A novel laboratory-scale fixed-bed reactor has been developed to investigate trace metal capture on selected sorbents for cleaning the hot raw gas in Integrated Gasification Combined Cycle (IGCC) power plants. The new reactor design is presented, together with initial results for mercury and arsenic capture on five sorbents. It was expected that the capture efficiency of sorbents would decrease with increasing temperature. However, a commercial activated carbon, Norit Darco "Hg", and a pyrolysis char prepared from scrap tire rubber exhibit similar efficiencies for arsenic at 200 and at 400 degrees C (70% and 50%, respectively). Meta-kaolinite and fly ash both exhibit an efficiency of around 50% at 200 degrees C, which then dropped as the test temperature was increased to 400 degrees C. Activated scrap tire char performed better at 200 degrees C than the pyrolysis char showing an arsenic capture capacity similar to that of commercial Nor-it Darco "Hg"; however, efficiency dropped to below 40% at 400 degrees C. These results suggest that the capture mechanism of arsenic (As-4) is more complex than purely physical adsorption onto the sorbents. Certain elements within the sorbents may have significant importance for chemical adsorption, in addition to the effect of surface area, as determined by the BET method. This was indeed the case for the mercury capture efficiency for all four sorbents tested. Three of the sorbents tested retained 90% of the mercury when operated at 100 degrees C. As the temperature increased, the efficiency of activated carbon and pyrolysis char reduced significantly. Curiously, despite having the smallest Brunauer-Emmet-Teller (BET) surface area, a pf-combustion ash was the most effective in capturing mercury over the temperature range studied. These observations suggest that the observed mercury capture was not purely physical adsorption but a combination of physical and chemical processes.