Abstract

The use of direct-reading instruments to measure the airborne fibre number concentration is on the increase. The response of two of these instruments (FM-7400 and Fibrecheck FC-2) was compared with the conventional method of sampling on filters and counting by phase contrast microscopy. Four types of fibres were studied at different concentrations and relative humidity levels. The FM-7400 can be calibrated by the manufacturer for two different levels of sensitivity (standard and high). For the tests where it was set to the sensitivity level with which it had been calibrated, the ratio of the concentration measured by the instrument to the concentration obtained by the conventional method varied in the range 0.5–1 for the different types of fibres studied (chrysotile, glass wool and ceramic fibres). The Fibrecheck FC-2 is a much less versatile instrument. On the basis of a calibration allowing correct detection of asbestos fibres, it greatly overestimated the concentration of man-made mineral fibres. In its normal calibration state a fine chrysotile aerosol was poorly detected. For man-made mineral fibres, the response was highly dependent on the nature of the fibres. These instruments require calibration with the type of fibres to be studied. Unfortunately, this operation is not always accessible to the user and may require the services of a specialized laboratory, as the manufacturer is not always in a position to carry this out.

Received 19 July 2002; in final form 20 January 2003

INTRODUCTION

There are many methods using different principles for the instantaneous measurement of aerosols. In general, these instruments are termed direct-reading or continuous measurement instruments (Görner and Fabriès, 1990). Among them, a number are dedicated to measuring the fibre number concentration in the air. With the strengthening of the legislation on asbestos, the use of these instruments for the rapid evaluation of concentrations appears to be increasing in France.

The automatic measurement of airborne fibre concentration is based on analysing the light scattered by the particles. The intensity of the light scattered is given by Mie’s theory for spherical particles (Mie, 1908). This theory was extended to the case of cylindrical particles by Kerker and Matijevic (1961). The fibres are aligned before reaching the measuring cell. In the case of conducting fibres, alignment is obtained by the action of an electric field (Lilienfeld et al., 1979; Lilienfeld, 1987; Cluff and Patitsas, 1992). Fibres with sufficient magnetic susceptibility can also be aligned by the action of a magnetic field, (Timbrell, 1975). In this case, the fibres are aligned either parallel or perpendicular to the field, although certain fibres are also aligned in intermediate directions. Alignment of fibres in parallel to the flow direction in a gradient flow field was also used by other authors (Prodi et al., 1982; Rood et al., 1992).

It emerges from an analysis of the literature that two instruments in particular have undergone extensive testing. The Vickers M88, based on magnetic alignment of fibres, appears no longer to be on sale. The measurements carried out with this instrument showed a close correlation with manual counting for samples taken in asbestos textile plants and a satisfactory correlation in the brake lining sector (Jones and Gale, 1982). Calibration was necessary for each variety of fibre (Du Toit, 1982). The other instrument is the FAM (fiber aerosol monitor) developed under the joint sponsorship of NIOSH (National Institute for Occupational Safety and Health), the US Bureau of Mines and the US Environmental Protection Agency (Baron, 1993). The current version of this instrument is the FM-7400. The fibres are aligned by the action of an electric field.

Two units are currently on sale in France: the FM-7400 of American origin, the first versions going back some 20 years; and the Fibrecheck FC-2, manufactured in Great Britain and of more recent development. The advantages of direct-reading instruments are well known (Baron, 1994). They allow concentration information to be obtained rapidly to estimate hazard levels, to evaluate control systems and to provide feedback so that exposed persons can modify their behaviour and thus reduce the risks run. They allow a much faster response than conventional methods, which require a sample to be taken and sent to a laboratory for analysis.

The preliminary results concerning these instruments were obtained primarily in the field in two plants producing man-made mineral fibres (ceramic and slagwool fibres) with a number of additional results on chrysotile asbestos obtained in the laboratory (Kauffer et al., 2000). More complete results obtained by varying the nature of the fibres tested, their concentration level and the degree of air humidity are the subject of the present article.

