Analyst, January 1994, Vol. 119 27 Performance Parameters for Assessing the Acceptability of Aerosol Sampling Equipment* Goran Lid& National Institute of Occupational Health, S-171 84 Solna, Sweden The European standardization organization (CEN) has been asked by the General Commission of the European Community to prepare standards for chemical agents in workplace atmospheres. The CEN Technical Committee 137 'Assessment of Workplace Exposure' has divided its work into working groups. Its main area of work is concentrated on sampling strategy; general performance requirements; sampling conventions for particulate matter; and performance requirements for gas and vapour sampling instruments, aerosol samplers, and personal sampling pumps. This paper presents the test procedure the CEN working group on particulate matter has proposed that shows how to evaluate the performance of aerosol samplers.The draft requirements on personal sampling pumps for dust sampling are also presented. The draft test procedure for aerosol samplers will be tested and evaluated in a pan-European test of samplers for inhalable dust, which is presented below (the study is ongoing). Keywords: Aerosol; performance; pumps; sampler; standardization Development of an Aerosol Sampler The modern approach of designing samplers from sampling conventions, rather than designating a specific existing sampler to be a reference sampler, irrespective of how it samples, is based on two factors: medicaVphysiologica1 knowledge of the human respiratory system and a conviction that it is technically possible to construct a sampler with the desired sampling characteristics.Although the human respiration system consists of many component parts, it may, from a medicaVphysiologica1 stand- point, be divided into three different regions with similar properties. The three regions are the head airways, the trachea and bronchi, and the alveolar region. These regions differ in function and surface tissues, air residence times, particle separation mechanisms and capacities, and clearance mechanisms.' The head airways act as an inlet and clean the inhaled air of large particles. Most deposited particles are cleared in a few minutes by sneezing or by swallowing, although some particles, e.g., those of hard wood can reside there for a long time and cause nasal cancer.2 The tracheo- bronchial region transports the air to the part of the lung where gaseous exchange takes place.The particle separation capacity here is low, and what has deposited is transported towards the mouth and swallowed within a few hours. In the alveolar region gas exchange between the air and the blood occurs. In this region only a small fraction of the air is changed with each breath, and hence the long residence time leads to considerable particle deposition. Deposited particles are either transported towards the bronchioles by macrophages, * Presented at the Conference on Modern Principles of Workplace Air Monitoring: Pumped and Diffusive Sampling for Contaminants, Geilo, Norway, February 15-18, 1993. dissolved in the blood, or moved into the blood or lymphatic systems.The removal time for these deposited particles may be up to a few months. From an aerosol sampling point of view, the human respiratory system may be considered as a sampling train consisting of an inlet, a pre-separator, a transport line with some losses, and a collection filter. For many years the regional particle deposition in the human respiratory system has been studied both experimen- tally and theoretically and the effects of particle size, breathing pattern, age, sex, individual variation, etc., have been measured.3-6 In the last 15 years experiments have also been performed to quantify the fraction of airborne particles that actually enter the respiratory system (the inhalability).7 Although deposition studies have been performed on humans, the inhalability studies have used breathing (in- and ex-haling) mannequins mounted in a wind tunnel.Using the results from the deposition studies it is possible to estimate an average human deposition as a function of particle size (and other human individual parameters) and to add on a safety margin in order to include a large fraction of the healthy working population. The inhalability studies on the other hand have mainly been assessed for a particular set of physiological conditions that correspond to light-to-medium work, however, the inhalability was found to have negligible dependence on small variations in breathing patterns. From a knowledge of the distribution of particle deposition in human breathing systems it is possible to formulate standards that represent a given fraction of the human population.Such a standard describes the fraction of particles in the ambient air that will reach a certain region in the respiratory system. The standards are called sampling conven- tions and the instruments used to sample any fraction ought to have a sampling efficiency as close to the sampling convention as possible. The European and international standardization organizations, and the American Conference of Governmen- tal Industrial Hygienists (ACGIH) have all agreed on identical sampling conventions for the three regions of the