Planarizing Cytosine: The S 1 State Structure, Vibrations and Nonradiative Dynamics of Jet-Cooled 5,6-Trimethylenecytosine

We measure the S 0 ! S 1 spectrum and time-resolved S 1 state nonradiative dynamics of the “clamped” cytosine derivative 5,6-trimethylenecytosine (TMCyt) in a supersonic jet, using two-color resonant two-photon ionization (R2PI), UV/UV holeburning and ns time-resolved pump/delayed ionization. The experiments are complemented with SCS-CC2, TD-CAMB3LYP, and MS-CASPT2 ab initio calculations. While the R2PI spectrum of cytosine breaks off (cid:24) 500 cm − 1 above its 0 00 band, that of TMCyt extends up to +4400 cm − 1 higher, with over a hundred resolved vibronic bands. Thus, clamping the cytosine C 5 -C 6 bond allows to explore the S 1 state vibrations and S 0 ! S 1 geometry changes in detail. The TMCyt S 1 state out-of-plane vibrations (cid:23) ′ 1 , (cid:23) ′ 3 and (cid:23) ′ 5 lie below 420 cm − 1 , the in-plane (cid:23) ′ 11 , (cid:23) ′ 12 , and (cid:23) ′ 23 vibrational fundamentals appear at 450 ; 470 and 944 cm − 1 . S 0 ! S 1 vibronic simulations based on SCS-CC2 calculations agree well with experiment if the calculated (cid:23) ′ 1 , (cid:23) ′ 3 and (cid:23) ′ 5 frequencies are reduced by a factor of 2 (cid:0) 3 . MS-CASPT2 calculations predict that the ethylene-type S 1 S 0 conical intersection (CI) increases from +366 cm − 1 in cytosine to > 6000 cm − 1 in TMCyt, explaining the long lifetime and extended S 0 ! S 1 spectrum. The lowest-energy S 1 S 0 CI of TMCyt is the “amino out-of-plane” ( OP X ) intersection, calculated at +4190 cm − 1 . The experimental S 1 S 0 internal conversion rate constant at the S 1 ( v ′ = 0) level is k IC = 0 : 98 (cid:0) 2 : 2 (cid:1) The T 1 state energy measured to lie 24580 (cid:6) 560 cm − 1 above the S 0 state. The S 1 ( v ′ = 0) lifetime is (cid:28) = 2 : 9 ns, resulting an estimated ﬂuorescence quantum yield of (cid:8) fl = 24 %. Intense two-color R2PI spectra of the TMCyt amino-enol tautomers appear above 36000 cm − 1 . A sharp S 1 ionization threshold is observed for amino-keto TMCyt, an ionization energy of 8 : 114 (cid:6) 0 : 002


Abstract
We measure the S 0 → S 1 spectrum and time-resolved S 1 state nonradiative dynamics of the "clamped" cytosine derivative 5,6-trimethylenecytosine (TMCyt) in a supersonic jet, using two-color resonant two-photon ionization (R2PI), UV/UV holeburning and ns time-resolved pump/delayed ionization. The experiments are complemented with SCS-CC2, TD-CAMB3LYP, and MS-CASPT2 ab initio calculations. While the R2PI spectrum of cytosine breaks off ∼ 500 cm −1 above its 0 0 0 band, that of TMCyt extends up to +4400 cm −1 higher, with over a hundred resolved vibronic bands. Thus, clamping the cytosine C 5 -C 6 bond allows to explore the S 1 state vibrations and  The T 1 state energy is measured to lie 24580 ± 560 cm −1 above the S 0 state. The S 1 (v ′ = 0) lifetime is τ = 2.9 ns, resulting in an estimated fluorescence quantum yield of Φ f l = 24 %. Intense two-color R2PI spectra of the TMCyt amino-enol tautomers appear above 36000 cm −1 . A sharp S 1 ionization threshold is observed for amino-keto TMCyt, yielding an adiabatic ionization energy of 8.114 ± 0.002 eV.
