High‐temperature carbon dioxide capture using metal oxides

Carbon dioxide (CO2) emissions to the atmosphere are considered a significant contributor to climate change due to the activity of the carbon dioxide molecule in the infrared spectrum. This causes solar radiation to be ‘trapped’ in the earth’s atmosphere if increased CO2 concentrations are present, leading to global warming. In order to decrease CO2 emissions from stationary sources, like fossil fuel-fired power plants, and halt global warming, carbon capture and storage is proposed as a viable option. Here, a big fraction of the CO2 normally being emitted to the atmosphere is to be separated from a flue gas stream of a conventional combustor facility (known as post-combustion) or from a synthesis gas (syngas) stream obtained by gasification of the fuel before its combustion (known as pre-combustion). The characteristics of the post-combustion flue gas and the pre-combustion syngas are substantially different, so that capture technologies must be developed for their respective application. The captured CO2 can then be either used in industrial applications like enhanced oil recovery, if a market exists, or else be stored in places separate from the atmosphere, so as to not cause global warming. Underground storage, consisting of saline aquifers or depleted natural gas caverns, is among the options considered for permanent and safe CO2 storage. This project investigated the use of metal oxide-based sorbents for pre-combustion capture of CO2 at 250 to 400 °C from syngas generated in the Integrated Gasification Combined Cycle (IGCC) process. At this temperature, syngas leaves the water gas shift reactor, which is the ultimate part of the gasification system within the IGCC process. Capturing the CO2 in-situ at this temperature is considered beneficial, as it eliminates the requirement to cool or heat the gas to CO2 sorption temperature, which would come with an energy penalty. In the first part, various metal oxides were screened for their suitability to capture CO2 in the temperature range of IGCC syngas. These were lanthanum oxide, magnesium oxide, zinc oxide and cadmium oxide. Based on a literature review, it was found that their carbonates decompose at temperatures considered here. A thermogravimetric screening test routine was outlined and the materials tested by exposure to carbon dioxide on a temperature ramp from 120 to 650 °C. Pure metal oxides and carbonates were considered as well as oxides doped with alkali metal compounds. It was found that cadmium oxide doped with various alkali halides sorbs and desorbs CO2 well in the screening test, so that more detailed studies were proposed. A magnesium oxide/cesium carbonate composite, which was analysed in detail in a previous project and synthesised using a variety of precursors here, was screened as well and found suitable. As the latter material exhibited a comparatively poor overall CO2 sorption capacity, it was hypothesised that improvements could be made by varying the synthesis method. In the second part, cadmium oxide/alkali halide mixtures were analysed for their optimal concentration of dopant required to give maximum CO2 sorption capacity. A variation of alkali halide dopants showed that 17.5 wt% of sodium iodide (based on cadmium carbonate used as synthesis precursor) used in synthesis yields the best-performing sorbent. This sorbent was tested in detail via thermogravimetric analysis (TGA) for single-cycle and multiple cycle CO2 sorption and desorption by partial pressure variation under isothermal conditions at various temperatures. The carbonation of cadmium oxide to carbonate and its decarbonation was confirmed as the reaction mechanism via powder X-ray diffraction (XRD) and Fourier-transform infrared (FTIR) spectroscopy. XRD also showed no significant change in the unit cell parameters of cadmium oxide due to sodium iodide doping. It was shown that alkali halides are necessary promoters for the carbonation reaction, as pure cadmium oxide did not exhibit any CO2 inclusion and conversion into cadmium carbonate. For the sodium iodide-doped cadmium oxide, a stoichiometric recarbonation was achieved, resulting in a mass gain of 26 % (based on cadmium oxide and considering the presence of dopant here). It was also shown in multicyclic sorption experiments, that a significant decay in working capacity (i. e., the amount of CO2 both captured and released per cycle) occurred if the initial decarbonation of the cadmium carbonate precursor was performed in air. This could be improved by carbonate decomposition in inert gases like nitrogen or argon. Elemental analysis performed by inductively coupled plasma mass spectrometry (ICP-MS) and atomic absorption spectrometry (AAS) showed that iodine is lost from the sample, which might be attributed to oxidation of iodide to iodine. The loss of iodine promoter is considered to be the cause of a decay in working capacity of the sorbent. Water