In order to confirm the accuracy of direct analysis of filter samples containing atmospheric aerosol particles collected on a polycarbonate membrane filter by PIXE, we carried out PIXE analysis on a polycarbonate membrane filter (Nuclepore®, pore size: 0.8 μm) on which was collected National Institute of Standards and Technology (NIST, USA) urban particulate matter (SRM 1648). We also investigated whether a polycarbonate membrane filter would possibly bias elemental quality determined by PIXE analysis. We found that the polycarbonate membrane filter did not bias values determined by PIXE. In the case of filter sample A in which the powder sample (urban particulate matter mixed with palladium-carbon powder (5% Pd )) was collected under dry conditions, elemental quantitative values were 80–110% relative to NIST certified and non-certified values except for Mn and Fe . On the other hand, in the case of filter sample B in which the powder sample was collected under wet conditions, elemental quantitative values were 70–110% relative to NIST certified and non-certified values except for Cl and Zn . Determined values for Mn and Fe in the case of filter sample A and Cl and Zn in the case of filter sample B were clearly lower than NIST certified and/or non-certified values, but this is believed to be attributable to the sampling procedure rather than a problem associated with PIXE analysis. Direct analysis of a filter sample containing aerosol particles by PIXE provided sufficient elemental quantity accuracy.