Annals of Occupational Hygiene Advance Access published online on July 7, 2008
Annals of Occupational Hygiene, doi:10.1093/annhyg/men039
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Generation and Homogeneity of Aerosols in a Human Whole-Body Inhalation Chamber
1 National Institute of Occupational Health, Department of Chemical and Biological Work Environment, PO Box 8149 Dep, NO-0033 Oslo, Norway
2 Department of Respiratory Medicine, Rikshospitalet University Hospital, NO-0027 Oslo, Norway
3 Department of Respiratory Medicine, Faculty Division Rikshospitalet, University of Oslo, Norway
* Author to whom correspondence should be addressed. Tel: + 47 23 14 53 24; fax: +47 23 19 52 06; e-mail: wijnand.eduard{at}stami.no
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Introduction: A 16 m3 whole-body exposure chamber for human exposure to aerosols is described. Several modifications of the aerosol generation and distribution system were needed to ensure a stable aerosol concentration in the chamber, especially when a cyclone pre-classifier was used.
Results: After these modifications, stable aerosol concentrations of aluminium oxide with a volume median diameter of 5.7 µm, and 
3 µm when the cyclone was used, could be achieved after 1 h of aerosol generation. Aerosol concentrations of 1–8 mg m–3 generated without the cyclone could be maintained for at least 2 h after the aerosol level had stabilized. The temporal variability [coefficient of variation (CV)] of the aerosol concentration was 4–6%, while concentrations <1 mg m–3 showed greater relative variability. The spatial variability at 3.8 mg m–3 without a volunteer in the chamber was 4.8%. With a volunteer in the chamber who performed 30 min of ergometric cycling during 2 h of aerosol exposure, the exposure estimated by personal sampling was 15–17% lower than monitored with an optical particle counter. The variability of personally measured exposure was higher than of stationary measurements showing CVs of 10–19%.
Conclusions: These results show that controlled exposure of human volunteers to a range of concentrations can be achieved with good accuracy in this inhalation chamber. The results compare favourably with other chambers described in the literature. Personal sampling showed lower aerosol concentrations than estimated in an empty chamber and the variability was significantly higher than measured stationary.
aerosols • aluminium oxide • fluidized bed generator • homogeneity • human exposure chamber
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Aerosol exposure in work and outdoor environments is an important cause of respiratory diseases such as asthma and chronic obstructive lung disease (Hendrick et al., 2002). Information on exposure–response relationships has mainly been obtained from animal experiments and epidemiological studies. Human challenge information may provide more reliable data on short-term effects because single agents can be tested at controlled exposure levels, and the effects from other agents that can be found in work environments are avoided.
Only a few human challenge studies in whole-body inhalation chambers involving single particulate agents have been described, e.g. with endotoxin (Taylor et al., 2000), flour dust (Gripenbäck et al., 2003), corn starch (Grunewald et al., 2003) and wood dust (Gripenbäck et al., 2005). This may be due to the methodological challenges of generating aerosols with stable concentrations over time and with even distribution in test chambers (Willeke, 1980).
We previously described a human inhalation chamber for gas exposure (Søstrand et al., 1997). Here, we report a whole-body inhalation chamber for aerosol exposure with focus on the aerosol generation and distribution system. The design criteria were that aerosols that mainly deposit in the thoracic region could be generated with a temporal and spatial variability of coefficient of variation (CV) <10% at concentrations of at least 4 mg m–3 and a duration of 2 h.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Design of the aerosol chamber and aerosol generation system
The main chamber is constructed in acid-proof steel (AISI 316 L) and has a volume of 16 m3, Fig. 1. It contains a 2 m3 entrance section in order to limit turbulence and air exchange when subjects enter the chamber. The floor was made of 2-mm aluminium plate perforated with 5-mm holes and 31% openness and supported by beams made of perforated plate. Four evenly spaced 100-mm diameter exhaust ducts with inlets facing upwards were mounted 200 mm beneath the floor to ensure an even vertical airflow inside the chamber. The chamber has two glass doors and two windows made of polycarbonate. A drain in the bottom facilitates cleaning of the chamber with water.
