Excited-State Structure and Dynamics of KetoâAmino Cytosine: The

Article pubs.acs.org/JPCB

Excited-State Structure and Dynamics of Keto−Amino Cytosine: The 1 ππ* State Is Nonplanar and Its Radiationless Decay Is Not Ultrafast Simon Lobsiger, Maria A. Trachsel, Hans-Martin Frey, and Samuel Leutwyler* Department of Chemistry and Biochemistry, University of Bern, Freiestrasse 3, CH-3012 Bern, Switzerland S Supporting Information *

ABSTRACT: We have measured the mass- and tautomerspecific S0 → S1 vibronic spectra and S1 state lifetimes of the keto−amino tautomer of cytosine cooled in supersonic jets, using two-color resonant two-photon ionization (R2PI) spectroscopy at 0.05 cm−1 resolution. The rotational contours of the 000 band and nine vibronic bands up to +437 cm−1 are polarized in the pyrimidinone plane, proving that the electronic excitation is to a 1ππ* state. All vibronic excitations up to +437 cm−1 are overtone and combination bands of the low-frequency out-of-plane ν1′ (butterfly), ν2′ (boat), and ν3′ (H−N−C6−H twist) vibrations. UV vibronic spectrum simulations based on approximate second-order coupled-cluster (CC2) calculations of the ground and 1ππ* states are in good agreement with the experimental R2PI spectrum, but only if the calculated ν1′ and ν2′ frequencies are reduced by a factor of 4 and anharmonicity is included. Together with the high intensity of the ν1′ and ν2′ overtone vibronic excitations, this implies that the 1ππ* potential energy surface is much softer and much more anharmonic in the out-of-plane directions than predicted by the CC2 method. The 1ππ* state lifetime is determined from the Lorentzian line broadening necessary to reproduce the rotational band contours: at the 000 band it is τ = 44 ps, remains at τ = 35−45 ps up to +205 cm−1, and decreases to 25−30 ps up to +437 cm−1. These lifetimes are 20−40 times longer than the 0.5−1.5 ps lifetimes previously measured with femtosecond pump−probe techniques at higher vibrational energies (1500−3800 cm−1). Thus, the nonradiative relaxation rate of keto−amino cytosine close to the 1ππ* state minimum is knr ∼ 2.5 × 1010 s−1, much smaller than at higher energies. An additional nonradiative decay channel opens at +500 cm−1 excess energy. Since high overtone bands of ν′1 and ν2′ are observed in the R2PI spectrum but only a single weak 2ν3′ band, we propose that ν3′ is a promoting mode for nonradiative decay, consistent with the observation that the ν3′ normal-mode eigenvector points toward the “C6-puckered” conical intersection geometry.

1. INTRODUCTION Cytosine (Cyt) is a canonical pyrimidine nucleobase that pairs with guanine within double-stranded DNA; its derivatives play an important role in epigenetics and in medicine. 1−9 Methylation of Cyt at the 5-position is one of the central gene regulation mechanisms and has a large impact on the epigenome.5−7 Regarding the photodynamics of isolated cytosine in aqueous solution and in the gas phase, excitedstate relaxation to the ground state has been experimentally observed on the picosecond or even subpicosecond time scale.10−15 This rapid nonradiative return to the ground state makes cytosine highly stable against UV irradiation, and the stability to UV photolysis has been taken to imply photochemical selection of cytosine as one of the molecular building blocks of life.16−19 In this context it is important to investigate the optical transitions, excited-state vibronic lifetimes, couplings between the excited states, and the rates of radiationless relaxation of keto−amino cytosine, which is the biologically relevant tautomer. De Vries and co-workers have measured the resonant two-photon ionization (R2PI) spectra of cytosine, 5-methylcytosine, and 1-methylcytosine.20−22 On the basis of the IR− © XXXX American Chemical Society

UV double-resonance spectrum of cytosine and supported by the close similarity of the R2PI spectra of cytosine and 1methylcytosine, which in the gas phase mainly exists as the keto−amino tautomer, they assigned the R2PI spectrum of cytosine near 32 000 cm−1 in the UV to the keto−amino tautomer. They observed five sharp vibronic bands up to ∼250 cm−1 above the origin.20−22 However, the nature of this excited state has so far not been experimentally identified, nor have the vibronic bands been analyzed. We have measured the rotational contours of the vibronic bands in the two-color resonant two-photon ionization (2CR2PI) spectrum of Cyt up to 500 cm−1 above the 000 band. These give detailed information on (i) the polarization of the electronic transition, which identifies the optically excited state, (ii) couplingsor the absence thereofbetween different excited states, and (iii) the radiationless deactivation rates of the excited vibronic levels as a function of excess vibrational energy. We compare the latter to the relaxation rates measured Received: February 22, 2013 Revised: April 24, 2013

A

dx.doi.org/10.1021/jp401881b | J. Phys. Chem. B XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry B

Article

importance.16,31−40 The minimum of the lowest optically accessible 1ππ* state of 1 has been shown to be connected to three different conical intersections (CIs) with the S0 surface; all three involve some degree of out-of-plane deformation of the pyrimidine ring. One CI has a semiplanar structure with sp3 hybridization of the C6 atom, shortening of the C2−N3 bond, and stretching of the C5C6 bond relative to the ground-state minimum.16,19,31−37,40,41 The second CI involves a N3 out-ofplane distortion, giving a “sofa” conformation and a large outof-plane amino deformation.32,37,40,41 The third conical intersection is characterized by a puckering of the C6 atom and a twist around the C5C6 double bond, with the H−C5− C6−H torsional angle of ∼120°.16,19,31−34,36,37,40,41 Three-state conical intersections have also been located and characterized.38,39,42 Tomic et al. have calculated the vertical and adiabatic absorption energies of the keto, enol, and keto−imino tautomers of cytosine.31 They calculated the DFT/MRCI energies along the TD-DFT reaction path that connect the excited-state 1ππ* minimum and this conical intersection and found that the 1ππ* minimum is separated by a 1600 cm−1 (0.2 eV) barrier from the CI.31 Femtosecond pump−probe time-resolved ionization and photoelectron spectroscopic measurements have been performed on thermally vaporized and jet-cooled cytosine.11−15 These show mono- or biexponential decay profiles, depending on the excitation wavelength (250 or 267 nm), experimental time resolution, and the fit procedure.11−13 Ullrich et al. then noted that both keto and enol tautomers of cytosine must contribute to the signals observed.12 Kosma et al. attempted to distinguish the decay of the biologically relevant keto−amino tautomer 1 from those of the enol−amino (2a, 2b) and keto− imino (3a, 3b) tautomers by exciting at 280−290 nm, where only 1 absorbs UV light.14 At 280 nm they fitted lifetimes of τ2 = 1.2 ps and τ3 = 55 ps; at 290 nm the lifetimes were τ2 = 1.1 ps and τ3 ≥ 150 ps. Recently, Ho et al. have performed fs pump− probe ionization measurements of Cyt at wavelengths up to 300 nm, employing low pump energies and ionization energies and checking for true one-photon excitation.15 Here we present the first detailed vibronic analysis of the low-energy part of the R2PI spectrum of keto−amino cytosine 1 and show that the 10 lowest bands up to 000 + 437 cm−1 can be systematically assigned to overtones and combinations of the three low-frequency out-of-plane vibrations. Analysis of the rotational band contours of the same vibronic bands reveals the orientation of the vibronic transition dipole moments in the molecule-fixed frame. All the rotational contours are broader than would be expected based on our laser resolution, which allows us to derive excited-state vibronic lifetimes via the Lorentzian line shape contribution to the bandwidths. The resulting lifetimes are 20−40 times longer than those measured