The lowest of these, which dominates the S 1 S 0 nonradiative decay, is called (Eth) X , since the intersection structure is similar to the CI structure of ethylene. This CI is characterized by a puckering of the C 6 atom and a twist around the C 5 -C 6 bond, with a H-C 5 -C 6 -H torsional angle of ∼ 120 • . [10][11][12][13]15,16,19,21,22 The next higher CI involves an N 3 out-of-plane bending and a large out-of-plane amino deformation and is called (OP ) X . 10,16,19,22 The third CI, called (n O , π * ) X , has a semi-planar structure with sp 3 hybridization of the C 6 atom, shortening of the C 2 -N 3 bond and stretching of the C 5 -C 6 bond relative to the ground state minimum. [10][11][12][13][14][15][16]19,21,22 Zgierski et al. have shown that covalently clamping the C 5 -C 6 bond of Cyt with a trimethylene bridge in 5,6-trimethylenecytosine (TMCyt) increases the S 1 state fluorescence lifetime and quantum yield in room-temperature aqueous solution by ∼ 1000 times relative to Cyt, [23][24][25] to τ = 1.2 ns and Φ f l ∼ 10 %. 26 Their configuration interaction singles (CIS) and second-order approximate coupled cluster (CC2) excited-state calculations predicted that this clamping shifts the (Eth) X conical intersection of cytosine to ∼ 1500 cm −1 above the S 1 minimum, making this CI energetically less accessible. 26 The trimethylene bridge in TMCyt hardly affects the π-electron framework of Cyt, so the S 0 → S 1 absorption band shifts from λ max = 267 nm for Cyt to 280 nm in TMCyt. 26 In the clamped cytosine derivative pyrrolocytosine (PC), the C 4 -amino group (see Figure 1) and the C 5 atom are covalently connected, resulting in a pyrrole ring fused to the Cyt chromophore. 27 This extension of the π-electron framework significantly shifts the S 0 → S 1 excitation maximum to λ max = 330 − 345 nm, or about 70 − 80 nm to the red, compared to Cyt. 28,29 For PC, Thompson and co-workers have measured a lifetime of τ = 2.9 ns and a quantum yield Φ f l ∼ 0.038 in pH 7 phosphate buffer. 29 Intrigued by these observations, we have measured and analyzed the S 0 → S 1 vibronic spectrum of supersonic jet-cooled TMCyt using two-color resonant two-photon ionization (2C-R2PI), UV/UV holeburning and depletion spectroscopies. We also measured the S 1 state lifetime and triplet-state formation kinetics as a function of E exc , using the nanosecond excitation/ionization delay technique, and report S 1 state nonradiative rate constants for internal conversion and intersystem crossing. In addition to the amino-keto tautomer 1 of TMCyt we have also observed an intense R2PI spectrum that we assign to the S 0 → S 1 transitions of the hydroxy-enol tautomers 2a/2b, see Figure 1 for the tautomer structures. The measurements are accompanied by calculations of the lowest excited singlet ( 1 ππ * ) and triplet ( 3 ππ * ) states of TMCyt using time-dependent density functional theory (TD-CAMB3LYP), spin-component scaled CC2 (SCS-CC2) , complete-active-state self-consistent field (CASSCF) and multi-state secondorder perturbation-theory (MS-CASPT2) methods.

A. Experimental Methods
TMCyt was synthesized in three steps from adiponitrile according to ref. 30

B. Computational Methods
A uniform theoretical treatment of the ground-and excited-state potential energy surfaces of TMCyt is difficult, and we have combined several methods following a similar approach to our recent work on 1-MCyt. 33 The electronic ground state of all 14 tautomers and rotamers of TMCyt was first optimized using density functional theory (B3LYP) with the TZVP basis set. The ground state structures of the six most stable tautomers are shown in Figure 1; these were re-optimized at the correlated level, using the Møller-Plesset (MP2) method in the resolution-of-identity (RI) approximation, the SCS-MP2 method and the CC2 method in the RI approximation, using the aug-cc-pVTZ basis set.
The adiabatic and vertical transition energies were calculated at the SCS-CC2 level of theory with the aug-cc-pVDZ basis set. Normal-mode calculations were performed for all geometry-optimized structures to assure that they correspond to true potential energy surface minima. These data were also calculated at the MS-CASPT2 level of theory. For the ππ * and 1 n O π * states we used TD-CAM-B3LYP/6-311G** optimized geometries, whose MS-CASPT2 energy is lower than that of their CASSCF and SCS-CC2 analogues. For the optimization of (OP ) M in we used the CASSCF(12,12)/6-311G** geometry because the other methods failed to converge a minimum for that state. To calculate the reaction path to (OP ) X we optimized the transition structure (TS) on S 1 , and obtained the path by combining the intrinsic reaction coordinate 36 and initial relaxation direction 37 techniques.
The calculated barrier includes the vibrational zero-point energy (ZPE) correction, which amounts to −475 cm −1 , based on CASSCF frequencies at ( 1 ππ * ) M in (with 3N − 6 vibrational modes) and at the corresponding TS (including 3N − 7 modes). The CI were optimized using the recently developed double Newton-Raphson algorithm. 38 The active space of the CASSCF and MS-CASPT2 calculations was specifically tailored for each path. for details see the supplementary material. We use (10,10) and (12,12) active spaces for the ethylene-and OP-type paths, respectively. With this approach, the MS-CASPT2 S 1 /S 0 energy gaps at the CI structures were 1973 and 2265 cm −1 (0.24 and 0.28 eV) at (Eth) X and (OP ) X , respectively. The path to (Eth) X , which has a sloped topology and does not involve a TS, was approximated with a linear interpolation in internal coordinates.
