Annals of Occupational Hygiene Advance Access originally published online on March 2, 2004
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Ann. occup. Hyg., Vol. 48, No. 4, pp. 351-368, 2004
© 2004 British Occupational Hygiene Society
Published by Oxford University Press
Performance of Personal Inhalable Aerosol Samplers in Very Slowly Moving Air When Facing the Aerosol Source
1 Institut National de Recherche et de Sécurité, INRS, Laboratoire de Métrologie des Aérosols, BP 27, 54501 Vandoeuvre Cedex, France; 2 Center for Health-Related Aerosol Studies, Department of Environmental Health, University of Cincinnati, PO Box 670056, Cincinnati, OH 45267-0056, USA; 3 Institut de Radioprotection et de Sûreté Nucléaire, IRSN/DSU/SERAC, Laboratoire de Physique et de Métrologie des Aérosols, BP 68, 91192 Gif-sur-Yvette Cedex, France
Received 3 January 2003; in final form 28 October 2003; published online on 2 March 2004
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
While personal aerosol samplers have been characterized primarily based on wind tunnel tests conducted at relatively high wind speeds, modern indoor occupational environments are usually represented by very slow moving air. Recent surveys suggest that elevated levels of occupational exposure to inhalable airborne particles are typically observed when the worker, operating in the vicinity of the dust source, faces the source. Thus, the first objective of this study was to design and test a new, low cost experimental protocol for measuring the sampling efficiency of personal inhalable aerosol samplers in the vicinity of the aerosol source when the samplers operate in very slowly moving air. In this system, an aerosol generator, which is located in the centre of a room-sized non-ventilated chamber, continuously rotates and omnidirectionally disperses test particles of a specific size. The test and reference samplers are equally distributed around the source at the same distance from the centre and operate in parallel (in most of our experiments, the total number of simultaneously operating samplers was 15). Radial aerosol transport is driven by turbulent diffusion and some natural convection. For each specific particle size and the sampler, the aerosol mass concentration is measured by weighing the collection filter. The second objective was to utilize the new protocol to evaluate three widely used aerosol samplers: the IOM Personal Inhalable Sampler, the Button Personal Inhalable Aerosol Sampler and the 25 mm Millipore filter holder (closed-face C25 cassette). The sampling efficiencies of each instrument were measured with six particle fractions, ranging from 6.9 to 76.9 µm in their mass median aerodynamic diameter. The Button Sampler efficiency data demonstrated a good agreement with the standard inhalable convention and especially with the low air movement inhalabilty curve. The 25 mm filter holder was found to considerably under-sample the particles larger than 10 µm; its efficiency did not exceed 7% for particles of 40–100 µm. The IOM Sampler facing the source was found to over-sample compared with the data obtained previously with a slowly rotating, freely suspended sampler in a low air movement environment. It was also found that the particle wall deposition in the IOM metallic cartridge was rather significant and particle size-dependent. For each sampler (IOM, Button and C25) the precision was characterized through the relative standard deviation (RSD) of the aerosol concentration obtained with identical samplers in a specific experiment. The average RSD was 14% for the IOM Sampler, 11% for the Button Sampler and 35% for the 25 mm filter cassette. A separate set of experiments, performed with the Simplified Torso showed that in very slowly moving air a personal sampler can be adequately evaluated even when it is not attached to a body but freely suspended (confirming the data reported previously).
Keywords: dust source; inhalable; personal sampler; sampling efficiency; slowly moving air
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Protection of workers against hazardous airborne dusts has received considerable attention as an important part of the overall objective to minimize occupational exposures. Airborne aerosol sampling in workplaces (personal and stationary) has become a key issue in occupational hygiene, since the data collected on the particle size and concentration are used by regulatory agencies for exposure and risk assessments in various industries. For example, the latest Council Directive 96/29/EURATOM (Council for the European Union, 1996) for the nuclear fuel handling industry recommends using the particle size distribution and concentration data to assess the effective dose of the worker’s inhalation of airborne radionuclides.
Among the strategies adopted throughout industries in different countries to adequately assess the true exposures of individual workers to occupational hazards, personal measurement (with an aerosol sampler mounted on the worker’s body) is frequently recommended. A wide variety of personal samplers capable of extracting the inhalable aerosol fraction have been developed over the last three decades (summarized by Vincent, 1995; Kenny et al., 1997; Hinds, 1999; Witschger, 2000; Maynard and Jensen, 2001). Resulting from the differences in their inlet design and operational parameters (e.g. the sampling flow rate), the samplers exhibit significantly different performance characteristics. Experimental and theoretical evaluations of the sampling efficiency of personal aerosol samplers revealed that, at least for the majority of available samplers, the sampling efficiency is a strong function of the particle size and the ambient air velocity.
The exposure level in indoor workplaces is largely dependent on the air characteristics, such as the wind speed and ventilation parameters. Exposure also depends on the aerosol source characteristics (e.g. the particle size and initial velocity). While available personal aerosol samplers have been primarily characterized based on wind tunnel tests conducted at relatively high wind speeds, modern indoor occupational environments represent very slow moving air, which is generally not achievable in a wind tunnel setting. The air speeds in indoor workplaces are usually 
0.2 m/s or less. An extensive air velocity survey performed in 55 work areas over a wide variety of indoor workplaces (Baldwin and Maynard, 1998) revealed that the vast majority of the air velocities did not exceed 0.3 m/s and in many situations were even <0.1 m/s (the velocity was defined for the worker’s motion relative to the air environment as measured during worker activity). The air speeds measured by Whicker et al. (2000) in a nuclear laboratory at the height representing a worker’s breathing zone showed the same trend for different tests with a median velocity of <0.2 m/s. The studies referred to above agreed well with data obtained by other investigators in similar workplaces and in residential indoor environments (Berry and Froude, 1989; Matthews et al., 1989; Wiasiolek et al., 1999). The effects of the aerosol source characteristics (its strength, geometry, initial particle velocity and direction of emission), as well as the source location relative to the worker, have been addressed in several experimental and numerical studies related to worker exposure (Lidén and Kenny, 1994; Kulmala et al., 1996; Welling et al., 2000; Guffey et al., 2001). These and other studies suggest that workers often operate in close proximity to a point aerosol source. Furthermore, in many work tasks and practices, the worker is exposed to air contaminants while facing their source most of the time, so that the sampler orientation angle is 0° relative to the particle emission direction.
Thus, the protocol for laboratory evaluation of personal aerosol samplers should involve (i) an environment with very slow air movement (Aitken et al., 1999) and (ii) samplers operating close to a point source while facing the source.
In perfectly calm air, the air motion is exclusively caused by the sampler’s aspiration action. In contrast, in actual occupational and residential indoor environments air currents of a few centimetres per second may have other causes, including natural convection. Therefore, various investigators used ‘very slowly moving air’ (Dunnett, 2002; Dunnett and Wen, 2002; Su and Vincent, 2002) or ‘nearly still air’ (Hinds, 1999) as an alternative term to ‘calm air’. Ogden (1983) introduced criteria to determine the maximum ambient air velocity U0 under which the calm air concept is relevant:
Here D is the inner diameter of the probe (in cm), US is the sampling velocity (in m/s), dae is the aerodynamic particle diameter (in µm) and EA is the minimum allowable aspiration efficiency (in %). Overall, however, the aerosol sampling in very slowly moving air remains poorly understood in spite of its considerable practical relevance (Gibson and Ogden, 1977; Kenny et al., 1999; Dunnett, 2002; Su and Vincent, 2002).
Similarly to Aitken et al. (1999), very slow moving air was defined in this study as a situation when the net air movement in the vicinity of the sampler is essentially 0 and the peak air velocity is <0.1 m/s. From a practical point of view, the 0 net air movement can be established if an omnidirectional hot wire anemometer detects no measurable air movements, i.e. the average air velocity is below its limit of detection. The latter was 0.05 m/s in the studies of Kenny et al. (1999) and Aitken et al. (1999).
Traditionally, personal samplers have been tested in moving air while mounted on a full-size human shaped manikin (some studies utilized rotating manikins to minimize the effect of the sampler orientation relative to the ambient air). This rather complex and time-consuming protocol requires a large cross-section wind tunnel. Witschger et al. (1998) and Ramachandran et al. (1998) proposed two different approaches to simplify the procedure and reduce the cost of the experiments. The Simplified Test Protocol proposed by Witschger et al. (1998) utilizes a 3-dimensional rectangular body (33 cm wide x 21 cm high x 21 cm deep) with rounded corners to simulate the effect of a full-size human body on the sampler. In this protocol, the samplers are attached to the torso at different orientations to the wind (0°, ±90° and 180°) so that the direction-averaged sampling efficiency can be determined without rotating the torso. A similar approach was used by Li et al. (2000) when evaluating six freely suspended inhalable samplers. The Simplified Test Protocol has been successfully adopted for evaluating several inhalable samplers (Aizenberg et al., 2000a, 2001; Kennedy et al., 2001). The wind tunnel tests with a full-size manikin and the Simplified Torso were conducted at relatively high air velocities (
0.5 m/s), which may be found, for example, in close proximity to a local exhaust with forced ventilation or near open doors. Alternatively, the study of Ramachandran et al. (1998) utilized a scaling approach that involves non-dimensional parametric analysis and transformation from a large wind tunnel setting to a small one. This approach has recently been used by Sreenath et al. (2002) to investigate the aspiration characteristics of blunt samplers.