by femtosecond (fs) laser pump−probe techniques at higher vibrational excess energies.11−15 We are able to identify the optically active vibrations by comparison to correlated ab initio ground- and excited-state calculations and modeling of the resulting vibronic Franck−Condon envelopes. This allows us to understand the role of these vibrations in promoting nonradiative relaxation of cytosine and its derivatives.23 The relative stabilities of the gas-phase tautomers of cytosine have been extensively investigated by theoretical calculations ranging from density functional (DFT) to highly correlated ab initio methods.24−31 Figure 1 shows the structures of the six

Figure 1. The six most stable tautomers of gas-phase cytosine. The tautomer numbering follows that of ref 24. The atom numbering is indicated for tautomer 1.

most stable cytosine tautomers. The highest-level correlated calculations [coupled cluster with singles, doubles, and iterated triple excitations CCSD(T)] consistently predict the trans enol−amino tautomer 2b to be the most stable, followed by the cis tautomer 2a and the keto−amino tautomer 1.24−30 This implies that the keto−amino tautomer investigated in this work is a minor species in the gas phase. Many computational studies of the excited-state dynamics and decay mechanisms of Cyt have been reported; most have focused on the keto−amino tautomer 1 because of its biological

Table 1. Calculated Relative Energies of Cytosine Tautomers (in kcal/mol) tautomer

B3LYP TZVP

SCS-MP2 aug-cc-pVTZ

QCISD(T) TZVP(2df, 2pd)a

CCSD(T) CBS extrapolatedb

CCSD(T) aug-cc-pVTZc

CCSD(T) cc-pVQZd

1 2a 2b 3b 3a 4

0.00 2.09 1.30 3.25 1.45 7.06

0.45 0.72 0.00 2.41 0.73 7.88

1.3

1.44 0.69 0.00 3.68 2.11

0.93 0.67 0.00

0.76 0.67 0.00

2.00

0.69

a

0.0 1.6

Piacenza and Grimme, ref 28. bTrygubenko et al., ref 27. cWolken et al., ref 30. dFogarasi, ref 24. B



Article

Table 2. Time-Dependent B3LYP and RI-CC2 Calculated Adiabatic and Vertical Transition Energies and Oscillator Strengths for Keto−Amino Cytosine 1 B3LYP/TZVP state (transition) 1

nπ* 1 ππ* 3 ππ* 3 ππ* 3 nπ* a

RI-CC2/aug-cc-pVTZ

adiab./eV

vert./eV

osc. strength

3.74 3.94 3.18

4.81 4.70 3.47 4.15 4.52

0.0009 0.0398

a

adiab./eV

vert./eV

osc. strengtha

exptl

3.26 3.77

4.83 4.61 3.83 4.67 4.76

0.002 0.052

3.95

Vertical excitation from S0 equilibrium geometry, length.

at 1500−4000 cm−1 higher excess energies with the fs pump− probe techniques. We complement the measurements by CC2 and TD-B3LYP calculations of the lowest excited singlet states.

ionizing photons,48 so the ionization wavelength was chosen near the short-wavelength end of the UV-OPO range (220− 230 nm). One-color resonant two-photon ionization by the OPO alone was found to be unavoidable, but is minimal around 226 nm. This gives rise to the small noise background in the two-color spectra. For the UV/UV hole-burning experiments a second independently tunable frequency-doubled dye laser with ∼1 mJ pulse energy was used for UV depletion; it was triggered 600 ns before the excitation and ionization laser pulses. Two 2C-R2PI spectra were measured successively, first without and then with previous depletion of the ground state. For the rotational contour measurements the energy of the excitation laser was reduced to ∼100 μJ. The bandwidth in the visible is ∼0.038 cm−1, as measured with a HighFinesse Ångstrom WS6 high-precision wavemeter. The UV bandwidth after frequency doubling is expected to be √2 wider (0.054 cm−1, 1610 MHz). The spectra were recorded at 0.012 cm−1 step size and calibrated by measuring the fundamental frequency with the WS6 wavemeter.

2. COMPUTATIONAL METHODS AND RESULTS The electronic ground states of the six most stable cytosine tautomers were optimized with the spin-component scale (SCS) MP2 method and the aug-cc-pVTZ basis set. Groundand excited-state calculations were also performed with the second-order approximate coupled-cluster (CC2) method and the aug-cc-pVTZ basis set and compared to the results of timedependent B3LYP density functional calculations with the TZVP basis set. Both CC2 and TD-B3LYP reproduce the excitation energies for the lowest 1ππ* and 1nπ* states calculated with high-level methods such as the equation-ofmotion excitation energy coupled cluster method EOMEECCSD(T).43 TD-B3LYP and CC2 excited-state harmonic frequencies were also calculated, the latter with the aug-ccpVDZ basis set. The calculations were performed using Turbomole 6.3.44,45 Table 1 summarizes the relative energies of the cytosine tautomers at different levels of theory. All the correlated calculations predict the trans enol−amino 2b tautomer to be the most stable, with the corresponding cis rotamer 2a about 0.6−0.7 kcal/mol higher. The keto−amino tautomer 1 investigated in this work is predicted to be the third most stable tautomer. Table 2 shows the calculated adiabatic transition energies for the keto−amino tautomer. While the lowest 1ππ* transition of the keto−amino tautomer 4 is predicted to be slightly higher than that of 1, tautomer 4 is not populated in the supersonic jet due to its high relative energy of +7.9 kcal/mol. The predicted and observed electronic transitions of the enol− and keto−imino tautomers are shifted to the blue, outside of the spectral range considered here.

