In recent years, more and more land has become contaminated with both organic and inorganic wastes. These contaminated lands pose a threat to the environment through volatilization of contaminants into the air, leaching of contaminants into the groundwater and accumulation of contaminants in plants and animals. These sites must be remediated. Supercritical Fluid Extraction (SFE) is a potential remediation technology for treating soils contaminated with mixed wastes. Before commercialization of this technology, however, feasibility studies that involve measuring and modelling solubilities of the compounds to be extracted, must be undertaken. In this research, the solubilities of two metal chelates, copper(II)acetylacetonate (Cu(acac)2) and copper(II)thenoyltrifluoroacetonate, (Cu(tta) 2) in supercritical carbon dioxide (SC CO2) were measured using a new piezoelectric quartz crystal technique. This new technique allows simple solubility measurements 'in-situ' and is applicable to compounds that are thermally labile or compounds that exhibit low solubilities in supercritical fluids. Cu(acac)2 solubilities, expressed as a mol fraction, were measured as 3.0 * 10-6 and 1.1 * 10-5 (±10.7%) at densities of 0.4965 (T = 45°C) and 0.6267 g/cm3 (T = 40°C), respectively. Cu(tta) 2 solubilities were measured as 6.8 * 10-7 (±71%) and 3.4 * 10-6 (±57%) at densities of 0.4573 (T = 45°C) and 0.6267 g/cm3 (T = 40°C), respectively. These solubilities take into account sorption of SC CO2 by the metal chelate. The higher standard deviations in the Cu(tta)2 solubilities were attributed to surface effects. Solubilities were then modelled using a density dependant solubility parameter approach. First, different methods were tested to correlate solvent solubility parameters d1 calculated from pressure-volume-temperature (PvT) data. A new two parameter linear relationship between d1 and the supercritical fluid (SCF) density was able to best correlate the PvT calculated values of d1 . Second, a model was developed to correlate solubilities of a variety of polar and non-polar compounds, including metal chelates, in SCFs. The proposed model was able to successfully correlate the solubilities using a two parameter linear relationship or a three parameter power relationship between the solute solubility parameter d2 and the SCF density. Correlation performances averaged 20% and 12% for the linear and power fits, respectively. Some initial success was obtained in using this model to predict solubilities, including experimentally measured Cu(tta)2 solubilities, in SCFs.