Historically, there has been limited success in the effort to experimentally investigate the behaviour of inhalable particles at lower air velocities in wind tunnels (<0.3 m/s). First, when the air velocity is comparable with the particle sedimentation velocity, it is a challenge to maintain a uniform aerosol concentration in the test zone of the wind tunnel. Second, it is difficult to ensure reasonably low inter-sample variability with several identical samplers in the test zone. The variability is defined through the relative standard deviation (RSD) (ratio of the standard deviation to the mean, expressed as a percentage of the sampling efficiency data). This characteristic is important in providing an adequate inter-laboratory comparison. The above difficulties have limited the use of wind tunnels for experimental studies of aerosol inhalability in low air movement environments, as well as for evaluation of the performance of personal aerosol samplers.
The definition of ‘inhalability’ in essentially calm air environments (with a wind speed <0.1 m/s) was recently addressed in two laboratories in the UK (the Health and Safety Laboratory in Sheffield and the Institute of Occupational Medicine in Edinburgh). In both studies the tests were carried out in test chambers with cross-sections of 1 m2 and an overall height of 3 m. In both test chambers the peak velocities were always <0.1 m/s, with average velocities below the anemometer’s limit of detection (0.05 m/s). The test particles with narrow size distributions were dispersed and mixed to produce homogeneous test aerosols in the vicinity of a manikin that was slowly rotated. The reference aerosol concentration was determined by an ingenious quasi-isokinetic method using sharp-edged probes mounted on a circling arm. The tests conducted with the manikin at various particle sizes revealed that the inhalability in calm air is consistently higher than predicted by the internationally standardized inhalable curve or the inhalable convention (CEN, 1993; ISO, 1995; ACGIH, 1999). The ACGIH–CEN–ISO inhalable convention was primarily based on the aspiration efficiency of mouth-breathing manikins tested in wind tunnels over a range of wind speeds from 0.5 to 4 m/s (Baldwin and Maynard, 1998; Aitken et al., 1999; Kenny, 2000); in addition, some limited data from Ogden and Birkett (1978) obtained in a calm air chamber were used. As an alternative to the inhalable convention, Aitken et al. (1999) proposed an empirical relationship for the low wind inhalability at an oral breathing rate of 20 l/min:
(dae, expressed in µm, < 100 µm). No definite explanation has been provided in the literature regarding the reason for the difference between the inhalable convention and equation (2).
Kenny et al. (1999) determined the sampling efficiencies of four existing personal inhalable/total dust samplers in low air movement environments and compared the results with equation (2). The experimental set-up was rotating slowly in a calm air aerosol chamber. The sampling efficiencies did not demonstrate a considerable difference when the sampler was isolated or operated on the manikin. This result is important for the testing of personal samplers as it implies that sampling tests in very slowly moving air could be carried out when the samplers are not mounted on a manikin.
Both UK teams indicated that RSDs in their experiments were considerably lower than those that had been obtained in previous studies using large cross-section wind tunnels with moving air flows. This has been attributed to an improved estimation of the reference concentration. It should be noted that the above tests attempted to create homogeneous aerosol concentrations across the chamber, but not a situation when the worker/sampler faces a point aerosol source.
To summarize, performance evaluations of personal inhalable aerosol samplers have been done mostly in moving air, while evidence suggests that this setting does not represent a typical workplace situation. Very few studies have been devoted to calm air (low air movement) environments, in which the majority of indoor occupational exposures occur. Exposure monitoring for workers facing an aerosol source in close proximity has also not been adequately studied. The above puts forward a clear need to evaluate performance characteristics of newly designed and existing personal inhalable aerosol samplers that operate while facing the dust source in very slowly moving air.
The present study had two objectives. The first was to design and validate a new experimental protocol and a test facility for measuring the sampling efficiency of personal inhalable samplers when the samplers face the dust source (0° orientation) in very slow air (essentially no measurable average velocity and a peak instantaneous velocity not exceeding 0.1 m/s). The test system was intended to be easily duplicable by other laboratories at minimum cost to allow effective inter-laboratory data comparison in the future. The second objective was to utilize the new protocol in order to determine the sampling efficiencies of three personal inhalable samplers commonly used by industrial hygienists worldwide: the IOM Inhalable Sampler at a flow rate of 2 l/min (SKC Inc., Eighty Four, PA), the Button Personal Inhalable Aerosol Sampler at a flow rate of 4 l/min (SKC Inc.) and the 25 mm Millipore filter holder (C25 cassette) at a flow rate of 1 l/min (Pall Corp., Ann Arbor, MI). At this time there is very little or no information concerning the sampling efficiency of the samplers in very slowly moving air when they face the aerosol source. Among the three above-listed samplers only the IOM Inhalable Sampler has been previously evaluated in low air movement (the slow rotating protocol of Kenny et al., 1999).
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Test system
Two issues were taken into consideration when designing the test system and the evaluation protocol for this study. First, the aerosol generation method must be chosen to achieve a homogeneous aerosol concentration and size distribution at a specific distance from the dust source and a specific height from the floor. Second, the air environment should be essentially calm, allowing only very slow air movement in the area of the sampling point.
The test system developed for the evaluation of personal aerosol samplers that face the dust source in very slowly moving air is shown in Fig. 1. The designed arrangement includes the aerosol generation system, which is placed in the centre of the test chamber and continuously rotates while dispersing particles in an omni-directional way. The aerosolized particles are transported from the source towards the test and reference samplers, which are located along the circumference around the generation point. Radial aerosol transport is driven primarily by turbulent diffusion and small natural convection. The set-up was designed so that no direct particle projection occurs. The radial distance between the aerosol injector and the samplers (marked as R, Fig. 1, Top View), as well as the elevations of the aerosol injector and sampling ports above the floor, were chosen based on validation tests. This configuration aimed at ensuring that (i) the generation system does not directly affect the microenvironment of the test and reference samplers and (ii) all the test and reference samplers are exposed to the same aerosol concentration during the test. Since the particles tested in this study cover very wide size range (the sedimentation velocity ranges from 0.15 to 16 cm/s), two distances R were established: R = 100 cm for the four smaller size fractions (dae = 6.9–38.7 µm) and R = 60 cm for the two larger ones (dae = 60.1–76.0 µm). The aerosol generation system consists of a powder disperser and a specially designed injector that continuously rotates at 2.3 r.p.m. driven by an electrical motor. A rotating brush generator (model RBG-1000; Palas GmbH, Karlsruhe, Germany) is capable of continuously aerosolizing the test powder at reproducible aerosol concentration levels. The output of the powder disperser is connected to a vertical stainless steel tube (4 mm internal diameter, 90° bend). The output of the tube, oriented horizontally, is connected to an aerosol injector made of a semi-spherical porous metal screen and positioned at a height of 115 cm above the floor (see Fig. 1, Front View). The porous screen, shaped like a hemisphere with identical small holes (diameter DS 
0.4 mm), minimizes the effect of direct air movement produced by the generation system on the sampler’s performance. It also improves the aerosol dispersion. The air flow rate in the powder generator is 37 l/min, which results in an exit air velocity of Uexit 2.7 m/s through the injector screen. The operating principle of the RBG-100 generator and a specially designed injector ensure an agglomerate-free aerosol from non-cohesive powders. The generator is placed inside a cylindrical wooden box with the aerosol injector outside.
|
Exiting the holes in the porous screen, the air jets quickly begin to expand and merge with one another over a very short distance. The area of contact between the air jets and the surrounding air increases, which consequently intensifies turbulent exchange. The directional air motion dies very quickly as the air velocity decreases considerably with increasing distance from the screen. Therefore, influence of the air jet does not extend a long distance. Generally, fluid flow through a finite hole into an infinite (or sufficiently large) space which is fully occupied by the same fluid in a relatively calm condition is regarded as a ‘free circular air jet’. There are several equations that allow assessment of the decay of free air jet velocity after passing the exhaust point (Baturin, 1972). The axial air velocity Uaxial in direction x in the main section of the air jet from a grille perforated with small holes can be reasonably approximated as follows (Baturin, 1972):
Here fS is the porosity of the screen (fS = 0.21), k is a porosity-dependent coefficient (k = 4 for fS = 0.05–0.3) and µ is the discharge coefficient (µ = 1 was chosen to estimate the highest possible air velocity at point x). Using the above equation, we calculated that if the relative distance x/DS 185, then the relative air velocity Uaxial/Uexit < 0.01, i.e. Uaxial < 0.01·Uexit 2.7 cm/s. In our experimental set-up, the relative distances x/DS were 2500 for x = R = 100 cm and 1500 for x = R = 60 cm. This suggests that the aerosol generator produces a very low air velocity at the sampling point (<<2.7 cm/s). Thus, we conclude that the jet of air through the porous screen of the aerosol injector does not extend to the microenvironment of the test or reference samplers.
