A SPECTROSCOPIC, KINETIC, AND PRODUCT STUDY OF THE (CH3)(2)C(OH)CH2O2RADICAL SELF-REACTION AND REACTION WITH HO2

A flash photolysis technique was used to measure the UV absorption spe ctrum of the peroxy radical (CH3)(2)C(OH)CH2O2 formed in the (CH3)(3)C OH/Cl/O-2 reaction system and to study the kinetics of its self reacti on and reaction with the HO2 radical at room temperature and above: 2( CH3)(2)C(OH)CH2O2 --> 2(CH3)(2)C(OH)CH2O + O-2 (5a); 2(CH3)(2)C(OH)CH2 O2 --> (CH3)(2)C(OH)CH2OH + (CH3)(2)C(OH)CHO + O-2 (5b); (CH3)(2)C(OH) CH2O2 + HO2 --> (CH3)(2)C(OH)CH2OOH + O-2 (8). The spectrum of the rad ical resembles that of other beta-hydroxyl substituted peroxy radicals in form and magnitude. Use of this and other known absorption cross s ections in an appropriate chemical model of the system allowed k(5), t he branching ratio alpha (=k(5a)/k(5)), and k(8) to be derived as a fu nction of temperature (T = 306-398 K) by an iterative procedure involv ing the simulation of experimental decay traces recorded at several wa velengths: k(5) = (1.4 +/- 0.6) x 10(-14) exp[(1740 +/- 150)K/T) cm(3) molecule(-1) s(-1); alpha = 0.59 +/- 0.15 (no discernible temperature dependence over this range); k(8) = (5.6 +/- 2.0) x 10(-14) exp[(1650 +/- 130)K/T) cm(3) molecule(-1) s(-1). These expressions yield values for k(5) and k(8) of 4.8 and 14 x 10(-12) cm(3) molecule(-1) s(-1) at 298 K, confirming the phenomena both of enhanced self reaction reacti vity upon beta-OH substitution and of a large rate coefficient for the reaction of all greater than or equal to C-2 peroxy radicals with HO2 . Product studies of reaction 5, using an FTIR-smog chamber system, co nfirmed the assumed reaction mechanism in 700 Torr of air at 296 K, na mely the unique formation of the peroxy radical of interest, rapid dec omposition of the alkoxy radical (formed in reaction 5a) through C-C b ond scission, and subsequent reaction with O-2 to yield formaldehyde a nd acetone (and HO2). A similar fate for the (CH3)(2)C(OH)CH2O radical is expected as part of the degradation of tert-butyl alcohol (TBA) an d isobutene under tropospheric conditions. Furthermore, the product di stribution results, when combined with the extrapolated k(5) and k(8) values, allow alpha be determined as 0.60 +/- 0.07 at 296 K, consisten t with the value obtained from the flash photolysis study. As part of the smog chamber work, a relative rate technique was used to measure t he rate coefficient for the reaction of Cl atoms with (CH3)(2)C(CH2Cl) OH as (9.2 +/- 1.1) x 10(-12) cm(3) molecule(-1) s(-1) at 296 K.