METHODOLOGY

Description of the instruments

FM-7400

As described in the user’s manual, the MIE model FM-7400 fibre monitor is a technologically advanced instrument designed to sample the air continuously, to detect (in real time) the presence of airborne fibres and to automatically count those fibres meeting user-selected fibre size criteria. The FM-7400 detects fibres by aligning the fibres in an oscillating electric field, illuminating the fibres with a polarized helium–neon laser operating at 632.8 nm and detecting the resulting pulses of light with a photomultiplier. The characteristic frequency, phase and shape of the pulses enable discrimination between the fibres of interest and particles. The instrument is calibrated by the manufacturer with asbestos fibres.

The instrument is designed for a flow rate of 2 l/min. Sampling times of 1, 10 and 30 min and 1, 4 and 8 h can be programmed. The fibres taken into account are those with diameters lying between 0.2 and 20 µm. A fibre length threshold facility allows values of 3, 5, 10 and 20 µm to be chosen, the 5 µm threshold being retained for this study. For correct operation, the relative humidity of the air must be between 30 and 90%. The measuring range extends from 104 to 25 fibres/ml. The instrument is transportable.

Fibrecheck FC-2

The Fibrecheck FC-2 is a portable instrument designed to measure the airborne concentration of fibrous aerosol samples. The sampling air flow rate is 2 l/min. A turbulent flow through a nozzle carries the particles, whatever their orientation, to the interior of the detection area. Electronic processing of the light signal emitted by a laser beam and scattered by the particles allows counting of the total number of particles and the number of fibres.

The total number of particles detected, the gravimetric concentration and the fibre number concentration in the range 0–5 fibres/ml are indicated on the control screen of the instrument. The fibres taken into account are those with a diameter of >0.2 µm and a length to diameter ratio exceeding 3:1.

Calibration of the instruments by the manufacturer

FM-7400

During the initial measurements taken with this instrument, the concentrations measured for chrysotile fibres generated in a dust collection chamber were much lower than those measured by the conventional method (Kauffer et al., 2000). This resulted in the manufacturer carrying out a new calibration of the instrument, the sensitivity being set to the high position (high sensitivity).

Two sensitivities are available on the unit: high sensitivity and standard sensitivity. It emerged during the study that for no apparent reason the instrument could change from the high sensitivity level at which it had been calibrated to the standard level. For this reason, certain experiments carried out during this study were with the unit set to high sensitivity mode and others with the unit set to standard sensitivity mode.

The existence of these two sensitivity levels is not mentioned in the user’s manual supplied with the instrument. The procedure to be followed to change from one mode to the other was forwarded to us by the seller (Instructions for special two-level sensitivity selection). Only the presence of the letter H opposite the Fiber Length line on the screen of the instrument or on the paper print-out indicates that the unit is in high sensitivity mode. No indication appears in standard sensitivity mode.

Fibrecheck FC-2

As supplied by the manufacturer, the instrument is calibrated with caffeine fibres generated by heat (initial state). As the preliminary results in this state (Kauffer et al., 2000) showed that the instrument was not sufficiently sensitive to detect chrysotile fibres generated in a dust generation chamber, the manufacturer carried out a modification of the initial calibration at INRS (modified state). Subsequently, it appeared that this calibration was not satisfactory to evaluate the concentration of man-made mineral fibres, which led us to request the manufacturer to return the unit to its initial state. Some of the experiments were therefore carried out with the instrument set for normal use (initial state) and others with a unit whose calibration had been modified (modified state).

History of the units

A historical overview of the instruments for the period April 1998–April 2000, including breakdowns as well as the various calibrations and modifications to calibrations, is given in Table 1.

Samples

The samples were taken in a dust generation chamber (Rihn et al., 1996). The aerosol was sampled simultaneously by three instruments (FM-7400, Fibrecheck FC-2 and a prototype instrument). The results of the prototype instrument will not be presented in this paper as it is still in the process of being improved and is as yet unavailable on the market. In parallel, three filters (25 mm diameter gridded Millipore®, pore diameter 0.8 µm, reference AAW 60250C) mounted on a Gelman® head made of Delrin were sampled for analysis by means of phase contrast optical microscopy. Two additional sampling points were used with a view to measuring the diameter and length distribution of the fibres present in the atmosphere. In this case, the sampling filter was a Nuclepore® polycarbonate filter with a pore diameter of 0.4 µm, also mounted on a Gelman® head. In the case of the filter samples, only one pump (Rietschle®, type YCA 15) was used. The flow rate (1 l/min) at each sampling head was ensured by critical orifices previously calibrated in the laboratory using membrane filters of the same nature.

