This paper is concerned with the sequestration of mercury, cadmium, lead, sodium and potentially other volatile and semi-volatile metals by high-temperature mineral, non-carbon based dispersed sorbents. The focus here is on kaolinite and lime powders (for Pb, Cd, and Na), and on intimate kaolinite/calcite/lime mixtures (for Hg), both of which undergo morphological and chemical changes when exposed to high temperatures. These changes play critical roles in the metal capture mechanisms, initially enhancing metal capture, either through a eutectic melt on the surface, or through some other transformation, but ultimately causing sorbent de-activation through a catastrophic melt that causes pore closure. Results and interpretation from two types of experiments are presented. The first employed a 20 kW downflow combustor, a low-pressure impactor and application of the aerosol fractionation method to quantitatively determine the fraction of metal sequestered by the sorbent at the sampling point. The second employed an externally heated quartz reactor and simulated flue gases containing mercury and injected sorbent. Quantitative, rate models have been extracted from the data to describe the global reactions of dispersed kaolinite with lead, sodium, and cadmium metal vapors. They have also been derived for the global reaction of cadmium with dispersed hydrated lime, which was ineffective for Na and Pb but very effective for Cd. Qualitative results showing the effective scavenging of Pb and Cd by the intimate kaolinite/calcite/lime mixture are also presented. The intimate kaolinite/calcite/lime mixture was also shown to capture metallic mercury through two mechanisms, namely, an in-flight mechanism and one involving interactions between Hg, the sorbent and the quartz reactor walls. The in-flight mechanism had similarities to those observed previously between kaolinite, lead, cadmium and sodium, in that mineral transformations played a key role in both sorbent activation and sorbent de-activation. At a constant residence time, the in-flight mechanism exhibited a maximum effectiveness at 900 degrees C, followed by apparent de-activation. Spent in-flight sorbent showed the presence of calcium silicates and calcium aluminosilicates. The wall deposit mechanism increased in efficacy with temperature and did not show a maximum, and showed the presence of substantial amounts of captured mercury. Adding to the complexity of this process is that Hg capture by this mixture required the presence of O(2) in the flue gas, albeit at low levels of 4000 ppm. (C) 2009 Elsevier Ltd. All rights reserved.