4. RESULTS 4.1. Resonant Two-Photon Ionization Spectra. Figure 2a shows the two-color resonant two-photon ionization spectrum of jet-cooled Cyt in the frequency range of 31 200−32 400 cm−1 when ionizing at 226 nm. The spectrum shows a sharp 000 band at 31 835 cm−1 (3.947 eV) and five medium to strong vibronic bands up to +205 cm−1. Between

3. EXPERIMENTAL METHODS Our experimental setup has been described previously.23,46−48 Briefly, neon carrier gas (Linde, ≥99.995%) at ∼1.8 bar backing pressure was passed through a pulsed nozzle (0.4 mm diameter) containing cytosine (Sigma, >99% purity) heated to 235 °C. Two-color resonant two-photon ionization spectra were measured by crossing the skimmed supersonic jet with the unfocused UV excitation and ionization laser beams in the source of a linear time-of-flight mass spectrometer (TOF-MS). Excitation was performed with 200 μJ UV pulses from a frequency-doubled Radiant Dyes NarrowScan D-R dye laser. Ionization pulses of ∼150−200 μJ were produced by an Ekspla NT342B ultraviolet optical parametric oscillator (UV-OPO). The ionization efficiency for keto−amino cytosine excited to its 1 ππ* origin increases steeply with increasing energy of the

Figure 2. (a) Two-color resonant two-photon ionization spectrum (with ionization at 226 nm) and (b) UV/UV hole-burning spectrum of supersonic jet-cooled keto−amino cytosine. The UV hole-burning laser was fixed at the intense band at 31 927 cm−1, marked by *. The wavenumber scale is relative to the 000 band at 31 835.2 cm−1. C



Article

+277 and +530 cm−1 there are about 10 very weak bands. The spectrum is similar to that reported by de Vries and coworkers.20−22 The lower trace part b in Figure 2 corresponds to the UV/UV hole-burning spectrum with the burn laser tuned to the most intense band at 31 927 cm−1, marked with an asterisk in Figure 2a. The intense and the weaker bands from 300 to 500 cm−1 are reproduced in the hole-burning spectrum, indicating excitation from the same ground state. The two weak bands appearing to the red of the 000 band are not observed in the UV/UV hole-burning spectrum, in agreement with the observation of Nir et al.20 However, given the lower signal to noise ratio (S/N) of our hole-burning spectra, a definite statement regarding these bands cannot be made. Below, we show that these bands can be assigned as hot bands of the same electronic transition. 4.2. Vibronic Band Assignments. We first attempted to assign the vibronic bands in the R2PI spectrum of cytosine in Figure 2 based on the CC2 and TD-B3LYP harmonic frequencies of the 1ππ* state given in Table 3. Experience with excited-state calculations of pyrimidinones shows that, while the in-plane vibrational frequencies are well-reproduced by CC2 and TD-B3LYP normal-mode calculations, this is not the case for the out-of-plane vibrations.23,46,47,49 The lowestfrequency in-plane vibration is predicted to lie at 527 cm−1 (B3LYP) or 486 cm−1 (CC2); thus, the vibronic bands below ∼450 cm−1 must arise from out-of-plane vibrational modes. Assuming a planar pyrimidinone ring in the ground state, only Δν = 2, 4, ..., overtone excitations in the out-of-plane vibrations are expected. The lowest two bands at 71 and 91 cm−1 cannot be part of one progression, and we assign them as 120 and 220, i.e., the overtones of the low-frequency out-of-plane vibrations ν1′ and ν2′ . This implies that the respective fundamentals lie at 34 and 39 cm−1, much lower than the CC2 and TD-B3LYP calculated harmonic values. Given this large difference between computation and experiment, we employed the vibronic band simulation module of the PGOPHER program.50 This module is based on a harmonic model for vibrations; anharmonic constants can be added for each mode. As inputs, the module uses the calculated geometries and the normal-mode l matrices of the S0 and excited states. The band intensities are based on full multidimensional Franck−Condon factors including both mode displacements and mixing between modes (Dushinsky effect). The simulated spectrum using the CC2 geometries and normal modes is shown in the lower half of Figure 3. We needed to decrease the ν′1 and ν′2 fundamental frequencies by about a factor of 4 in order to fit the experimentally observed 120 and 220 band frequencies. The band assignments of the ν1′ and ν2′ progressions are shown above Figure 3a, the assignments of weaker combination bands are shown in Figure 3b. The ν′1, ν′2, and ν′3 normal-mode eigenvectors are shown in Figure 4. The simulation predicts the ν′ = 4 and 6 overtones of ν1′ at 177 and 308 cm−1 and the ν′ = 4 and 6 overtones of ν2′ at 205 and 344 cm−1, all in excellent agreement with the observed frequencies. The band at 161 cm−1 is predicted by the simulation to be the overtone combination 2ν1′ + 2ν2′ . The lowest-energy band that could not be assigned as an overtone or combination of ν′1 and/or ν′2 is a weak band at 437 cm−1, see Figure 3, parts a and b. We assign it as the first overtone of the ν3′ out-of-plane vibration. The overtone combination band 2ν1′ + 2ν3′ is also observed at 530 cm−1. However, the CC2 calculation predicts the lowest-

Table 3. CC2/aug-cc-pVDZ and TD-B3LYP/TZVP Calculated 1ππ* Harmonic Vibrational Frequencies (in cm−1) of Keto−Amino Cytosine 1 mode

irrep.

ν1 ν2

a′′ a′′

ν3

a′′

ν4 ν5 ν6 ν7 ν8 ν9 ν10

a′′ a′′ a′′ a′′ a′ a′ a′′

ν11 ν12 ν13 ν14

a′ a′′ a′′ a′

ν15 ν16

a′′ a′

ν17

a′

ν18

a′

ν19

a′

ν20

a′

ν21

a′

ν22

a′

ν23

a′

ν24

a′

ν25

a′

ν26

a′

ν27 ν28 ν29 ν30 ν31 ν32 ν33

a′ a′ a′ a′ a′ a′ a′

a

description

b

butterfly γCO/γC5/γC5H/ γC6H γC6H/γN1H/ γC6/γN7Hb γasNH2/δC2O γasNH2 γC6H γN1H 3 6b/γN1H γC5H/γC4/ γsNH2 6a NH2 inversion 4 δN1C6C5/ δC2N3C4/ νC2O γC5H νN3C4/νC2O/ δN1H/δC5H δC6H/δNH/ δC5H βasNH2/δC6H/ δC5H βasNH2/δC5H/ δC6H δNH/δN1C2N3/ δC5H/δC6H δNH/δC6H/ δC5H δC5H/δC6H/ νC4C7 δC6H/δC5H/ νN3C4/βasNH2 δNH/δC5H/ νC6N1/δN7Hb δC6H/δN1H/ νC5C6 βsNH2/νC4C5/ νC2N3 βsNH2/νC5C6 βsNH2/νC4C5 νC5H νC6H νsNH2 νNH νasNH2