To verify the above estimation, the air velocity was measured at various distances from the aerosol injector using a hot wire anemometry (IFS-100 system based on a 54T21 Omnidirectional Transducer for low velocity; Dantec Dynamics, Mahway, NJ). In these measurements the aerosol generation system operated with no powder inside (it was blowing air but did not disperse powder). As expected, the air velocity decreased rapidly as the radial coordinate increased. The average values, measured at R = 60 and 100 cm, were below the anemometer’s detection limit of 5 cm/s. The turbulence intensity was 30% (the same level as that referred to by Baldwin and Maynard, 1998). Thus, the air velocity measurements confirmed that there were no measurable air movements caused by the disperser within the microenvironment of the test or reference samplers.
To examine whether the dispersed particles may be projected towards the microenvironment of the test or reference samplers, the stopping distance was calculated as the product of the relaxation time and the particle emission velocity. Assuming that the latter was equal to the exit air velocity (Uexit 2.7 m/s), the stopping distance for the test particles ranged from 0.04 to 3.2 cm, depending on the particle size. As the stopping distance was much less than R, the projection effect did not extend up to the sampler locations (see also Lidén and Kenny, 1994).
Since the tests were conducted in a calm air chamber and the sampling process was not directly affected by the aerosol generation system, it was concluded that the samplers in our study operated in a low air movement environment, which can be characterized as calm air.
In a typical experimental run, 12–15 samplers were placed along the circumference, including three replicates of each of the three personal samplers under test (total of nine), and 3–6 reference samplers (see Fig. 1, Top View). The samplers were distributed randomly along the circumference at even distances. Sharp-edged, thin-walled, cylindrical probes made of stainless steel were used as the reference probes. The particles were collected on a fibreglass filter (Vitro-DiskTM; Omega Speciality Instrument Co., Chelmsford, MA). The inner diameter, sampling flow rate and the orientation of the reference probes in the field of gravity were determined in separate experiments to ensure a sampling efficiency of 100%. Three orientations of the reference samplers were tested: horizontal, upward and downward (see details below). In most of the experiments, the test samplers were attached to small square plates (8 x 8 cm), which were oriented towards the centre. The Button and the IOM personal samplers were oriented horizontally, also facing the centre. The 25 mm Millipore filter holder was positioned on the square plate with the cassette entry orifice oriented downwards at an angle of 45° (see Fig. 1, Front View) (Buchan et al., 1986). In separate experiments the Simplified Torso (Witschger et al., 1998) was also used with one IOM Sampler attached to its front, thus facing the aerosol source (orientation angle 0°). A high capacity pump supplied the vacuum needed to create the flow rates through critical flow orifices designed to ensure an error of 5% or less. This pump was located outside the chamber to avoid air movements caused by thermal convection near the heated pump. Reinforced silicone tubes were used to connect the samplers to critical orifices and pumps. Air flow calibration was performed prior to each run using a portable bubble flow meter (mini-BUCKTM model M-30; A.P. Buck, Orlando, FL).
The entire test system (including the generation system and the samplers) was grounded and placed in the middle of a non-ventilated environmental chamber hereafter referred to as the CEPIA (Chambre d’Etudes et de Recherches sur les Préleveurs Individuels et d’Ambiance). The chamber dimensions were L x W x H = 400 x 300 x 300 cm. The chamber walls were thermally insulated. The CEPIA was placed inside an air-conditioned laboratory at IRSN (Institut de Radioprotection et de Sûreté Nucléaire, Gif-sur-Yvette, France). Therefore, the only air disturbance that affected the sampler microenvironments occurred due to the turbulence created near the aerosol disperser by merging air jets and possibly due to low intensity natural convection.
Sampling efficiency
The sampling efficiency ES of the test sampler was obtained using the ‘indirect method’ (according to the terminology suggested by Vincent, 1989):
where CS and C0 are the particle mass concentrations measured by the test and reference samplers, respectively. For the IOM Sampler, if CS is measured in accordance with the protocol (filter and cartridge as a single assembly), the sampling efficiency is essentially equal to the inlet entry efficiency (Vincent, 1989). Since the sampling efficiency is particle size-dependent, the sampler evaluation tests were conducted using close-to-monodisperse particles of six sizes within the inhalable particle size range. In our method, each sampling run utilized at least three identical test samplers of each type and 3–6 reference samplers operated simultaneously. To determine the aerosol concentrations, each filter was weighed before and after the test using an electronic balance (model MC210P; Sartorius AG, Goettingen, Germany) and the ratio of the gained mass to the sampling volume was obtained for each sampler. For each particle size fraction (characterized by the mass median aerodynamic diameter) and for each type of the sampler under the test, the average values of CS and C0 and the standard deviations of these values were determined.
Test particles
Six narrowly distributed Al2O3 powders (SPM nos 102, 95, 91, 84, 71 and 52; Alumines Durmax, Chatellerault, France), which range from 7 to 100 µm aerodynamic diameter in the original powder and thus represent the inhalable aerosol fraction, were utilized in this study. These powders have been widely used worldwide for the evaluation of personal inhalable samplers in wind tunnels and calm air chambers.
Particle gravitational settling affects the aerosol transport, especially for larger particles, i.e. the particle size distribution measured at a distance from the source may differ from that of the original powder. Therefore, in addition to microscopic analysis of the original powder samples, characterization of each powder was performed after it was aerosolized and subsequently collected at a distance R from the aerosolization point on Nuclepore® polycarbonate membrane filters in sharp-edged reference probes.
The aerodynamic size of each powder was calculated from the particle volume equivalent diameter distribution measured by the Coulter technique (Multisizer IIe; Beckman Coulter, Villepinte, France), similar to the procedure used in the study by Witschger et al. (1997). Aluminium particles are irregularly shaped and have a density of 3.95 g/cm3, as was determined using a centrifuge method (Association Française de Normalisation, 1995). The calculation of the aerodynamic particle diameter required information on the Al2O3 particle dynamic shape factor 
, which was available from Mark et al. (1985). As a part of our method validation, we collected one sample of each of the aerosolized powders and analysed the samples using a scanning electron microscope (JEOL JSM 840a; JEOL USA Peerbody, MA). The scanning electron micrographs revealed significantly irregularly shaped particles, very similar to those presented by Mark et al. In addition, we collected four replicate samples of the finest powder, SPM 102, with an eight-stage cascade impactor (Andersen Mark II; Thermo Andersen, Smyrna, GA) that allowed us to directly determine the aerodynamic particle size distribution. This distribution was then compared with the result of our calculation utilizing the Coulter counter data. It was found that the mean value of the dynamic shape factor was = 1.5, which is in full agreement with the data reported by Mark et al. (1985).
The volumetric particle size distributions were obtained, respectively, from the original alumina powder samples (before they were dispersed) and from the aerosol samples collected with reference probes at a distance R from the source. Figure 2 shows the data for each of the six fractions. Measured volume distributions are presented versus the aerodynamic diameter dae. It is seen that the difference between the powder and aerosol particle size distributions is more pronounced for larger particles because some larger particles may not reach point R due to their rapid gravitational settling. This results in a shift in the particle size distribution towards smaller sizes. Since the test system was designed to evaluate samplers operating at a fixed distance from the source, the powder characteristics indicated hereafter represent aerosol sampling data obtained with reference probes located at distance R from the disperser.
|
Table 1 summarizes the characteristics of the six test powders, including the particle dynamic shape factors (from Mark et al., 1985), as well as the mass median aerodynamic diameters (MMAD) and the geometric standard deviation (GSD) (as measured in this study with reference probes).