|
The airflow in the ventilation duct could be varied between 70 m3 h–1 (1 m s–1) and 280 m3 h–1 (4 m s–1) and the air exchange rate in the chamber from 4.6 to 18 times per hour. However, the tests were performed at the lowest flow rate in order to maximize the aerosol concentration in the chamber. A circular disk baffle upstream of the outlet of the neutralizer and perpendicular to the air stream induced turbulence to ensure mixing of the aerosol and the supply air. The ventilation duct was connected to the aerosol distribution section of the inhalation chamber by a horizontally flat and vertically expanding inlet (Fig. 1). The aerosol entered the inhalation section through a 50-mm thick honeycomb roof (Core type: 5.2
25, Aviation Ltd, Cambridge, UK) in order to remove turbulence. The air temperature and relative humidity were recorded with electronic sensors (Johnson Controls, Inc., Milwaukee, WI, USA). The aerosol generation system is outlined in Fig. 2. Dust was generated with a fluidized bed aerosol generator (Model 3400 TSI Inc., St Paul, MN, USA), which can generate particles in the size range from 0.5 to 40 µm. Bronze beads with size 40 µm were used as bed material. The generator was operated with or without a cyclone (optional equipment from TSI Inc.). The cyclone removes particles with 3.5 µm aerodynamic diameter with 50% efficiency at a flow rate of 9 l min–1 according to the producer. Clean dry pressurized air was applied with flow rate of 20 l min–1 and a pressure of 2 bar. A Kr-85 aerosol charge neutralizer (Model 3054, TSI Inc.) was mounted on the top of the fluidized bed, which injected the aerosol centrally into the supply air duct at a distance of 550 mm from the aerosol inlet of the chamber.
|
Modifications
Initial tests showed that dust accumulated on the surface of the fluidized bed, which indicated that the fluidized bed did not mix properly. This was ascribed to the shape of a flow obstructor beneath the porous plate that directs the air stream to the region where particles enter the bed. Different shapes of the flow obstructer were tested, and the geometry shown in Fig. 3 was found to give the best mixing.
|
In the original version of the aerosol generation system problems were encountered with deposition of particles in the upper section of the elutriator and in the inlet of the cyclone when concentrations >1 mg m–3 were generated, see Fig. 4. Sudden increases in aerosol output were observed within an hour of operation, probably occurring when the deposit dislodged, Fig. 5. Dust deposition was significantly reduced when vibration was applied during aerosol generation with a pneumatic vibrator that was welded on the outside of the elutriator (Model FP-12-S with silencer, Houston Vibrator Inc., Houston, TX, USA) (recommendations by Göran Lidén).
|
|
Particle deposition also occurred in the lower section of the charge neutralizer when the cyclone was used which was solved by extending the cyclone outlet with a tube with the same internal diameter and a conical expansion to the end of the inlet tube of the aerosol neutralizer, Fig. 4. Without the cyclone, particle deposition in the charge neutralizer was minimal.
After these modifications, it was possible to obtain stable concentrations of 1 mg m–3 after aerosol had been generated for 1 h, and this level could be maintained for up to 1.5 h, Fig. 5. Since higher concentration levels for longer time periods were needed in human challenge experiments and we wanted to study effects in the thoracic region of the airways, the cyclone was excluded in further experiments.
Characterization of the aerosol in the inhalation chamber
Fused aluminium oxide with particle size <20 µm (5F/FE, Washington Mills, Electro Minerals Ltd, Manchester, UK) was used in the experiments because we wanted to study aerosols with low toxicity in the initial human challenges. The temporal variability, spatial distribution and particle size distribution of the aerosol generated without the cyclone were evaluated without human volunteers in the chamber after the aerosol concentration had reached a stable level, typically after 1 h. One test was performed per day.
