Recovery of gold from spent matrices using supercritical carbon dioxide
Van Zyl, Pieter Gideon
MetadataShow full item record
The feasibility of recovering gold preloaded onto activated carbon by supercritical carbon dioxide (sc-CO2) was investigated in view of the actuality of a more environmentally friendly process for the retrieval of gold from real-world matrices (ion-exchange resin, cellulose, activated carbon) than the harsh elution methods currently employed. Several complexes of gold(I) and gold(III) were synthesised and characterised (AA, uv/visible), and their solubility in sc-CO2 investigated, to identify a suitable compound for loading onto activated carbon and subsequent retrieval by sc-CO2. KAu(CN)2 was successfully loaded onto activated carbon. The mass percentage of gold relative to that of other elements on the carbon surface was monitored (SEM-EDS) before and after exposure to sc-CO2 using a laboratory-size supercritical extractor of the latest design. A decrease in ca. 3.6 % in the mass percentage of gold on the carbon surface, and the presence of gold in the resulting extracts (ICP-MS), confirmed that a small amount (< 1 %) of the loaded gold could be recovered. Since the same results were obtained regardless of whether sc- CO2 extraction was performed in static mode (to establish equilibrium conditions for solubility measurement) or dynamic mode (to warrant exhaustive extraction), it could be concluded that the limiting factor was not the solubility of the adsorbate in the fluid but the amount of removable gold in the outer layer(s) of the adsorbed material on the surface of the activated carbon. The extent of gold recovery was improved to ca. 10 % of the loaded gold by entraining sc-CO2 with tributyl phosphate (TBP) or TBP-HNO3 (a Lewis-acid-Lewis-base-adduct supposed to improve recovery by oxidising gold(I) to gold(III) species), but the results revealed, just as with pure sc-CO2, that there was a surface related limitation and not a TBP related limitation to the amount of recoverable gold. This conclusion could be substantiated by enlarging the surface either by larger samples of loaded carbon or by crushing samples of loaded carbon to very small particles, and in both cases correspondingly higher recoveries were obtained. The influence of different variables on the recovery of KAu(CN)2 from activated carbon by TBP-HNO3 entrained sc-CO2 was studied by surface response analysis based on a statistical design. The independence of density and thus solvent strength suggested that sc-CO2 was not the primary extractant, and a comparison between gold recovery in the absence and presence of sc-CO2 revealed that the supercritical fluid neither improved nor deteriorated the extent of recovery by TBP-HNO3. The slight positive effect of temperature allowed a small positive enthalpy of recovery to be calculated which could be interpreted in terms of physical desorption of gold from the carbon surface. The recoverable gold probably resided in the outer or secondary layer(s) attached to the primary monolayer on the carbon surface from where it could be desorbed by ion-dipole interaction between opposite charges (Au(CN)2- and positive end of polar TBP), explaining why the process was found to be surface-limited. The pore size distribution of the activated carbon was determined (mercury porosimetry) to find an explanation why gold was removed only from the surface. Although the activated carbon consisted for 95 % of micropores (8 - 100 A), the molecular diameter of Au(CN)2.pTBP was found to be ca. 20 A so that it could easily exit a micropore and thereby warranted that, in principle, gold could be removed from the inner pore structure as well. In the course of investigating the central hypothesis, valuable additions to existing knowledge were achieved. The spectral data acquired for a few gold complexes either confirmed existing literature data or represent new data published for the first time. The uv-visible spectra of the complexes were not only recorded but also interpreted in terms of different types of transitions (d-d, charge-transfer, ∏* <-- ∏, ∏* <-- n). The solubility in sc-CO2 of an organophosphorous compound (TBP) was determined with the available supercritical extractor using a self-invented procedure which prevented liquid samples from being swept away instantaneously by sc-CO2 on entering the extraction thimble. The solubility so obtained was inexcellent agreement with values reported in the literature for similar type compounds at comparable conditions. The proved solubility of 1.10-phenanthroline in sc-CO2 opened up the possibility of purifying a phenanthroline complex of gold(III) containing some unreacted ligand as an impurity by sc-CO2 instead of recrystallising the substance from solution. The investigation drew the attention to the decisive role played by activated carbon in limiting the recovery of gold regardless of which extractant (sc-CO2, TBP or TBP-HNO3) was used. This encourages investigation of alternative matrices (ionexchange resins, cellulose membranes) with weaker attractive forces in future if an efficient gold recovery process need to be developed.