Acetylacetone Photolysis at 280 nm Studied by Velocity-Map Ion Imaging

The photolysis of acetylacetone (AcAc) has been studied using velocity-map ion imaging with pulsed nanosecond lasers. The enolone tautomer of AcAc (CH3C(O)CH=C(OH)CH3) was excited in the strong UV absorption band by UV pulses at 280 nm, preparing the S2(ππ*) state, and products were probed after a short time delay by single-photon VUV ionization at 118.2 nm. Two-color UV + VUV time-of-flight mass spectra show enhancement of fragments at m/z = 15, 42, 43, 58, and 85 at the lowest UV pulse energies and depletion of the parent ion at m/z = 100. Ion images of the five major fragments are all isotropic, indicating dissociation lifetimes that are long on the timescale of molecular rotation but shorter than the laser pulse duration (<6 ns). The m/z = 15 and 85 fragments have identical momentum distributions with moderate translational energy release, suggesting that they are formed as a neutral product pair and likely via a Norrish type I dissociation of the enolone to form CH3 + C(O)CH=C(OH)CH3 over a barrier on a triplet surface. The m/z = 43 fragment may be tentatively assigned to the alternative Norrish type I pathway that produces CH3CO + CH2C(O)CH3 on S0 following phototautomerization to the diketone, although alternative mechanisms involving dissociative ionization of a larger primary photoproduct cannot be conclusively ruled out. The m/z = 42 and 58 fragments are not momentum-matched and consequently are not formed as a neutral pair via a unimolecular dissociation pathway on S0. They also likely originate from the dissociative ionization of primary photofragments. RRKM calculations suggest that unimolecular dissociation pathways that lead to molecular products on S0 are generally slow, implying an upper-limit lifetime of <46 ns after excitation at 280 nm. Time-dependent measurements suggest that the observed photofragments likely do not arise from dissociative ionization of energized AcAc S0*.


Spectrometer calibration
The velocity-map imaging spectrometer was calibrated using the well-known UV photolysis of CH3I at 280 nm with VUV ionization detection of the I( 2 P3/2) and I*( 2 P1/2) atomic products. [1][2][3][4] CH3I → CH3 + I D0 = 2.39 eV (S1a) CH3I → CH3 + I* D0 = 3.33 eV (S1b) The uncertainty in the dissociation energies (1σ) is 0.03 eV (~1%). 5 Both I and I* products can be readily ionized at 118 nm. The photoionization cross section is approximately 20 times greater for I than I*, as a result of an accidental resonance with an autoionizing Rydberg state. 6 Figure S1 shows a typical symmetrized DC slice ion image. Ion counts near the center of the image are an experimental artefact. The ion image shows three distinct rings. The outermost and innermost rings can be assigned to I and I* atoms, respectively, formed in conjunction with CH3 radicals in the vibrational ground sate. The intermediate feature corresponds to formation of I atoms with vibrationally excited CH3(ν1 = 1). The angular distributions are consistent with a parallel transition to the 3 Q0 state of the parent CH3I molecule.
The radial distribution, also shown in Figure S1 is obtained by conversion to polar coordinates and direct integration over the polar angle, weighted by an r sin θ Jacobian. Photofragment speeds v are related to the image radius r in pixels by the equation where τTOF is the flight time for the m/z = 127 ion (11.34 μs) and k is a calibration factor. In order to account for internal (rotational) excitation of the undetected CH3I fragment, we determine the cutoff in the radial distributions as r0+2σ, where r0 is the center of a Gaussian fit and σ the standard deviation. A pixel-to-speed calibration factor k = (36.6±0.6) m s -1 μs pixel -1 results in excellent S3 agreement with the expected maximum speeds based on the known dissociation energies for all three features present in the radial distribution.
Figure S1 Velocity-map ion image and normalized radial (speed) distribution for I and I* products formed from the photolysis of CH3I at 280 nm. Shaded areas represent 1σ uncertainties from repeated measurements. Dashed and solid red lines are the fitted individual Gaussian components and overall fit, respectively.

Fragment time profiles
Time profiles for the heights of the m/z = 15, 42, 43, 58, 85, and 100 peaks in the mass spectra as a function of the time delay between the UV and VUV beams (Δt -tUV -tVUV) are shown in Figure S2.

Photolysis power dependence
Ion images of the m/z = 15, 43, 58, and 85 photofragments resulting from excitation of AcAc at 280 nm were acquired using different pulse energies to test for effects of multiphoton excitation. The speed distributions P(v) obtained using UV pulse energies that are a factor of 10 different are shown in Figure S3 below. The speed distributions are identical within measurement uncertainties. Power dependences were also explored using time-of-flight mass spectra acquired with UV pulse energies spanning the range 40-420 μJ (fluence 5-54 mJ cm -2 ). The double logarithmic plots of peak height versus pulse energy are showin in Figure S4 and the gradients are compiled in Table S1.  Table S1 Table S1 Gradients of the log-log plots of peak height versus UV pulse energy derived from time-of-flight mass spectra shown in Figure S4. Uncertainties are 1σ.

Fragment momentum distributions
The momentum distributions P(p) for the m/z = 42 and 58 fragments are shown in Figure S5. uncertainties derived from repeated measurements.

Ab initio calculations
Ab initio calculations were performed at the B3LYP/cc-pVDZ level of theory using the GAMESS package (version 2020 R2). 7 Optimized geometries and harmonic frequencies of the enolone and diketone tautomers, the tautomerization transitions state TS1, and the molecular dissociation transitions states TS2-TS8 were calculated. The TS labels are the same as those used for the equivalent stationary points characterized at the CBS-QB3 level by Antonov et al. 8 Cartesian coordinates are listed in Table S2 and the electronic energies and harmonic frequencies are compiled in Table S3.