The aerosol volume concentration was measured in eight optical size fractions, ranging from 0.5 to 40 µm with cut points at 1, 2, 3.5, 5, 7.5, 10 and 15 µm, by an optical particle counter (GRIMM 1.105, GRIMM Aerosol Technik GmbH & Co. KG, Ainring, Germany). The temporal variability was estimated by the standard deviation of the aerosol concentrations measured with 1-min intervals in the centre of the chamber (position 5, Fig. 6). The spatial distribution in the chamber was monitored sequentially at five locations 1.5 m above the floor by arithmetic mean concentrations of 5-min periods and the measurement cycle was repeated 15 times, Fig. 6. A remote-controlled mobile platform was used in order to move the particle counter in the chamber with minor disturbances of the aerosol distribution in the chamber.
|
The size of particles >0.25 µm was also measured with a Jeol (Akishima, Tokyo, Japan) JSM-6400 scanning electron microscope. For this purpose, a sample was collected on a polycarbonate filter with pore size 0.8 µm in a 25-mm diameter standard aerosol cassette made of graphite-filled polypropylene (Air Monitoring Cassette, PALL Co., Ann Arbor, MI, USA) at a flow rate of 2 l min–1. The sample was prepared and analysed by scanning electron microscopy as described previously (Skogstad et al., 1999).
The volume of each particle was estimated from the projected area and the largest diameter assuming a prolate ellipsoid shape (Lorenzo et al., 2006) and the particle size by the diameter of a sphere with size equal to the estimated volume. A shape factor was computed for each particle by the formula 
where values >1 indicate deviation from circular shape.
Aerosol measurements during human challenge
Aerosol concentrations measured by personal and stationary sampling were compared with a human volunteer in the chamber in order to evaluate the representativity of stationary samples for personal exposure. The Regional Ethics Committee, Oslo, Norway, approved the human exposure study.
All human challenges were performed with a target concentration of 4 mg m–3 and a duration of 2 h. The volunteer entered the chamber when the concentration had stabilized, performed two 15-min periods of ergometric cycling and was sitting during the remaining time. The locations in the chamber are shown in Fig. 6.
Stationary measurements were performed in the centre of the chamber with the optical particle counter and by sampling on 25-mm diameter PVC filters with 5-µm pore size (Type PVC502500, Millipore, Billerica, MA, USA) in vertically oriented closed-faced 25-mm diameter standard aerosol cassettes (Millipore) and inhalable PAS6 cassettes (van der Wal, 1983) facing downwards. A flow rate of 2 l min–1 was maintained with portable pumps (PS101; National Institute of Occupational Health, Oslo, Norway). Personal samples were collected in the breathing zone using the same instrumentation. One of each cassette type was attached to each shoulder and the left and right positions were exchanged between experiments. All exposed filters were desiccated and conditioned in the weighing room for at least 24 h before weighing on a Sartorius Micro MC5 balance under controlled conditions, relative humidity 39 ± 1% and temperature 21 ± 1°C. The detection limit was estimated to 12 µg based on 20 filter blanks that had been mounted in cassettes.
Data analysis
Standard measures of central tendency and distributions of normally distributed data, arithmetic means and standard deviations were calculated. The shape factor was not normally distributed and described by the median and 10 and 90 percentiles. The spatial and temporal variability of the exposure was shown by CV. The particle size distributions were shown by log-probability plots and characterized by the volume median diameters (VMDs) and 10 and 90 percentiles. One-way analysis of variance (ANOVA) was used in the statistical analyses with the Levene's test for testing of homoscedasticy and Dunnett's T3 test for multiple comparisons when variances were not equal. All data analyses were performed with SPSS version 14.0 for Windows (SPSS Inc., Chicago, IL, USA) except that the particle size data obtained with the optical particle counter were analysed by GRIMM software 1.177, version 3.10 build 2 (GRIMM Aerosol Technik GmbH & Co. KG, 2001).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
The particle size distribution of the aluminium oxide aerosol generated without a cyclone estimated by the optical particle counter showed that the VMD was 5.7 µm with 10 and 90 percentiles at 2.7 and 11 µm. The particle size distribution of the aerosol generated with a cyclone showed a VMD of 2.9 µm with 10 and 90 percentiles at <1 and 6.6 µm estimated with the optical particle counter.