The DFT and CC2 calculations were performed using Turbomole 6.4. 39,40 The CASSCF optimizations were performed with a modified version of Gaussian09, 41 and the MS-CASPT2 calculations with Molcas 7.8. 42,43 Vibronic band simulations were done with the PGOPHER program. 44 As inputs, we used the SCS-CC2 calculated S 0 ground and S 1 excited state geometries and the corresponding normal-mode l matrices, employing conformer 1a. Additional diagonal anharmonicity constants 44 were included for some modes. The vibronic band intensities are based on full multidimensional Franck-Condon factors, including both mode displacements and mixing between modes (Dushinsky effect). 44 The vibronic simulations for conformer 1b are very similar to those for 1a.

A. Computational Results
Tautomers and Relative Energies: Figure 1 shows the six most stable calculated tautomers and rotamers of TMCyt, and Table I summarizes their relative energies calculated at different levels of theory. All the correlated wave function methods predict the trans-amino-enol 2b tautomer to be the most stable one, with the corresponding cis-rotamer 2a ∼ 0.6 kcal/mol higher. The amino-keto N1H tautomer 1 that is experimentally investigated below, exists in conformer 1a, where the amino group and trimethylene ring are out-of-plane in the same direction, denoted Up-up (or Down-down), where the first (capitalized) orientation refers to the NH 2 group. In conformer 1b, the NH 2 group and trimethylene ring are arranged in opposite directions (Up-down or Down-up). The 1a and 1b forms are close in energy with 1b calculated to lie 3 − 4 cm −1 above 1a. In the gas phase, both 1a and 1b are less stable than amino-enol conformers by 1.08 kcal/mol (CC2) or 1.54 kcal/mol (SCS-MP2). The B3LYP density functional method predicts the amino-keto N1H tautomer to be the most stable tautomer; however, it is known that this method predicts the order of the cytosine tautomers incorrectly. 6,45 The other TMCyt tautomers 2b, 2a and 4 also exist as pairs of conformers analogous to 1a/1b, but only one form was calculated since the energy difference is expected to be very small. All the imino-enol forms lie > 13 kcal/mol above the most stable tautomer 2b at the B3LYP/TZVP level, hence we do not consider them any further. and 3a conformers predicted by the SCS-CC2 method, together with the MS-CASPT2 transitions for 1a.

Electronic Transition Energies:
Both methods are in good agreement, which validates our computational approach. They predict that the state did not converge with SCS-CC2 because it reached a region of S 2 /S 1 degeneracy, which is consistent with the small S 2 /S 1 energy gap found at the 1 n O π * minimum at the MS-CASPT2 level. Optimization of the 1 n N π * state at the CASSCF level leads to (OP ) M in , with an adiabatic energy of 33017 cm −1 . The electronic configuration at this structure is analogous to that described in our previous work on 1-MCyt. 33 The adiabatic transition energy of conformer 1a is calculated to lie slightly above that of 1b, differing by 51 cm −1 at the SCS-CC2 level. With this method, the S 0 → S 1 transitions of the major tautomers 2b and 2a are calculated to be 1 ππ * and to lie at ∼ 35000 and ∼ 34500 cm −1 , respectively, or about 3500 cm −1 further to the blue than the transitions of the 1a/1b conformers. The lowest-energy electronic transition of the imino-keto tautomers 3a and 3b are predicted at 41400 cm −1 and 40970 cm −1 , respectively. This is above the experimental spectral range covered in this work. On the other hand, the lowest 1 ππ * transition of the 4 (N3H) tautomer is predicted to lie very close to that of the 1 (N1H) tautomer.
However, tautomer 4 is calculated to be 5.5 − 5.8 kcal/mol less stable than tautomer 1, hence we do not expect this tautomer to be observable in the supersonic jet.

Ground-and Excited-State Structures:
In the SCS-CC2 S 0 optimized structure of 1a the pyrimidinone framework is C s symmetric, and the amino group and the trimethylene ring are bent slightly out of the ring plane. In the 1 ππ * excited state, the SCS-CC2 and TD-CAMB3LYP methods predict (i) a stronger pyramidalization of the amino group, (ii) an in-plane deformation of the pyrimidinone framework and (iii) an out-of-plane bend at the C 6 atom (see Figure 1 for the atom numbering). Figure 2 shows the SCS-CC2/aug-cc-pVDZ calculated geometries and geometry changes of TMCyt for both amino-keto N1H conformers. The TD-CAM-B3LYP optimized structure has similar out-of-plane deformations, see Figure S1 in the supplementary material. This is in line with previous results for 1-MCyt 33 for which both methods predict a substantial deplanarization at the 1 ππ * state minimum.