respiratory system ,8-10 based on Soderholm's review of deposition data." However, the sampling conventions are not defined as one convention per region. Instead they are defined cumulatively: the respirable convention for the alveolar region; the thoracic convention for the tracheo-bronchial and the alveolar regions; and the inhalable fraction for all three regions.The three sampling conventions are shown in Fig. 1 as functions of particle aerodynamic size. The main difference between the inhalable convention and the two other conventions, is that due to experimental difficulties the inhalable sampling con- vention is only known up to 100 pm (where it has not come down to zero sampling efficiency), whereas the respirable and thoracic conventions have a virtually zero sampling efficiency above 10 and 30 pm, respectively. The new respirable sampling convention is a compromise between the older American ACGIH convention,12 and European convention, originally defined by the British Medical Research Council (BMRC) and given international recognition at The Second Pneumoconiosis Conference in Johannesburg in 1959.1328 Analyst, January 1994, VoE.119 The sampling conventions were not formulated with the object of enabling sampling of doses for specific individuals. The statistical nature of the definition implies that the sampling conventions are to be used for measuring the quality of the air that a worker inhales, rather than the dose a specific worker has received. Although there is wide variation in individual deposition doses, the air quality (or contamination as measured according to the sampling convention) is a unique property. Therefore, the wide human dose variation may not be used as an argument for allowing a wide variation among samplers.However, it might be argued, that any sampler that only poorly matches the sampling convention, most probably corresponds to the deposition of at least one worker in the world. If a wide variation among sampler efficiencies were allowed, a deplorable situation can occur where the vested interest of management, workers, or factory inspectors could determine the choice of sampler. If a specific individual’s dose is to be measured, then that individual’s deposition must first be determined under hisher working conditions. Aerosol Sampler Performance To evaluate a sampler’s performance it must be compared with either another sampler or a sampling convention, in order to ensure that different samplers using different inlets and different means of size-selection measure similar aerosol concentrations.Some samplers are in themselves implied standards. This is often the case with those samplers devel- oped in the 1960s and 70s and presently used for sampling total and respirable dust. Total dust, for example, is measured world-wide, but many countries have their own definition of it, i.e., their national version of a total dust sampling head. In US Coal Mines a specific cyclone is required for sampling respirable coal dust, i.e., the 10 mm nylon cyclone.14 In British coal mines, coal dust is measured in the return air with a stationary horizontal elutriator, which in calm air perfectly follows the BMRC sampling convention for respirable dust. 15 The latter sampling convention was itself designed as the theoretical penetration of a horizontal elutriator.Presently, only cyclones used to sample respirable dust (excluding coal mines), and samplers for inhalable dust used in some countries in Europe (e.g., United Kingdom16) can be said to be designed to follow a specific sampling convention. However, because a virtual one-design monopoly exists on which sampler to use in any particular country, the sampler commonly used may be viewed as a standard in itself. The aim of the CEN standards on aerosol sampling is to ensure that different samplers may be used, provided they give similar results. This is only 1.0 I 1 o-2 I 01 +- -...__ 1 1 10 100 Particle aerodynamic diametedpm Fig. 1 CEN-ISO-ACGIH sampling conventions for: A, inhalable; B, thoracic; and C respirable fractions possible if the samplers follow a pre-determined sampling convention, rather than all being standards in themselves.To elevate any of the national samplers presently used in Europe to a European standard was deemed not to be politically possible, and furthermore was contrary to the General Performance Requirements of CEN.17 In order to better follow the new international sampling conventions, national (standard-in-itself) samplers may have to be optimized if they deviate too much from the sampling convention. CEN’s approach stimulates the development of new and hopefully better, samplers, as opposed to the present system. In principle, three different methods exist to establish whether a new sampler follows a sampling convention.The first, and simplest, method is possible only if a reference sampler that perfectly follows the sampling conven- tion exists. In such a case the mass concentrations sampled simultaneously both with the new and the reference sampler may be compared. Such a comparison must be repeated for several different aerosols with different size distributions. A possible requirement is that for each size distribution both the bias and variance of the