Analyses

The fibres on the Millipore® filters were counted by phase contrast optical microscopy in accordance with Standard X43-269 (AFNOR, 1991). To increase the accuracy of counting, the stopping rules criteria were modified (100 fibres or 200 fields instead of 100 fibres or 100 fields). The fibres taken into account were those >5 µm long, <3 µm in diameter and with a length to diameter ratio exceeding 3. Each filter was counted by two operators.

The length and diameter distribution of the fibres was determined on Nuclepore® filters by scanning electronic microscopy in accordance with a European protocol (WHO, 1985), with a magnification of 5000× for chrysotile and 2000× for man-made mineral fibres.

Execution of the tests

The most important specifications given by the manufacturers for correct use of the equipment concern the relative humidity of the air and the fibre number concentration (see Description of the instruments, above). In the case of the FM-7400 for example, the relative humidity of the air must be within the range 30–90%. This is not difficult to understand, as the relative humidity of the air can modify the conductivity of the surface of the fibres and prevent correct alignment of fibres in the electric field of the device.

This is the reason why the different tests were carried out according to an experimental design in which the relative humidity of the air and the fibre number concentration were varied for the different types of fibres studied (chrysotile, glass wool, rock wool and ceramic fibres).

Three levels were retained for relative humidity (1 = 20%, 2 = 45%, 3 = 70%). Four levels were retained for fibre number concentration. The concentration range targeted, but not always achieved, extended from 0.1 fibres/ml for level 1 to 4 fibres/ml for level 4. In fact, as the FM-7400 was used to help assess whether the concentration was on target, the concentration range targeted was achieved to a varying degree depending on the setting of the sensitivity mode (see Calibration of the instruments by the manufacturer). As it is known that microscopists tend to overcount when filters are very lightly loaded and undercount when the samples are very highly loaded, the sampling duration was varied according to the concentration range targeted, from 10 min for level 4 to 380 min for level 1.

The experimental design was the same for each fibre and is given in Table 2. All the combinations of relative humidity and concentration level were tested, four combinations being tested twice.

Data processing

For each experiment, the reference concentration was the mean of the concentrations measured by the two operators. The calculation of the concentration and the effects studied employed the Bayesian method using the BUGS software, which also yielded the standard deviation and the confidence interval of the parameters (Spiegelhalter et al., 1996).

In this simulation-based approach, it was assumed that the number of fibres counted by counter c (1–2) for sample p (1–3) of line m (1–16) of the experimental design followed a Poisson distribution with a mean equal to:

[(n × s × V)/S] × Cm,p × exp(αc)

where n is the number of fields observed, s is the surface area of the graticule, S is the useful surface area of the sampling filter, V is the volume of air sampled, Cm,p is the mean concentration measured by the two counters for sample p of line m of the experimental design and exp (αc) represents the ratio of the concentration measured by counter c to the mean geometric concentration measured by counters 1 and 2.

It was also assumed that the three concentrations Cm,p for each line m of the experimental design follow a log-normal distribution with a mean log(Cm) and variance σ2p. The coefficient of variation expressed as a percentage characterizing the homogeneity of the simultaneous samples in the chamber is equal to 100 × [exp(σ2p) – 1]1/2.

The effects linked to the different instruments, different relative air humidity levels and different levels of fibre number concentration were determined assuming that:

