Residential indoor and outdoor ultrafine particles in Windsor, Ontario

UFPs in the 20–100 nm size range were measured for 10 mins every hour for 5 consecutive days in 45 homes of nonsmoking adults in summer 2005 and in 49 homes of asthmatic children in summer and winter 2006. Median hourly outdoor levels across all measurements were 10,800, 12,000 and 6300 cm−3 and median indoor levels were 2700, 3730 and 2580 cm−3 in summer 2005, winter 2006 and summer 2006, respectively. Outdoor levels generally exceeded indoor levels but indoor concentrations were higher around 5–7pm, suggesting a strong influence of cooking. Daily and weekly infiltration factors were estimated for each home using three methods. Weekly infiltration factors (Finf’s) based on a censored indoor/outdoor ratio method varied widely across homes; median Finf’s across homes were 0.16 (summer 2005), 0.21 (winter 2006) and 0.26 (summer 2006). Large indoor peaks and low infiltration of ambient PM resulted in the indoor sources generally contributing more than infiltrated ambient UFPs to indoor concentrations. Median estimates of the percentage of indoor-generated contribution to total indoor levels were 58% (summer 2005), 65% (winter 2006) and 69% (summer 2006). The proportion of homes having more than half of their indoor concentrations provided by indoor sources was 66% (summer 2005), 67% (winter 2006) and 79% (summer 2006). Median deposition rates ranged from 0.61–0.79 h−1 across the 3 sampling sessions. Spatial variability was higher for outdoor UFPs than concurrently measured PM2.5. The median correlations of hourly averaged outdoor UFPs between pairs of homes were moderate in the three sampling sessions (0.56–0.65), but were considerably lower than corresponding PM2.5 correlations. The wide range of infiltration factors across homes as well as the spatial variability and moderate between- home correlations of outdoor UFPs could cause measurement error in epidemiology studies that use central site UFP measurements as a surrogate for personal exposure to ambient UFPs.