new sampler is within some specified limits. References 18 and 19 describe tests based on some of these principles. In the second method the sampling efficiency of the new sampler is compared with that of the prescribed sampling convention. Usually the efficiency data for the new sampler are plotted against those of the sampling convention to construct a sampling efficiency curve.A possible requirement here is to establish tolerance bands on how much the sampling efficiency may deviate from the convention, together with some requirements on the particle size corresponding to 50% efficiency and the slope of the new sampler’s efficiency curve .20,21 In the third method the efficiency data for the new sampler are also used to construct a sampling efficiency curve. However, in this case the data are used to compute the sampled mass fraction of the new sampler for a set of log- normal size distributions, and then compare these computed data with a similar set based on an ideal hypothetical sampler that perfectly follows the sampling convention. A possible requirement in this case is that some combination of bias and variance, somehow determined over the whole set of size distributions, should be below a specified value.22 All three methods have their pros and cons.The first method is easy and cheap to use, but requires several different test aerosols. It is only possible when a reference instrument exists, as for example the horizontal elutriator for the BMRC convention in calm air. The second method gives more information, but the relationship between the tolerance bands and measured concentrations is not apparent.15.23 The third method gives results that are more related to the needs of the end users. However, the validity of the calculations of sampled concen- trations have, at least in one case, been questioned.24 In order to be able to test and evaluate a sampler the test configuration has to be determined.The ideal configuration is one which tests a sampler for all possible situations likely to occur in field sampling. It is, however, neither possible to anticipate all possible situations, nor would it be economically feasible. An example of a field sampling situation is shown in Fig. 2. Both the wind velocity and the aerosol concentrations far from the worker are low. The dust source, possibly with simultaneous emission of air jets, is local and is thus related to the worker’s actions. There is, therefore, a preferred direction relative to the source. Hopefully a local exhaust is also present. In order to test a sampler an ideal (simplified) model that emulates only the most significant aspects of field sampling, must therefore be designed.The concept of inhalable dust is based on such a simplified ideal model. The inhalability is determined in a wind tunnel using a rotatingAnalyst, January 1994, Vol. 119 29 life-sized breathing mannequin and homogeneously dis- tributed particles released remotely upstream from the mannequin and thus in a steady-state with respect to the external air flow. Fig. 3 shows a schematic diagram of this set-up. In this ideal model no objects disturb the air flow, and there is no local emission of particles or air jets. For this ideal situation it is possible to design experiments to determine how well a sampler emulates a sampling convention. Workplace reality must therefore, not be confused with the ideal test configuration.In workplaces that differ significantly from the ideal model one cannot assume that the measured inhalable fraction approximates what the worker inhaled. This is exemplified when the exposure is extremely unidirectional, or the particles are released into the air in such a way that they are not in a steady-state relative to the external air flow, when sampled or aspirated by the worker. CEN Sampler Performance Test Procedure Using a sampler that emulates a sampling convention, we need to be sure that the measured concentrations when used for workplace sampling are correct within acceptable tolerances. In order to determine if a sampler is good enough it must be evaluated. Preferably this should be done in a simple, cheap, quick, generally applicable, and easily performed experiment that for a standardized test aerosol gives results that are valid for all aerosols encountered in the work environment.Such a test is not possible to design, and the performance test that is proposed by the CEN working group is concerned mainly with the fact that the results obtained should be valid in most work environments. This paper presents a broad outline of the draft performance test standard.25.26 Readers interested in the full text may contact the working group’s convenor. The object of the working group was that the test procedure should follow the Technical Committee’s General Performance Require- ments.17 However, because of the special nature of aerosol sampling, some deviations were necessary and these will be presented explicitly later in the text. The CEN working group on particulate matter has pro- posed two different test procedures to evaluate the perfor- I IY Fig.2 Model of a workplace with worker close to emissions of particles and air jets. A local exhaust is used. Far from the source, both the wind speed and