RI-CC2/ aug-ccpVDZ

B3LYP/ TZVP

exptl fundamentalsc

135.1 178.6

142.2 222.3

(34) (39)

221.1

189.2

(218)

299.6 327.4 345.0 483.8 485.9 497.1 524.5

327.0 415.5 364.8 512.2 530.6 527.5 612.1

544.6 629.5 678.8 724.2

570.4 420.3 748.2 778.8

815.0 834.0

822.8 916.7

952.8

1018.1

976.4

1006.8

1110.2

1122.3

1024.2

1163.6

1159.9

1218.1

1281.8

1333.1

1320.6

1340.7

1382.8

1438.2

1431.8

1504.2

1581.5

1623.6

1594.3 1640.8 3205.1 3266.2 3509.5 3593.5 3640.7

1549.1 1663.6 3193.1 3284.2 3575.8 3643.9 3699.2

530

a

Irreducible representations in the Cs point group. bBased on CC2/ TZVP calculation; ν = stretching vibration; δ = in-plane deformation vibration; γ = out-of-plane deformation vibration; βs = scissoring vibration; βas = in-plane rocking vibration; γs = out-of-plane wagging vibration; γas = torsion vibration. cValues in parentheses are half the observed ν′ = 2 overtone frequency.

frequency in-plane excitation 1110 at 545 cm−1, within 20 cm−1 of this band. The in-plane mode ν11 ′ in cytosine corresponds to the ν6a ′ mode in benzene, and since this is one of the most intense vibrations in the vibronic spectra of benzene derivatives, D



Article

Figure 3. (a) Two-color resonant two-photon ionization spectrum of jet-cooled cytosine including vibronic band assignments of the ν1 and ν2 overtone bands. (b) Simulated vibronic spectrum (with PGOPHER, plotted in the negative direction) based on the CC2/aug-cc-pVDZ ground and 1 ππ* calculations. Additional assignments of hot bands, sequence bands, and weak combination bands are indicated.

Table 4. Experimental 1ππ* State Vibrational Frequencies (in cm−1) of Keto−Amino Cytosine 1, cf. Figure 3 assignment

frequencya

111 131 000 120 220 120 140 140 120 160 260 280 320

−145 −59 (31835.2) 71 91 161 177 205 277 308 344 430 437 530

+ 220

+ 240

1110/220 + 320

In vacuum; these values are shifted by +10 cm−1 relative to those of refs 20−22. a

Figure 4. CC2 and TD-B3LYP calculated normal-mode eigenvectors of the three lowest-frequency out-of-plane vibrations that dominate the S1 → S0 spectrum of keto−amino cytosine.

Figure 5. The rotational contours of the lowest 10 vibronic bands in the R2PI spectrum up to +437 cm−1 were measured at 0.054 cm−1 resolution and are shown in Figures 6−8. The 000 contour in Figure 6 shows a double-wing shape that is characteristic for a/b-hybrid in-plane transitions and lacks a central peak that would indicate c-type polarization.51 This implies that the electronic transition dipole moment (TDM) lies in the pyrimidinone plane corresponding to a ππ* transition. The rotational contour of the 000 band was simulated with a rigid-rotor asymmetric top Hamiltonian using the PGOPHER program.50 Table 5 summarizes the CC2 and B3LYP calculated rotational constants. The B3LYP S0 state rotational constants

the 530 cm−1 band might also be assigned as 1110, see Table 4. A list of the vibronic assignments is given in Table 4. We note that the calculated harmonic ν′1 and ν′2 frequencies in Table 3 differ from the observed frequencies (Table 4) by a factor of 4. This indicates that the corresponding excited-state potential-energy surface is much flatter and more anharmonic along these coordinates than predicted by the excited-state calculations. 4.3. Rotational Band Contours. Cytosine is an asymmetric top with its c inertial axis perpendicular to the pyrimidinone plane and its a/b axes within this plane, see E



Article

Table 5. S0 and 1ππ* State CC2 and (TD-)B3LYP Calculated Rotational Constants of Keto−Amino Cytosine, Transition Dipole Moment Components, and Comparison to Rotational Contour Simulation Values for the 000 Band

Figure 5. B3LYP calculated (blue) and experimental (full and dashed line) electronic transition dipole moment vectors for the 1ππ* transition of cytosine. The calculated S0 state inertial a and b axes are drawn in red. A sign ambiguity exists for the experimental TDM direction, see the text; the preferred orientation is drawn as a full-line arrow.

parameters

CC2/aug-cc-pVTZ

B3LYP/TZVP

A′′ (MHz) B′′ (MHz) C′′ (MHz) A′−A′′ (MHz) B′−B′′ (MHz) C′−C′′ (MHz) |μa|2:|μb|2:|μc|21 ΔLor (MHz) ΔGauss (MHz) Trot (K)

3865 2012 1324 −105 −33 −26

3870 2023 1329 −63 −14 −13 16:84:0

000 band

−96 (80)

5(6):95(9):0(5) 3600 (500) 1600 2.0 (3)

A′′, B′′, C′′ were used as fixed inputs. The TD-B3LYP 1ππ* state rotational constants A′, B′, C′ and the 1ππ* TDM vector components μa, μb, and μc were used as starting values for the simulations. The calculated TDM vector is plotted in blue in Figure 5. First the relative TDM magnitudes were fitted, followed by the 1ππ* state A′ rotational constant. The fitted B′ and C′ values differed insignificantly from the TD-B3LYP starting values; hence, the latter were retained for the simulations. The ground-state rotational temperature Trot was optimized for best agreement in the wings of the contour, which are most sensitive to the high-J contributions. We found optimum agreement with experiment for Trot = 2.0 ± 0.3 K. This is in good agreement with the Trot = 3−5 K found for other pyrimidines and purines in our supersonic jets.46,47,51 After the first round of fitting, all the parameters listed in Table 5 were optimized. The uncertainty was estimated by varying each parameter independently until the simulated contour changed significantly. The sign of the angle between the TDM vector and the a/b axes cannot be experimentally determined, giving the two possible TDM orientations shown as full-line and dashed black double-headed arrows in Figure 5. We prefer the full-line orientation, because (a) it lies close to the TDM for the closely related 5-methylcytosine23 and (b) it is closer to the TDB3LYP calculated TDM given by the blue arrow in Figure 5. The calculated TDM has a |μa|2:|μb|2 ratio of 15:85. The contour simulation yields |μa|2:|μb|2 = 5:95, with an estimated error of ±10%. Each rovibronic line is convoluted with a Lorentzian profile that reflects the excited-state lifetime τ of that vibronic level, resulting in a Lorentz line width contribution ΔLor = (2πτ)−1 (in Hz). Upon convolution with the Gaussian laser line profile of 1610 MHz, the resulting rovibronic line shape becomes a Voigt profile. Figure 6 compares the observed 000 band contour to simulated contours with Lorentz line width contributions that successively increase from 900 to 53 000 MHz. The best agreement of simulation with experiment is observed for Figure 6c, giving a lifetime of τ = 44 (±5) ps at the 000 band. Simulations with longer lifetimes, Figure 6, parts a and b, predict a central dip in the contour that is deeper than observed, while those with shorter lifetimes, Figure 6d−f, predict central dips that are clearly too small. This lifetime of 44 ps is about 40 times longer than those measured by fs pump− probe and fs time-resolved photoelectron experiments at high vibronic excess energies,11−15 see the next section.