The size of 1206 particles was also measured by scanning electron microscopy. A VMD of 3.2 µm and 10 and 90 percentiles at 1.1 and 4.6 µm were found with the scanning electron microscope, Fig. 7. The shape factor of the particles had a median of 1.28 with 10 and 90 percentiles at 1.10 and 1.82, respectively. For comparison, the shape factor of a square is 4/
1.27, indicating that most particles were isometric (all three dimension of similar size). The aluminium oxide particles appeared as compact particles in the scanning electron microscope except for some aggregation, especially of small particles, Fig. 8.
|
|
Without the cyclone, the aerosol concentration in the chamber reached a stable level from 0.5 to 1 h after generation started depending on the target concentration. It was possible to maintain a level of 8 mg m–3 for >2 h, but this was not explored systematically. The temporal variability decreased with increased mean aerosol concentrations from a CV of 19% at 0.3 mg m–3 to 4.1% at 8 mg m–3 during 1-h test periods, Table 1. The results in Table 1 represent one experiment at each concentration. The air temperature during these experiments was 24°C and the relative humidity 48%.
|
The results of the spatial distribution tests are shown in Table 2. The aerosol concentration was 5.4% lower in the periphery of the chamber than in the centre while there were minor differences between positions 1 and 4. The differences between the positions were significant, ANOVA, P = 0.04 but variances were unequal, Levene's test, P < 0.01. Dunett's T3 test for multiple comparisons showed that positions 1 and 2 were significantly different from position 5, P < 0.01 and 0.05, respectively, but not positions 3 and 4 due to higher standard deviations. The spatial variability in the chamber was estimated to a CV of 9.1% from the mean square of the positions and the temporal variability to a CV of 5.6% by the mean square error. The air temperature was 23°C and the relative humidity 15–27%.
|
A total of 14 human provocation tests were performed on different days. Measurements with the optical particle counter showed an arithmetic mean of 3.81 mg m–3 with a CV of 4.1% between tests and the temporal variability within tests pooled over all tests showed a CV of 9.1%. Table 3 shows the results for eight tests including personal and stationary filter sampling. The largest difference was observed between the PAS6 cassettes of which the stationary collected samples were 15% higher than the personally collected samples. The differences between four sampler/sampling combinations were of borderline significance, ANOVA with the four sampler/sampling combinations as independent variable, P = 0.08, with inhomogeneous variances, Levene's test P < 0.01. The variability of the personal measurements was higher than of the stationary measurements. Especially, the CV of 19% for personal samples collected with the PAS8 cassette was high. Dunnett's T3 test for multiple comparisons showed that samples collected stationary with the PAS6 cassette were close to significantly different from personal samples collected with both cassette types, P = 0.07. The personal samples collected with the PAS6 and standard aerosol cassettes were 17 and 15% lower, respectively, than measured with the optical particle counter and 22 and 19% lower than the target concentration of 4 mg m–3. The air temperature was 22–24°C and the relative humidity 12–34%.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Several problems with deposition of dust in the aerosol generating system were encountered. After two modifications, stable concentrations of 1.3–8 mg m–3 of an aerosol of aluminium oxide with a VMD of 5.7 could be generated without the use of the cyclone, while further modifications were needed when the cyclone was used. Stable aerosol levels in the inhalation chamber were reached within 1 h and could be maintained when the cyclone was not used for at least 2 h with a temporal variability of the aerosol concentration of 4–6%. Concentrations <1 mg m–3 showed higher variability with CV
10%. The spatial variability was 5% at a concentration of 3.8 mg m–3. The aerosol concentration with a volunteer in the chamber measured by personal sampling was 20% lower than expected from experiments without a volunteer in the chamber.