aerosol concentrations are low Fig. 3 Ideal model of a workplace used for determining inhalability in wind tunnels. All particles are generated far from the worker, where both the aerosol concentration (C,) and wind velocity (w) are constant. No large objects close to worker, and therefore all local air movement is caused by wind divergence around the worker mance of aerosol samplers. The first is a laboratory evaluation of personal or static reference samplers.This test is only intended for samplers with aerodynamic separation of par- ticles and subsequent collection of the sampled particles on a substrate, e.g., a filter. The classification obtained as a result of the test will be a property of the sampler. The second test is a field comparison of any type of sampler, e.g., a light- scattering instrument, with a reference sampler. In this case the classification will be specific to the test situation and will not be valid under other circumstances. Reference Sampler Clussifcation The laboratory evaluation of sampler performance is a comprehensive test that measures the sampling efficiency as a function of particle size and other influencing variables, such as wind speed. The test should preferably be performed on used commercial samples, not prototypes.The test is com- prised of several stages. The first stage is a critical review of the physics of the sampling process, including all the steps in the process by which the instrument aspirates, transports, separates and collects particles. In the laboratory experiment all influencing variables considered to be of importance for the sampling process are investigated. From the measured effi- ciency data, sampling efficiency curves are determined. There is no specific requirement on which type of curve should be used to model the sampling efficiency. The only requirement is that the efficiency curve must have a shape consistent with the physics of the sampling process and not have unrealistic asymptotics outside the range of the data points.Using the efficiency curves the mass collected by the sampler is calculated for a set of log-normally distributed aerosols. These masses are compared to what would have been collected by an ideal hypothetical sampler perfectly following the sampling convention. Both the sampler bias and variance are calcu- lated. Finally the sampler is classified by combining the bias and variance. CEN assumes in its General Performance Requirements17 that the performance test is specific for each substance, that is, the same sample is used for the determination of both sampling and analytical errors, expressed as the total overall uncertainty. This implies that one test should be performed for each substance on the threshold limit value list, for each of its solifliquid occurrences, for each sampling method and for each analytical method.It is not feasible to perform such a long list of performance tests. In addition, monodisperse test particles, which are necessary for a test against a sampling convention, cannot be generated from all kinds of substances. Therefore, the laboratory evaluation of a reference sampler using only one type of aerosol which may not even have a threshold limit value is the first departure from the scheme laid out by the CEN General Performance Requirements. Critical Review of Sampling Process The purpose of the critical review is to identify instrumental or instrumentaknvironmental interaction factors that influence the sampler performance relative to the sampling convention.The critical review may analyse the following steps of the sampling process: sampler preparation (e. g., cleaning); collec- tion and substrate preparation (e.g. , selection and application of impaction media); airflow adjustment; aerosol aspiration; internal aerosol separation; aerosol collection; transport of collected material to analytical laboratory; and chemical analysis. The critical review will determine the design of the test to quantify the effects of the most important factors. The draft standard lists typical factors that may be important (see Table 1). It also specifies some potential problems, as for example particle de-agglomeration and particles flying into30 Analyst, January 1994, Vol. 119 the inlet under their own momentum, rather than being aspirated into the sampler by air velocities caused by a combination of external air movement and sampler airflow.For the factors most likely to have strong influences, the draft standard specifies the ranges over which the influencing variables should be tested. A summary of these variables is presented in Table 2. Laboratory Experiment Based on the critical review an experimental plan is designed. The draft standard provides some efficient statistical designs that allow the effects of particle size, one environmental factor (e.g., wind speed), sampler specimen variability and personal sampler position to be determined simultaneously. Instruments that are to be used as personal samplers, must be tested as such, i.e., mounted on a lifesize mannequin in a wind tunnel.The internal penetration of samplers for the respirable and thoracic fractions may be tested separately in a calm air chamber or by introducing the aerosol directly into the