log(Ya,m) = log(Cm) + βa + γah + δac + eam

where Ya,m is the concentration measured by instrument a for line m of the experimental design, Cm is the mean concentration in the chamber determined by the conventional method and exp(βa) is the ratio of the concentration measured by instrument a to the mean concentration measured by the two operators. This is termed the instrument effect in Tables 6–10. A ratio equal to r means that the concentration measured by instrument a after correction for humidity and concentration effects is equal to r × the concentration measured by the conventional method. exp(βa + γah) is the ratio of the concentration measured by instrument a to the mean concentration measured by the two operators for humidity level h (instrument effect per humidity level, Tables 610). A ratio equal to r means that the concentration measured by instrument a at humidity level h after correction for the concentration effect is equal to r × the concentration measured by the conventional method. exp(γah1 – γah2) represents the ratio of the concentrations measured by instrument a for two humidity levels h1 and h2 (humidity effect, Tables 610). A ratio equal to r means that the concentration measured by instrument a at humidity level h2 after correction for the concentration effect is equal to r × the concentration measured at humidity level h1 (h1/h2). exp(δac1 – δac2) represents the ratio of the concentrations measured by instrument a for two fibre number concentration levels c1 and c2 (concentration effect, Tables 610). A ratio equal to r means that the concentration measured by instrument at concentration level c1 after correction for the humidity effect is equal to r × the concentration measured at concentration level c1 (c1/c2). eam is the residual error which was assumed to follow a normal centred distribution of variance σ2a + σ2p. The coefficient of variation expressed as a percentage characterizing the reproducibility of each instrument is equal to 100 × [exp(σ2a) – 1]1/2. Reproducibility measures the ability of the instrument to produce identical results under identical experimental conditions.

RESULTS

For the four types of fibres generated (chrysotile, glass wool, rock wool and ceramic fibres), Table 3 indicates the five parameters (geometric mean and geometric standard deviation of the lengths and diameters, coefficient of correlation between the logarithm of the diameter and the logarithm of the length) which characterize the fibre size distribution in the case of a bivariate log-normal distribution (Schneider and Holst, 1983). The length and diameter distributions of the fibres generated are also represented in the form of histograms in Figs 14.

For the different lines of the experimental design and the different fibres studied the results obtained by the three instruments as well as the mean concentration determined by the conventional method (mean of three sampling points and two counters) are presented in Table 4. It sometimes appeared, in the case of low concentrations, that the response of the instrument did not seem to be correlated with the concentrations measured by the conventional method, as if a limit of detection existed. These values, which were not taken into account in the analysis of the results, are presented in bold italic in the tables. Also in bold italic in the tables are the values measured with the Fibrecheck FC-2 for which the mean concentration measured by the conventional method was >4 fibres/ml. These values were also not taken into account in the analysis of the results as the range of use advised by the manufacturer of the Fibrecheck FC-2 extends from 0 to 5 fibres/ml. The data shown in Table 4 is also presented in Fig. 5. This figure shows the concentration measured by the two direct-reading instruments for all the fibres tested in relation to the concentration measured by phase contrast optical microscopy. The results obtained at different values of relative humidity are distinguished. The slope of the straight line is equal to 1.

The data allowing quantification of the counter effect and the homogeneity of the dust generation chamber is shown in Table 5 for the different experiments. The counter effect is the ratio of the concentration measured by counter 1 to the mean geometric concentration measured by counters 1 and 2. The homogeneity of the chamber corresponds to the coefficient of variation of the concentrations measured at the different sampling points of the dust generation cell after correction for the counter effect. The counter effect and its confidence interval are also given in Fig. 6.

Instrument effects, instrument reproducibility, instrument effects per relative humidity level, humidity and concentration effects are given for the different types of fibres in Tables 610. The values in bold italic in these tables correspond to a significant effect (at a 5% risk threshold) for humidity and concentration effects. Instrument effect per humidity level is illustrated in Figs 7 and 8 for the FM-7400 and Fibrecheck FC-2 machines, respectively.

DISCUSSION

To compare the concentrations measured by the direct-reading instruments with those obtained by the conventional method, it was necessary to have a stable aerosol in the dust generation chamber and as reliable an evaluation as possible of the reference concentration. The values indicated in Table 5 show that the coefficient of variation characterizing the dispersion of the concentrations measured at the three sampling points was <20% for all the experiments. The counts carried out on the different sampling filters were done by a laboratory accredited for this type of analysis. All the filters were counted by two counters. As can be seen in Fig. 6 and Table 5, the average deviation for any fibre type between the concentration measured by one counter and the mean concentration never exceeded 16%.

The range of fibre densities corresponding to the mean concentrations given in Table 4, according to the sample durations for the different concentration levels, ranged from 49 to 620 fibres/mm2, which is very close to the recommended density range laid down in the standard (50–650 fibres/mm2). This assumed that the counts of the microscopists increased linearly as filter loading increased.

