Environ. Sci. Technol. 2009, 43, 4659–4664
Experimentally Determined Human Respiratory Tract Deposition of Airborne Particles at a Busy Street ¨ N D A H L , * ,† JAKOB LO ANDREAS MASSLING,† ERIK SWIETLICKI,† ¨ UNER,‡ ELVIRA VACLAVIK BRA § MATTHIAS KETZEL, JOAKIM PAGELS,| AND STEFFEN LOFT‡ Division of Nuclear Physics, Department of Physics, Lund University, PO Box 118, SE-221 00 Lund, Sweden, Department of Environmental Health, Institute of Public Health, University of Copenhagen, Denmark, Department of Atmospheric Environment, National Environmental Research Institute, Aarhus University, Roskilde, Denmark, and Division of Ergonomics and Aerosol Technology, Lund University, Lund, Sweden
Received October 28, 2008. Revised manuscript received February 17, 2009. Accepted February 27, 2009.
Traffic is one of the major sources of harmful airborne particles worldwide. To relate exposure to adverse health effects it is important to determine the deposition probability of the inhaled particles in the human respiratory tract. The sizedependent deposition of 12-580 nm particles was measured with a novel setup in 9 healthy subjects breathing by mouth on the windward side of a busy street in Copenhagen, Denmark. The aerosol was characterized both at the curbside and, to obtain the background concentration, at rooftop level. Particle hygroscopicity, a key parameter affecting respiratory tract deposition, was also measured at the same time of exposure. The total deposition fraction of the curbside particles in the range 12-580 nm was 0.60 by number, 0.29 by surface area, and 0.23 by mass. The deposition fractions of the “traffic exhaust” contribution, calculated as the hydrophobic fraction of the curbside particles, was 0.68, 0.35, and 0.28 by number, surface area, and mass, respectively. The deposited amount of traffic exhaust particles was 16 times higher by number and 3 times higher by surface area compared to the deposition of residential biofuel combustion particles investigated previously (equal inhaled mass concentrations). This was because the traffic exhaust particles had both a higher deposition probability and a higher number and surface area concentration per unit mass. To validate the results, the respiratory tract deposition was estimated by using the well-established ICRP model. Predictions were in agreement with experimental results when the effects of particle hygroscopicity were considered in the model.
* Corresponding author phone: +46-46-2228113or +46-735518636; e-mail:
[email protected]. † Department of Physics, Lund University. ‡ University of Copenhagen. § Aarhus University. | Division of Ergonomics and Aerosol Technology, Lund University. 10.1021/es803029b CCC: $40.75
Published on Web 03/24/2009
2009 American Chemical Society
Introduction Traffic is one of the sources that contributes most to air pollution and has been associated with adverse health effects in numerous epidemiological and toxicological studies (1, 2). The negative outcomes are mainly caused by the aerosol particles deposited in the respiratory tract during breathing (3). Thus, in order to understand the mechanisms behind the health responses, it is vital to determine the deposition probability of the inhaled particles and to calculate the deposited particle dose. Few measurements exist of respiratory tract deposition of real-world particles and to the best of the authors’ knowledge none of them have been carried out in an environment where traffic is shown to be the dominating source. The size of the airborne particles in the respiratory tract during inhalation is the most important property determining the deposition probability. The size depends primarily on the particle dry diameter and hygroscopicity, the latter of which is the ability of the particle to grow by absorption of water vapor. The relative humidity (RH) in the lungs has been estimated to be 99.5% (4). At this RH ambient particles may grow by a factor 1 (no growth) to 5 in diameter depending on dry size and chemical composition. Although a number of studies have examined the hygroscopicity of aerosol particles in urban environments (5), whereof at least two were at ground level with streets in the vicinity (6, 7), none were found that measured closer than 40 m from the traffic source. According to the available studies, a majority of the urban background particles are nearly hydrophobic and contain large fractions of soot and water-insoluble organic compounds (particularly as sulfur concentrations in the fuel is decreased). However, these particles presumably become more hygroscopic by transformation processes in the atmosphere such as condensation of vapors. Aged particles, more prone to grow by water uptake, are also observed close to streets. There is at present no publicly available deposition model that accounts for the complex structure of atmospheric particles, their varying mixing states and chemical composition, and their evaporation or restructuring in the elevated temperature and humidity conditions in the lungs. Thus, it has been unfeasible so far to derive the respiratory tract deposition of particles in traffic environments from available models and experiments. The objectives of this study were to (1) measure the respiratory tract deposition of a typical urban street aerosol for a group of healthy adults, (2) experimentally determine the hygroscopic properties of the particles, and (3) based on the measurements provide a tool to estimate particle deposition in other similar environments. The size-dependent particle deposition was determined with a recently developed setup (8) that was extended with a novel processing unit to account for particle size changes in the lungs.