sampler inlet, if the aspiration efficiency of a sampler in a wind tunnel and in calm air are both known. The flow dependence of the sampler must be known, or experimentally determined, for subsequent analysis. The draft standard does not specify exactly how the aerosol experiments are to be carried out. Instead a great deal is left to the discretion of the testing laboratory, but with requirements on the allowed experimental sampling and analytical errors. The laboratory tests may be performed with either mono- or poly-disperse aerosols. The choice of test aerosol depends on the availability of suitable methods for measuring particle aerodynamic size with unique and monotonic calibration functions.Correction factors for particle density and shape Table 1 Factors influencing sampler performance Environmental Particle size Wind speed and direction Aerosol composition Aerosol concentration Aerosol charge Temperature, pressure, humidity Vibration, orientation Instrumental Specimen variability How variations Surface treatments Sampler position Sampled aerosol mass Sample transportation Electromagnetic susceptibility may be used. The test aerosol concentration and size distribution should be spatially homogeneous. The allowed analytical errors in determining the sampling efficiency must be less than 1-2%. The ambient aerosol concentration should be sampled with thin-walled sharp-edged probes, operating iso-kinetically in a wind tunnel.The experimental uncertain- ties should be evaluated. Several aspects of the wind tunnel test are still provisional and will be studied over the next few years. Among these answers are required to the following questions: Should the mannequin be breathing? May the mannequin be replaced by a simple wood board? What is the effect of turbulence? Data Treatment How the experimental sampling efficiency data is to be treated is extensively described in the draft standard. Only a broad outline will be presented here. From the measured sampling efficiency data, an average sampling efficiency curve, Eff, (dae), is drawn as a function of particle aerodynamic diameter, d,,, for each environmental factor tested.This curve is then used to calculate the sampled mass fraction for a set of log-normal aerosol size distributions. For all three sampling conventions the set is based on the following size distribution parameters: aerodynamic mass median diameter (MMAD) in the range 1-25 pm in steps of 1 pm, geometric standard deviation (og) in the range 1.75-3.50 in steps of 0.25. Size distributions for which the fraction is less than 5% of the total ambient are excluded (this occurs only with the respir- able fraction for narrow size distributions with large MMADs). The average sampled mass fraction, F,, is calcu- lated by 00 Fe (MMAD, og) = J Effe(dae) A(dae I MMAD, o g ) ddae (1) 0 where A(d,, I MMAD, og) is a log-normal aerosol size dis- tribution with parameters MMAD and og.The mass fraction that would have been sampled by an ideal hypothetical sampler perfectly following the sampling conven- tion is calculated using the same equation, but exchanging Eff,(d,,) for the sampling convention curve. From these two sets of mass fraction data, the bias of the sampler is calculated as the percentage difference between the mass fractions sampled by the sampler and the sampling convention. It is a function of log-normal aerosol size distributions (with parameters MMAD and og) and environmental factors. For Table 2 Range and number of influencing variables to be tested Variable Particle size Wind speed Wind direction Range Inhalable 1-100 pm Thoracic 0.1-35 pm Respirable 0.1-15 pm Indoor workplaces 0-1.5 m s- Outdoor workplaces 0-4 m s- Omnidirectional average Aerosol composition Sampled mass Aerosol charge Sampler specimen variability Flow rate variations Particle collection surfaces Position of use (personal samplers) Phase: solid andor liquid; particles of known shape Up to mass corresponding to: maximum concentration x design flow rate X sampling time Charged or neutralized aerosol; conducting or insulating sampler Test group to be as large as possible and always at least three specimens Design flow rate f5% Choice of collection materials (e.g., filters and foams) and details of any surface treatments to be stated Within area stated in the user instructions Number of values 9-12 3 3 Continuous revolution or 24 Choose suitable materials values stepwise Choose and document 3 1-531 Analyst, January 1994, Vol.I1 9 each tested influencing variable, a bias map is drawn showing the sampler bias relative to the sampling convention as a function of aerosol size distribution parameters. An example is shown in Fig. 4. In the diagram the area marked with horizontal lines represents a bias between f O and + 10%. The thick solid lines represent the region in which the performance will be classified. If the average bias is similar for aerosol size distributions of interest and all tested relevant influencing variables, the bias may be reduced by multiplying all concentrations with a correction factor. A correction factor that significantly decreases the bias is allowed and shall then always be applied when the sampler is used. Using all the experimental efficiency data