vapour addition (1 vol%) to the sorption gases lead to an increase in stability in terms of the working capacities over the number of cycles, but had no noticeable effect on the iodine loss over sorption cycles. Powder samples were considered for multicyclic sorption as well as pelletised samples. It was shown that pelletisation via mechanical compression under vacuum resulted in materials with lower working capacities than powder equivalents, being attributed to the larger pellets being less accessible to the gas than small powder grains. Pellets made from carbonate lost their mechanical strength after 25 CO2 sorption cycles, whereas pellets made in the oxide state virtually exhibited no working capacity. Introduction of SBA-15 (porous silica) as spacer made the oxide pellets more accessible to the gas so that they showed a certain working capacity. However, this again led to partial loss of mechanical strength, but to a lower extent as observed for the carbonate pellets. In the third part of the thesis, the physico-chemical properties of the cadmium oxide/sodium iodide material were studied in detail. Transmission electron microscopy (TEM) showed morphological changes if the samples underwent multiple sorption cycles. Initially, spherical cadmium oxide nanoparticles of approx. 200 nm width were observed, which partially broke into irregular pieces after 25 sorption cycles. It was also confirmed by scanning electron microscopy (SEM) and elemental mapping that sodium iodide is interspersed into the cadmium oxide as discrete particles. In-situ powder XRD showed that sodium iodide contained in the cadmium oxide/sodium halide mixture is amorphous at room temperature after initial calcination, but becomes crystalline upon heating to CO2 sorption temperature (in this case 325 °C). These results lead to the conclusion that the cadmium oxide/sodium halide mixed sorbent is not a mixed metal oxide, but rather a mixture of halide and an oxide particles. The initiation of the CO2 sorption by the dopant halide must thus be due to a mechanistic contribution of the crystalline halide, rather than the formation of a cadmium-alkali-halide-oxide mixed phase under CO2 uptake after sorption. Analysis of the exit gas from CO2 desorption in a fixed bed by mass spectrometry confirmed the loss of iodine by showing mass-to-charge values of 126 in the mass spectrum, whereas X-ray photoelectron spectroscopy (XPS) revealed two species of iodine on the material’s surface, possibly due to different oxidation states. These results support the idea that iodine might be lost during multicyclic sorption by oxidation of iodide to iodine. In the fourth part of the thesis, improvements in the sorption capacities of magnesium oxide/cesium carbonate sorbents are shown. Using a solvothermal process involving hydration of magnesium methoxide and cesium carbonate methanolic solution mixed with toluene and subsequent treatment in an autoclave at 265 °C and flash-evaporation of the solvent, higher surface areas were obtained compared to the material made from commercial magnesium oxide as substrate. The activated solvothermally-made magnesium oxide/cesium carbonate sorbent appeared chemically similar to the one made from commercial templates as shown by XRD, but exhibited a higher working capacity (5 wt% instead of 4 wt%) than the latter. TEM and elemental mapping techniques showed that the solvothermal method leads to smaller particle sizes and also showed a mostly uniform distribution of the cesium throughout the magnesium, with occasional clustering of cesium being observed. The fifth part of the thesis investigates the effect of syngas components on the sorbents. As syngas contains hydrogen, which is a strong reducing agent, it was deemed necessary to assess if the sorbents are stable if hydrogen is present during CO2 sorption. It was also analysed if hydrogen sulphide, which is a trace component in syngas, has an effect on the sorbents. It was shown that hydrogen tends to lower the working capacity of the cadmium sorbents, possibly due to reduction of the cadmium oxide to cadmium metal. Lowering the sorption temperature 20 °C below the one achieving the highest multicyclic working capacity (i. e., to 285 °C), was able to significantly reduce capacity loss in a CO2 sorption experiment containing hydrogen. The magnesium-cesium sorbent appeared stable in the presence of hydrogen. Hydrogen sulphide, however, had a significant impact on both the cadmium and magnesium-based materials by reducing their working capacities. In summary, a cadmium oxide based CO2 sorbent was developed and its physical and chemical properties examined. A magnesium-based mixed oxide developed in a previous study was improved in sorption capacity. The stability of both these materials under simulated syngas conditions was examined and future work is proposed to focus on the application of real syngas to these sorbents, with the goal of further improvements in their stability