Experimental Section The study consists of three main parts: (1) characterization of the aerosol close to a busy street, (2) measurement of the respiratory tract deposition of the particles for a group of healthy subjects, and (3) modeling of the deposition. The measurements were performed March 27-31, 2006 on H.C. Andersens Boulevard in Copenhagen, capital of Denmarks a 6-lane street with about 65 000 vehicles passing per day on weekdays, of which 4-5% were heavy duty vehicles and about 18% were light-duty delivery vans and taxis, mainly diesel powered. Of the 78% private cars on the street, 17% were diesel powered. Diesel contained less than 0.001% (10 ppm) VOL. 43, NO. 13, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4659
sulfur in Denmark in 2006. The speed limit on the street was 60 km/h, but the actual pace varied because of traffic lights. Studded tires were not used in Copenhagen. The respiratory tract deposition measurements took place during daytime on the northeastern side of the street. Due to the prevailing wind direction (south to west) the measuring station was affected by high concentrations of traffic emissions. The study was approved by the local ethics committee in accordance with the Declaration of Helsinki and an informed written consent was obtained from all subjects. Aerosol Characterization. The aerosol was characterized simultaneously at the curbside and at an urban background station on the roof of the 20 m high H.C. Ørsted Institute 3 km north (9). Measured parameters were particle number size distributions (with custom built Vienna type differential mobility particle sizers, 6-700 nm (DMPS) (Vienna type medium, inner radius 25.0 mm, outer radius 32.4 mm, aerosol flow 1.0 L/min, sheath flow 7.0 L/min), hygroscopic growth factors (H-TDMA, described below, curbside only), particle mass concentration, PM10 and PM2.5 (by a tapered element oscillating microbalance, TEOM series 1400A, Ruprecht & Patashnik Inc., filter temperature 50 °C), NOx (monitor model M 200A, API Inc., CA), CO (monitor model M 300, API Inc., CA), O3, (monitor model M 400A, API Inc., CA), and SO2 (monitor model M 100A, API Inc., CA, curbside only). The scan time of the two DMPS systems was 3 min. Weather data were provided by the Danish Meteorological Institute. The hygroscopic growth factor (Gf) is a vital parameter that enables modeling of the respiratory tract deposition. It is defined as the diameter of a particle at a given RH divided by its dry diameter. During the inhalation experiments, Gf was measured at the curbside using a hygroscopicity-tandem differential mobility analyzer (H-TDMA) (5). In an H-TDMA, monodisperse dry particles are selected from the polydisperse aerosol with a differential mobility analyzer (DMA). Thereafter the particles are humidified to a well-defined RH and the humidified size distribution is determined with a second DMA in combination with a condensation particle counter (CPC). Because of difficulties in maintaining a stable RH close to saturation, the H-TDMA was operated at 91% RH for particles in the dry (95%) RH reached by this procedure. The losses during the procedure were about 10% for 20 nm particles and 5% for particles larger than 100 nm. Particle number size distribution was measured after the processing unit both with and without heating/humidification. According to these measurements there was a negative size shift of 2-3% for the curbside particles. Modeling Deposition. The ICRP model (14) was used to calculate the size-dependent respiratory tract deposition fraction (DFmodeled). The experimentally determined breathing parameters for the subjects were used in the model to derive a mean DFmodeled. The ICRP model does not explicitly include a correction for the hygroscopicity of the particles, although there is an outline for such calculations. The first step in order to apply the ICRP model with consideration of particle water absorption is to derive the hygroscopic growth of the particles. Hygroscopic growth factors (Gf) were measured at 91% RH and hence needed to be calculated at the RH prevailing in the lung (i.e., 99.5%). The curbside aerosol was observed to separate into two
TABLE 1. Characteristics of the Aerosol at the Curbside Measurement Site on H.C. Andersens Boulevard and at an Urban Background Station at Rooftop Levela
NO (ppb) NOx (ppb) CO (ppm) O3 (ppb) SO2 (ppb) PM2.5 (µg/m3) PM10 (µg/m3) PNC (1/cm3)
curbside
rooftop
50 ( 16 83 ( 21 0.75 ( 0.22 12 ( 4 1.02 ( 0.36
4(3 19 ( 7 0.31 ( 0.04 25 ( 4
40 ( 25 17900 ( 7800
10 ( 2 5700 ( 200
Values are mean ( SD during exposure time between 8 a.m. and 6 p.m. PNC is the total particle number concentration (6-700 nm). a
particle groups with different hygroscopic properties, namely one “hydrophobic” and one “hygroscopic” group. For each group, a simple model of hygroscopic growth was applied assuming that all uptake of water at 91% RH was attributed to one single inorganic salt component, in this case ammonium sulfate (5). Based on this assumption the fraction of soluble particle volume was derived from the H-TDMA measurements. Subsequently the Gf at 99.5% RH was calculated using parameterizations for ammonium sulfate droplets (15). For each of the two particle growth groups the original DFmodeled curve was adjusted based on the derived Gf values at 99.5% RH. For example, 100 nm “hygroscopic” particles with a Gf ) 3.0 at 99.5% RH were considered to deposit with the same probability as 300 nm hydrophobic particles with a Gf ) 1.0. Finally DFmodeled for the curbside aerosol was determined as the mean value of the DFmodeled for the two growth groups weighted by the observed number fraction of particles in each group.
Results and Discussion Aerosol Characteristics. There are several indicators of the plausibility that the major part of the aerosol inhaled by the subjects originated from traffic. The aerosol characteristics measured are summarized in Table 1. As can be seen, NO, which is a common indicator of fresh traffic exhaust, was more than 10 times higher at the exposure location than at the rooftop station. The O3 level was lower near the street because of reactions with NO to form NO2. Both CO, a common tracer for combustion, and particle number concentration were higher at the curbside compared to the rooftop. The characterization and exposures took place on the windward side of the street. Figure 2 shows the average dry particle number size distributions at the curbside and at the urban background station at rooftop level. Three log-normal modes were fitted to the mean curbside distribution with count median diameters (CMD) of 20, 58, and 100 nm, number concentrations of 13 000, 3800, and 3000 cm-3, and geometric standard deviations (σg) of 1.8, 1.5, and 2.0. Using these three modes for the description of the entire aerosol population, the size interval studied, 12-580 nm, comprises more than 90% of the particles by number and surface area and 68% of the particles by mass. However, it is evident from the measured PM values that a coarse particle mode outside the measurement range contributes considerably to the total particle mass. The use of the mixing container from which subjects inhaled decreased the concentration variability over time by a factor of 3-4, but the shape of the particle number size distribution was essentially unchanged apart from a substantial loss of particles
