A standard pyrolytlc graphite L'vov platform was compared to one treated by soaking In a 10% thorlum nitrate solution.The half-wldhs d fwnace atomk absorptbn peak profiles for Ai were reduced by as much as a factor of 4 by the modlfled surlace, and the maxlmum pennWble drarrlng temperature was extended by more than 800 O C . The peak area preckh (1.5% relative standard devlatbn for 200 pg of AI) and sen-sltivHy were comparable to those obtained with untreated platforms. Thls behavlor was stlli observed afler 500 flrlngs of a treated platform at 2450 O C .Improvements in trace-metal analyses using graphite furnace atomic absorption spectrometry (GFAAS) have been hindered by the lack of fundamental knowledge concerning the high-temperature chemistry taking place in the furnace itself. Due to this uncertainty many analysts have resorted to an empirical approach in their attempts to modify the behavior of analyte in the furnace. Such research necessarily has a trial-and-error appearance-various reagents and conditions examined until success is reached. This has been particularly true in the area of graphite atomizer surface modification.The treatment of graphite with various metals or solutions of their salts has produced new surfaces exhibiting properties markedly different than those of the original graphite. The chief pattern that has emerged is the fact that most metals used to modify graphite are known to form metal carbides.It is not known whether the carbide-forming metals not reported in the literature as surface modifiers have yet to be tried or have been found ineffective.The most commonly reported advantage of a metal-treated surface is an increase in analyte sensitivity (1-8). It should be pointed out that in some instances various authors explicitly mentioned that peak heights were used in preference to integrated peak areas or did not mention which mode of quantification was used. It is possible in some of these cases that the reported sensitivity improvement for peak heights was in fact a peak sharpening, leaving integrated area un-changed. Theory (9) predicts that integrated peak area is the better indicator of the extent of free atom formaton. It should also be pointed out that most of the reported sensitivity enhancements were in comparison to nonpyrolytic graphite tubes, which have now been largely supplanted by pyrolytically coated graphite. In some instances the advantages indicated might not have been realized over the newer graphite.The benefits of metal-modified graphite surfaces have extended beyond signal enhancement. It was reported by Norval et al. (10) that a tungsten-or tantalum-treated surface resulting from sputtering of the metal led to an improved resistance of the graphite surface toward oxidizing acids. A similar result was found by Sotera et al. (11) for a borontreated surface. Vickrey et al. found that a zirconium carbide surface allowed organolead (12) and organotin (13) compounds to be determined by using aqueous standards in the furnace while this could not be accomplished on ...