- Title
- A fundamental study on membrane integrated chemical looping air separation process
- Creator
- Paymooni, Khadijeh
- Relation
- University of Newcastle Research Higher Degree Thesis
- Resource Type
- thesis
- Date
- 2016
- Description
- Research Doctorate - Doctor of Philosophy (PhD)
- Description
- Oxygen constitutes 30% share of the global industrial gas market and is the second largest-volume chemical produced in the world after sulfuric acid. Commercial applications of oxygen can be found in industry sectors as diverse as metallurgical industry, chemical synthesis, glass manufacturing, pulp and paper industry, petroleum recovery / refining, and health services. Advanced power generation systems, such as integrated gasification combined cycle (IGCC), Oxy-fuel combustion and solid oxide fuel cells, SOFC represent emerging markets for oxygen. Oxygen is commonly produced at industrial scales by air separation using cryogenic distillation and adsorption based technologies. Advanced technologies such as membrane separation (e.g. ion-transport membrane, ITM) and in-situ air separation are also being developed for small-volume point-of-use oxygen generation. While conventional cryogenic and adsorption air separation methods are matured technologies, their energy intensity and high costs can no longer be tolerated under the current economic, energy, and environmental crises. Membrane separation methods whilst less energy intensive remain expensive due to challenges associated with their fabrication, installation and integration. In view of the above, a team of researchers led by Prof Moghtaderi at the University of Newcastle devised the Chemical Looping Air Separation (CLAS) process and its variants including: (i) Integrated Chemical Looping Air Separation (ICLAS) process for oxy-fuel applications and (ii) Redox Energy Storage (RES) process for thermo-chemical energy storage. Depending on the operating temperature and frontend conditions, the energy input into the CLAS process is 50% - 80% lower than that required for the cryogenic process and, as such, the CLAS family of processes has a relatively small energy footprint. CLAS works in a cyclic fashion by continuous recirculation of metal oxide particles between a set of two interconnected reactors, where oxidation (O₂ coupling) and reduction (O₂ decoupling) of carrier particles take place, respectively. In the original version of the CLAS process the reduction half cycle is carried out in presence of steam so that the desired oxygen product can be obtained by condensing out the steam. However, the production and condensation of steam is energy consuming and if it can be replaced by an alternative method the overall energy footprint of the CLAS process can be further reduced. Motivated by this, Prof Moghtaderi and his team turned their attention on to yet another variant of CLAS named Membrane Integrated Chemical Looping Air Separation (MICLAS) where the oxygen product from the reduction reactor is separated from steam or other suitable reducing agents (e.g. N₂) using an oxygen transport membrane (OTM) system. The combination of the OTM and CLAS in the MICLAS process creates an air separation platform which is far more cost effective than the standalone OTM based processes because of the smaller volume of gases involved, thereby, smaller physical dimensions and lower capital / operating costs. The principal vision in this PhD project was to determine the fundamental science underpinning the operation of the above integrated membrane systems using a combined theoretical and experimental approach. While both steam and nitrogen compatible OTMs studied theoretically, the primary focus of the experimental investigations was on the use of OTMs compatible with nitrogen as past experience has shown that by and large steam detrimentally impacts on the operation of membranes suitable for oxygen transport. As part of the theoretical component of the project the feasibility of integrating an OTM into the CLAS process was examined using Aspen Plus simulations, in which the net amount of energy required for each process was determined. An empirical model was then developed for calculating the oxygen permeation fluxes of the membranes using both simulation results and literature-based experimental data. It was found that the energy saving of MICLAS over the conventional CLAS was approximately 30% if a 100% oxygen recovery was assumed. When the actual oxygen recovery data were employed the energy saving of MICLAS over CLAS was 12% for BSCF type perovskite membranes reported in the literature and 16% for BSCF and 22% for LSCF type perovskite membranes specifically developed as part of this PhD study. The experimental component of the project was designed to address several important research questions fundamental to the effective operation of the MICLAS process. Of these, an in-depth understanding of the mechanisms underpinning the particle deposition on a membrane and its impact on the membrane performance (i.e., the oxygen flux across the membrane and thereby the reactions taking place within the CLAS system) was the first research question examined in this study. For this purpose, a fluidised bed chamber was designed and constructed to study particle deposition on surrogate ceramic membranes. Images of the surface coverage of each membrane were taken with a Phantom V5 high speed camera and analysed using Image J software. Preliminary deposition experiments using the experimental setup with the Perspex fluidised bed created a severe electrostatic charging issue which led to the redesign and construction of the fluidised bed chamber. The modified deposition setup consisted of a combination of a fluorine-doped tin oxide (FTO) coated glass window and a complete metallic body machined from aluminium. Grounding the experimental setup, including the new fluidised bed chamber, almost completely eliminated the electrostatic charging problem. The results of this set of experiments showed that the degree of particle deposition and electrostatic forces acting upon the particle deposition in the two phase flow were minimised. The experimental results showed that the surface coverage of the surrogate membranes with particles was less than 3%. Therefore, the impact of particle deposition which was initially considered to be detrimental to the effectiveness of the membrane was found to be insignificant. The other research questions addressed in the experimental component of the project were related to the effectiveness of membrane materials and identifying the most suitable candidate for MICLAS based processes. In this context, experimental studies were directed toward the synthesis of two types of perovskite membranes and their characterisation. These were namely Ba
