The feasibility was studied of a process for removal of sulfur oxides from flue gas using a reversible dry absorbent. Sulfur dioxide in the flue gas is first oxidized to SO3 and then absorbed on a fluidized solid. The sorbent consists of Na2SO4 impregnated on an inert carrier and takes up SO3 through the formation of sodium pyrosulfate. Oxidation of SO2 was not studied, however, catalytic oxidation was used to prepare an SO3 feed stream for the experimental program. An experimental program was carried out to develop kinetic and equilibrium data for the reaction of Na2SO4 supported on silica gel with SO3. The data were obtained under conditions suitable for a cyclic absorption and desorption process. Nitrogen oxides and water vapor in the concentration normally found in flue gas were found to have little effect on the SO3 absorption rate. Desorption of SO3 at a concentration of 6% in the gas was found to be very rapid at 930 F, and essentially quantitative recovery of SO3 was obtained at 1020 F. A process design was developed based on the reaction rates and temperature conditions determined in the experimental program. Economic studies pointed up the importance of using an inexpensive catalyst and sorbent and minimizing the loss of these materials. Therefore, the concept of using a combined catalyst and sorbent consisting of fly ash impregnated with V2O5 and Na2SO4 was developed. Preliminary experiments carried out using a fly ash based catalyst and fly ash sorbent were promising. The desorbed SO3 can be absorbed in sulfuric acid to produce concentrated H2SO4 or oleum for sale, or by maintaining a slightly reducing atmosphere in the desorber, a concentrated stream of SO2 (44%) can be produced. The estimated capital and operating costs show that the process is competitive with alternative processes. (Author)