Infrared identification of the Criegee intermediates syn- and anti-CH3CHOO, and their distinct conformation-dependent reactivity

The Criegee intermediates are carbonyl oxides that play critical roles in ozonolysis of alkenes in the atmosphere. So far, the mid-infrared spectrum of only the simplest Criegee intermediate CH2OO has been reported. Methyl substitution of CH2OO produces two conformers of CH3CHOO and consequently complicates the infrared spectrum. Here we report the transient infrared spectrum of syn- and anti-CH3CHOO, produced from CH3CHI + O2 in a flow reactor, using a step-scan Fourier-transform spectrometer. Guided and supported by high-level full-dimensional quantum calculations, rotational contours of the four observed bands are simulated successfully and provide definitive identification of both conformers. Furthermore, anti-CH3CHOO shows a reactivity greater than syn-CH3CHOO towards NO/NO2; at the later period of reaction, the spectrum can be simulated with only syn-CH3CHOO. Without NO/NO2, anti-CH3CHOO also decays much faster than syn-CH3CHOO. The direct infrared detection of syn- and anti-CH3CHOO should prove useful for field measurements and laboratory investigations of the Criegee mechanism.


Supplementary Figure 7 | Plot of the reciprocal integrated absorbance versus reaction time. a,
reciprocal integrated absorbance of band A 1 versus reaction time t. Data for t ≥ 12 μs, assumed to be mainly due to syn-CH 3 CHOO, are fitted with a line (red) with a slope 0.23±0.02 cm μs -1 , whereas data for t ≤ 4 μs are fitted with a line (blue) with a slope 0.44±0.02 cm μs -1 . b, Reciprocal integrated absorbance of anti-CH 3 CHOO versus reaction time t estimated by assuming that syn-CH 3 CHOO dominates after 10 μs and that the fitted (red) line in Fig. 5a represents I -1 for syn-CH 3 CHOO. Data are fitted with a line with a slope 6.3±0.6 cm μs -1 . 3 (1)

Supplementary Table 5 | Comparison of relative energy, anharmonic vibrational wavenumbers (cm 1 ) and IR intensities of vibrational modes of syn-CH 3 CHIOO and anti-CH 3 CHIOO predicted with various methods.
Mode (2)   c Hot band results are problematic, see text for discussion.

calculations. SP refers to the torsion saddle point.
Minimum  1 we simulated the spectrum of each band using experimental rotational parameters A", B", and C" reported by Nakajima and Endo, 2 ratios of rotational parameters A'/A", B'/B", and C'/C" and a-type/b-type ratios of each mode predicted with the MULTIMODE method, as listed in Table S7  Additional computational details. MULTIMODE is a program for vibrational calculation, which has been discussed in detailed and applied to various systems. 3 Generally speaking, the MULTIMODE calculation uses Watson Hamiltonian with mass-scaled normal coordinates. The key feature of MUTLMODE is to expand the full potential to a hierarchy n-mode representation, which greatly reduces the multi-dimensional integration. The CH 3 CHOO molecule has 18 degrees of freedom, and we use 4-mode representation in our study.
The Watson Hamiltonian is rigorous for semi-rigid molecules, which is not obviously the case for syn-and anti-CH 3 CHOO with the internal torsion motion. However, from a recent Diffusion Monte Carlo calculation, 4 the ground state wave functions are localized, and can be applied with the semi-rigid approach. Since the energies of highly excited states of torsion mode exceed the torsion barrier height, 730 and 427 cm 1 for syn-and anti-CH 3 CHOO, respectively, the excitation of torsion mode is restricted to 2 to avoid any unreasonable couplings with torsion mode. The semi-rigid treatment for anti-CH 3 CHOO is more problematic, especially for "hot bands", than for syn-CH 3 CHOO because the barrier to internal rotation for the anti-conformer is roughly 300 cm -1 lower than for the syn-conformer. 4 Thirteen harmonic oscillator basis are used for each mode in the first vibrational self-consistent field (VSCF) calculation. Then the VSCF virtual states are used as the basis-function for the following vibrational configuration-interaction (VCI) step. In addition, we performed another smaller set of calculations which only includes 13 modes of syn-and 14 modes of anti-conformer to compare with the calculations that includes all 18 modes.
For syn-conformer, the frequencies of nine vibration modes are in the measured spectral range from 800 to 1500 cm -1 , and ten such modes for anti-conformer. The smaller set calculation of 13 modes for syn-CH 3 CHOO includes such nine modes and four CH stretch modes, and similarly for calculation of the 14 modes for anti-CH 3 CHOO.
The hot-band transition can be calculated from MULTIMODE calculation as well. We present two sets of calculation results in Supplementary Table 6. One is from the calculation with all 18 modes coupling; another is from the smaller set of calculation with selected 13 modes for syn-conformer (14 modes for anti) as well as another two lowest frequency modes. As seen, relatively large deviation of the hot-band transition shift is observed between the experiment and MULTIMODE prediction. One main reason is the treatment of the torsional mode of CH 3 CHOO.
As mentioned above, the maximum excitation of torsion mode is restricted to 2. Such restriction affects the calculated energies of torsion states and other modes with small energies. In addition, including the low-frequency modes enhances the coupling among states. For the ground state, the torsion splitting is small and such restriction has negligible effect to calculation. However, as the torsion mode is excited to the fundamental or overtone states, our "single-reference" treatment of torsion motion may be problematic. Nevertheless, the blue-shifted hot-band transition of these bands agrees with experiment simulation.