Hexafluoroisopropanol (HFIP) is a key intermediate used in the production of Sevoflurane, an inhalation anesthetic. Typically, HFIP is produced by the hydrogenation of hexafluoroacetone (HFA) in a batch or a continuous vapor phase reactor in the presence of a supported nickel catalyst. The batch process has limited production capacity due to long cycle times and the reaction rate is limited by the heat removal capacity. Furthermore, a batch reactor contains substantial inventory of HFA, a toxic and volatile chemical, necessitating stringent process hazard and containment requirements. While an improvement over a batch process, vapor phase hydrogenation process also has limitations. The latter process has a high temperature gradient across the catalyst bed due to the heat released by the exothermic hydrogenation reaction. This temperature rise can cause catalyst bed hot spots, resulting in byproduct formation, including the formation of hydrofluoric acid, and catalyst deactivation. Based on the drawbacks encountered in both of these processes, it was deemed necessary to design a process, that would reduce mass transfer limitation, eliminate temperature gradient across the catalyst bed, and improve the efficacy and life of the catalyst.

Several technology options were considered for the continuous hydrogenation of HFA including, various three-phase (gas-liquid-solid) reactors (such as trickle bed, slurry bubble column, continuous stirred tank and monolith reactor configurations). After careful evaluation of all the options, a novel approach was selected wherein the hydrogen absorption from gas into the liquid phase was decoupled from the reaction of the dissolved H₂ and HFA on the catalyst in contact with the liquid. This allowed elimination of the H₂ gas/liquid mass transfer limitation and simplified the scale-up of the reaction zone from the conceptually difficult three-phase gas-liquid-solid trickle bed to a two-phase gas-liquid bed with liquid phase saturated with H₂. The reactor design included a recycle loop comprising of an autoclave acting as a hydrogen absorber, followed by a fixed bed reactor, as shown in the schematic below. The reactor is fed directly from the autoclave with a hydrogen-saturated liquid stream. The recycle stream consists primarily of the product HFIP, a small concentration of HFA, and excess dissolved hydrogen. The low conversion per pass across the reactor allows stable adiabatic reaction, with a small temperature rise and elimination of hot spots. Furthermore, the low inventory of toxic HFA in the reactor also provides an inherently safer process.

For scale-up estimates, intrinsic reaction kinetics were obtained initially from slurry laboratory gas-liquid-solid autoclave data without intraparticle and liquid-solid diffusion limitations, but accounting for gas-liquid H₂ mass transfer. The intrinsic kinetics were subsequently updated with data from a lab reaction system with fixed bed catalyst of the same catalyst composition as that used in the slurry system, emulating the commercial recycle scheme, and accounting for the effectiveness factor resulting from intraparticle diffusion limitations and liquid-solid mass transfer. A detailed VLE for the HFA-HFIP-H₂ system was also developed. Laboratory scale fixed bed pressure drop measurements were performed in a cold flow system with HFIP to validate/update pressure drop correlations with the commercial fixed bed catalyst. A comprehensive reactor system model based on Aspen Plus, including a user routine for the mass transfer limited reaction was developed to provide commercial scale estimates, including desired recycle rates, temperature, pressure, and optimal reactor geometry (height/diameter) and volume. The results were used to design and successfully start up the commercial plant with a productivity 3X that of the original batch process.