the uncertainty in the sampler bias and the variance due to differences among sampler specimens is calculated. The total sampler variance has two components, the specimen variability and the flow setting errors.The draft standard presents two performance indexes, the overall uncertainty as defined by the CEN General Perfor- mance Requirements,l7 and the sampler accuracy (defined in the draft standard). The overall uncertainty is calculated as the sum of the absolute bias and twice the relative standard deviation, incorporating any error from a subsequent chem- ical analysis (which therefore has to be added to the total sampler variance). The sampler accuracy is defined for the sampler itself, excluding any subsequent chemical or gravi- metric analysis.It is the maximum error relative to the true value that 90% of all samples taken with the sampler will have, with 90% confidence (incorporating the uncertainties in measured bias and total sampler variance). The CEN working group prefers the accuracy criteria for sampler classification because it classifies the sampler itself, even though it is at variance with the CEN General Performance Requirements. Similar to bias maps, accuracy maps should be drawn for each tested relevant influencing variable, presenting the accuracy as a function of the aerosol size distribution parameters. See Fig. 5 for an example of an accuracy map obtained for one value of a tested influencing variable. The thick solid lines represent the region in which the performance is classified.ClassiJicaton The draft standard suggests that a sampler is deemed to be satisfactory for any environment where its accuracy for the calculated sampled mass fraction is within or equal to +30%. This figure of 30% is regulated in the CEN General Performance Requirements for samples taken in the concen- tration range from one half to twice the limit value.17 It is believed that there would not be many samplers having an accuracy better than 30% for all size distributions of interest and all tested environmental factors. Instead of rejecting those samplers with an accuracy >f30% which is obtained in some of these cases, a classification scheme is designed so that a sampler is assigned to a higher class the wider the range of conditions for which it has an accuracy S230%.This is the third major deviation from the CEN General Performance Requirements. The accuracy is calculated for the set of size distributions defined above, and in the classification scheme the individual size distributions used are termed classification points. Three classes were designed, labelled 1,2 and 3. The draft classification scheme has been arbitrarily decided and will be reviewed in light of the experience gained from the pan- European test of inhalable samplers. Samplers failing to meet any of the class requirements and untested samplers will be termed unclassified. These shall not be used for health-related aerosol sampling in workplace air. Class I sampler The average sampling efficiency for this sampler at each tested particle size should be in the range 8&125% of the ideal efficiency given by the sampling convention, for particle aerodynamic diameters larger than 1 pm and where the value of the sampling convention is greater than 10% it meets the required accuracy for 100% of the classification points, for all tested values of relevant influencing variables.and Class 2 sampler This sampler meets the required accuracy for 270% of the classification points, for all tested values of relevant influenc- ing variables it meets the required accuracy for 100% of the classification points, for only two out of three tested values of relevant influencing variables. or Class 3 sampler This meets the required accuracy for 230% of the classifica- tion points, for all tested values of relevant influencing variables it meets the required accuracy for 100% of the classification or Fig.4 Example of bias map, showing contours of equal bias (%) in Fig. 5 Example of accuracy map, showing contours of equal the thoracic fraction of a tested sampler, relative to the sampling accuracy (%) in the thoracic fraction of a tested sampler, relative to convention for one tested influencing variable, for log-normal aerosol the sampling convention for one tested influencing variable, for size distributions, as a function of MMAD and up log-normal aerosol size distributions, as a function of MMAD and ug32 Analyst, January 1994, Vol. I19 points, for only one out of three tested values of relevant influencing variables. In the case of a thoracic sampler as shown in Fig.5, the accuracy is less than 30% in all classification points for this tested relevant influencing variable. Field Sampler Comparison The draft standard presents a method for field comparisons between any type of aerosol sampler (e.g., a light-scattering instrument) and a reference sampler having an accuracy better than or equal to f30% under the circumstances of the field test. The purpose of the field test is to obtain a correction function relating the tested sampler concentrations to those of the reference sampler. If an acceptable correction function is obtained, the tested sampler is placed in class 4, otherwise it remains unclassified. A