Interconversion between the 1a and 1b isomers: As shown in Table I Table II), we should observe two spectra that are mutually shifted by about 50 cm −1 . However, the R2PI and UV/UV holeburning spectra discussed below show only a single ground-state species. The reason for this is the large-amplitude amino-inversion of TMCyt, which interconverts the conformers 1a and 1b.
We calculated the one-dimensional (1D) inversion potential at the same level by incrementing the H-N- were taken to represent the harmonic potential at this angle, and µ red,θ for the 1D calculation was fixed such that the calculated normal-mode and 1D frequencies in this harmonic potential were the same. Figure 3 shows that lowest-energy v inv = 0 level lies ∼ 130 cm −1 above the barrier. Its wave function is delocalized over both the 1a and 1b geometries with its maximum near planarity (θ inv = 0 • ). That the vibrational ground state of TMCyt is quasiplanar (delocalized over both 1a and 1b) explains why the UV/UV holeburning spectra, discussed in the next section, reflect the presence of a single ground-state species only. The second amino-inversion level v inv = 1 lies 380 cm −1 higher. It will be collisionally cooled out in the supersonic expansion and will not be considered further.
In the S 0 state, the planar (C s symmetric) structure of TMCyt is an index-2 stationary point. Normalmode analysis at this point yields imaginary frequencies for both the NH 2 inversion and trimethylene-ring out-of-plane vibrations. The S 0 -state barrier to planarity is 307 cm −1 at the SCS-CC2 level. In the 1 ππ * excited state the barrier to planarity is much higher, 1297 cm −1 . Four imaginary frequencies are obtained at the C s stationary point.  Detailed vibronic assignments are given in the next section. A high-resolution UV/UV holeburning spectrum is shown in Figure 5(b) and was recorded with the burn laser at the intense band at 0 0 0 + 59 cm −1 , marked by with an asterisk in Figure 5(a). It reproduces the 2C-R2PI spectrum in Figure 5(a) in great detail. From this we conclude that all the observed vibronic bands originate from the ground-state level that gives rise to the transition at 0 0 0 + 59 cm −1 . Figure 5(c) shows the corresponding UV/UV depletion spectrum in which the holeburning laser is scanned with the detection laser fixed at the intense 0 0 0 + 59 cm −1 band. The UV/UV depletion spectrum also reproduces the R2PI spectrum, although the signal/noise ratio is lower than in the UV holeburning spectrum. At 900 cm −1 above the electronic origin the widths of the vibronic bands begin to increase, which indicates the onset of rapid non-radiative processes, see section III F. Although no further bands can be observed in the depletion spectrum above +1000 cm −1 , the signal remains slightly below the baseline, indicating a constant depletion of the ion signal.

C. Vibronic Band Assignments
We first attempted to assign the vibronic bands in the R2PI spectrum of TMCyt in Figure 5(a) based on the SCS-CC2, CC2 and TD-B3LYP harmonic frequencies of the 1 ππ * state given in Table III. The lowest-frequency in-plane vibration is predicted to be ν ′ , hence the vibronic bands below ∼ 250 cm −1 must arise from outof-plane vibrations. Experimentally, the two lowest-frequency bands at 38 cm −1 and 59 cm −1 cannot belong to the same progression, so we assign these as fundamentals of the lowest-frequency out-of-plane vibrations ν ′ 1 and ν ′ 2 (that is, as 1 1 0 and 2 1 0 ). Table III shows that the lowest two frequencies calculated with the SCS-CC2, CC2 and TD-B3LYP methods are two to three times larger. Previous experience with SCS-CC2, CC2 and TD-B3LYP excited-state calculations of cytosine derivatives and pyrimidinones has shown that while the in-plane S 1 state vibrational frequencies are well reproduced, the calculated out-ofplane vibrational frequencies are often 2-3 times higher than observed experimentally. [6][7][8]31,33 For the PGOPHER 44 vibronic band simulations (see section II B), we therefore decreased the out-ofplane frequencies to the experimental values. Figure 6(a)-(c) shows the simulated vibronic bands in red for the 0 − 420, 420 − 870 and 870 − 1320 cm −1 sections of the spectrum and compares these to the high-resolution 2C-R2PI spectrum in black. We first fitted the S 1 state in-plane vibrational frequencies.