Marple et al. (1978), who described the fluidized bed generator used in the present study, found that several hours of aerosol generation were needed before the aerosol output was stable when new bed material is used. However, when the bed material had been used before, 1 h of aerosol generation was required after adjustments had been made of the dust-feeding rate. This is similar to our findings. This indicates that the time needed to fill the exposure chamber with a stable aerosol concentration had limited influence on the start-up time.
As the particles in the aerosol were of fused aluminium oxide and appeared as compact particles in the scanning electron microscope, they are expected to have a density close to the density of 3.9 g cm–3 of crystalline aluminium oxide. This is substantially greater than the density of 1 g cm–3 that is specified in the definition of the aerodynamic diameter and indicates an aerodynamic diameter about twice the size estimated by the optical particle counter and the scanning electron microscope. Some aggregation was observed, especially of smaller sized particles, which will decrease the apparent density of the particles, and the non-spherical shape of the particles will further reduce the aerodynamic diameter by stronger aerodynamic drag forces than on spherical particles (Baron and Willeke, 2001). These effects are not expected to fully compensate for the density difference because aggregation was limited and most particles were isometric.
The temporal variability of 4–6% found in this study for aerosol concentrations >1 mg m–3 was relatively low compared to CV values of 5–19% found in other studies of animal and human aerosol exposure chambers (Yeh et al., 1986; Lidén et al., 1998; Taylor et al., 2000; O'Shaughnessy et al., 2004; Lundgren et al., 2006a,b). The highest CV values were found in studies that did not apply charge neutralizer (Taylor et al., 2000; O'Shaughnessy et al., 2004). Lundgren et al., (2006b) further reported in a comparative study that omission of the charge neutralizer resulted in higher temporal variability. The good temporal stability of the aerosol concentration found in the present study is probably due to several factors. Initially, the dust did not mix properly in the fluidized bed, which was solved by a modification of the flow obstructor beneath the bottom plate of the fluidized bed. The importance of the neutralization device in the generation set-up has earlier been stressed by Hinds (1980) and Liu et al. (1986). However, formation of dust deposits in the fluidized bed generator, the cyclone and the charge neutralizer was not prevented. This was subsequently solved by attaching a vibrator to the generator (Spurny et al., 1975) and by extension of the cyclone outlet with a tube with the same inner diameter. Furthermore, the turbulent mixing of the aerosol in the air supply of the chamber may have contributed as well because supply of an inhomogeneous aerosol may lead to fluctuations at fixed locations in the chamber.
The spatial variability of an aerosol concentration of 4 mg m–3 was 5%, which also was low compared to CV values of 3–16% found in other studies of animal and human aerosol exposure chambers (Moss et al., 1982; Yeh et al., 1986; Lidén et al., 1998; O'Shaughnessy et al., 2003; Lundgren et al., 2006a,b). Small particle size was associated with lower variability of aerosols with mass median diameters of 1–3 µm (Yeh et al., 1986; O'Shaughnessy et al., 2003). As the mass median aerodynamic size in the present study was probably >6 µm, our results compare favourably with other studies. Lidén et al. (1998) reported a spatial variability of 15% for aerosols with mass median aerodynamic size of ca. 10 µm. In this chamber, the aerosol entered the chamber through a conical duct that induced turbulent mixing with supply air. Lundgren et al. (2006a) reported relative low variability of 3–8% in a chamber with the same design as described by Lidén et al. (1998) for aerosols with similar particle size, but these measurements were done close to the breathing zone of a volunteer or mannequin and were not representative for the whole chamber. The highest CV values were found in studies that did not apply charge neutralization (Taylor et al., 2000; O'Shaughnessy et al., 2003). Lundgren et al. (2006b) reported in a comparative study that omission of the charge neutralizer also resulted in higher spatial variability. Several factors may have contributed to the good spatial distribution of the aerosol in our study. The mixing of the generated aerosol in the supply air was improved by turbulence induced by a round disk baffle placed upstream of the aerosol inlet in the ventilation duct. After the mixing stage, rapid changes in the air stream were minimized in order to avoid formation of inhomogeneities by the inertia of the particles; by (i) the gradually expanding air inlet of the chamber; (ii) the honeycomb roof and (iii) the evenly spaced exhaust ducts in the bottom of the chamber.
