Protection of the human respiratory system from exposure to nanoparticles is becoming an emerging issue in occupational hygiene. The potential adverse health effects associated with particles of approximately 1-100 nm are probably greater than submicron or micron-sized particles. The performance of two models of N95 half-facepiece-filtering respirators against nano-sized particles was evaluated at two inhalation flow rates, 30 and 85 l min(-1), following a manikin-based protocol. The aerosol concentration was measured outside and inside the facepiece using the Wide-Range Particle Spectrometer. Sodium chloride particles, conventionally used to certify N-series respirators under NIOSH 42 CFR 84 regulations, were utilized as the challenge aerosol. The targeted particle sizes ranged from 10 to 600 nm, although the standard certification tests are performed with particles of approximately 300 nm, which is assumed to be the most penetrating size. The results indicate that the nanoparticle penetration through a face-sealed N95 respirator may be in excess of the 5% threshold, particularly at high respiratory flow rates. Thus, N95 respirators may not always provide the expected respiratory protection for workers. The highest penetration values representing the poorest respirator protection conditions were observed in the particle diameter range of approximately 30-70 nm. Based on the theoretical simulation, we have concluded that for respirators utilizing mechanical filters, the peak penetration indeed occurs at the particle diameter of approximately 300 nm; however, for pre-charged fiber filters, which are commonly used for N95 respirators, the peak shifts toward nano-sizes. This study has confirmed that the neutralization of particles is a crucial element in evaluating the efficiency of a respirator. The variability of the respirator's performance was determined for both models and both flow rates. The analysis revealed that the coefficient of variation of the penetration ranged from 0.10 to 0.54 for particles of 20-100 nm in diameter. The fraction of N95 respirators for which the performance test at 85 l min(-1) demonstrated excessive (>5%) penetration of nanoparticles was as high as 9/10. The test results obtained in a relatively small (0.096 m(3)) test chamber and in a large (24.3 m(3)) walk-in chamber were found essentially the same, thus, suggesting that laboratory-based evaluations have a good potential to adequately represent the respirator field performance.