Supplementary Information for Direct observation of unimolecular decay of CH3CH2CHOO Criegee intermediates to OH radical products Yi Fang,1 Fang Liu,1 Stephen J. Klippenstein,2 and Marsha I. Lester 1* Department of Chemistry, University of Pennsylvania, Philadelphia, PA 19104-6323 Chemical Sciences and Engineering Division, Argonne National Laboratory, Argonne, IL, 60439 1

2

*

Correspondence to: [email protected]

1

Representative geometric structures and potential energy along an intrinsic reaction coordinate (IRC) from the minimum energy configuration of the CH3CH2CHOO Criegee intermediate to the transition state (TS) leading to OH products have been evaluated using B3LYP/6-311G+(2d,p) density functional theory. The structural changes along the IRC for CH3CH2CHOO, shown in Figure S1, are very similar to those found previously for syn-CH3CHOO and (CH3)2COO in Refs. 1, 2. The influence of tunneling on the RRKM rates and lifetimes of CH3CH2CHOO is shown in Figure S2. Tunneling through the H-atom transfer barrier makes a very significant contribution to the unimolecular dissociation rate both above and clearly below the zero-point corrected barrier of 16.46 kcal mol-1 for conformer A of CH3CH2CHOO. Tunneling also contributes significantly to the RRKM rates predicted for the less stable conformer B (see main text). Figure S3 provides a comparison of thermal dissociation rates predicted from master equation modeling of the unimolecular decay of syn-CH3CHOO, (CH3)2COO, and CH3CH2CHOO to OH products in the high pressure limit over the 200 to 350 K temperature range relevant to the troposphere.3 The thermal rates of the alkyl-substituted Criegee intermediates are principally controlled by the TS barrier heights (including zero-point corrections), which are predicted to be 17.05 kcal mol-1 for syn-CH3CHOO,3 16.16 kcal mol-1 for (CH3)2COO,3 and 16.46 kcal mol-1 for the lowest energy conformer of CH3CH2CHOO. Figure S4 provides an Arrhenius plot of the thermal dissociation rates for syn-CH3CHOO, (CH3)2COO, and CH3CH2CHOO in the high pressure limit over the temperature range from 200-700 K. Within the 200 to 350 K temperature range, the high pressure rate constants for unimolecular decay are reasonably well reproduced by modified Arrhenius expressions given previously for syn-CH3CHOO and (CH3)2COO,3 and kd(T)= 2

3.59 × 10-70 T26.8 exp(3815/T) s-1 for CH3CH2CHOO. Over the 350 K to 1000 K temperature range, the high pressure rate constant predictions are reasonably well fit by expressions given previously for syn-CH3CHOO and (CH3)2COO,3 and kd(T)= 0.783 T3.77 exp(-4850/T) for CH3CH2CHOO. The sensitivity of the predicted thermal dissociation rate coefficients to some of the key theoretical parameters is illustrated in Figure S5. The barrier height variations by 0.4 kcal mol-1 correlate with a variation in the rate coefficient by factors of 2.1, 1.9, 1.5, and 1.2 at 200, 300, 500, and 1000 K, respectively. Meanwhile, the variation in the tunneling frequency by 100 cm-1 yields rate coefficient variations of a factor of 3.6, 1.9, 1.2, and 1.03 at 200, 300, 500, and 1000 K, respectively. Sensitivities to other theoretical parameters, such as the vibrational partition functions, may become significant at the higher temperatures (e.g., at 1000 K), but the effect of uncertainties in these quantities is likely quite modest (e.g. < 20%). These sensitivities clearly suggest considerable uncertainties in the theoretical predictions for the thermal rate coefficient. However, the energy resolved dissociation rate data provide considerable further constraints on the parameters in the theoretical model. These constraints show up in correlations between any errors in the barrier height predictions and corresponding errors in the tunneling frequencies (which serve as a surrogate for errors in the prediction of the tunneling rate). In particular, any underestimate of the barrier height, must be balanced by a corresponding underestimate of the tunneling frequency in order to obtain a predicted microcanonical dissociation rate that agrees with the experimentally observed value. Notably, such correlations between barrier height and imaginary frequency are routinely observed in ab initio electronic structure calculations. Thus, the agreement 3