The personal measurements of the aerosol concentration with a volunteer in the chamber were 20% lower than expected from experiments without a volunteer in the chamber. This may partly be due to the location of the stationary sampler in the centre of the chamber, but the volunteer may also be expected to influence the aerosol level by introducing air turbulence that may increase particle deposition on clothes and walls. The deposition may also vary due to the changing activity of the volunteer and may be the cause of the increased temporal variability with volunteers in the chamber. The difference can be easily adjusted for by changing the settings of the generator. However, the variability of the exposure during different human challenges when measured by personal sampling was also higher than when measured stationary. Especially, the PAS6 cassette showed a two to four times higher CV during repeated experiments. It can be speculated that this is due to more variable conditions when a sampling cassette is mounted on a volunteer that is cycling during 25% of the challenge time compared to stationary sampling. It is not understood why a PAS6 cassette is more affected than a standard aerosol cassette as the collection efficiency of both cassette types is fairly similar for the particle sizes used in the challenges (Kenny, 1996). The number of replicates in this series is rather small and the possibility of a chance observation cannot be ruled out. However, it is an interesting question if personal sampling increases the variability of exposure measurements even in an homogeneous environment like that can be achieved in exposure chamber as described here.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
After two modifications of a fluidized bed aerosol generator without a cyclone pre-separator concentrations of 1–8 mg m–3 of an aluminium oxide aerosol with a VMD 5.7 µm could be maintained for at least 2 h in a whole-body exposure chamber for human exposure with a temporal variability of 4–6%. Aerosol concentrations <1 mg m–3 showed larger temporal variability, 10–19%. As it takes
1 h to reach a stable concentration in the chamber and to ensure correct exposure conditions, continuous monitoring of the aerosol concentration is needed in human challenge studies. The spatial variability at 3.8 mg m–3 was 4.8% without volunteers in the chamber. These results compare favourable with other chambers described in the literature. The exposure level during human challenge was 20% lower than estimated from experiments with an empty chamber but this can be regulated. The variability of exposure measured by personal sampling during different challenges was higher than of stationary measurements and exceeds 10%. Thus, all design criteria were met when no cyclone was used and exposure was assessed by stationary measurements. However, personal measurements showed greater variability. Preliminary experiments with the use of a cyclone in the aerosol generator showed that further modifications were needed but after these changes, the period with stable aerosol concentration was still limited to 1–1.5 h. |
ACKNOWLEDGEMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
We gratefully acknowledge Per Søstrand and colleagues at the Rikshospitalet University hospital in Oslo, for information, discussions and practical help with the experiments. We especially acknowledge Neil Alexis (University of North Carolina, Chapel Hill), James A. Vincent (University of Minnesota), Göran Liden (University of Stockholm), Lennart Lundgren (University of Stockholm) and Bruce Urch (University of Toronto) for helpful scientific discussions.
Received December 7, 2007; in final form May 24, 2008
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Baron PA, Willeke K. Aerosol measurements (2001) New York, NY: John Wiley.
Gripenbäck S, Lundgren L, Eklund A, et al. Wheat flour exposure results in recruitment of inflammatory cells in the lungs of healthy individuals. Am J Ind Med (2003) 44:75–82.[CrossRef][Web of Science][Medline]
Gripenbäck S, Lundgren L, Eklund A, et al. Accumulation of eosinophils and T-lymphocytes in the lungs after exposure to pinewood dust. Eur Respir J (2005) 25:118–24.
Grunewald J, Eklund A, Katchar K, et al. Lung accumulations of eosinophil granulocytes after exposure to cornstarch glove powder. Eur Respir J (2003) 21:646–51.