|
The definition of the reference concentration C0 is particularly problematical when the sampling is performed in a calm or very slowly moving air environment. Generally, a sharp-edged/thin-walled cylindrical probe operating isokinetically (with the sampling velocity, US, equal to the ambient air velocity, U0) and isoaxially (facing the ambient air flow) is used as a reference sampler since its aspiration efficiency is 100% (Vincent, 1989). The traditional approach involving isokinetic probes is not applicable when the ambient air flow is very low or non-existent. However, even if the sharp-edged/thin-walled probe operates non-isokinetically (e.g. US >> U0), its aspiration efficiency may still be close to 100%, depending on the Stokes number (Stk, which represents inertial effects) and the particle sedimentation factor (v, which represents the gravitational settling effect):
Here 
0 is the particle density (designated 1.0 g/cm3), C is the slip correction factor, 
is the air viscosity, VS is the particle settling velocity and g is the acceleration due to gravity. Governed by particle inertia and gravitational settling, the aspiration into a sharp-edged/thin-walled reference probe may lead to an overestimation (ES > 100%) or an underestimation (ES < 100%) of the true aerosol concentration. The data reported in several experimental, semi-empirical and theoretical studies conducted in calm or low air movement environments suggest that (i) the aspiration efficiency is 100% if Stk << 1 and v << 1 and (ii) the critical values of these two non-dimensional parameters, at which ES 100%, may depend on the probe orientation (Davies, 1968; Gibson and Ogden, 1977; Yoshida et al., 1978; Agarwal and Liu, 1980; Grinshpun et al., 1989, 1990, 1993; Vincent, 1989; Aitken et al., 1999; Dunnett, 2002; Su and Vincent, 2002). In this study the inner diameter and sampling flow rate of the cylindrical horizontally oriented reference probes were determined utilizing the non-dimensional criteria analysis of the aerosol aspiration of inhalable particles (<100 µm) in calm air (Vincent, 1989). Postulating that the aspiration efficiency of a reference probe is 100 ± 10%, we determined that D = 2 cm and Q = 10 l/min. This inlet diameter and sampling flow rate result in a sampling velocity of US = 53 cm/s. To minimize the wall losses inside the reference probe, its length was chosen to be as low as 2 cm (Brockmann, 2001).
Using Ogden’s (1983) criteria (equation 1), it was calculated that the calm air concept is applicable as long as U0 < 8.6 cm/s. The following parameters were used in this calculation: D = 2 cm, US = 53 cm/s, ES = 90% and dae = 76.0 µm (the largest tested). Thus, according to Ogden’s criteria, the reference samplers operated in an essentially calm atmosphere.
As part of the method validation, comparative measurements were carried out at different orientations of reference probes: horizontal, vertical facing upwards and vertical facing downwards. In each test, three probes were utilized for each different orientation (nine in total) and these were located along the circumference around the aerosolization point (R = 100 cm for dae = 6.9, 14.1, 28.4 and 38.7 µm and R = 60 cm for dae = 60.1 and 76.0 µm). The probe orifices were 15 cm below the height of the aerosolization point (see Fig. 1, Front View). Ideally, the concentrations measured by identical probes at these orientations should be the same as long as the samplers are exposed to the same aerosol concentration. However, since the particle velocity has a vertical component due to gravitational settling, some differences associated with the orifice orientation were expected to occur, especially for larger particles.
Figure 3 demonstrates the ratios of the particle mass concentrations measured with the vertically oriented reference probe (upward and downward, respectively) to that measured with the horizontal probe (up/horizontal and down/horizontal). The concentration ratios are plotted against the particle aerodynamic diameter. Each experimental data point represents the average value obtained within one test and determined with 95% confidence interval. Two data points shown for some particle sizes represent replicates. In addition, the ratio of up/horizontal was calculated using the semi-empirical model of Grinshpun et al. (1993) for the aspiration efficiency of sharp-edged probes (this is the only model available at this time that works for upward-oriented and horizontal samplers operating in calm or very slow moving air environments; no adequate model is available to determine the down/horizontal ratio). The curve in Fig. 3 shows the calculated up/horizontal ratio as a function of dae. Both the experimental data and the model suggest that the aerosol concentration measured by the reference probe oriented horizontally is about the same as that obtained with the upward-oriented reference probe when sampling particles of dae < 80 µm. For very large particles, the model predicts that the upward-oriented probe should measure considerably greater concentrations than the horizontal one (up/horizontal increases from 1–1.5 to >2 when dae increases from 80 to 100 µm). This illustrates that particle gravitational settling affects the sampling efficiency of horizontal and vertical samplers in different ways. Since the upward-oriented probe over-samples by a factor of (1 + v) (Grinshpun et al., 1990, 1993), the overestimation is negligibly small when VS << US, but becomes more pronounced when the particle settling velocity VS is comparable to the sampling velocity US. The reference sampler oriented downwards was found to significantly under-sample, especially large particles. As a result of the above considerations, the horizontal orientation of reference probes was chosen for this study.
|
Generally, the concentration obtained with the reference sampler (Cref) differs from the true aerosol concentration in the environment (C0) by a factor of
. This factor should be taken into account when equation (4) is being used to determine the sampling efficiency from the data obtained with the horizontally oriented reference sampler. The factor represents the effect of the particle settling velocity (Brockmann, 2001). Hence, in this study the sampling efficiency ES of the test sampler was calculated as:
The correction factor , calculated using the semi-empirical model of Grinshpun et al. (1993) for the aspiration efficiency, is 1 for dae = 6.9, 14.1, 28.4 and 38.7 µm, is 0.90 for dae = 60.1 µm and is 0.76 for dae = 76.0 µm.
The particle wall deposition inside the reference probes was measured using a washing technique (Witschger et al., 1997). For all six powders, the deposited mass did not exceed 1% of the total sampled particulate mass.
Performance evaluation of the test system with reference samplers
The test system was evaluated with reference probes oriented horizontally. The spatial non-uniformity of the aerosol concentration measured in the test zone is caused by its variability along the circumference and its vertical gradient. The spatial concentration non-uniformity was determined for each of the six particle size fractions used in this study. The temporal variability of the aerosol concentration was also assessed.
First, n reference probes (n = 4–9, depending on the test) were equally distributed along the circumference at a distance R from the source and exposed to the same aerosol for 40 min. For each size fraction, the aerosol concentrations were measured at n different angular coordinates 360° k/n (where k is an integer 
[0;n]). From these measurements, the mean values of the aerosol concentration and the standard deviations were determined and the spatial variability, with respect to the angular coordinate, was characterized by the relative standard deviation.
Second, gravitational particle settling generally causes a vertical gradient of the aerosol concentration. Since the inhalable samplers were placed at a level of 100 cm from the floor, the vertical gradient was assessed within the area of ±10 cm of that level (taking into account the dimensions of the samplers). Thus, the aerosol concentration was measured for each particle size fraction at 110 (C110) and 90 cm (C90) from the floor and then compared with that measured at a height of 100 cm from the floor. The non-dimensional differences (bias) [(C110 – C100)/C100 and (C90 – C100)/C100, expressed as percentages] which characterize the vertical gradient of aerosol concentration in the test zone were calculated. The data on the spatial variability are summarized in Table 2. It is seen that the results obtained at different angular coordinates along the circumference are very consistent: RSD 
12%, which does not depend monotonically on the particle size. The average RSD of the pool was 8%, which is approximately equal to the one reported by Aitken et al. (1999) for the same particle sizes. The vertical concentration bias was found to be relatively small even for large particles.
|
Third, the aerosolization rate must be sufficiently high to perform an accurate gravimetric analysis and stable enough to ensure low temporal variability during each experimental run (up to 40 min). The temporal variability of the particle concentration was measured at one location (angular coordinate 0°; horizontal coordinate R = 100 cm; vertical coordinate = 100 cm from the floor). An optical particle size spectrometer (model 1.108; Grimm Technologies, Douglasville, GA) was used for these measurements. This device counts the particles ranging from 0.3 to
20 µm and records them in 15 channels. Due to the upper particle size limit of the Grimm monitor, the tests were carried out only with the two finest powders (dae = 6.9 and 14.1 µm). The temporal variability, characterized by the RSD of the mean particle concentration, was <15%. Fourth, although our theoretical assessment has proven that the particles do not project from the rotating disperser towards the sampling point, this was verified experimentally. The probes, located at a distance R = 60 or 100 cm from the source, were exposed to powders aerosolized by the disperser, while operating as passive samplers (no suction through the sampling orifice was applied). Gravimetric analysis detected no particulate mass gain on the collection filters or the internal walls of the reference probes facing the source. This confirms our conclusion (made earlier from the particle stopping distance calculation) that the projection effect is negligible. As a consequence, no sampling bias occurred due to the proximity of the samplers to the source.
Fifth, as a part of our study design validation, the test system was used with CALTOOL, for which the sampling efficiency measured in very slow moving air has been made available by the Institute of Occupational Medicine (Edinburgh, UK) and the National Institute for Working Life (Solna, Sweden) (reported in Witschger et al., 2001). CALTOOL is a calibration tool, which is currently being developed within European project SMT4-CT98-2254 to test personal inhalable samplers on a simplified full-scale manikin in the workplace (Mark et al., 2000). The sampling efficiencies of the CALTOOL mouth inlet (D = 15 mm and Us 190 cm/s) were measured for particle sizes ranging from 5 to 40 µm, using the newly developed test system. The results were compared with those obtained in the two above-mentioned studies (reported in Witschger et al., 2001), which followed the Aitken et al. (1999) protocol for measuring inhalability in a low air movement environment. No statistically significant difference, with respect to the CALTOOL sampling efficiency, was found (the findings are described in detail in a separate publication by Fauvel et al., 2003). Thus, our test method and facility were found to be adequate for performance evaluation of inhalable samplers in very slowly moving air.
