Temperature-Dependent Kinetics of the Reactions of the Criegee Intermediate CH2OO with Hydroxyketones

Though there is a growing body of literature on the kinetics of CIs with simple carbonyls, CI reactions with functionalized carbonyls such as hydroxyketones remain unexplored. In this work, the temperature-dependent kinetics of the reactions of CH2OO with two hydroxyketones, hydroxyacetone (AcOH) and 4-hydroxy-2-butanone (4H2B), have been studied using a laser flash photolysis transient absorption spectroscopy technique and complementary quantum chemistry calculations. Bimolecular rate constants were determined from CH2OO loss rates observed under pseudo-first-order conditions across the temperature range 275–335 K. Arrhenius plots were linear and yielded T-dependent bimolecular rate constants: kAcOH(T) = (4.3 ± 1.7) × 10–15 exp[(1630 ± 120)/T] and k4H2B(T) = (3.5 ± 2.6) × 10–15 exp[(1700 ± 200)/T]. Both reactions show negative temperature dependences and overall very similar rate constants. Stationary points on the reaction energy surfaces were characterized using the composite CBS-QB3 method. Transition states were identified for both 1,3-dipolar cycloaddition reactions across the carbonyl and 1,2-insertion/addition at the hydroxyl group. The free-energy barriers for the latter reaction pathways are higher by ∼4–5 kcal mol–1, and their contributions are presumed to be negligible for both AcOH and 4H2B. The cycloaddition reactions are highly exothermic and form cyclic secondary ozonides that are the typical primary products of Criegee intermediate reactions with carbonyl compounds. The reactivity of the hydroxyketones toward CH2OO appears to be similar to that of acetaldehyde, which can be rationalized by consideration of the energies of the frontier molecular orbitals involved in the cycloaddition. The CH2OO + hydroxyketone reactions are likely too slow to be of significance in the atmosphere, except at very low temperatures.


Reactant Concentration Measurements
The AcOH and 4H2B samples were used as provided by the manufacturers, with reported purities of 90% and 95%, respectively.IR spectra of the headspace above liquid samples of both hydroxyketones were recorded in a 10 cm cell with ZnSe windows in a JASCO 4700 FT-IR spectrometer to identify any impurities present.The spectra are shown in Figure S1.The AcOH spectrum is the same as the that obtained from the PNNL database, 1 with no obvious features present that could be attributed to impurities.To the best of our knowledge, the gas-phase spectrum of 4H2B has not been reported in the literature, but it appears very similar to that of AcOH, although the absorbance is weaker as a result of its lower vapor pressure.
The concentrations (cm -3 ) of the hydroxyketone reactants present in the flow reactor during measurements can be estimated based on reported vapor pressures 2,3 and the gas flow rates using the equation where χ is the mole fraction of X in the X/N2 flow, F is its fractional contribution to the total gas flow, Ptot is the total pressure (Torr), NA is the Avogadro constant (6.022×10 23 mol -1 ), T is the reactor temperature (K), and R is the gas constant (62.364×10 3 cm 3 Torr -1 mol -1 K -1 ).The actual concentrations [X]exp are determined from absorption spectra recorded in the 275-298 nm range using a pulsed LED nominally centered at 280 nm under identical conditions to those used in the kinetics measurements.5][6][7] The JPL recommendation was used for AcOH.The calibration plots of [X]exp versus [X]est shown in Figure S3 and Figure S4 were used to determine temperature-independent scaling factors that allow the concentration estimates to be corrected.
To extract kloss, a pseudo-1 st order rate constant that accounts for background losses and reaction with the hydroxyketone X kself is the self-reaction rate constant, which is held fixed at 7.8×10 -11 cm 3 s -1 . 8gure S6 and Figure S7 show the pseudo-1 st order plots of CH2OO loss rates as a function of the AcOH and 4H2B concentrations, respectively, at temperatures in the range 275-335 K.
Figure S8 shows plots of ln (k/T2) against 1/T for the CH2OO + AcOH and CH2OO + 4H2B reactions that are used to obtain the standard entropy, enthalpy, and Gibbs energy of activation, summarized in Table S1, according to the equation where p° is standard pressure of 10 5 Pa.
Figure S9 shows CH2OO loss rates in the presence of AcOH and 4H2B as a function of total pressure.
Experimental rate constants for a range of CH2OO + R1R2CO reactions at room temperature are compiled in Table S2.

Figure 1 Figure
Figure S1Headspace FT-IR spectra of AcOH (red) and 4H2B (blue).Also shown is the AcOH spectrum (black) obtained from the PNNL IR spectral database.1

Figure
Figure S3 Calibration plots of experimental versus estimated AcOH concentrations in the range 275-335 K.

Figure
Figure S4 Calibration plots of experimental versus estimated 4H2B concentrations in the range 275-335 K.

Figure
Figure S5 Broadband transient absorption spectra recorded at various time delays after photolysis at 295 K. (a) Spectra recorded in the absence of any additional reactive species, (b) spectra recorded with [AcOH = 5.1×10 15 cm -3 .(c) [CH2OO] time profiles resulting from fitting the experimental transient spectra.

Figure
Figure S6 Global pseudo-1 st order plots for the reaction of CH2OO with AcOH in the temperature range 275-335 K. Three independent kinetic runs are compiled at each temperature.

Figure
Figure S9 Pressure dependence of CH2OO loss rates in the presence of AcOH (red, [AcOH] = (5.1±0.3)×10 15 cm -3 ) and 4H2B (blue, [4H2B] = (8.0±1.1)×10 14cm -3 ) at 295 K. Total pressure was varied by changing only the flow rate of the N2 buffer, with all other gas flows held constant.The solid lines indicate the average loss rates determined from the kinetics measurements.Shaded areas represent 1σ uncertainties.