The ν ′ 11 and ν ′ 12 normal-modes correspond to the ν ′ 6a and ν ′ 6b in-plane vibrations that are characteristic of the S 0 → S 1 spectra of benzene and its derivatives. We therefore assigned the bands at 449 cm −1 and 471 cm −1 as 12 1 0 and 11 1 0 , respectively, see Figure 6(b); the order of these two vibrations was interchanged to obtain a better fit with the experimental R2PI spectrum. The 6 1 0 transition was fitted to the band at 257 cm −1 ; its intensity is rather small and it does not contribute further to the spectrum. The band at 615 cm −1 was assigned to the 15 1 0 fundamental. The band at 944 cm −1 is assigned as the in-plane fundamental ν ′ 23 , as the overtone 11 2 0 had no intensity, see Figure 6(c). We then fitted the out-of-plane vibrations, see Figure 6(a). The weak band at 38 cm −1 is assigned as the ν ′ 1 fundamental. Since the ν ′ 2 and ν ′ 4 vibrations involve structural changes of the trimethylene ring, see Table III and hardly appear in the simulation, the intense 59 cm −1 band is assigned as the 3 1 0 "butterfly" vibrational fundamental. The 3 2 0 overtone was fitted to the band at 126 cm −1 . The fundamentals of ν ′ 5 and ν ′ 7 were fitted to the bands at 93 cm −1 and 221 cm −1 . The out-of-plane normal-mode eigenvectors ν ′ 1 , ν ′ 3 , ν ′ 5 and ν ′ 7 are shown in Figure 8. Note that the SCS-CC2, CC2 and TD-B3LYP harmonic frequencies in Table III differ from the fitted frequencies (Table IV) by a factor of 2 − 3, indicating that the S 1 state potential-energy surface is much flatter and more anharmonic along these coordinates than predicted by the excited-state calculations. Figure 9 shows the photoionization efficiency (PIE) curves of the S 1 ( 1 ππ * ) state, which were recorded at 0 ns delay of the ionization laser, and of a long-lived state, which was recorded at 50 ns delay. The PIE curve of the long-lived state shown in Figure 9 is scaled according to the relative signal heights discussed in the next section, where the T 1 ion signal reaches 25 % of the S 1 signal when ionizing at 225 nm.

D. Photoionization Efficiency Curves
The PIE curve of the S 1 ( 1 ππ * ) state in Figure 9 exhibits a steplike ionization threshold at 33930 ± 20 cm −1 , indicating that the geometry change between the v ′ = 0 level of the S 1 ( 1 ππ * ) state and the TMCyt + ion ground state D 0 is small. The Franck-Condon factor for adiabatic ionization from the S 1 state is sufficiently large so the adiabatic ionization energy (AIE) threshold can be observed. The sum of the S 0 → S 1 0 0 0 excitation energy of 31510 cm −1 and the PIE threshold in Figure 9 is 65440 ± 20 cm −1 , giving an AIE= 8.114 ± 0.002 eV. The SCS-CC2 calculated AIE= 8.18 eV of tautomer 1a is in good agreement with this value (see Table II).
The delayed-ionization PIE curve of the long-lived state shown in Figure 9(b) is relatively noisy; since the UV spectrum of the TMCyt amino-enol forms begins around ∼ 36000 cm −1 , this contribution to the signal had to be subtracted. The PIE curve exhibits a gradual signal onset at 40320 cm −1 followed by a slow rise. We interpret the long-lived state as the lowest triplet state T 1 , and this slow onset as photoionization of the hot vibrational levels of T 1 that are formed by S 1 T 1 intersystem crossing (ISC); the S 1 ↔ T 1 energy difference is converted to vibrational energy of the T 1 state during the ISC process. The signal onset at 40320 cm −1 is thus interpreted as the lower limit to the AIE of the T 1 state. The upper limit to the AIE is estimated by back-extrapolation of the linear part of the PIE curve to the zero-signal line at 41400 cm −1 . Subtracting these two values from the AIE of the S 1 ( 1 ππ * ) state (65440 ± 20 cm −1 ) places the T 1 state between 24020 cm −1 and 25140 cm −1 above the S 0 ground state.
Our calculation supports the assignment of the T 1 state. The calculated adiabatic energy of this state is ∼ 27800 cm −1 , whereas the alternative of a dark 1 n O π * state can be discarded because its estimated energy is much higher, 37597 cm −1 , see Table II.

E. Ns Pump/ Ionization Delay Scan Measurements and Nonradiative Kinetics
We measured the excited-state lifetime and nonradiative kinetics of TMCyt using ns laser pump/delayed ionization measurements by ionizing at 225 nm. The convolution of the pulse widths of the pump and ionization laser yields a Gaussian instrument response function (IRF) with a full width at half-maximum (FWHM) of 4.2 ns. We modeled the S 1 ( 1 ππ * ) state kinetics as where k rad is the S 1 → S 0 radiative decay rate. The SCS-CC2 calculated oscillator strength of TMCyt is f el = 0.0918, giving τ rad ∼ 12 ns or k rad = 8.3 · 10 7 s −1 . This value is in good agreement with the τ rad = 13 ns that Zgierski et al. estimated from the integrated S 0 → S 1 absorption spectrum of TMCyt in aqueous solution. 26 . The S 1 state is assumed to decay nonradiatively to S 0 by internal conversion (IC) with the rate constant k S 1 IC and by intersystem crossing (ISC) to the T 1 state with the rate constant k S 1 ISC . T 1 is assumed to relax to S 0 by T 1 S 0 reverse ISC and also by phosphorescence; these two pathways are combined into a single rate constant k T : However, k T is very low (< 5 · 10 6 s) and we cannot determine it by delay measurements on the ∼ 50 ns time scale, so it is set to zero. The simulated time-dependent concentrations [S 1 ] and [T 1 ] were convoluted with the IRF and were least-square fitted to the experimental pump/ionization signal traces.