Inhalable personal samplers selected for the performance evaluation
The IOM inhalable sampler was selected because it is commonly used worldwide for personal inhalable sampling. At the same time, information on its performance characteristics in calm air or very low air movement environments is limited to a single investigation by Kenny et al. (1999). In the IOM Sampler used in this study, a stainless steel cassette houses a 25 mm collection filter. The sampler body is made of a conductive black plastic. The aerosol particles are aspirated through a 15 mm circular protruding inlet at a flow rate of 2 l/min. The IOM Sampler operating instructions specify that the collection filter and the lightweight cassette should be weighted as a single unit. Hence, the entire particulate mass aspirated into the sampler is accounted for (Mark and Vincent, 1986). This protocol, however, is incompatible with some analytical methods, in which only the particles collected on the filter surface constitute the sample to be quantified and no additional procedure involving the washing of the cassette internal walls is advised. For instance, in the analysis of bioaerosol samples, microbial enumeration by fluorescence microscopy is conducted directly on the collection filter (Eduard et al., 2001); the quantitative analysis of radioactive aerosol samples by counting is also performed directly on the filter (Hoover and Newton, 2001). Several studies have shown that an appreciable fraction of the aspirated particulate mass is deposited on the inner walls of the IOM cartridge and thus does not reach the sampling filter. As a result of his wind- tunnel study, Mark (1990) reported that the deposition efficiency ranged from 20 to 45% for a particle size range dae = 6–34 µm. Kenny et al. (1997) referred to a deposition efficiency range of 0–75% for particle sizes between 7 and 100 µm. Both studies were performed in moving air (>0.5 m/s) using a rotating manikin. The field study of Lidén et al. (2000), which involved various dusts (generated from thermosetting plastics, wood, paper and animal breeding), showed that the wall losses in the IOM Sampler ranged from 24 to 36%. Some studies indicate lower values, e.g. when analysing metals from field samples obtained with the IOM Sampler, Demange et al. (2002) found that inner wall deposition inside the cassette was between 2 and 10%. Although no clear particle size dependency of the deposition efficiency was shown in the above investigations, a low entry velocity of the IOM Sampler (19 cm/s) suggests that wall deposition should be greater for larger particles. While the IOM has been extensively evaluated in wind tunnels, very limited information is available on its performance in a calm atmosphere or low air movement environments. There is also a lack of laboratory data on the performance of the IOM Sampler when gravimetric analysis was conducted only with a filter (as opposed to a combination of a filter with a cassette). Thus, in this study the IOM filters were weighted separately from the cassette assembly, which enabled us to address the internal particle deposition issue.
The Button Personal Inhalable Aerosol Sampler with a porous curved inlet has been increasingly used in occupational environments because of several features, such as the low sensitivity of its sampling efficiency to ambient conditions, low inter-sample variability and good collection uniformity (Kalatoor et al., 1995; Aizenberg et al., 1998, 2000b). This sampler has also been found to be suitable for total enumeration of bacterial and fungal airborne spores (Aizenberg et al., 2000c). Inhalable aerosol particles aspirated through 381 µm orifices of the Button Sampler’s porous inlet operating at 4 l/min are deposited on a 25 mm filter located directly behind the inlet (this design allows elimination of wall losses). While the Button Sampler has been extensively evaluated in wind tunnels, no laboratory data have been published on its performance in calm air or low air movement environments.
The closed-face 25 mm Millipore filter holder (C25) has been widely utilized in Europe. Being the only personal dust sampler listed in the French standard (AFNOR, 1988) for ‘inspirable’ (inhalable) particles, it has been used in France to characterize industrial exposure to uranium compounds in the nuclear industry (Ansoborlo et al., 1989). In Norway and Denmark the C25 sampler has been utilized as a ‘total dust’ sampler. Some sampling cassettes common in other countries, e.g. the 37 mm filter cassette, are essentially similar to the 25 mm Millipore filter holder. The C25 sampler is a three piece cassette made of a non-conductive plastic. In this sampler the aerosol particles are aspirated through a 4 mm diameter orifice with a flow rate of 1 l/min and collected on a 25 mm filter. When attached to the worker’s collar bone, the inlet is always facing downward with its axis at an angle of 45° to the vertical (Buchan et al., 1986). Although this personal sampler is widely used, it has a number of limitations (Paskar et al., 1991; Hinds, 1999; Demange et al., 2002). First, the non-conductive plastic material enhances electrostatic losses of charged particles. Second, the cassette assembly often results in misalignment between parts, causing bypass leakage that may affect sampling efficiency (Baron et al., 2002). Third, non-uniform deposition on the collection filter imposes a considerable limit on the methods that can be used to analyse samples obtained with the C25 cassette. Fourth, the orientation of the C25 inlet orifice, when it is attached to the worker’s body (pointing 45° downward), is a factor causing under-sampling of large inhalable particles. The sampling efficiency of the C25 cassette in calm air or low air movement environments has not been reported.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Sampling efficiency
Figure 4 presents the average sampling efficiencies of the IOM, Button and C25 Samplers operating in a low air movement environment facing the aerosol source. The points represent our experimental data obtained for six aerodynamic particle diameters: 6.9, 14.1, 28.4, 38.7, 60.1 and 76.0 µm. Each data point was determined as an average value of at least three replicates with the standard deviation calculated for the 95% confidence interval (shown as error bars). The internationally standardized ISO–ACGIH–CEN inhalability convention (CEN, 1993; ISO, 1995; ACGIH, 1999) and the recently proposed ‘low wind inhalability’ curve (equation 2; Aitken et al., 1999) are also plotted in Fig. 4.
|
The C25 cassette shows rather poor sampling efficiency, which rapidly decreases from 72% at 6.9 µm to <7% at 38.7 µm. It almost reaches 0 at dae
60 µm. These results are comparable with the data obtained in the study of Kenny et al. (1999), which followed the slowly rotating protocol when testing a freely suspended 37 mm closed-face cassette operating in calm air at 2 l/min. Comparison of the 25 and 37 mm closed face cassettes is legitimate since both devices sample through a 4 mm inlet and operate in a downward pointing orientation. Having a 4 mm opening and 25 mm body, the downward-oriented C25 cassette does not create a sufficient flow field in the vicinity of its body to ‘turn around’ the large particles that settle rapidly due to gravity. The Button Sampler exhibits a sampling efficiency that slowly decreases with particle size: from 99% at 6.9 µm to 59% at 76.0 µm. As seen from Fig. 4, the data points are slightly above the inhalability convention (CEN, 1993; ISO, 1995; ACGIH, 1999) and close enough to the low wind inhalability curve proposed by Aitken et al. (1999) (see equation 2). These data suggest that the Button Sampler is suitable for inhalable aerosol sampling in very slowly moving air when it operates facing the particle source.
The IOM sampling efficiency presented in Fig. 4 was obtained by weighing the filter cassette assembly as a single unit (according to the manufacturer’s protocol). The data show that the IOM samples overestimate the aerosol concentration: the sampling efficiency was found to be >100% for all particles sizes, increasing from 126% at dae = 6.9 µm to 178% at 28.4 µm and then decreasing to 100% at dae = 76.0 µm. These values are greater than those reported by Kenny et al. (1999) for a freely suspended IOM Sampler operating in a low air environment.
The IOM sampling efficiencies determined in our study following the traditional protocol (filter + cartridge) and a non-traditional protocol (filter only), respectively, are presented in Fig. 5. The data obtained in two other studies (HSL and IOM), which were performed with a freely suspended (off-manikin) IOM Sampler following the traditional protocol over the particle size range 6–90 µm, are also presented in Fig. 5. For particle sizes ranging from 5 to 50 µm our data set, obtained following a non-traditional (filter only) protocol, is in better agreement with the Kenny et al. data than for larger particles. Particle deposition inside the cassette cartridge is considerable. The deposition efficiency was found to be particle size dependent and increased from 20% at dae = 6.9 µm to 55% at dae = 76.0 µm.
|
Since direct particle projection from the disperser was shown to have no significant effect on the sampling efficiency of reference probes, projection was not believed to be a cause of the over-sampling by the IOM Sampler (used according to the traditional protocol). Nevertheless, a special experiment was conducted to prove this point. Similarly to the tests with reference probes, which were operated without suction, three unpumped IOM Samplers were exposed to the aerosolized powder of dae = 14.1 µm. The IOM filter and the cassette assembly were weighted as a single unit before and after the test. As expected, no measurable mass gain was found.
