Removal of carbon dioxide from gas streams with metal oxides at high temperature

At present over 85 percent of world energy demand is supplied by fossil fuel combustion (coal, oil, natural gas). Fossil fuel combustion is always associated with the emission ofC02, a major greenhouse gas contributing significantly to global warming. During PC combustion, the dominant power generation technology, a typical 500 MWe conventional power station emits approximately 10,000- 12,000 t of C02 per day. A typical flue gas composition in a PC combustor includes 15-16 vol % C02. The flue gas is hot, ~400°C after the economizer, and is at atmospheric pressure. Any process that depends on large pressure gradients and/or cooling of the entire flue gas stream for C02 recovery is likely to be expensive. All of the current C02 separation technologies necessitate the use of relatively low temperatures and/or high pressures to achieve sufficient sorption capacities. Reaction based processes, involving formation of carbonates from metal oxides at ~400°C in the presence ofC02, are one possible, economical approach for separating C02 from the hot flue gas. Once the metal oxide has reached its ultimate conversion, it can be thermally regenerated to the metal oxide and C02 by heating it beyond the carbonation temperature. Thus a cyclic CDR (carbonation/decarbonation reaction) process to remove and concentrate the C02 from high temperature flue gas streams should be possible. A variety of prospective metal oxides identified in the literature survey were screened by thermogravimetric analysis for carbonation -decarbonation reactivity. MgO was considered to be the most suitable on the basis of its reasonable C02 capacity and low decarbonation temperature. Pure MgO reacted with C02 even at room temperature but the carbonation temperature range for MgO doped with Cs or K increased to the range required for removing C02 from flue gas. Cs doped MgO had a higher C02 capacity than K doped MgO at 200-400°C so that further work was restricted to Cs/MgO. Comparison of Cs forms used for doping showed that Cs2C03/MgO had a higher C02 capacity than CsOH/MgO so that Cs2C03/Mg0 was used in further experiments. In order to remove sorbed gases and water the Cs/MgO (Cs2C03/Mg0) sample needed xxii to be pre-treated and it was found that 550°C was the optimum pre-treatment temperature for Cs/MgO. Among the different mixture compositions ranging from 10mol% to 50mol%, 15-20mol% Cs/MgO showed the maximum CO2 uptake when all the samples were kept in CO2 at their maximum carbonation temperature. For further study on large scale separation 20mol% Cs/MgO was chosen because the durability/pellet forming ability of 20mol% would be expected to be greater than that of 15mol% Cs/MgO. The 20mol% Cs/MgO sample needed to be pre-treated the first time at 550°C in Ar. Subsequently just changing the temperature and gas, i.e., carbonation at 350 °C in C02 and decarbonation at 550 °C in Ar, gave nearly reproducible weight changes provided that the sample was not cooled i.e the could be cycled reproducibly in the TGA. Cooling the sample will change its C02 capacity and it would be necessary to pre-treat before every cycle. The procedure of making 20mol% Cs/MgO gave product reproducibility in the sense that different batches showed the same C02 uptake capacity. Durable pellets with similar CO2 capacity to the powder were made by pressing in an IR press and crushing and on a large scale for GSR experiments, by extrusion. The effect of water was tested by adding 1.2 vol% H2O to the CO2 stream. Adding the water had no effect on the CO2 uptake capacity. When MgO was mixed with CS2CO3 to prepare the sample Cs/MgO by wet impregnation, during wet impregnation MgO was completely converted to Mg(OH)2 and CS2CO3 was converted to CsOH. This was deduced from stoichiometric calculations and TGA thermograms supplemented by XRD analysis. The same techniques suggested that during pre-treatment at 550°C in Ar for 1 hour, Mg(OH)2 decomposed to MgO and CsOH remained unchanged but could not exclude decomposition ofCsOH to CS2O. XPS showed that the pre-treated sample contained MgO and Cs20. The 20mol% Cs/MgO pretreated sample was subjected to carbonation at 350°C (in 90% CO2/10% Ar) for 1 hour. In the TGA % mass increased by up to -12% by taking up CO2. Stoichiometric calculation, XRD and XPS analysis showed that some of the MgO and CS2O took part in carbonation. Crystalline CS2CO3 was identified by XRD. xxiii Both MgC03 and CS2CO3 were identified by XPS. The MgC03 must have been amorphous because it was not detected by XRD. Some unconverted MgO and CszO were detected by XPS. The carbonated sample was decarbonated in Ar at 550°C for 1 hour, The TGA/ XRD showed that its weight decreased almost to the original level of the pre-treated sample and that the original concentration o fMgO was restored. XPS showed that Cs2C03 decomposed to CS2O, which must have been amorphous because it was not detected by XRD. 20mol% Cs/MgO sorbed more CO2 by itself than as a dispersion on SBA-15, possibly because some Cs/MgO filled the pores of SBA-15 so that formation of carbonates blocked the entrance of further C02 into the pore. The carbonation temperature of Cs/MgO dropped from 350°C to 30°C when dispersed on SBA-15 so that the dispersion could not be used to remove CO2 from hot flue gas. A gas separation rig (GSR) for larger scale experiments was set up and preliminary problems were overcome. It was successfully tested with an adsorbent of known properties, zeolite 13X. The Cs/MgO sample had to be pelletized to be used in the GSR, and this could be done without any added binder. It was possible to load/use a large amount of pelleted sample made by extrusion (62g) for C02 separation in the GSR. The results obtained from the GSR were similar to those from the TGA. 20mol% Cs/MgO could be cycled reproducibly, whether as powder or pellets in the TGA or as pellets in the GSR. The CDR cycles in the GSR gave similar results to the CDR cycles in the TGA. 20mol% Cs/MgO sorbed -2.5moles/kg CO2 in a 90%CO2 atmosphere, ~2 moles/kg in a 50% CO2 atmosphere and ~1.2 moles/kg in a 20%CO2 atmosphere at 350°C. The isotherm was of Langmuir form, as often found for chemical sorption. In both the GSR and the TGA decarbonation at 550°C in each cycle always brought the sample back to its starting weight. In summary a sorbent was produced that can successfully separate C02 at high temperatures and atmospheric pressure. The sorbent can be produced reproducibly and was cycled repeatedly on a 60g scale without losing its activity.