Method 1 (which is described in the literature, 1'2) yielded MAGSORB pellets with the highest reactivity, but uneven distribution of potassium iii resulted in Mg:K atomic ratios that varied widely among individual pellets. This introduced a large potential error into thermobalance test results. Method 2 yielded denser MAGSORB pellets with acceptable Mg:K intradlstribution, but sharply reduced reactivity. These pellets were approximately 2.4 times as dense as the Method 1 pellets, and had 73% lower pore volume. Method 3 yielded pellets with good homogeneity and reactivity close to that: of the Method 1 formulation. These pellets were 1.5 times as dense as Method 1 pellets and had 40t less total pore volume. MAGSORB pellets made by method 3 were used in most of the parametric thermobalance tests and all of the packed-bed cycle tests. MAGSORB pellets were tested for reaction with COz in a high-pressure, high-temperature thel_obalance for temperature ranges from 650"F to 950"F and system pressures from atmospheric to 600 psig. Thermobalance and differential scanning calorimeter testing of Method 1 pellets showed that COz absorption and desorption rates were negligible below a minimum temperature of about 750"F, but increased rapidly as temperature approached 800°F. Various temperature and pressure combinations for desorption of the CO2 were studied. A sweep gas was required to enact complete desorption at 800°F and atmospheric pressure, lt was found that desorption could occur without a sweep gas if reduction of CO2 partial pressure was accompanied by a temperature increase. For example, CO2 was desorbed by reducing CO2 partial pressure from 139 to 46 psi while increasing temperature from 800°to 900°F. A series of thermobalance tests was performed to show reproducibility in a standard gas mixture averaging 42 volt CO2, 42 volt He, and 16 volt H20. These tests were conducted at 820°to 830"F and 300 psig system pressure, with a CO2 partial pressure of 131 to 134 psi. These tests gave typical asymptotic weight gain curves with a mean CO2 absorption of 70.1% ± 12.2% of the stoichlometric limit after 30 minutes. The stoichiometrlc weight gain limit, based on the Mg, K, and CO 2 analyses of the MAGSORB, represented an absolute weight gain of 76.9%. The error limits given above are for 95% confidence. Thermobalance tests under similar conditions as above, but containing syngas components (H2, CO, N2), steam, H2S, and NH 3 proved that these components did not react with the sorbent or reduce its react _ity. In fact, iv sv the sorbent prepared by Method 3 showed higher CO 2 absorption rates and capacities in the presence of syngas, H2S, or NH 3, with 30-minute weight gains ranging from 86.4% to 96.7% of the stolchlometric limit. Desorption was essentially complete in less than 12 minutes with standard gas with or without HzS or NH 3, but took about 30 minutes in the syngas mixture. A packed-bed reactor was constructed to conduct cycle testing of a selected MAGSORB formulation with a 50 volt CO2/He mixture. The conditions sele...