The over-sampling by the IOM Sampler observed in this study was attributed to air turbulence. In perfectly calm air with no turbulence, the force of gravity primarily governs particle motion. In relatively fast moving air, the primary force is drag due to the motion of the moving air. In very slowly moving air, with pronounced turbulence, the drag caused by the air turbulence itself and the force of gravity both determine particle motion from the aerosol generation point to the sampling zone. The ability of a particle to respond to the turbulent motion of the air that surrounds it can be assessed through the dimensionless inertial parameter (Vincent, 1989):
where 
is the particle relaxation time, u' is the fluctuating component of the air velocity and L is the characteristic length scale of the air turbulence. The larger the value of K, the weaker is the particle response to chaotic air motion.
There is a region in close proximity to the inlet where the ambient air turbulence interacts with the flow field, which is formed of aspiration and pre-inlet turbulence. The turbulent motion allows particles to penetrate through the interaction region and become captured by the suction air flow near the inlet. The capture is more pronounced as the value of K increases. Once aspirated into the inlet, these particles primarily deposit on the inner walls of the inlet due to turbulent diffusion (which results in a positive net gain, as was observed in our experiments). However, if the particle size increases further, gravity becomes predominant over turbulent motion. The effect of air turbulence on sampling efficiency is rather complex. It has been investigated in very few studies, all of which were conducted in moving air (reviewed by Vincent, 1989) and not in a calm atmosphere or low air movement environments.
Air turbulence in the microenvironment of a sampler seems to be more pronounced for the IOM Sampler than for the Button Sampler, since the latter has a curved inlet with small holes that reduces the intensity of pre-inlet air turbulence. Comparison with the C25 cassette, with respect to the turbulence effect, is not representative because the sampling efficiency of the cassette decreases rapidly to almost 0 for particle sizes of 40–100 µm.
The data in Fig. 5 suggest that the application of a non-traditional protocol (weighing the IOM filter only) results in a better fit of the sampling efficiency data to the low wind inhalability curve and a lower data variability. This finding is particularly useful in practical situations where sample analysis must be performed directly on the collection filter.
Precision
The sampler precision was evaluated by determining the relative standard deviation of the aerosol concentration obtained by this sampler in a specific experiment (at least three replicates). A total of 16 experiments were carried out with each sampler type (including the three types of test samplers and the reference probes) to characterize sampler precision. Table 3 presents the maximum, minimum and average (pooled) RSD values. The reference (REF) sampler and the Button Sampler showed average RSD values of 10 and 11%, respectively, with RSDmax < 25%. The IOM Sampler and the 25 mm closed-face cassette exhibited average RSD values of 14 and 35%, respectively, with relatively high RSDmax values (64 and 58%). The data presented in Table 3 suggest that the Button Sampler has the highest precision among the three personal samplers tested in this study. This finding agrees with the wind-tunnel data reported by Aizenberg et al. (2000b).
|
IOM Sampler in two configurations: freely suspended and attached to the torso
Kenny et al. (1999) suggested that the performance tests of a personal sampler in a calm air environment can be conducted with a freely suspended (isolated) sampler. In other words, in contrast to testing in moving air environments, it is not necessary to attach the personal sampler to a manikin. This finding was verified using our newly developed experimental facility, in which the IOM Sampler was evaluated while freely suspended and while mounted on a Simplified Torso (Witschger et al., 1998), respectively. The experiments were performed with the six SPM powders listed in Table 1. In each experiment at least seven samplers operated simultaneously, including at least three reference probes, three isolated IOM Samplers and one IOM Sampler mounted on the Simplified Torso. The mounted sampler was attached to the front of the torso, thus facing the aerosol source (orientation angle 0°). All samplers where positioned at the same height from the floor (100 cm) with their entry orifices at a distance R from the source (see Fig. 1, Top View).
The IOM sampling efficiency data obtained with isolated (freely suspended) samplers are plotted in Fig. 6 against the data obtained with the sampler mounted on the Simplified Torso facing the source (0°). In both configurations the samplers’ orientation with respect to the source and gravity were the same. Two sets of data are presented: when the samples were analysed following the traditional (white diamonds) and non-traditional (black diamonds) protocols, respectively. Each data point represents the efficiency values obtained for a specific aerodynamic particle diameter. A Mann–Whitney (Wilcoxon) test was run to compare the medians of the two series of data. We found that in our experimental facility the IOM Sampler performs in the same way regardless of whether it is freely suspended or mounted on the Simplified Torso (P = 0.47 at the 95% confidence interval). This confirms the finding of Kenny et al. that it is not necessary to attach the personal sampler to a body to evaluate the sampling efficiency in very slowly moving air.
|
To summarize, the sampler evaluation methodology developed in this study for low air movement environments allows the investigator to use a direction-specific approach, which represents a common situation when the worker faces the source during exposure. The direction-specific approach (0°) seems to be suitable to represent many occupational exposure situations (Lidén and Kenny, 1994).
Sampling efficiency deviation from the ‘target’ levels
Based on the data presented in Fig. 4, we calculated the difference between the aerosol mass concentration measured by a specific sampler for a specific particle size fraction and the ‘target’ concentration of this fraction. The three target levels represent: (i) a 100% efficient sampler, (ii) a sampler that has an efficiency that follows the inhalability convention and (iii) a sampler with an efficiency that fits the low wind inhalablity curve of Aitken et al. (1999). Table 4 presents these deviations as percentages determined for three log-normal particle size distributions with MMAD of 5, 10 and 15 µm, respectively, and GSD of 2. The MMADs represent the ‘lower end’ of the inhalable particle size range and are typical for a number of occupational settings, including work environments with radioactive aerosols, as reviewed by Dorrian and Bailey (1995). The International Commission on Radiological Protection (ICRP, 1994) has selected MMAD = 5 µm as a default MMAD value for occupational exposure in the nuclear industry. To determine the sampling efficiencies of the IOM, Button and C25 Samplers for the selected particle sizes, the experimental data presented in Fig. 4 were approximated by polynomial functions.
|
A comparison of the sampling efficiency of a specific sampler with the 100% level is meaningful to the nuclear industry since the effective dose of radionuclides is usually assessed through the ‘total’ aerosol concentration measured by a stationary sampler, which is supposed to exhibit ES = 100% for all particle sizes. Comparison with this ‘target’ level, the IOM Sampler over-samples the particles of MMAD = 5–15 µm by 29–47%, while the C25 under-samples them by 33–67%. The efficiency of the Button Inhalable Aerosol Sampler is only slightly below 100%: the bias ranges from –3 to –12%.
The sampling efficiencies were also compared with the international inhalable convention. The legitimacy of this comparison is questionable since the convention is based primarily on measurements of the aspiration efficiency of mouth breathing manikins tested in wind tunnels over a range of wind speeds from 0.5 to 4 m/s (Kenny et al., 1999). However, as mentioned earlier, this convention was widely accepted as a standard and is presently being used for the evaluation of inhalable aerosol samplers in various environments, irrespective of the air flow characteristics (the only limitation on wind speed mentioned in the standard is 4 m/s). The efficiencies of both the IOM and the Button Samplers were above the inhalability standard. The deviation was found to be very significant for the IOM Sampler (up to 110%), while it did not exceed 26% for the Button Sampler (MMAD = 15 µm). The data obtained with the 25 mm closed-face cassette were below the inhalable convention by 54% for MMAD = 15 µm. As discussed above, this deviation becomes dramatic with a further increase in particle size.
As compared with the low-wind inhalability curve specifically developed for low air movement environments (Aitken et al., 1999), the IOM Sampler considerably over-samples, with a deviation well above 30% for MMAD = 5–15 µm. The C25 cassette under-samples relative to the low-wind inhalability curve, with the bias being >30% for MMAD 5 µm. The Button Sampler shows a very good agreement with the curve: the bias lies between –1 and –5%.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
In many occupational settings workers operate in very slowly moving air environments (essentially calm atmosphere) and face the aerosol source in close proximity to it. Personal aerosol samplers (used to measure human exposure in these environments) are usually characterized based on wind tunnel testing. However, the air flow in a wind tunnel is significantly higher than that in a typical indoor occupational environment. Furthermore, most of the personal sampler evaluation techniques are based on a direction-averaged approach that eliminates sampling efficiency dependence on sampler orientation with respect to the particle source. This integrated approach is valid only when the worker is located far from the source and exposure is not direction-dependent.
To address the above contradictions, a new, low cost method and facility were developed. The method allows evaluation of the performance characteristics of personal aerosol samplers in a low air movement environment when the samplers face the point source and operate in close proximity to the source. In the new system, an aerosol generator, which is located in the centre of a room-sized, non-ventilated chamber continuously rotates and omnidirectionally aerosolizes test particles of a specific size that are subsequently transported by turbulent diffusion and natural convection. The test and reference samplers, operating in parallel, are equally distributed along a circumference around the source. For each specific particle size and each specific sampler type, the aerosol mass concentration is measured through gravimetric analysis of the collection filter before and after sampling.
