Breathing Zone Exposure Assessment - Aerosols: Measurement, Dosimetry, and Health Effects - page 35

2.2.2.2 Confounding from Secondhand Smoke
The single largest confounder to personal level contaminant exposures during nonoccupational
studies is likely to be secondhand smoke (SHS) from tobacco. Paoletti et al. (2006) reported the
substantial inluences of SHS on PM 10 in indoor air (and personal exposures), noting the huge
range of gas and particle-phase chemicals found in this smoke than can confound exposure studies.
Slezakova et al. (2009) similarly reported on the wide range of metallic elements from SHS and
their potential to confound PM 2.5 exposure measurements. Rodes et al. (2010) reported signiicant
SHS levels in a general population personal exposure study of nonsmoking households that resulted
from both study protocol violations and passive smoking exposures when away from home. These
levels were quantiied by applying the optical absorbance signature methodology of Lawless et al.
(2004). Data were presented on the number of personal exposure samples that exceeded a 1.5 μg/m 3
level, deined as excessively confounding the target mass concentrations from all other sources. For
these participants in presumed nonsmoking households, tobacco smoke contributes to the mean
24 h PM 2.5 mass concentrations that typically exceeded 3.0 μg/m 3 for 7.9%-30.3% of the time. Since
the average personal PM 2.5 levels across the study averaged only 20.3 μg/m 3 , this was a substantial
contribution. Without applying some adjustment method for SHS confounding, these contributions
could seriously bias the resultant data interpretations.

2.2.2.3 “Bluff Body” Biases
Studies of aerosol trajectories around objects have long suggested that the presence or absence of the
human body shape can affect particle capture methodologies (e.g., Wood and Birkett, 1979; Mark
and Vincent, 1986). They describe test methodologies that evaluate the performance of BZ samplers
incorporating the presence of a manikin behind the inlet to simulate the human body. Trajectories
of particles approaching the body (and the PEM inlet) must navigate through the low streamlines
decelerating and diverging around the body shape. A mathematical study by Ingham and Yan (1994)
attempted to estimate the level of sampling bias using a particle trajectory model and suggested that
the presence of the human “bluff body” behind an aspirated personal sampling inlet could bias the
aerosol collection by as much as factors of 2 or more. The models showed that important inluenc-
ing variables included factors such as (a) the particle Stokes number (particle size), (b) the level of
particle bounce from inlet and clothing surfaces (and subsequent total capture by the inlet), (c) the
distance the sampling inlet projects away from the body, (d) the sampling inlet velocity relative to
the local air velocity, and (e) the width of the inlet relative to that of the human body. The relative
sampling eficiency data in Table 2.1 illustrate the predicted inluences of particle size, airspeed,
and presence or absence of a bluff body behind the BZ inlet (compared with an inlet alone in the

TABLE 2.1
Predicted Fractional Overall Sampling Eficiencies, Illustrating the Inluence of
Airspeed, Particle Size, and Scenario, with and without a Bluff Body Present

Indoor (15 cm/s) Walking (100 cm/s) Outdoor (400 cm/s)

Particle
Size ( μ m) With Body Without Body With Body Without Body With Body Without Body

0.5 1.0 1.0 1.2 1.1 2.0 1.5

1.0 1.5 1.5 2.0 2.0 3.5 3.5

2.5 1.6 2.5 2.2 3.0 4.0 6.0

10 1.7 3.7 2.5 4.2 4.2 7.5

Source: Ingham, D.B. and Yan, B., J. Aerosol Sci ., 25(3), 535, 1994.