HCl-Catalyzed Decomposition of Di-Tertiary Butyl Peroxide in the Gas Phase

The homogeneous, gas‐phase decomposition of di‐tertiary butyl peroxide (dtBP), catalyzed by small amounts (1–16 mole %) of HCl, has been studied in the temperature range 90°—120°C and about 50 mm Hg pressure. Under these conditions, the rate of the uncatalyzed process is negligibly slow. The principal products may be represented by the two stoichiometric equations: $$\begin{array}{c}\mathrm{dtBP}\to \mathrm{isobutene\; oxide}\mathrm{+t}\mathrm{-BuOH}\text{(20\%}\mathrm{-}\text{50\%)}\\ \to \mathrm{isobutene\; oxide}+\mathrm{acetone}+{\mathrm{CH}}_4\text{(80\%}\mathrm{-}\text{50\%).}\end{array}$$ Very small amounts of C₃H₈, CO, isobutyraldehyde, and C₂H₆ are also observed. HCl is not consumed. The over‐all kinetics may be represented by a simple scheme in which t‐BuO and Me radicals abstract H from HCl, and the HCl is regenerated by Cl‐atom attack on peroxide. Acetone and isobutene oxide (i‐BuOx) act as inhibitors for the Cl‐atom chain, yielding the inert CH₃COCH₂ and Me₂CCHO radicals, respectively. The latter is probably responsible for a slow radical isomerization of i‐BuOx to the isobutanal. This reaction provides a very fast isomerization path for the conversion of epoxides to aldehydes. A derived rate law consistent with the data is given by $$-\frac{\mathrm{d}(\mathrm{dtBP})}{\mathrm{dt}}\sim \frac{{\text{2k}}_1{k}_5(\mathrm{dtBP}{)}^2}{k_6(\mathrm{acetone})},$$ in which k₁ is the rate constant for the splitting of dtBP into two t‐BuO radicals, and k₅/k₆ is the ratio of rate constants for attack of Cl atoms on dtBP or on acetone (and/or i‐BuOx), respectively. As predicted, the rate is independent of HCl. The chain length is given by λ=k₅(dtBP)/k₆(acetone) and is equal to about 25 at 10% decomposition and is almost independent of temperature. The ratio of yields of acetone to t‐BuOH is a direct measure of the kinetics of spontaneous decomposition of the t‐BuO radical, and the data show that it is a pressure‐dependent process. Unfortunately, the scatter in the data is too great to derive quantitative step constants for the relevant energy‐transfer processes.