Figure 6. Calculated asymmetric rotor band contours (in black) of the 000 band of jet-cooled cytosine, as a function of the Lorentz (lifetime) broadening ΔLor (in MHz, left-hand corner), equivalent to the excitedstate lifetime τ (in ps, right-hand corner). The experimental two-color R2PI contour (at 0.054 cm−1 laser resolution) is plotted in gray. The other fit parameters are listed in Table 5. The optimum fit to the contour is at τ = 44 ± 5 ps.

F



Article

Figure 7 shows the contours of the 000 and the first five vibronic bands 120, 220, 120 + 220, 140, and 240. All exhibit a/b-type

Figure 8. Experimental rotational contours (in gray) of the vibronic bands of keto−amino cytosine in the range of +308 to +437 cm−1. Superimposed in black are the fitted asymmetric rotor contours for an in-plane ππ* transition. The lifetimes τ derived from the fitted Lorentzian widths ΔLor (MHz) are given for each band.

about 40% of the 000 band. In cytosine the 1110 band has a FCF of only 3%, due to a reduced lifetime, which we estimate as τ = 3−6 ps; this value is compatible with the Lorentzian width measured in the R2PI spectrum of Figure 3. The disappearance of the R2PI signal above 530 cm−1 implies that the vibronic lifetimes decreases to ∼1 ps. Figure 7. Experimental rotational contours (in gray) of the 000 and first five vibronic bands of jet-cooled keto−amino cytosine, cf. Figure 6. Superimposed in black are the asymmetric rotor contours for an inplane 1ππ* transition. The lifetimes τ derived from the fitted Lorentzian widths ΔLor (MHz) are given for each band.

5. DISCUSSION The excited-state lifetimes of jet-cooled cytosine have been measured by fs pump−probe ionization and threshold photoelectron analysis measurements.11−15 The earlier experiments11−13 were performed by exciting at 250 and 267 nm in a spectral region where at least four different tautomers of cytosine absorb UV radiation.14,31 These decay profiles were not tautomer-specific, and the multiple lifetimes observed were partially due to the different tautomers.14 Kosma et al. pointed out this problem and undertook fs pump−probe measurements at longer wavelengths (267−290 nm).14 According to all excited-state calculations, only the keto−amino tautomers 1 and 4 absorb beyond λ = 290 nm.14 Since tautomer 4 is calculated to lie at least 7 kcal/mol above tautomer 1, see Table 1, the measurements in ref 14 characterize the decay of the keto−amino tautomer 1. More recently, Ho et al. have extended fs lifetime measurements of keto−amino cytosine out to 310 nm.15 They found that the long τ3 = 15−25 ps decays reported by Kosma et al.14 could only be observed when using high pulse energy (∼100 μJ/pulse) for ionization; since the long lifetimes were more pronounced in the fragment ion channels, they concluded that the τ3 lifetimes contributions arise from dissociation of cytosine in its lowest ionic state.15 Ho et al. also concluded that the τ1 < 1 ps decay component observed in earlier work11−14 were due to nonresonant

contours, implying that the respective vibronic TDMs also lie in the pyrimidinone plane and that the respective excited-state levels are totally symmetric (a′ in Cs). Therefore, the vibronic excitations must be even overtones with ν′i = 2, 4, ..., or combination bands with νi′ + νj′ = 2, 4, ..., in agreement with the band assignments given above. The simulations were calculated with the same ground-state rotational temperature and TDM orientation as for the 000 band. The fitted lifetimes lie between τ = 33 ps and τ = 44 ps. The contours of four weaker vibronic bands at 308, 328, 430, and 435/437 cm−1 vibrational excess energy are shown in Figure 8. Since these transitions are 5−20 times weaker than the bands in Figure 7, the S/N ratio is smaller and the quality of the fits is worse. However, we can specify a lower lifetime limit of τ > 25 ps. The highest-energy band at +530 cm−1 probably corresponds to excitation of the 6a vibration, which is denoted ν11 in keto− amino cytosine. In the R2PI spectrum of 5-methylcytosine,23 the corresponding band has a Franck−Condon factor (FCF) of G



Article

intersection and have found a 1600 cm−1 (0.2 eV) barrier along this path.31 Since the geometry of the CI is nonplanar and if we assume that a superposition of displacements in the lowest eight out-of-plane vibrational modes ν1′ to ν8′ is necessary to cross the barrier, we need to add the zero-point vibrational energy (ZPVE) of these eight modes to the energy of the minimum. Combining the experimental ν′1 to ν′3 frequencies and the CC2 calculated frequencies for ν4 to ν8 gives an estimated ZPVE of 1200 cm−1, leaving an estimated 400 cm−1 to the top of the barrier. This value is perfectly compatible with the observed increase of the nonradiative relaxation rate by 2 orders of magnitude around ∼ +500 cm−1 vibrational excess energy. Since the R2PI spectrum exhibits high overtone bands of ν′1 and ν2′ with lifetimes τ = 40−45 ps, but only a single weak 2ν3′ band with a shorter lifetime of τ = 25 ps, we speculate that, while all three vibrations are promoting modes for S1 → S0 nonradiative decay, the ν′3 mode is the strongest promoter. This is consistent with the CC2 and TD-B3LYP calculated ν3′ normal-mode eigenvector (shown in Figure 4), which is dominantly an H−N1−C6−H twist out of the pyrimidinone plane. Of the ν′1, ν′2, and ν′3 vibrational eigenvectors, the latter points most directly toward the “C6-puckered” conical intersection geometry, which has a twisted C5−C6 bond, a puckered C6 atom, and a wave function with dominantly ππ* character.16,19,31−34,36,37,40,41

multiphoton ionization around time zero, when the pump and probe pulses temporally overlap.15 These results have led to the general perception that the excited-state relaxation of gas-phase keto−amino cytosine is ultrafast with τ ∼ 1 ps.16,31,33−40 Our measurements show that this is not the case for excitation to the 000 band and levels up to about 500 cm−1 above, in other words for near-adiabatic excitation to the vicinity of the 1ππ* state minimum. Figure 9