The new protocol was validated and subsequently utilized to evaluate three widely used aerosol samplers: the IOM Personal Inhalable Sampler, the Button Personal Inhalable Aerosol Sampler and the 25 mm Millipore filter holder (closed-face C25 cassette). The sampling efficiencies of each sampler were measured with six particle fractions, MMAD = 6.9–76.0 µm. For each particle size, the data were compared with the three ‘target’ levels: the 100% efficiency level, the efficiency indicated by the international ISO–ACGIH–CEN inhalability convention, and the efficiency predicted by the recently proposed low wind inhalability curve (Aitken et al., 1999).
The IOM Sampler facing the source (Fig. 5) was found to over-sample compared with the data of Kenny et al. (1999), which were obtained using a slowly rotating protocol for the sampler freely suspended in a low air movement environment. It was found that particle wall deposition in the IOM cassette cartridge was rather high and particle size-dependent. On the one hand, according to the traditional protocol for gravimetric analysis of IOM samples, the filter is weighed together with the entire cassette assembly. On the other hand, the data suggest that the IOM Sampler efficiency fits the low wind inhalability curve and is characterized by a higher precision when applying a non-traditional analysis protocol (weighting the filter only). This finding is of particular importance in situations when it is either necessary or preferable to perform the analysis directly on its collection filter without washing the cassette and eluting the filter deposit (e.g. scanning microscopy, X-ray analysis, fibre counting, radioactive particle analysis and total spore enumeration). A separate set of experiments, performed with one IOM Sampler mounted on the Simplified Torso (Witschger et al., 1998), showed that its sampling efficiency was the same regardless of whether the sampler was on the Simplified Torso facing the source or freely suspended (Fig. 6). This result confirms the findings of Kenny et al. obtained in low air movement environments.
The Button Sampler data demonstrated good agreement with the standard inhalability convention and especially with the low wind inhalability curve. The Button Sampler showed the highest precision (the average RSD was 11%) among the three personal samplers tested in this study.
The 25 mm Millipore filter holder (C25 cassette) was found to considerably under-sample particles >10 µm; its efficiency did not exceed 7% for particles of 40–100 µm (reaching essentially 0 at dae 60 µm). The C25 sampler showed the lowest precision (the average RSD was 35%).
Over-sampling by the IOM Sampler, as well as the discrepancy between the IOM sampling efficiencies obtained in this study and those reported by Kenny et al. (1999), were attributed to an air turbulence effect.
In summary, the methodology and the experimental set-up developed in this study for low air movement environments were found to be suitable for testing personal aerosol samplers. The new method allows the investigators to use a direction-specific approach, which represents a common situation when the worker primarily faces the source during exposure. The newly developed sampler evaluation system is easily duplicable and can be used to test a number of samplers simultaneously in a time- and cost-effective way.
Acknowledgements—the authors expressed their sincere appreciation to Ms Alexandra Appatova for her valuable help in editing this paper. This work was partially performed while Dr Witschger was working at the Institut de Radioprotection et de Sûreté Nucléaire (IRSN).
|
FOOTNOTES |
|---|
* Author to whom correspondence should be addressed. Tel: +33-3-83-50-98-38; fax: +33-3-83-50-20-60; e-mail: olivier.witschger{at}inrs.fr
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
ACGIH. (1999) Threshold limit values for chemical substances and physical agents and biological exposure indices. Cincinnati, OH: American Conference of Governmental Industrial Hygienists. p. 85.
Agarwal JK, Liu BYH. (1980) A criterion for accurate aerosol sampling in calm air. Am Ind Hyg Assoc J; 41: 191–7.[Web of Science][Medline]
AFNOR. (1988) Qualité de l’air—air des lieux de travail—prélèvement individuel de la fraction inspirable de la pollution particulaire, NF X 43-257. Paris La Défense: Association Française de Normalisation.
Aitken RJ, Baldwin PEJ, Beaumont GC, Kenny LC, Maynard AD. (1999) Aerosol inhalability in low air movement environments. J Aerosol Sci; 30: 613–26.[CrossRef]
Aizenberg V, Bidinger E, Grinshpun SA, Willeke K, Hamed A, Tabakoff W. (1998) Air flow and particle velocities near a personal aerosol sampler with Curved, porous aerosol sampling surface. Aerosol Sci Technol; 28: 247–58.
Aizenberg V, Grinshpun SA, Willeke K, Smith J, Baron PA. (2000a) Measurement of the sampling efficiency of personal inhalable aerosol samplers using a simplified protocol. J Aerosol Sci; 31: 169–79.[CrossRef]
Aizenberg V, Grinshpun SA, Willeke K, Smith J, Baron PA. (2000b) Performance characteristics of the button personal inhalable aerosol sampler. Am Ind Hyg Assoc J; 61: 398–404.
Aizenberg V, Reponen T, Grinshpun SA, Willeke K. (2000c) Performance of Air-O-Cell, Buckard, and Button samplers for total enumeration of airborne spores. Am Ind Hyg Assoc J; 61: 855–64.
Aizenberg V, Choe K, Grinshpun SA, Willeke K, Baron PA. (2001) Evaluation of personal aerosol samplers challenged with large particles. J Aerosol Sci; 32: 779–93.[CrossRef]
Ansoborlo E, Bérard P, Chalabreysse J. (1989) Study of industrial exposure to class Y uranium compounds: methods and results. Radiat Protect Dosim; 26: 101–5.[Abstract]
Association Française de Normalisation. (1995) General methods of tests for pigments and extenders—Part 23: Determination of density (using a centrifuge to remove entrained air), EN ISO 787-23. Paris La Defense: Association Française de Normalisation.
Baldwin PEJ, Maynard AD. (1998) A survey of wind speed in indoor workplaces. Ann Occup Hyg; 20: 303–13.
Baron PA, Khanina A, Bennett JS, Martinez AB, Grinshpun SA. (2002) Investigation of filter bypass leakage and a test for aerosol sampling cassettes. Aerosol Sci Technol; 36: 857–65.[CrossRef]
Baturin VV. (1972) Fundamentals of industrial ventilation, 3rd edn. Oxford: Pergamon Press.
Berry RD, Froude S. (1989) An investigation of wind conditions in the workplace to assess their effect on the quantity of dust inhaled, Internal Report IR/L/DS/89/3. London: Health and Safety Executive.
Brockmann JE. (2001) Sampling and transport of aerosols. In Baron PA, Willeke K, editors. Aerosol measurement: principles, techniques, and applications, 2nd edn. New York: Wiley Interscience. pp. 143–95.
Buchan RM, Soderholm SC, Tillery MI. (1986) Aerosol sampling efficiency of 37 mm filter cassettes. Am Ind Hyg Assoc J; 47: 825–31.[Web of Science][Medline]
CEN. (1993) Workplaces atmospheres: size fraction definitions for measurements of airborne particles in the workplaces, CEN Standard EN 481. Brussels: Comité Européen de Normalisation.
Chung W, Vincent JH. (2002) New experimental studies to directly measure aspiration efficiencies of aerosol samplers in calm air. J Aerosol Sci; 33: 103–18.[CrossRef]
Council of the European Union. (1996) Basic safety standards for the protection of the health of workers and the general public against the danger arising from ionising radiation, Council Directive 96/29/Euratom. Off J Eur Union; L159: 1–114.
Davies CN. (1968) The entry of aerosols into sampling tubes and heads. Br J Appl Phys; 25: 921–32.
Demange M, Görner P, Elcabache JM, Wrobel R. (2002) Field comparison of 37-mm closed-face cassettes and IOM samplers. Appl Occup Environ Hyg; 17: 200–8.[CrossRef][Medline]
Dorrian MD, Bailey MR. (1995) Particle size distributions of radioactive aerosols measured in workplaces. Radiat Protect Dosim; 60: 119–33.[Abstract]
Dunnett SJ. (2002) Particle motion in the vicinity of a bulky sampling head operating in calm air. Aerosol Sci Technol; 36: 306–17.
Dunnett SJ, Wen X. (2002) A numerical study of the sampling efficiency of a tube sampler operating in calm air facing both vertically upwards and downwards. J Aerosol Sci; 33: 1653–65.[CrossRef]
Eduard W, Blomquist G, Nielsen BH, Heldal KK. (2001) Recognition errors of micro-organisms by fluorescence microscopy. Ann Occup Hyg; 46: 493–8.[CrossRef]
Fauvel S, Basso G, Witschger O. (2003) Laboratory and field testing of a calibration tool (CALTOOL) for evaluating personal aerosol samplers performances. J Aerosol Sci; 34: S1177–8.