Note that because of the 4.2 ns width of the IRF, which is similar to the inverse of the k IC and k ISC rate constants, the ratio of the ionization efficiencies of molecules in the S 1 and T 1 states, σ ion (S 1 ):σ ion (T 1 ), can only be estimated within certain limits discussed below. If the width of the two laser pulses were significantly shorter than the inverse of the k IC and k ISC rates, then the experimental pump/ionization transient would exhibit a much more intense S 1 signal that would peak close to 100 % on the scale of Figure 10, and the observed S 1 : T 1 signal ratio would be correspondingly larger.
In Figure 10(a-c) we show the experimental pump/ionization transient with excitation at the 0 0 0 band and ionization at 225 nm, marked by a dashed vertical line in Figure 9. This transient is fitted for three different assumptions for the ionization efficiency ratio σ ion (S 1 ):σ ion (T 1 ). In Figure 10(a) we assume σ ion (S 1 ):σ ion (T 1 )= 1, giving the nonradiative rate constants k IC = 2.2 · 10 8 s −1 and k ISC = 4.1 · 10 7 s −1 .
Note, however, that this ratio is unrealistically low, since ionization at 225 nm is 10000 cm −1 above the S 1 ionization threshold but only 2700 cm −1 above the T 1 ionization threshold. For the fit in Figure 10(b) we assume that the ionization efficiency ratio σ ion (S 1 ):σ ion (T 1 )= 4, which is the apparent experimental ratio between the S 1 and T 1 ion signals at 225 nm shown in Figure 9, and between the ion signals at 0 ns delay and 40 ns delay shown in Figure 10. This fit gives the nonradiative rate constants k IC = 9.8·10 7 s −1 and k ISC = 1.6 · 10 8 s −1 . If -as the other limiting case -we assume k IC to be zero and fit k ISC and the σ ion (S 1 ):σ ion (T 1 ) ratio, we obtain the fit curves shown in Figure 10(c). The resulting σ ion (S 1 ):σ ion (T 1 ) = 6.4 is the maximum possible ratio, and the fitted k ISC = 2.6 · 10 8 s −1 is an upper limit for the ISC rate.
These IC and ISC rate constants of TMCyt can be compared to those of 1-MCyt, which are k IC = 2 · 10 9 s −1 and k ISC = 2 · 10 8 s −1 near the S 1 (v ′ = 0) level. 33 The main difference lies in the decrease of k IC by a factor of 10 − 20. The ISC rate constant probably changes little upon rigidization of the pyrimidinone, but the uncertainty is large. Thus the increase in excited-state lifetime at the 0 0 0 band upon clamping the C 5 -C 6 bond originates mainly from the decrease of the IC rate.
The pump/ionization transients were also measured at an ionization wavelength of 245 nm, which is the same wavelength as used to record the 2C-R2PI spectra. The measured 0 0 0 band transient was well fitted with the three sets of k IC and k ISC constants that correspond to Figure 10(a-c). However, Figure 9 shows that ionization of the T 1 state at 245 nm is very inefficient; thus, the σ ion (S 1 ):σ ion (T 1 ) ratio was re-fitted and is 15.5 times larger that for ionization at 225 nm. These fits are shown in the supplementary Figure S3(a-c). Ns pump/ionization transients were also measured for the bands at 0 0 0 +530, 0 0 0 +1174 and 0 0 0 + 1646 cm −1 , but only with ionization at 245 nm, see supplementary Figure S3(d-f). These transients were fitted with a fixed σ ion (S 1 ):σ ion (T 1 )= 15.5, which corresponds to assuming σ ion (S 1 ):σ ion (T 1 ) = 1 and T 1 states at 225 nm. All fitted k IC and k ISC values assuming σ ion (S 1 ):σ ion (T 1 ) = 1 at 225 nm are collected in Table V.
Summarizing, one sees that although the ns time resolution of the pump/ionization transient measurement and the unknown ratio σ ion (S 1 ):σ ion (T 1 ) lead to considerable uncertainty, k IC is determined within a factor of 2.5 between k IC = 9.8 · 10 7 and 2.2 · 10 8 s −1 . Similarly, the limits of the ISC rate constant are determined within a factor of four as k ISC = 4.1 · 10 7 to 1.6 · 10 8 s −1 . For all three fits, the lifetime at the 0 0 0 band is τ = 2.9 ns. Given the calculated radiative rate constant k rad = 8.3 · 10 7 s −1 and that τ = 1/(k rad + k ISC + k IC ), one finds that the fluorescence quantum yield of TMCyt is Φ f l = 24 %.
