Process intensification of nitrous gas absorption

Abstract

The absorption of nitrogen oxides in water has important applications in nitric acid manufacture and pollution control. The design for optimum absorption efficiency and air pollution control has made necessary the installation of large reaction chambers and absorption towers for the adequate oxidation and absorption of nitrous gases. The worldwide production of weak acid has seen the progression of the process from the use of low through medium- to high-pressure technology in the efforts of achieving a more compact construction and avoiding the need for catalytic tail-gas treatment in plants with ever increasing capacities. Even at high pressures (8 bar), absorption columns employing sieve plates can reach up to 40 m in height for large-tonnage plants, and the relatively large pressure drop across the plates at these pressures leads to high power consumption and increased costs. As the dimensions of the absorption tower are typically governed by the conditions required for NO oxidation and thermal design, intensification of the process via miniaturisation can address the issues above through the high surface area to volume ratio offered by microreactor technology. The substantial improvement in heat and mass transfer due to the increase in effective exchange surface leads to an acceleration of the slow NO oxidation reaction and the enhancement of absorption rates. In addition to the development of such novel equipment for process intensification, the flow of the process can also be structured to improve process efficiency. An interesting method would be the replacement of the nitrogen ballast typically used in industry with steam, as the concentration of the gases upon condensation can lead to improved gas phase reaction rates. Furthermore, the provision of increased residence times due to the decrease in gas velocity upon condensation also makes the process more efficient. In this way, the size of the absorber can be significantly reduced and the high capital and operating costs associated with the employment of compressors in high pressure plants can be reduced. The objective of this thesis is to gain a fundamental understanding of the complex behaviour of nitric acid production in microchannels and obtain data for the development of a model used for process design and optimisation. Experiments on the oxidation and absorption of xviii nitrogen oxides have been conducted for a wide range of nominal residence times (0.03 – 1.4 s), gas compositions (5 – 10% NO, 5 – 49% O2, 46 – 82% H2O, balance argon), system pressures (2 – 10 bar absolute), mass fluxes (1.5 – 30 kg m-2 s-1), coolant mass fluxes (66 kg m-2 s-1 and 341 kg m-2 s-1), and coolant temperatures (23 – 51ºC) in circular tubes with internal diameters of 1.4 and 3.9 mm. Absorption efficiencies of up to 99% have been achieved without the use of counter-current flow typically employed in conventional nitric acid plants. The use high steam fractions was shown to cause significant improvements in gas phase reaction rates such that the usual industrial practice of applying high system pressures to enhance the NO oxidation reaction becomes unnecessary. Absorption efficiency can also be increased by increasing system pressures, but there are certain limits to which this can be done; a decline in performance may result when pressures are increased sufficiently high such that mass transfer becomes limiting. In addition to decreasing the tube diameter, increasing both the NO concentration and cooling duty also led to improved nitric acid yields. A simple model of condensing two-phase shear-driven annular flows, in which both laminar and turbulent regimes are valid and the vapour-liquid interface is assumed to be smooth, have been constructed and compared against experimental data. The model qualitatively captures most of the effects observed, but the presence of uncertainties in model parameters and the use of particular assumptions on the flow pattern and structure of the interface had to be compensated for through the use of a model fitted parameter iAθ. Larger corrections to the model were required in cases where the fluid was tending towards slug or plug flow, such as systems employing high H2O/NO ratios, since the interfacial area between vapour and liquid would be larger than that obtained if annular flow was assumed to occur under the same conditions. Higher values of iAθwere also found to give better fit to the experimental data at short nominal residence times (< 0.10 s) for absorption carried out under high system pressures, high oxygen partial pressures or high NO partial pressures, presumably due to incorrect representation of the overall heat and mass transfer flux under these conditions, among other things such as the parameter uncertainties, the presence of interfacial waves and the possibility of a flow regime transition from annular to intermittent flow. On the other hand, interfacial area multipliers less than unity were better suited to larger xix channels due to the possibility of flow stratification which acts to decrease the interfacial area and hence the nitric acid yield. The predictions of the model were subsequently used to determine the operating conditions optimal for the production of nitric acid in microreactors on a larger scale. In most of the cases considered, the pressure drop across the absorber length was found to be relatively small, hence much smaller channels can be utilised for increased absorption efficiency without considerable loss in pressure. It was also shown that most of the heat liberated near the inlet of the absorber stems from the release of latent heat of condensation, while chemical reactions account for most of the heat released downstream of the reactor. The absorption volume required for the commercial production of nitric acid in microchannels was compared against that typically employed by current industrial absorbers. The volume of the microreactor system was found to be about 2 orders of magnitude smaller than its larger counterpart. Although additional volume may be required for distillation of the weaker acid produced from the smaller system, substantial reduction in plant size can still be achieved since the volume of the cooler-condenser was excluded from the industrial plant calculations while the physical and chemical reactions involved in the cooler-condenser are already inherent in the microreactor system. In summary, the results of the experiments and model simulations have demonstrated that the absorption of nitrous gases in microchannels with the use of a steam ballast and close-to-stoichiometric quantities of oxygen can lead to intensification of the process, thus presenting an opportunity for a paradigm shift in nitric acid production.