6. CONCLUSIONS We have measured the UV vibronic spectra of jet-cooled keto− amino cytosine (Cyt) using two-color resonant two-photon ionization spectroscopy. The orientation of the electronic TDM is determined from the 000 rotational band contour, measured at 0.05 cm−1 resolution. The TDM lies completely in the pyrimidinone plane, implying that the transition is 1ππ*. The experimental TDM ratio along the in-plane a/b axes is |μa|2:|μb|2 = 5%:95%. The TDM lies along the N3−C6 direction, within a few degrees of the calculated 1ππ* TDM direction. The low-lying part of the 1ππ* spectrum20−22 is dominated by low-frequency out-of-plane vibronic excitations, similar to the 1ππ* spectrum of the closely related 5-methylcytosine.23 All bands up to 000 + 437 cm−1 are assigned to overtones of the lowfrequency out-of-plane butterfly (ν1′ ), boat (ν2′ ), and HN1−C6H twist (ν′3) vibrations. The high intensity of the 120 and 220 bands indicates that the 1 ππ* excitation induces a dramatic change of the potential energy surface along the ν′1 and ν′2 coordinates. The respective fundamental frequencies are ν1′ = 34 cm−1 and ν2′ = 39 cm−1, which is 4−5 times lower than the harmonic 1ππ* state frequencies predicted by the CC2 and TD-B3LYP calculations. This poor agreement is due to the very flat (possibly doubleminimum) potential energy surface along the ν1′ and ν2′ coordinates, which limits the value of normal-mode frequencies. This concurs with the prediction of Tomic et al. that the 1ππ* excited-state minimum of keto−amino cytosine is very flat and anharmonic.31 The rotational contours of the vibronic bands up to +437 cm−1 exhibit the same shape as the 000 band, implying that the excited-state vibrational levels are totally symmetric (a′ in Cs). The absence of c-type contributions to the rotational contours of these 10 vibronic bands implies that no significant 1ππ* ↔ 1 nπ* mixing occurs in this energy range.

Figure 9. Excited-state vibronic lifetimes of keto−amino cytosine in its 1 ππ* state from the 000 band to excess energies of +4000 cm−1. Lifetimes marked by Δ and □ are from refs 14 and 15, respectively.

plots the experimental lifetimes of keto−amino cytosine versus the vibrational excess energy in the 1ππ* state and compares our values in the 0−530 cm−1 range to those measured at higher energies in refs 14 and 15. The lifetimes of the lowest six levels are in the range of τ = 33−44 ps. Those of the next four levels decrease slightly to τ = 25−30 ps; hence, the respective band intensities in the R2PI spectrum are not small because their lifetimes are short, but because their Franck−Condon factors are small. At an excess energy of ∼530 cm−1, the lifetime abruptly drops from 35 to 45 ps to a few picoseconds. Although there is a gap in the data between 530 and 1500 cm−1, the lifetimes of keto−amino cytosine measured in this work extrapolate nicely toward the τ = 0.5−1.1 ps lifetimes measured by fs pump−probe techniques by Kosma et al.14 and Ho et al.,15 measured at excess energies of 1500−3700 cm−1, showing that the lifetimes measured in three different experiments are mutually consistent. Figure 9 implies first that at zero or small vibrational excess energy the nonradiative relaxation rate of 1ππ* state keto− amino cytosine is knr ∼ 2.5 × 1010 s−1 which is fast, but not ultrafast. Around +500 cm−1 (0.065 eV) excess vibrational energy, knr increases by nearly 2 orders of magnitude. This may correspond to the access to the lowest-lying conical intersection (CI) of keto−amino cytosine. As discussed above, most computational studies agree that the lowest-energy CI is the C6-twist CI. Tomic et al. have calculated the TD-DFT reaction path connecting the excited-state 1ππ* minimum to this conical H



Article

The lifetimes of the low-lying 1ππ* state levels are determined via the Lorentzian broadening of the respective band contours. At the 000 band the lifetime is τ = 44 ± 5 ps; it remains between 30 and 45 ps up to +205 cm−1, decreases to τ ∼ 25 ps for five weaker bands between 308 and 447 cm−1 and further to τ ∼ 25 ps for the weak band at 530 cm−1. No bands are observed above +530 cm−1, putting a lifetime limit of τ ∼ 1−2 ps on the higher vibronic levels. The drop in lifetime above 530 cm−1 is nicely compatible with fs pump−probe lifetimes measured at vibrational excess energies of 1500−3800 cm−1, which give τ = 0.5−1.5 ps.14,15 We speculate that the sharp drop in lifetime above 000 + 530 cm−1 is a result of the molecule being able to access the lowest C6-twist conical intersection, which is driven by the excitation of the out-of-plane vibrations ν1′ (butterfly), ν2′ (boat), and ν3′ (H−N−C6−H twist) that couple to the lowest 1ππ* excitation. Especially the ν′3 normal-mode eigenvector points directly toward this intersection, and ν3′ may be the main promoting mode for internal conversion.