class 4 award following a field comparison is only valid under the exact conditions under which the test was carried out.That is, it depends on the properties of the aerosol sampled, and on the environmental conditions of the test. It cannot be assumed that the correction function may apply to other circumstances. At the test a personal sampler should be compared with a personal reference sampler, and a static sampler compared with a static reference sampler. The number of measurement pairs should be as large as possible and never less than 10. The measurements should cover the range of aerosol properties, including concentration and environmental conditions occurring at the sampling sites. To check the stability of the correction function two separate comparison exercises should be carried out, each consisting of at least 10 measurement pairs. The correction function between the tested sampler and the reference sampler concentrations may be calculated by any statistical procedure. The data and the curve should be plotted onto a graph, and the residuals of the curve should be examined.The tested sampler is classified according to whether it passes some preliminary statistical tests which will later be reviewed. The first requirement is that the average relative residual bias differs insignificantly from zero. The second requirement is that the average relative residual biases from the two comparison exercises must not differ significantly. The third requirement is that no more than 10% of the absolute residual biases exceed half of the average concentration. Sampling Pump Requirements Another CEN working group has drafted a standard on personal sampling pumps, both for aerosol and/or gas-vapour sampling.27 This standard is applicable to any personal sampling pump with a flow rate in the range 5 ml min-* to 5 I min-1.The standard gives both the performance require- ments and the procedure to test the pumps. Table 3 lists the performance requirements. Pumps which have passed the test should be specially labelled. The performance of a personal sampling pump is evaluated in a laboratory test. The test of the influence of back pressure is made at the minimum and maximum nominal flow rates claimed by the manufacturer. The test of the pump's capability to withstand long operating times is made at two flow rates, 2 1 min-1 and maximum nominal flow rate, with a flow resistance of 1.6 kPa (corresponding to an unused 37 mm cellulose acetate filter of 0.8 pm pore size).For the test of the influence of temperature, and orientation, a flow rate of 2 1 min-1 and a flow resistance of 0.5 kPa (corresponding to an unused 37 mm cellulose acetate filter of 8 pm pore size) is used. In the test for mechanical strength, a flow rate of 2 1 min-1 and a flow resistance of 3.2 kPa is used. The pulsation test is carried out at a flow rate of 2 I min-* using a flow resistance of 0.75 kPa. The test does not require flow resistances as large as the pressure drop at approximately 2 1 min-1 over 25 mm cellulose acetate filters with 0.8 pm pore size (=4 kPa) or 25 mm polycarbonate filters with 0.4 pm pore size (=5 kPa). The problem of accurate airflow measurements is not addressed by this standard.European Test of Samplers for Inhalable Dust As mentioned earlier, the aerosol sampler test protocol will be used in a European test (which began in spring 1993) of personal samplers used for sampling inhalabk dust or total dust.28 This sampler test will also be a test of several assumptions behind the test procedure, and hopefully, as a result future tests of samplers for inhalable dust may be carried out according to a more simple procedure. The following samplers are included in the test: The SKC IOM personal sampler for inhalable dust, the British seven- hole sampler, the French Arelco CIP-10 with the internal respirable separator removed, the German Strohlein GSP Table 3 Performance requirements for personal sampling pumps for aerosol sampling Features Mass Mechanical safety Electromagnetic compatibility Operating time Flow-rate stability due to back pressure Flow-rate stability over operating time Emergency indicator/ shut-down Temperature Mechanical strength dependence Flow-rate pulsation Orientation Explosion resistance Clock accuracy Instructions for use Labelling Charger Holder to attach pump to wearer Malfunction indicator or automatic Fuse or electrical current restrictor in case Inadvertent flow adjustment made difficult Including batteries and integral holders Avoid sharp corners or uncomfortable parts Meet requirements in EN 50 081-1 and At 2 1 min-1 and maximum nominal Flow rate within +5% if pressure varies emergency stop of short circuit S1.2 kg EN 50 082-1 capacity, at least 2 hand preferably 8 h 10 times, e.g., for a pump with nominal capacity of 2 1 min-1 the pressure range is 0.34.0 kPa Within k5%0 at 20 and 5 "C If flow is reduced by 50% for 1 min, indicator must activate and remain until reset, or pump must stop and not start by itself In range 540 "C flow rate may deviate G +5% from value set at 20 "C No general pump function impaired by shock treatment.The flow rate may deviate <+5% S10% For pump tilted go", flow rate must deviate If claimed by manufacturer, according to