Measurements and modeling of deposition rates from near-supercritical, aqueous, sodium sulfate and potassium sulfate solutions to a heated cylinder

Thesis (Ph.D.)--Massachusetts Institute of Technology, Dept. of Mechanical Engineering, 1998.Includes bibliographical references (leaves 365-373).In the Supercritical Water Oxidation (SCWO) process, a technology emerging for the disposal of hazardous organic wastes, organic compounds containing heteroatoms such as S, Cl and P are oxidized to the corresponding acid. In order to reduce corrosion, bases are therefore often injected into the reactor. The salts that are formed upon neutralization (sulfates, chlorides, phosphates, etc.) have low solubility in supercritical water (SCW) and consequently precipitate as solid phases. These salts can form agglomerates and coat internal surfaces, leading to plugging of transport lines and inhibition of heat transfer. The purpose of this dissertation is to develop an understanding of salt deposition kinetics and nucleation phenomena at conditions relevant to SCWO. Solubility and deposition experiments were performed with aqueous sodium sulfate and aqueous potassium sulfate solutions at elevated temperatures and pressures typical of the SCWO process. For both the solubility and deposition experiments, the test cell is a six-port chamber which was fabricated by modifying a 1.91 cm (3/4 inch) diameter Swagelok cross. One port was used to mount a 0.200 inch (5.08 mm) diameter internally heated cylinder into the center of the chamber and the remaining ports provided fluid cross flow, visual observation capability and instrumentation access. Aqueous salt solutions containing up to 10 wt% salt were pumped at about 250 bar through preheaters that brought the solution to a temperature close to that at which precipitation occurs. Inside the test cell, the heated cylinder raised the temperature of the nearby solution above this precipitation temperature, thus limiting deposition almost exclusively to the heated cylinder. Experimental deposition rate data from sodium sulfate- and potassium sulfate-containing SCW streams to the heated cylinder were obtained by removing the heated cylinder from the cell following each run and measuring the mass of salt deposited on it. The deposition rate data were obtained as a function of time and concentration of salt in the solution entering the test cell. Salt concentration and time in the deposition experiments were varied between 2 and 8 wt% and 6 and 12 minutes respectively. In the solubility experiments, the solubility temperature of sodium sulfate and potassium sulfate in water at a pressure of 250 bar was measured for salt concentrations up to 10 wt%. Natural convection dominates transport at all of the conditions investigated in the solubility and deposition rate experiments. The equations governing the transport of salt to the interface that develops between the salt layer which forms on the heated cylinder/hot finger and the solution in the adjoining boundary layer are developed in a fairly rigorous context. Then, the equations governing transport are scaled to determine a set of criteria which, when satisfied or partially satisfied, allows various terms, e.g., those accounting for the Soret effect and Dufour effect in the species and energy conservation equations respectively, to be neglected or simplified. All of the criteria are evaluated for the conditions in the deposition experiments, justifying a relatively simple set of ordinary differential equations which govern the deposition rate of salt at the salt layer-solution interface. The simplified model is numerically solved to predict the rate of transport of salt to the salt layersolution interface for all the conditions investigated in the deposition experiments as a function of time. The theoretical deposition rate predictions are then compared to the deposition rate data in the context of a sensitivity analysis. For the deposition experiments in which the sodium sulfate and potassium sulfate concentrations in the solution entering the test cell were 4 wt% or less, the theory and data compare well. In fact most, but not all, of the experimental data for these experiments fall within the bounds predicted by a sensitivity analysis accounting for uncertainties in thermophysical properties and experimentally measured variables input into the theoretical deposition rate formulation. For higher salt concentrations, however, the model underpredicts the experimental data by up to a factor of about two. It is shown that, as the concentration of salt in the solution entering the test cell increases, deposition of salt within the porous salt layer formed on the hot finger is likely to become more significant. Thus it is hypothesized that the theory and deposition data do not compare as well at higher salt concentrations because the predictive model accounts for deposition at the salt layer-solution interface, but not deposition within the porous salt layer. In the model developed for the deposition rate of salt at the salt layer-solution interface it is assumed that salt nucleation occurs exclusively at the salt layer-solution interface, i.e., there is no homogeneous nucleation of salt in the boundary layer. This assumption is validated, to some extent, by visual observations during the experiments. Additionally, a major chapter of this dissertation is devoted to nucleation modeling. A model to predict whether or not homogeneous nucleation and/or supersaturation will occur in the boundary layer formed around the salt layer-solution interface is developed. It predicts that homogeneous nucleation and/or supersaturation are unlikely, if not impossible, at the conditions investigated in the deposition experiments. This justifies the formulation used to solve for the rate of mass transfer at the salt layer-solution interface a posteriori.by Mark S. Hodes.Ph.D