■

(7) Saitou, M.; Kagiwada, S.; Kurimoto, K. Epigenetic Reprogramming in Mouse Pre-Implementation Development and Primordial Germ Cells. Development 2012, 139, 15−31. (8) Pai, M. P.; Bruce, H.; Felton, L. A. Clinical Pharmacokinetics of Oral Controlled-Release 5-Fluorocytosine. Antimicrob. Agents Chemother. 2010, 54, 1237−1241. (9) Johnson, A. J.; Ardiani, A.; Sanchez-Bonilla, M.; Black, M. E. Comparative Analysis of Enzyme and Pathway Engineering Strategies for 5FC-Mediated Suicide Gene Therapy Applications. Cancer Gene Ther. 2011, 18, 533−542. (10) Crespo-Hernández, C. E.; Cohen, B.; Hare, P. M.; Kohler, B. Ultrafast Excited-State Dynamics in Nucleic Acids. Chem. Rev. 2004, 104, 1977−2019. (11) Kang, H.; Lee, K. T.; Jung, B.; Ko, Y. J.; Kim, S. K. Intrinsic Lifetimes of the Excited State of DNA and RNA Bases. J. Am. Chem. Soc. 2002, 124, 12958−12959. (12) Ullrich, S.; Schultz, T.; Zgierski, M. Z.; Stolow, A. Electronic Relaxation Dynamics in DNA and RNA Bases Studied by TimeResolved Photoelectron Spectroscopy. Phys. Chem. Chem. Phys. 2004, 6, 2796−2801. (13) Canuel, C.; Mons, M.; Piuzzi, F.; Tardivel, B.; Dimicoli, I.; Elhanine, M. Excited States Dynamics of DNA and RNA Bases: Characterization of a Stepwise Deactivation Pathway in the Gas Phase. J. Chem. Phys. 2005, 122, 074316−74321. (14) Kosma, K.; Schröter, C.; Samoylova, E.; Hertel, I. V.; Schultz, T. Excited-State Dynamics of Cytosine Tautomers. J. Am. Chem. Soc. 2009, 131, 16939−16943. (15) Ho, J. W.; Yen, H.-C.; Chou, W.-K.; Weng, C.-N.; Cheng, L.-H.; Shi, H.-Q.; Lai, S.-H.; Cheng, P.-Y. Disentangling Intrinsic Ultrafast Excited-State Dynamics of Cytosine Tautomers. J. Phys. Chem. A 2011, 115, 8406−8418. (16) Merchan, M.; Serrano-Andrés, L. Ultrafast Internal Conversion of Excited Cytosine via the Lowest ππ* Electronic Singlet States. J. Am. Chem. Soc. 2003, 125, 8108−8109. (17) Brauer, B.; Gerber, R. B.; Kabelac̆, M.; Hobza, P.; Bakker, J. M.; Riziq, A. G. A.; de Vries, M. S. Vibrational Spectroscopy of the Guanine-Cytosine Base Pair: Experiment, Harmonic and Anharmonic Calculations and the Nature of the Anharmonic Couplings. J. Phys. Chem. A 2005, 109, 6974−6984. (18) Shukla, M. K.; Leszczynski, J. In Radiation Induced Molecular Phenomena in Nucleic Acids; Shukla, M. K., Leszczynski, J., Eds.; Springer: The Netherlands, 2008; pp 1−14. (19) Serrano-Andrés, L.; Merchan, M. Are the Five Natural DNA/ RNA Base Monomers a Good Choice from Natural Selection? J. Photochem. Photobiol., C 2009, 10, 21−32. (20) Nir, E.; Janzen, C.; Imhof, P.; Kleinermanns, K.; de Vries, M. S. Pairing of the Nucleobases Guanine and Cytosine in the Gas Phase Studied by IR−UV Double-Resonance Spectroscopy and Ab Initio Calculations. Phys. Chem. Chem. Phys. 2002, 4, 732−739. (21) Nir, E.; Müller, M.; Grace, L. I.; de Vries, M. S. REMPI Spectroscopy of Cytosine. Chem. Phys. Lett. 2002, 355, 59−64. (22) Nir, E.; Hünig, I.; Kleinermanns, K.; de Vries, M. S. The Nucleobase Cytosine and the Cytosine Dimer Investigated by DoubleResonance Laser Spectroscopy and Ab initio Calculations. Phys. Chem. Chem. Phys. 2003, 5, 4780−4785. (23) Trachsel, M. A.; Lobsiger, S.; Leutwyler, S. Out-of-Plane LowFrequency Vibrations and Nonradiative Decay in the 1ππ* State of JetCooled 5-Methylcytosine. J. Phys. Chem. B 2012, 116, 11081−11091. (24) Fogarasi, G. High-Level Electron Correlation Calculations on some Tautomers of Cytosine. J. Mol. Struct. 1997, 413, 271−278. (25) Kobayashi, R. A CCSD(T) Study of the Relative Stabilities of Cytosine Tautomers. J. Phys. Chem. A 1998, 102, 10813−10817. (26) Fogarasi, G. Relative Stabilities of Three Low-Energy Tautomers of Cytosine: A Coupled Cluster Electron Correlation Study. J. Phys. Chem. A 2002, 106, 1381−1390. (27) Trygubenko, S. A.; Bogdan, T. V.; Rueda, M.; Orozco, M.; Luque, F. J.; Sponer, J.; Slavicek, P.; Hobza, P. Correlated Ab Initio Study of Cytosine and Its Tautomers in the Gas Phase, in a

ASSOCIATED CONTENT

S Supporting Information *

Five tables with the TD-B3LYP and CC2 calculated adiabatic and vertical transition energies of the other cytosine tautomers 2a, 2b, 3a, 3b, and 4 and four tables with the TD-B3LYP and CC2 calculated S0 and S1 state equilibrium coordinates of tautomer 1. 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 We thank Dr. Colin Western for helpful information regarding the use of PGOPHER. Financial support by the Schweiz. Nationalfonds (project no. 200020-121993) is gratefully acknowledged.

■

REFERENCES

(1) Vermes, A.; Guchelaar, H.-J.; Dankert, J. Flucytosine: A Review of its Pharmacology, Clinical Indications, Pharmacokinetics, Toxicity and Drug Interactions. J. Antimicrob. Chemother. 2000, 46, 171−179. (2) Nakamura, H.; Mullen, J. T.; Chandrasekhar, S.; Pawlik, T. M.; Yoon, S. S.; Tanabe, K. K. Multimodality Therapy with a ReplicationConditional Herpes Simplex Virus 1 Mutant that Expresses Yeast Cytosine Deaminase for Intratumoral Conversion of 5-Fluorocytosine to 5-Fluorouracil. Cancer Res. 2001, 61, 5447−5452. (3) Boucher, P. D.; Im, M. M.; Freytag, S. O.; Shewach, D. S. A Novel Mechanism of Synergistic Cytotoxicity with 5-Fluorocytosine and Ganciclovir in Double Suicide Gene Therapy. Cancer Res. 2006, 66, 3230−3237. (4) Lu, Y.-C.; Luo, Y.-P.; Wang, Y.-W.; Tai, C.-K. Highly Efficient Gene Transfer to Solid Tumors in Vivo by Tumor Selective Replicating Retrovirus Vectors. Int. J. Mol. Med. 2010, 25, 769−775. (5) Iqbal, K.; Jin, S.-G.; Pfeifer, G. P.; Szabó, P. E. Reprogramming of the Paternal Genome upon Fertilization Involves Genome-Wide Oxidation of 5-Methylcytosine. Proc. Natl. Acad. Sci. U.S.A. 2011, 108, 3642−3648. (6) Nabel, C. S.; Kohli, R. M. Demystifying DNA Demethylation. Science 2011, 333, 1229−1230. I