Hendrick D, Beckett W, Burge SP, Churg A. Occupational disorders of the Lung (2002) London: Saunders Ltd.
Hinds WC. Dry-dispersion aerosol generators. In: Generation of aerosols and facilities for exposure experiments—Willeke K, ed. (1980) Ann Arbor, MI: Ann Arbor Science.
Kenny LC. Developments in workplace aerosol sampling. A review. Analyst (1996) 121:1233–9.[CrossRef][Medline]
Lidén C, Lundgren L, Skare L, et al. A new whole-body exposure chamber for human skin and lung challenge experiments—the generation of wheat flour aerosols. Ann Occup Hyg (1998) 42:541–7.
Liu BY, Pui DYH, Lin BY. Aerosol charge neutralization by a radioactive alpha source. Part Part Syst Char (1986) 3:111–6.[CrossRef]
Lorenzo R, Kaegi R, Gehrig R, et al. Particle emissions of a railway line determined by detailed single particle analysis. Atmos Environ (2006) 40:7831–41.
Lundgren L, Skare L, Lidén C, et al. Large organic aerosols in a dynamic and continuous whole-body exposure chamber tested on humans and on a heated mannequin. Ann Occup Hyg (2006a) 50:705–15.
Lundgren L, Kongerud J, Kruse K, et al. Airborne pinewood dust for controlled inhalation studies—performance of two different dynamic human whole-body exposure chambers. In: Large organic aerosols in a human exposure chamber. Applications in occupational dermatology and lung medicine—Lundgren L, ed. (2006b) PhD Thesis. Stockholm, Sweden: Karolinska Instututet.
Marple VA, Liu BYH, Rubow KL. A dust generator for laboratory use. Am Ind Hyg Assoc J (1978) 39:26–32.[Web of Science][Medline]
Moss OR, Decker JR, Cannon WC. Aerosol mixing in an animal exposure chamber having three levels of caging with excreta pans. Am Ind Hyg Assoc J (1982) 43:224–49.
O'Shaughnessy PT, Achutan C, O'Neill ME, et al. A small whole-body exposure chamber for laboratory use. Inhal Toxicol (2003) 15:251–63.[CrossRef][Web of Science][Medline]
O'Shaughnessy PT, Mehaffy J, Watt J, et al. Characterization of a hooded human exposure apparatus for inhalation of gases and aerosols. J Occup Environ Hyg (2004) 1:161–6.[CrossRef][Web of Science][Medline]
Skogstad A, Madsø L, Edward W. Classification of particles from the farm environment by automated sizing, counting and chemical characterisation with scanning electron microscopy-energy dispersive spectrometry. J Environ Monit (1999) 1:379–82.[CrossRef][Web of Science][Medline]
Spurny K, Boose C, Hochrainer D. On the aerosolization of asbestos fibers in a fluidized bed aerosol generator. Staub Reinhalt Luft (1975) 35:440–5. [in German].
Søstrand P, Kongerud J, Eduard W, et al. A test chamber for experimental hydrogen fluoride exposure in humans. Am Ind Hyg Assoc J (1997) 58:521–5.[Web of Science][Medline]
Taylor L, Reist PC, Boehlecke BA, et al. Characterization of an aerosol chamber for human exposures to endotoxin. Appl Occup Environ Hyg (2000) 15:303–12.[CrossRef][Medline]
van der Wal J. Comparative measurements of the total dust concentration at the work place with different samplers—part I. Staub Reinhalt Luft (1983) 43:291–4. [in German].
Willeke K, ed. Generation of aerosols and facilities for exposure experiments (1980) Ann Arbor, MI: Ann Arbor Science.
Yeh HC, Newton GJ, Barr EB, et al. Studies of the temporal and spatial distribution of aerosols in multi-tiered inhalation exposure chambers. Am Ind Hyg Assoc J (1986) 47:540–5.[Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||