Open pore polyurethane foam (PUF) ran be used effectively as a substrate for conventional inertial impactors with both high particle collection efficiency and minimal vaporization of semi-volatile particle components. The collection characteristics of PUF as an impaction substrate were studied as a function of PUF density, Reynolds number, impaction substrate diameter, and nozzle-to-plate distance. The conventional impaction substrate of the PM2.5 Harvard Impactor sampler was replaced with the PUF substrate, The use of PUF resulted in significant changes in the collection efficiency curve, with the 50% cut-off size (d(50)) decreasing from 2.48 to 1.12 mum, corresponding to root Stk = 0.24, While the theory for conventional hat impaction substrates accurately predicts d(50) values (at root Stk = 0.49), for PUF substrates this same theory predicts d(50) values much larger than the experimentally determined values. After the collision of the particles with the PUF, a greater fraction of their excess kinetic energy may be absorbed by the substrate than is absorbed by conventional substrates, reducing the amount of particles that would otherwise bounce off or be reentrained, Furthermore, qualitatively similar results were obtained for PUF densities between 1.9 x 10(4) and 5.0 x 10(4) g/m(3), Results obtained for varying Reynolds numbers also suggest that the difference in collection efficiency curves between PUF and oil-coated substrates is due to different flow patterns, In addition, tests shelved that the overall impactor performance was better for larger impaction plate diameters for both PUF and conventional substrates. Finally, significant distortion of the collection efficiency curve was observed for larger nozzle-to-plate distance.