0.5Sr0.5Co0.8Fe0.2O3−δ (BSCF5582) and La0.5Sr0.5Co0.8Fe0.2O3−δ (LSCF5582) which were prepared at different sintering temperatures and characterised. The BSCF5582 membranes were also synthesised using a modified Pechini method in collaboration with the FIMLab at the University of Queensland. The BSCF5582 and LSCF5582 membranes were sintered in the temperature ranges of 1000 °C - 1150 °C for 8 h and 1050 °C - 1350 °C for 5 h, respectively. The sintered membranes were characterised using an array of complementary techniques to determine the optimum sintering temperature in order to optimise their desired properties and oxygen permeability. The surface morphology of the membranes were examined using field emission scanning electron microscopy, which revealed that the segregation of the cobalt oxide phase occurring at the sintering temperatures of 1250 °C and 1100 °C for the LSCF5582 and BSCF5582 membranes, respectively; impacts on the oxygen permeability of the membranes. The sintering experiments also revealed that the sintering temperatures of 1350 °C and 1150 °C were impractical for the sintering of LSCF5582 and BSCF5582 membranes due to an excessive liquid phase forming at these temperatures. The phase and crystal structure of the membranes were identified using the X-ray diffraction technique. The presence of segregated phases for the membranes at high sintering temperatures was also verified using XRD and analysed using EDS techniques. The surface areas of the membranes were measured using the BET method in order to study their oxygen adsorption and desorption characteristics. A significantly lower surface area was acquired for the membranes sintered at elevated temperatures due to the formation of larger grain sizes at these temperatures. For example, the LSCF5582 fresh powder and LSCF5582 membrane sintered at 1350 °C had the highest (5.84 m².g-1) and the lowest (0.556 m².g-1) surface areas, respectively. Studies of the oxygen desorption/adsorption characteristics of the membranes showed a significantly higher oxygen adsorption rate over the oxygen desorption rate. The characterisation results demonstrated that the optimum sintering temperatures of the LSCF5582 and BSCF5582 membranes were 1200 °C and 1050 °C, respectively. The oxygen permeation experiments were conducted in three permeation cells with: Concentric Tube configuration, Double-Tube configuration and Multi-Tube configuration. The adjustable variables in the Concentric Tube setup were the pressure and flow rate of the sweep gas and the permeation temperature. The Double-Tube setup was used to determine the oxygen recovery of the membranes. However, sealing was the major challenge in the Double-Tube setup; therefore the Multi-Tube setup was designed and constructed. The adjustable parameters in the Multi-Tube setup were the flow rates and pressures of the feed and the sweep gas, as well as the permeation temperature. Oxygen permeation fluxes increased significantly with temperature due to the enhanced formation of oxygen vacancy concentrations in the crystal structure of the membranes with increased temperatures. As oxygen ions migrate through the vacancies, the oxygen permeation fluxes increased at elevated temperatures due to the increased oxygen vacancies. Considerably higher oxygen permeation fluxes were acquired using the Multi Tube setup than with the other permeation setups as the oxygen partial pressure gradient was adjustable in this particular setup. In order to investigate the feasibility of the integration of an OTM into the CLAS process, the reduction reactions of the oxygen carrier particles was initially examined using the TGA instrument and a fixed bed reactor. The equilibrium conversion of the metallic oxides and the concentrations of oxygen produced at different reduction temperatures were determined. Subsequently, the metallic oxides were packed in the vicinity of the LSCF and BSCF membranes using the Multi-Tube apparatus. The permeation fluxes and recoveries of oxygen obtained with the membranes were measured in the course of the reduction of the metallic oxide particles containing 17 weight percent active CuO. Finally, in a set of experiments, hypothetical oxygen carriers containing 10%, 15% and 21% oxygen concentrations were considered as future oxygen carriers for the CLAS process. The permeability and oxygen recovery of the LSCF5582 and BSCF5582 membranes were measured under simulated experimental conditions using hypothetical oxygen carriers. A mixture of oxygen and nitrogen was used to simulate the oxygen content of the hypothetical oxygen carriers. The experimental results showed that significantly higher oxygen permeation fluxes and recoveries were obtained for membranes at a 10% oxygen concentration, compared with a 21% oxygen concentration. In summary, it has been demonstrated that the OTM integrated CLAS process, even with the existing oxygen carriers which contain 0.3% oxygen, is a viable technology that can be usefully applied to a future air separation system. - Subject
- oxygen; looping air; integrated chemical looping air; separation process; gas
- Identifier
- http://hdl.handle.net/1959.13/1322432
- Identifier
- uon:24584
- Rights
- Copyright 2016 Khadijeh Paymooni
- Language
- eng
- Full Text
- Hits: 707
- Visitors: 1443
- Downloads: 358
| Thumbnail | File | Description | Size | Format | |||
|---|---|---|---|---|---|---|---|
| View Details Download | ATTACHMENT01 | Thesis | 15 MB | Adobe Acrobat PDF | View Details Download | ||
| View Details Download | ATTACHMENT02 | Abstract | 446 KB | Adobe Acrobat PDF | View Details Download |