Gibson H, Ogden TL. (1977) Some entry efficiencies for sharp-edged samplers in calm air. J Aerosol Sci; 8: 361–5.[CrossRef]
Grinshpun SA, Lipatov GN, Semenyuk TI. (1989) A study of sampling of aerosol particles from calm air into thin-walled cylindrical probes. J Aerosol Sci; 20: S1561–4.[CrossRef]
Grinshpun SA, Lipatov GN, Sutugin AG. (1990) Sampling errors in cylindrical nozzles. Aerosol Sci Technol; 12: 716–40.
Grinshpun SA, Willeke K, Kalatoor S. (1993) A general equation for aerosol aspiration by thin-walled probes in calm and moving air. Atmos Environ; 27A: 1459–70. Corrigendum (1994) Atmos Environ; 28: 375.[CrossRef]
Guffey SE, Flanagan ME, van Belle G. (2001) Air sampling at the chest and ear as representative of the breathing zone. Am Ind Hyg Assoc J; 62: 416–27.
Hinds WC. (1999) Sampling for inhalable aerosols. In Particle size-selective sampling for particulate air contaminants. Cincinnati, OH: American Conference of Governmental Industrial Hygienists. pp. 119–40.
Hoover MD, Newton GJ. (2001) Radioactive aerosols. In Baron PA, Willeke K, editors. Aerosol measurement: principles, techniques, and applications, 2nd edn. New York: Wiley Interscience. pp. 979–1009.
ICRP. (1994) Human respiratory tract model for radiologocal protection. ICRP Publication 66. Oxford: Elsevier Science Ltd. p. 42.
ISO. (1995) Air quality—particle size fraction definition for health-related sampling, ISO Standard 7708. Geneva: International Organization for Standardization.
Kalatoor S, Grinshpun SA, Willeke K. (1995) Aerosol sampling from fluctuating flows into sharp-edged tubular inlets. J Aerosol Sci; 26: 387–98.[CrossRef]
Kennedy NJ, Tatyan K, Hinds WC. (2001) Comparison of a simplified and full-sized mannequin for the evaluation on inhalable sampler performance. Aerosol Sci Technol; 35: 564–8.[CrossRef]
Kenny LC. (2000) The international conventions for health-related sampling of aerosols—a review of current status and future evolution. Appl Occup Environ Hyg; 15: 68–71.[CrossRef][Medline]
Kenny LC, Aitken R, Chalmers CP, et al. (1997) A collaborative European study of personal inhalable aerosol sampler performance. Ann Occup Hyg; 41: 135–53.
Kenny LC, Aitken RJ, Baldwin PEJ, Beaumont GC, Maynard AD. (1999) The sampling efficiency of personal inhalable aerosol samplers in low air movement environments. J Aerosol Sci; 30: 627–38.[CrossRef]
Kulmala I, Säämänen A, Enbom S. (1996) The effect of contaminant source location on worker exposure in the near wake region. Ann Occup Hyg; 40: 511–23.
Li SN, Lundgren DA, Rovell-Rixx D. (2000) Evaluation of six inhalable aerosol samplers. Am Ind Hyg Assoc J; 61: 506–16.
Lidén G, Kenny LC. (1994) Errors in inhalable dust sampling for particles exceeding 100 µm. Ann Occup Hyg; 38: 373–84.
Lidén G, Melin B, Lidboln A, Lindberg K, Norén J-O. (2000) Personal sampling in parallel with open-face filter cassettes and IOM samplers for inhalable dust—implications for occupational exposure limits. Appl Occup Environ Hyg; 15: 263–76.[CrossRef][Medline]
Mark D. (1990) The use of dust-collecting cassettes in dust samplers. Ann Occup Hyg; 34: 281–91.
Mark D, Vincent JH. (1986) A new personal sampler for airborne total dust in workplaces. Ann Occup Hyg; 30: 89–102.
Mark D, Vincent JH, Gibson H, Whitherspoon WA. (1985) Applications of closely graded powders of fused alumina as test dusts for aerosol studies. J Aerosol Sci; 16: 125–31.
Mark D, Aitken RJ, Beaumont G, et al. (2000) development of a novel calibration tool for workplace aerosol samplers. Review of progress of EU project. J Aerosol Sci; 31: S392–3.[CrossRef]
Matthews TG, Thompson CV, Wilson DL, Hawthorne AR, Mage DT. (1989) Air velocities inside domestic environments: an important parameter in the study of indoor air quality and climate. Environ Int; 15: 545–50.[CrossRef]
Maynard AD, Jensen PA. (2001) Aerosol measurement in the workplace. In Baron PA, Willeke K, editors. Aerosol measurement: principles, techniques, and applications, 2nd edn. New York: Wiley Interscience. pp. 779–99.
Ogden TL. (1983) Inhalable, inspirable and total dust. In Marple VA, Liu BYH, editors. Aerosols in the mining and industrial work environments, volume 1. Ann Arbor, MI: Science. pp. 185–95.
Ogden TL, Birkett JL. (1978) An inhalable dust sampler for measuring the hazard from total airborne particulate. Ann Occup Hyg; 21: 41–50.
Paskar MA, Harkins JM, Moomey JD, Hecker LH. (1991) Internal wall losses of pharmaceutical dusts during closed-face, 37-mm polystyrene cassette sampling. Am Ind Hyg Assoc J; 52: 280–6.[Web of Science][Medline]
Ramachandran G, Sreenath A, Vincent JH. (1998) Towards a new method for experimental determination of aerosol sampler aspiration efficiency in small wind tunnels. J Aerosol Sci; 29: 875–91.[CrossRef]
Sreenath A, Ramachandran G, Vincent JH. (2002) Aspiration characteristics of idealized blunt aerosol samplers at large angle to the wind. J Aerosol Sci; 33: 871–81.[CrossRef]
Su WC, Vincent JH. (2002) New experimental studies to directly measure aspiration efficiencies of aerosol samplers in calm air. J Aerosol Sci; 33: 103–18.
Vincent JH. (1989) Aerosol sampling: science and practice. Chichester: Wiley.
Vincent JH. (1995) Aerosol science for industrial hygienists. Oxford: Elsevier Science Ltd.
Welling I, Andersson IM, Rosen G, Räisänen I-M, Mielo T, Marttinen K, Niemela R. (2000) Contaminant dispersion in the vicinity of a worker in a uniform velocity field; Ann Occup Hyg; 44: 219–25.
Whicker JJ, Baker GD, Wasiolek PT. (2000) Quantative measurements of airflow inside a nuclear laboratory. Health Phys; 79: 716–21.
Wiasiolek PT, Whicker JJ, Gong H, Rodgers JC. (1999) Room airflow studies using sonic anemometers. Indoor Air; 9: 125–33.[CrossRef][Web of Science][Medline]
Witschger O. (2000) Sampling of airborne dusts in workplaces atmospheres. Kerntechnik; 65: 28–33.
Witschger O, Wrobel R, Fabriès JF, Görner P, Renoux A. (1997) A new experimental wind tunnel facility for aerosol sampling investigation. J Aerosol Sci; 28: 833–51.[CrossRef]
Witschger O, Willeke K, Grinshpun SA, Aizenberg V, Smith J, Baron PA. (1998) Simplified method for testing personal inhalable aerosol samplers. J Aerosol Sci; 29: 855–74.[CrossRef]
Witschger O, Mark D, Aitken RJ, et al. (2001) Development of a novel calibration tool for workplace aerosol samplers. In Abstracts of the 17th French Aerosol Conference, CFA 2001. pp. 148–154.
Yoshida H, Uragami M, Masuda H, Iinoya K. (1978) Particle sampling efficiency in still air. Kagaku Kogaku Ronbunshu; 4: 123–8.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  S. J. Reynolds, J. Nakatsu, M. Tillery, T. Keefe, J. Mehaffy, P. S. Thorne, K. Donham, M. Nonnenmann, V. Golla, and P. O'shaughnessy Field and Wind Tunnel Comparison of Four Aerosol Samplers Using Agricultural Dusts Ann. Hyg., August 1, 2009; 53(6): 585 - 594. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Gori, M. Carrieri, M. L. Scapellato, G. Parvoli, D. Ferrara, R. Rella, A. Sturaro, and G. B. Bartolucci 2-Methylanthraquinone as a Marker of Occupational Exposure to Teak Wood Dust in Boatyards Ann. Hyg., January 1, 2009; 53(1): 27 - 32. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. M. PETERS, D. OTT, and P. T. O'SHAUGHNESSY Comparison of the Grimm 1.108 and 1.109 Portable Aerosol Spectrometer to the TSI 3321 Aerodynamic Particle Sizer for Dry Particles Ann. Hyg., November 1, 2006; 50(8): 843 - 850. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||