A Monte Carlo Simulation for Graphite Furnace Atomization of Copper

Publisher Website
Google Scholar
Abstract
Monte Carlo simulation techniques are used to model the release of Cu from a graphite furnace atomizer during the heating pulse. The simulation allows for control of nonuniform axial temperature distribution of the furnace, surface readsorption of the analyte, and various furnace geometries and initial sample location. Spatial maps of the particles are obtained as a function of time during the simulations, and axial and radial spatial distributions of the analyte particles are presented for Cu. Spatially integrated and spatially resolved absorbance profiles from experimental measurements are also presented for comparison. Explanations of the Monte Carlo approach are given and the equations used to govern release and loss are presented. The approach relies heavily on physical/chemical constants rather than curve fitting approaches. The release of Cu is a first-order process proceeding from submonolayer coverages on the surface with an activation energy of 29 ± 3 kcal. With the use of a kinetic approach for the release and readsorption it was found that the rate of release and rate of readsorption were nearly identical. This suggests that equilibrium is approached at the gas/graphite interface for Cu under these conditions. At the peak approximately 10% of the original sample is within the furnace volume as a vapor. At 900 K/s heating rate, convective expulsion of the analyte is negligible in comparison to diffusive loss. However, in a mini-Massman-type furnace the presence of the sample introduction hole is responsible for 35% of the analyte lost. This is in contrast to the 11% predicted by area considerations alone.