This value does not depend on the exact k IC and k ISC rate constants. For TMCyt in room-temperature aqueous solution, Zgierski et al. determined Φ f l ∼ 10 % from the lifetime of τ = 1.2 ns. 26 That the fluorescence quantum yield at room temperature is lower than at the low temperature in the supersonic jet is very reasonable and to be expected from the increase of k IC with increasing vibrational energy, as is documented in Table V.
In contrast to the S 0 → S 1 vibronic spectra of Cyt and its derivatives 1-MCyt, 5-MCyt and 5-FCyt, [6][7][8]33 which exhibit sharp break-offs at 450 − 1200 cm −1 above the 0 0 0 bands, indicating the onset of an ultrafast process, the S 0 → S 1 2C-R2PI spectrum of TMCyt 1a/ 1b extends up to 4400 cm −1 above the 0 0 0 band and does not show a spectral break-off. The vibronic bands either merge or become diffuse at ∼ 2100 cm −1 above the 0 0 0 band of the amino-keto tautomer. To investigate the reason for the broadening, we modeled the complete vibronic spectrum for TMCyt using PGOPHER 8.0; 44 the simulated spectrum is shown in Figure 4(b). In addition to the nine optically active vibrational modes ν 12 , ν ′ 15 and ν ′ 23 that were employed for the simulation in Figure 6 in section III C, we included the fundamental excitations of all vibrations with calculated Franck-Condon factors > 15% of the 0 0 0 band. These are the in-plane vibrations ν ′ 33 , ν ′ 39 , ν ′ 43 , ν ′ 44 and ν ′ 45 and the ν ′ 8 out-of-plane vibration. These frequencies were not fitted to experimental transitions but were taken from the SCS-CC2 calculations. The overtones and combination tones of these six vibrations could not be included because of the limited array sizes of PGOPHER.
A Gaussian line shape with a FWHM of 5 cm −1 was employed, reflecting the bandwidth of the UV-OPO. When setting the Lorentzian linewidth contribution ∆ Lor to zero, the simulated spectrum exhibits resolved vibronic bands up to +4400 cm −1 . If we include a Lorentzian linewidth contribution ∆ Lor = 5 cm −1 in the simulation, which corresponds to a lifetime of 1 ps, we see in Figure 4(b) that the bands broaden and merge into a semi-continuous background that is similar to the experimental spectrum in Figure 4(a). This suggests that the broadening of the spectrum at excess energies above +2100 cm −1 does not reflect just spectral congestion, but arises from a decrease in the excited-state lifetime.
To account for the additional broadening observed in the experimental spectrum, we have calculated the two most energetically favorable excited-state decay paths that are analogous to those for cytosine and 1-MCyt. 10,16,18,19,[46][47][48][49] According to expectations and in line with the calculations of Zgierski et al., 26 the access to the ethylene-type intersection is hindered by the trimethylene modification. The calculated energy of the (Eth) X CI is approximately 6800 cm −1 relative to the 0 0 0 transition. The path from 1 ππ *

M in
to that CI has a sloped topology, and the barrier for the decay is given by the energy of the CI itself, see Figure S2 in the supplementary material.
The energetically favored decay path involves out-of-plane deformation of N 3 and the amino group.
The calculated energy profile along this path is shown in Figure 11. The path leads from the S 1 ( 1 ππ * M in ) structure through a transition state (TS) to a second minimum, (OP ) M in , which is similar to that previously characterized for Cyt and 1-MCyt. 10,16,18,19,[46][47][48][49] The MS-CASPT2 barrier over the TS is 1935 cm −1 , and the energy of (OP ) M in relative to 1 Figure 12), which is consistent with the decay path where (OP ) M in lies before the (OP ) X CI.
The fact that broadening of the vibronic bands in the R2PI spectrum sets in at around ∼ 2100 cm −1 , but a semi-continuous spectrum continues up to 4400 cm −1 above the electronic origin is in qualitative agreement with the calculated decay path topology. We interpret the additional broadening beyond ∼ 2100 cm −1 as due to the coupling between the vibrations belonging to the S 1 (ππ * )and the OP M inminima below the barrier. The density of vibronic states belonging to both minima rises enormously when the energy exceeds this barrier (MS-CASPT2 barrier 1935 cm −1 ). The semi-continuous spectrum that reaches up to at least 4400 cm −1 is also in good agreement with the calculated CI at 4300 cm −1 .
