The production of pure hydrogen with simultaneous capture of carbon dioxide

Bohn, Christopher (2010-10-12)


The need to stabilise or even reduce the production of anthropogenic CO2 makes the capture of CO2 during energy generation from carbonaceous fuels, e.g. coal or biomass, necessary for the future. For hydrogen, an environmentally-benign energy vector whose sole combustion product is water, to become a major energy source, it must be produced in an efficient, CO2-neutral manner. A process, which uses a packed bed of iron and its oxides, viz. Fe, Fe0:947O, Fe3O4 and Fe2O3, has been formulated to produce separate, pure streams of H2 and CO2. The process is exothermic and has the following stages: 1. Reduction of Fe2O3 to Fe0:947O or Fe in syngas (CO + H2) from gasifying coal or biomass. This stage generates pure CO2 for sequestration, once the water has been condensed. 2. Subsequent oxidation of Fe or Fe0:947O to Fe3O4 using steam. This stage generates H2 of sufficient purity for use in polymeric membrane fuel cells. 3. Further oxidation of Fe3O4 to Fe2O3 using air to return the oxide to step (1). It was shown that reduction to Fe0:947O in step (1) gave stable yields of H2 in step (2) after 40 cycles, near those predicted from reaction stoichiometry. By contrast, reduction to Fe, rather than Fe0:947O, in step (1) gave low levels of H2 in step (2) after just 10 cycles. This demonstrates that modifying the iron oxide is unnecessary unless reduction to Fe is performed. Wet-impregnation of Fe2O3 was performed with salts of Al, Cr and Mg or with tetraethyl orthosilicate for Si to give loadings of 1-30 mol % of the additive element. The addition of Al stabilised the quantity of H2 produced when the sample was reduced to Fe. Using a sol-gel method, composite particles with diff erent mass ratios of Fe2O3 and Al2O3 were prepared. For reduction to Fe over 40 cycles, 40 wt. % Al2O3 was required to give stable conversions near 75 % of that expected from reaction stoichiometry. Prior to this research, it had been assumed that the alumina acted as an inert support. However, this was shown to be incorrect since the formation of FeO.Al2O3 was quantitatively confirmed using X-ray diffraction. The presence of the compound, FeO.Al2O3, is significant since it reduces the loss in internal surface area but binds reactive iron, two contradictory e ects for the production of H2. The production of separate streams of pure H2 and CO2 from solid fuels, lignite and subbituminous coal, was demonstrated. Pure H2 with [CO] ~< 50 ppmv and [SO2] ~~ 0 ppmv was produced from a low-rank coal, showing that the process is e cacious with an impure fuel. Contaminants found in syngas which are gaseous above 273 K apparently do not adversely affect the iron oxide material or purity of the hydrogen. Subsequent oxidation of the Fe3O4 with air, step (3), removed sulphurous and carbonaceous contaminants deposited during reduction, generated useful heat and did not lead to a decrease in the H2 yield in step (2). It is therefore recommended that step (3) be included in the process. Rates of reaction are reported for the reduction of iron oxide particles by a mixture of CO, CO2 and N2. Importantly, rates were investigated over multiple cycles. Reduction of either Fe2O3 to Fe3O4 or of Fe3O4 to Fe0:947O was found to be first-order in CO. With the particle sizes used, the rates of reduction were controlled by intrinsic chemical kinetics. Activation energies and pre-exponential factors are reported. The rates were used to simulate, satisfactorily, the reduction of a packed bed of iron oxide. The rate of reduction was doubled by the addition of 1 mol. % Ce to the granulated iron oxide. The overall rate was shown to be dependent on the active surface area of the iron oxide. A lattice Boltzmann model, which incorporates hydrodynamics, mass transport and reaction, was developed. The composition of the solid changed with time. Quantitative agreement between the model and experiments for the reduction of a single particle of Fe2O3 to Fe3O4 in CO was achieved. Additionally, the model correctly predicted a sharp front in the CO concentration for reduction of a packed bed of Fe2O3 to Fe3O4.