Article

for the maximum gradient element, 4 × 10−6 au for the RMS displacement, and 10−6 au for the RMS gradient. (46) Lobsiger, S.; Frey, H. M.; Leutwyler, S.; Morgan, P.; Pratt, D. S0 and S1 State Structure, Methyl Torsional Barrier Heights, and Fast Intersystem Crossing Dynamics of 5-Methyl-2-hydroxypyrimidine. J. Phys. Chem. A 2011, 115, 13281−13290. (47) Lobsiger, S.; Sinha, R. K.; Trachsel, M.; Leutwyler, S. Low-Lying Excited-States and Nonradiative Processes of the Adenine Analogues 7H- and 9H-2-Aminopurine. J. Chem. Phys. 2011, 134, 114307− 114320. (48) Lobsiger, S.; Leutwyler, S. The Adiabatic Ionization Energy and Triplet T1 Energy of Jet-Cooled Keto−Amino Cytosine. J. Phys. Chem. Lett. 2012, 3, 3576−3580. (49) Frey, J. A.; Leist, R.; Tanner, C.; Frey, H. M.; Leutwyler, S. 2Pyridone: The Role of Out-of-Plane Vibrations on the S1 ↔ S0 Spectra and S1 State Reactivity. J. Chem. Phys. 2006, 125, 114308−114321. (50) Western, C. M. PGOPHER 6.0, a Program for Simulating Rotational Structure. http://pgopher.chm.bris.ac.uk. (51) Lobsiger, S.; Frey, H. M.; Leutwyler, S. Supersonic Jet UV Spectrum and Nonradiative Processes of the Thymine Analogue 5Methyl-2-hydroxypyrimidine. Phys. Chem. Chem. Phys. 2010, 12, 5032−5040.

Microhydrated Environment and in Aqueous Solution. Phys. Chem. Chem. Phys. 2002, 4, 4192−4203. (28) Piacenza, M.; Grimme, S. Systematic Quantum Chemical Study of DNA−Base Tautomers. J. Comput. Chem. 2004, 25, 83−98. (29) Yang, Z.; Rodgers, M. T. Theoretical Studies of the Unimolecular and Bimolecular Tautomerization of Cytosine. Phys. Chem. Chem. Phys. 2004, 6, 2749−2757. (30) Wolken, J. K.; Yao, C.; Turec̆ek, F.; Polce, M. J.; Wesdemiotis, C. Cytosine Neutral Molecules and Cation-Radicals in the GasPhaseStructures, Energetics, Ion Chemistry, and Neutralization− Reionization Mass Spectrometry. Int. J. Mass Spectrom. 2007, 267, 30− 42. (31) Tomic, K.; Tatchen, J.; Marian, C. M. Quantum Chemical Investigation of the Electronic Spectra of the Keto, Enol and Keto− Imine Tautomers of Cytosine. J. Phys. Chem. A 2005, 109, 8410−8418. (32) Ismail, N.; Blancafort, L.; Olivucci, M.; Kohler, B.; Robb, M. Ultrafast Decay of Electronically Excited Singlet Cytosine via a ππ* to nOπ* State Switch. J. Am. Chem. Soc. 2002, 124, 6818−6819. (33) Zgierski, M. Z.; Patchkovskii, S.; Fujiwara, T.; Lim, E. C. On the Origin of the Ultrafast Internal Conversion of Electronically Excited Pyrimidine Bases. J. Phys. Chem. A 2005, 109, 9384−9387. (34) Zgierski, M. Z.; Patchkovskii, S.; Fujiwara, T.; Lim, E. C. Ab Initio Study of a Biradical Radiationless Decay Channel of the Lowest Excited Electronic State of Cytosine and its Derivatives. J. Chem. Phys. 2005, 123, 081101−081105. (35) Blancafort, L.; Cohen, B.; Hare, P. M.; Kohler, B.; Robb, M. A. Singlet Excited-State Dynamics of 5-Fluorocytosine and Cytosine: An Experimental and Computational Study. J. Phys. Chem. A 2005, 109, 4431−4436. (36) Merchán, M.; González-Luque, R.; Climent, T.; Serrano-Andrés, L.; Rodríguez, E.; Reguero, M.; Peláez, D. Unified Model for the Ultrafast Decay of Pyrimidine Nucleobases. J. Phys. Chem. B 2006, 110, 26471−26476. (37) Kistler, K. A.; Matsika, S. Radiationless Decay Mechanism of Cytosine: An Ab Initio Study with Comparisons to the Fluorescent Analogue 5-Methyl-2-pyrimidinone. J. Phys. Chem. A 2007, 111, 2650−2661. (38) Kistler, K. A.; Matsika, S. Three-State Conical Intersections in Cytosine and Pyrimidinone Bases. J. Chem. Phys. 2008, 128, 215102− 215116. (39) Hudock, H. R.; Martínez, T. J. Excited-State Dynamics of Cytosine Reveal Multiple Intrinsic Subpicosecond Pathways. ChemPhysChem 2008, 9, 2486−2490. (40) Barbatti, M.; Aquino, A. J. A.; Szymczak, J. J.; Nachtigallova, D.; Lischka, H. Photodynamical Simulations of Cytosine: Characterization of the Ultrafast Bi-Exponential UV Deactivation. Phys. Chem. Chem. Phys. 2011, 13, 6145−6155. (41) Kotur, M.; Weilnach, T. C.; Zhou, C.; Kistler, K. A.; Matsika, S. Distinguishing between Relaxation Pathways by Combining Dissociative Ionization Pump Probe Spectroscopy and Ab Initio Calculations: A Case Study of Cytosine. J. Chem. Phys. 2011, 134, 184309−184318. (42) Blancafort, L.; Robb, M. A. Key Role of a Threefold State Crossing in the Ultrafast Decay of Electronically Excited Cytosine. J. Phys. Chem. A 2004, 108, 10609−10614. (43) Szalay, P. G.; Watson, T. J.; Perera, A.; Lotrich, V. F.; Bartlett, R. J. Benchmark Studies on the Building Blocks of DNA. 1. Superiority of Coupled Cluster Methods in Describing the Excited States of Nucleobases in the Franck-Condon Region. J. Phys. Chem. A 2012, 116, 6702−6710. (44) TURBOMOLE V6.3 2011, a development of Universität Karlsruhe (TH) and Forschungszentrum Karlsruhe GmbH, 1989− 2007, TURBOMOLE GmbH; available from http://www.turbomole. com. (45) The thresholds for SCF and one-electron density convergence were set to 10−9 and 10−8 au, respectively. The convergence thresholds for all structure optimizations were set to 10−8 au for the energy change, 6 × 10−6 au for the maximum displacement element, 10−6 au J


Excited-State Structure and Dynamics of KetoâAmino Cytosine: The

Recommend Documents

Excited-State Structure and Dynamics of KetoâAmino Cytosine: The