For pumps with internal clock, after 8 h The minimum requirement for the <?5%0 EN 50 014 use, deviation <1 min instruction manual is listed in the draft standard manufacturer, type designation, serial number, the European standard, and whether it may be used in explosive areas battery type, and meets requirements of The pump shall be labelled with Designated by manufacturer, tuned to EN 60 335-133 Analyst, January 1994, Vol.I1 9 sampler for inhalable dust, the Italian Lavoro e Ambiente PERSPEC, the Dutch PAS-6 sampler, and the 37 mm cassette in open and closed face versions. The laboratory experiments use a breathing mannequin in a wind tunnel. The test aerosol are nine monodisperse particle sizes, made of fused aluminium oxide and ranging from 5 to 90 p,m. The samplers are tested in three positions on the front of the mannequin. The wind speeds tested are 0.3, 1.5 and 4 m s-1. The draft standards described are the result of the collective work of the CEN working groups. 1 2 3 4 5 6 7 8 9 10 11 12 13 References Phalen, R.F., in Particle Size-Selective Sampling in the Workplace, Report of the ACGIH Technical Committee on Air Sampling Procedures, ACGIH, Cincinnati, USA, 1985. Nordiska Expertgruppen for Gransvardesdokumentation 77, ‘Troestgv’ (Nordic Expert Group Basis for an Occupational Health Standard 77, Wood Dust), Arbete och Halsa, 1987: 36. Hatch, T., and Gross, P., Pulmonary Deposition and Retention of Inhaled Particles, Academic Press, New York, 1964. Chan, T. L., and Lippmann, M., Am. Ind. Hyg. Assoc. J., 1980, 41, 399. Heyder, J., Gebhard, J., Rudolph, G., Schiller, C. F., and Stahlhofen, W., J. Aerosol Sci., 1986, 17, 811. Miller, F. J., Martonen, T. B., Menache, M. G., Graham, R. C., Spektor, D. M., and Lippmann, M., in Inhaled Particles V I , eds.Dodgson, J . , McCall, R. I., Bailey, M. R., and Fisher, D. R., Pergamon Press, Oxford, 1988, pp. 3-10. Vincent, J. H., Mark, D., Miller, B. G., Armbruster, L., and Ogden, T. L., J. Aerosol Sci., 1990, 21 (4), 577. ComitC EuropCen de Normalisation, EN481 Workplace Atmos- pheres. Size Fraction Definitions for Measurement of Airborne Particles, Brussels, Belgium. International Organisation for Standardisation, Technical Com- mittee 146, DIS 7708 Air Quality-Particle Size Definitions for Health-related Sampling, Geneva, Switzerland. American Conference of Governmental Industrial Hygienists, 1992-1993 Threshold Limit Values for Chemical Substances and Physical Agents and Biological Exposure Indices, ACGIH, Cincinnati, OH, 1992. Soderholm, S. C., Ann. Occup.Hyg., 1989, 33 (3), 301. Committee on Threshold Limit Values, Threshold Limit Values of Airborne Contaminants for 1968, American Conference of Governmental Industrial Hygienists, Cincinnati, OH, 1968. Orenstein, A. J . , Proc. Pneumoconiosis Conference, Johannes- burg 1959, J & A Churchill, London, 1960. 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 30 Code of US Federal Regulation, part 74, 1982. LidCn, G., and Kenny, L. C., Ann. Occup. Hyg., 1991,35 (5), 485. Health and Safety Executive, MDHSl4 General Methods for Gravimetric Determination of Respirable and Total Inhalable Dust, London. ComitC EuropCen de Normalisation, prEN482 General Require- ments for the Performance of Procedures for Workplace Measurements, Brussels, Belgium. Tomb, T. F., Treaftis, H. N., Mundell, R. L., and Parobeck, P. S., Comparison of Respirable Dust Concentrations Measured with MRE and Modified Personal Gravimetric Sampling Equip- ment, Report RI7772, US Bureau of Mines, Pittsburg, USA, 1973. Caplan, K. J., Doemeny, L. J., and Sorenson, S. D., Am. Ind. Hyg. Assoc. J., 1977,38 (4), 162. Luftbeschaffenheit am Arbeitplatz- Einatembarer und alveolen- gangiger Staub, (Workplace Air Quality-Inhalable and Respir- able Dust), NLuft/AA 11 AK No. 8-88, Deutsches Institute fur Normung e.v., Berlin, 1988. McCawley , M. A., in Particle Size-Selective Sampling in the Workplace, Report of the ACGIH Technical Committee on Air Sampling Procedures, ACGIH, Cincinnati, OH, 1985. Bowman, J. D., Bartley, D. L., Breuer, G. M., Doemeny, L. J., and Murdock, D. J., Accuracy Criteria Recommended for the Certification of Gravimetric Coal Mine Dust Samplers, National Institute of Occupational Safety and Health, Cincinnati, OH, 1984. Bartley, D. L., Doemeny, L. J., Am. Ind. Hyg. Assoc. J . , 1986, 47 (8), 443 and A498. Treaftis, H. N., Gero, A. J., Kacsmar, P. M., and Tomb, T. F., Am. Ind. Hyg. Assoc. J., 1984, 45, 826. CEN/TC137/WG3/N125 Workplace Atrnospheres-Assessment of Performance of Instruments for Measurement of Airborne Particles (draft 5). Available from the CEN working group convenor, Ogden, T., Health and Safety Executive, London. Kenny, L. C., 1. Aerosol Sci., 1992, 23 (7), 773. CENITC137lN90 Workplace Atmospheres-Pumps for Per- sonal Sampling of Chemical Agents. Requirements and Test Methods. Available from the CEN working group convenor, Siekmann, H., Berufsgenossenschaftliches Institut fur Arbeits- sicherheit, Germany. Measurement and Testing Programme of the Council of Ministers of the European Community, Contract CT920047, Pilot Study of CEN Protocols for the Performance Testing of Workplace Aerosol Sampling Instruments. Paper 3103207H Received June 4, 1993 Accepted September 9, 1993