its S1 structure and the other in its S0 structure. (See Figure 5.) The S1 to S0 transition is a HOMO-LUMO transition, and the orbital plots (Figure 6) show that the excitation is localized on one monomer. The S1 to S0 transition has a finite oscillator strength of 0.09 (BHandH-LYP), roughly of the magnitude of the 2PY monomer S1 to S0 transition. In the excited monomer, the HOMO-LUMO transition removes electron density from the oxygen atom. This elongates one of the two hydrogen bonds, consistent with the observed symmetry breaking in the S1 structure. For (2HP)2, the ground-state energy is similar to that of (2PY)2, whereas the excitation energies are much higher, consistent with the monomer results. As in the case of (2PY)2, the (2HP)2 S1 state is Cs and is not dark, and the computed S1 structure and molecular orbitals plots (Figure 7) are consistent with an exciplex with one monomer in its S0 state and the other in its S1 state. The S1 to S0 transition in the 2PY · 2HP mixed dimer is a HOMO-LUMO transition localized on the 2PY monomer (Figure 8), in agreement with the conclusions of ref 15. The computed 2PY · 2HP S1 excitation energy is ∼150 cm-1 below the computed (2PY)2 S1 excitation energy, in reasonable agreement with the 121 cm-1 found experimentally by Mu¨ller et al.14,15 In Table 5, we compare the computed rotational parameters of the monodeuterated dimer with those found from rotationally resolved fluorescence excitation spectra by Held and Pratt.13 The computed change in rotational constants upon excitation is mostly consistent with the data found by Held and Pratt. Whereas in their case B changed by -7.6 MHz and C changes by -6.4 MHz, our computed B changes -8 MHz and C changes
-6 MHz. The A parameter increases by 9.1 MHz, whereas the experimental value is almost unchanged by excitation. The intermolecular S1 normal modes are shown in Table 6 for the undeuterated case and the two variants of the singledeuterated case. Also displayed are the normal modes of the undeuterated S2 state. As can be observed, the difference in vibrational frequencies is very small between the three. Not surprisingly, the biggest variation was found in the stretching mode. However, this variation was still