The 0 0 0 band is identified at 31510 cm −1 . The lowest 400 cm −1 of the S 0 → S 1 spectrum is dominated by fundamentals, overtone excitations and combination bands of four out-of-plane vibrations. Based on the energetic sequence of the SCS-CC2 calculated vibrational frequencies and on their predicted Franck-Condon factors, we assign these as ν ′ 1 , ν ′ 3 , ν ′ 5 and ν ′ 7 . Similar to the spectra of cytosine, 5-methyl-and 5-fluorocytosine, 6-8 the longest vibronic progression is observed for the butterfly vibration ν ′ 3 . Combination progressions in ν ′ 3 are also built on the in-plane vibrational fundamentals ν ′ 11 , ν ′ 12 , ν ′ 15 and ν ′ 23 . In contrast to unsubstituted cytosine, whose S 0 → S 1 spectrum breaks off above ∼ 500 cm −1 , the R2PI spectrum of TMCyt extends to ∼ 4400 cm −1 above, with more than 100 resolved vibronic bands. This is the most extended S 0 → S 1 spectrum of any cytosine derivative measured so far. We have also observed the R2PI spectra of the amino-enol tautomers 2a and 2b starting at ∼ 36000 cm −1 , but these will be discussed elsewhere.
Sharp vibronic bands can be observed up to +2100 cm −1 above the 0 0 0 band. Hence, bridging of the C 5 -C 6 bond with the trimethylene ring strongly raises the barrier to the ethylene-type (Eth) X conical intersection. Above +2100 cm −1 a semi-continuous R2PI spectrum is observed up to at least +4400 cm −1 .
The vibronic band simulation performed with the PGOPHER program 44 nicely reproduces the vibronic band structure and intensities of the R2PI spectrum up to +1320 cm −1 . Towards higher frequencies the simulations predict resolved vibronic transitions, whereas in the R2PI spectrum an increased density of bands leads to an intense continuous signal.
From a mechanistic perspective, our computational work shows that by blocking the twist of the C 5 -C 6 bond we not only change the energetically favored decay path, but also the topology. In Cyt and 1-MCyt, the decay path leads from the 1 ππ * minimum via a TS to the ethylene-type (Eth) X CI. The CI can be reached as soon as enough energy is available to go over the TS, and this is observed as a sharp break-off of the R2PI spectrum above ∼ 500 cm −1 in these systems. In contrast, in TMCyt the lowest CI is the amino out-of-plane bend (OP ) X and the (Eth) X CI is raised ∼ 6800 cm −1 above the S 1 vibrationless level. The path to (OP ) X involves an additional minimum that lies before the intersection.
As a consequence, the R2PI spectrum does not completely break off when enough energy is available to go over the TS. This suggests that the broad, shapeless spectral region between 2100 and 4300 cm −1 is a signature of the calculated topology.
The excited-state lifetime of amino-keto TMCyt at the 0 0 0 band is τ = 2.9 ns, which is a fourfold increase relative to that of cytosine at its 0 0 0 band. Additionally, the lifetime τ drops off much more slowly with increasing vibrational excess energy, being τ = 1.6 ns even at a vibrational excess energy E exc = 1174 cm −1 . From the calculated S 1 state radiative lifetime and experimental lifetime τ , we infer that the fluorescence quantum yield at the v'=0 level is Φ f l = 24 %, which makes TMCyt the strongest fluorescing cytosine derivative in the gas phase known to date. Φ f l drops to ∼ 6 % at E exc = 1646 cm −1 . These fluorescence lifetimes and quantum yields are in qualitative agreement with the τ = 1.2 ns and Φ f l ∼ 10 % values that Zgierski et al. determined for TMCyt in room-temperature aqueous solution. 26 The availability of a strongly fluorescent gas-phase cytosine derivative opens exciting new research opportunities based on fluorescence measurements.   TABLE II. SCS-CC2 and MS-CASPT2 calculated adiabatic and vertical transition energies (in cm −1 ) and electronic oscillator strengths f el for five tautomers of 5,6-trimethylenecytosine (see Figure 1).    V. Internal conversion (IC) and intersystem crossing (ISC) rate constants, decay lifetimes, fluorescence quantum yields Φ f l and ISC quantum yields Φ ISC , from fits to the ns excitation/ionization transients in Figure 10 and Figure S3 (supporting information), assuming the relative ionization efficiencies of the S 1 and T 1 state at 225 nm to be equal.  IG. 2. SCS-CC2/aug-cc-pVDZ calculated geometries and geometry changes of amino-keto 5,6trimethylenecytosine upon 1 ππ * excitation (ground state is light-colored and the 1 ππ * state is darker). Bond length changes ≥0.05Å and bond angle changes ≥3 • are indicated.   . Photoionization efficiency curves of 5,6-trimethylenecytosine following excitation at the S 1 0 0 0 band (a) with prompt ionization (0 ns delay), the steplike adiabatic photoionization threshold is shown in the insert (5x). (b) PIE curve with the ionization laser delayed by 50 ns, relative to the excitation. The uncertainty of the T 1 photoionization threshold is indicated with a blue bar.