between theory and experiment for the reference model should not be taken as definitively determining the theoretical model. For the CH3CH2CHOO dissociation we find that increasing the barrier height by 0.14 kcal mol-1 while also increasing the imaginary frequency by 100 cm-1, yields similar agreement between the predicted and measured microcanonical dissociation rate constants as that of our reference calculation. The effect of such variations on the predicted thermal dissociation rate coefficients is illustrated in Figure S6. The predicted thermal rate coefficients demonstrate sensitivities of factors of 2.8, 1.5, 1.00, and 1.04 at temperatures of 200, 300, 500, and 1000 K, respectively. The minimum in this sensitivity near 500 K indicates that at this temperature the maximum contribution to the thermal rate arises from an energy corresponding to that of the microcanonical rate constant determinations.

4

Figure S1. Intrinsic reaction coordinate (IRC) showing the reaction pathway from the minimum energy configuration of the most stable conformer A of the CH3CH2CHOO Criegee intermediate to the transition state leading to OH products. The IRC is displayed as a function of distance between the terminal oxygen and α-hydrogen of the CH2 group (ROH). Representative structures along the IRC path are indicated (red) and four coordinates are specified: CH2 carbon to terminal oxygen distance (RCO), α-hydrogen to CH2 carbon distance (RCH), CCCO dihedral angle (τCCCO) and terminal methyl group torsional angle (τCH3).

5

Figure S2. RRKM rates and corresponding lifetimes calculated using Eckart (solid lines) and SCTST (dashed lines) models for tunneling as well as without tunneling (dotted lines) are shown for conformers A (black) and B (blue) of CH3CH2CHOO at various excitation energies. The experimental rates and corresponding lifetimes observed at two excitation energies (gray symbols) are also shown.

6

Figure S3. Thermal dissociation rates predicted from master equation modeling of the unimolecular decay of syn-CH3CHOO (red), (CH3)2COO (blue), and CH3CH2CHOO (orange) to OH products in the high pressure limit as a function of altitude and corresponding temperature in the Earth’s troposphere.

7

Figure S4. Thermal rates predicted from master equation modeling of the unimolecular decay of syn-CH3CHOO (red), (CH3)2COO (blue) and CH3CH2CHOO (orange) to OH products in the high pressure limit over a wide range of temperature from 200-700 K.

8

Figure S5. Sensitivity plot of thermal rate coefficients predicted from master equation modeling of the unimolecular decay of CH3CH2CHOO in the high pressure limit. The increase/decrease in rate coefficient resulting from decreasing/increasing the barrier height (E0) by 0.4 kcal mol-1 is shown by the dashed/dotted red lines, while the increase/decrease in rate coefficient resulting from increasing/decreasing the imaginary frequency (νimag) at the transition state by 100 cm-1 is shown by the dashed/dotted blue lines.

9

Figure S6. Sensitivity plot of thermal rate coefficients predicted from master equation modeling of the unimolecular decay of CH3CH2CHOO in the high pressure limit. The increase/decrease in rate coefficient resulting from decreasing/increasing the barrier height (E0) by 0.14 kcal mol-1, while simultaneously decreasing/increasing the imaginary frequency (νimag) at the transition state by 100 cm-1 is shown by the dashed blue and dotted red lines, respectively.

10

References 1 F. Liu, J. M. Beames, A. S. Petit, A. B. McCoy, and M. I. Lester, Science 345, 1596 (2014). 2 F. Liu, J. M. Beames, and M. I. Lester, J. Chem. Phys. 141, 234312 (2014). 3 Y. Fang, F. Liu, V. P. Barber, S. J. Klippenstein, A. B. McCoy, and M. I. Lester, J. Chem. Phys. 144, 061102 (2016).

11