Fundamentally Based Reaction Mechanism for Oxidation of Isooctane

ion of the tertiary hydrogen atom by peroxy group located at A-position, viz., IO-AQJ IO-AQCJ is the lowest energy process among the H-transfer reactions involving any isooctane peroxy radicals. The activation barrier ΔH (TS-C1A) is as low as ca.17 kcal mol (16.74, 17.09, 16.95 kcal mol, correspondingly, at B3LYP/6-31G(d), G3MP2B3, and CBS-QB3 levels of theory in agreement with the unified parameter of 17.05 kcal mol used by Curran et al. for all 77 US Combustion Meeting – Paper # RK26 Topic: Reaction Kinetics 6 memberred ring TS transitions [10]). This value is lower by 7-15 kcal mol than those for other isomerization reactions. Importantly, the difference between two predicted values at B3LYP/631G(d) single level DFT and the G3MP2B3 composite method employing the same level of geometry calculation, is only 0.35 kcal mol. In this case it is an exceptionally good agreement. This is also consonant with literature data (comparative analysis of all relevant H-transfer processes will be presented elsewhere [17]). It should be highlighted that the easiest to form IO-CQJ peroxy radical has no options involving 6-membered ring TS to isomerize, where six membered ring H atom transfer is commonly believed to be the kinetically most suitable isomerization channel for linear hydrocarbons. TSA1C (Fig.2b) included also in PE diagram in Fig.3, deals with a 7-membered TS-ring and involves higher activation energy at 24.4 kcal moldue to endothermicity. The barrier for a typical 6-membered ring TS isomerization of IO-BQJ IO-BQAJ is calculated as 25.2, 23.2, 23.7 and 22.2 kcal mol, correspondingly, at B3LYP/6-31G(d), B3LYP/6-31G(d,p), G3MP2B3 and CBS-QB3 levels of theory vs. the generalized 24.0 kcal mol estimate of Curran et al. [10]. In original version of Curran’s mechanism [10], the value of 24.4 kcal mol was used, however in the recent update [16] it is somewhat reduced to 24.00 kcal mol which according to authors speeds up substantially second O2-addition process and increases efficiency of chainbranching and OH + ketohydroperoxide formation. The above calculated values show the barriers and reaction enthalpies for the IO-BQJ, intramolecular transfer of H-atoms (cf. TS-A1C in Fig.3) are somewhat overestimated at the B3LYP/6-31G(d) level. This is particularly due to the lack of p-polarization extension functions in the basis sets for hydrogen atoms, although the energies of other types TS dealing with heavy atoms, are in fairly good agreement with higher level results (vide infra). Judging from benchmark calculations of Zheng and Truhlar [22], the B3LYP functional is of good performance in butoxyl radical reactions compared to CCSD(T)/CBS method including the core-valence correlation contributions and scalar relativistic corrections. From a wide variety of tabulated results in [22] one could extract that the stability of radicals with spin located on Oatoms is somewhat underestimated by B3LYP functional. This is in full agreement with the underprediction of well depths for association reaction R•+O2 RO2• observed here. Well Depths: The only and consistently persistent exception in performance of B3LYP/6-31G(d) method is the (rather expected) underprediction of well depths, the reaction enthalpy of initial association reaction of alkyl radicals with oxygen. Consequently, well depths were evaluated additionally at higher computational levels employing either atomization [25] or work reaction procedures. The values calculated at CBS-QB3 level using a more complex isodesmic reaction scheme [17] is in close agreement with the value arrived from the direct dissociation limit for the reaction RO2 R+O2 (34.1 vs. 34.7 kcal mol, respectively). Another systematic (additive) error in B3LYP energetics can be traced to the overestimation of the enthalpy of O2-molecule; this has been corrected with bond additivity corrections. In the PEdiagram shown in Fig.3, we show the non-corrected value in blue to be consistent with relative energies of other stationary point. Some remaining discrepancy is likely due to the underestimation of dispersion interactions and electron correlation in B3LYP calculation scheme [17]. 7 US Combustion Meeting – Paper # RK26 Topic: Reaction Kinetics 7 Comprehensive PES for IO-CJ+O2 Reaction System: As mentioned above, it is a priori expected that the initial reactions of IO-CJ+O2 intermediates, will have a dominant role in low temperature processes, especially at higher pressures. Association of an oxygen molecule with the only tertiary isooctyl radical forms a peroxy radical-adduct IO-CQJ (via TS-A1C, Fig.3). Further isomerization of this peroxy radical occurs via 5 and 7 membered ring transition states, correspondingly to form IO-CQBJ and IO-CQDJ hydroperoxy radicals via abstraction of B secondary and D primary hydrogen atoms, as well as a IO-CQAJ radical via the transfer of any one of the nine A-primary hydrogen atoms to C-peroxy radical center (TS-A1C, Fig.2). Fig.3 details a PE diagram for the formation and further transformations of the IO-CQAJ hydroperoxy alkyl intermediate. There are two additional isomerization processes (with further reaction steps) for the chemically activated IO-CQJ peroxy radical (from B and D positions) not presented here [17].