The fibre size distribution (cf. Figs 1–4) shows that the fibres generated were quite fine. This is the case for chrysotile in particular, as more or less all the fibres have a diameter of <0.5 µm. As regards the glass wool and rock wool fibres, the diameters are more often not less than 1 µm, the diameter of the ceramic fibres being most often <3 µm.

Knowledge of fibre size distribution, particularly in terms of diameter, is important. Indeed, as the direct-reading instruments count fibres whatever their diameter and as the conventional method takes into account only fibres with diameter of <3 µm, it is clear that this particular fibre size distribution creates the optimal conditions for close agreement between the results of the two techniques.

An effect linked to the relative humidity of the air was highlighted for the concentration measured by the FM-7400 (cf. Fig. 7 and Tables 7, 8 and 10) and an effect linked to the concentration in the dust generation chamber (cf. Table 7). The tendency was to underestimate the concentrations at low relative air humidity (20%). Although this effect is not surprising in that the manufacturer advises against using the instrument for man-made mineral fibres when the relative humidity of the air is less than 30%, it should be pointed out that this was observed primarily when the instrument was set to the standard sensitivity level although it had been calibrated at the high sensitivity level. If the results in which the instrument was set to the standard sensitivity level are excluded, the ratio of the concentration measured by the FM-7400 (all humidity levels together) to the conventional method varied from 1 to 0.53 (cf. Tables 6, 9 and 10, instrument effect). The best agreement between the two techniques was obtained for the chrysotile aerosol, which is consistent with the fact that this instrument was calibrated with this type of fibre. The reproducibility of the FM-7400 was better than 20% if the test on rock wool, where it was set to the standard position, is excluded (cf. Tables 6, 7, 9 and 10, instrument reproducibility).

No data were found in the literature concerning the use of the FM-7400 in the man-made mineral fibre industry. As regards the evaluation of asbestos fibres, one recent article has been published (Besson et al., 1999). Unfortunately, it does not provide elements of comparison with the conventional method using phase contrast optical microscopy. It would appear that it is the oldest versions, namely the FAM-1 and FAM-2, that have been studied most. The instrument was tested by comparison with counts carried out with optical microscopy in different situations (Droz, 1982). A reasonable agreement was found between the two methods, particularly for high concentrations. Another study is more critical. The verdict was good for the comparisons carried out in the laboratory, but less so for field studies (Iles and Shenton-Taylor, 1986). Finally, certain authors recommend using the FAM-1 as a screening method to evaluate the asbestos fibre concentration, but exclude its use to replace the conventional membrane filter method with phase contrast optical microscopy for compliance with regulatory standards (Phanprasit et al., 1988).

Regarding the Fibrecheck FC-2, only one significant effect on the concentration measured due to the relative humidity level of the air was highlighted (cf. Table 10), along with only one effect linked to the fibre concentration level in the dust generation chamber (cf. Table 6). If, however, the results presented in Fig. 8 are examined, it can be seen that the Fibrecheck FC-2 has an overall tendency to undercount when the relative humidity increases. This effect was observed for man-made mineral fibres (glass wool, ceramic and rock wool fibres) whether the instrument was in its initial or modified state. The ratio of the concentration measured by the instrument to that measured by the conventional method, all humidity levels taken together, ranges from 0.43 to 16.9 (cf. Tables 610, instrument effect). It would thus appear that in its modified state (i.e. after calibration with a chrysotile aerosol generated at INRS), the instrument is much too sensitive (very high overestimation of the concentrations). If only the results obtained with the instrument in the state where it is normally calibrated are considered, the divergences are smaller. Contrasting results were, however, noted in relation to fibre type (2.3 times higher for ceramic fibres and 2.3 times lower for glass wool fibres). Here again, the best results were obtained with the chrysotile aerosol, with which, it must be said in all fairness, the instrument had been calibrated. Concerning reproducibility, the values for the Fibrecheck FC-2 range between 9.7 and 56.5 (cf. Tables 610, instrument reproducibility).

If the results obtained with the instrument in its initial state are considered (cf. Table 4, experiments 4 and 5), it can be seen that it was unable to estimate the fibre number concentration in three cases where the concentrations measured by the conventional method were >4 fibres/cm3 (cf. Table 4, lines 8, 13 and 15, experiment 4). This is consistent with the range of use specified for the instrument, which covers 0–5 fibres/ml. In contrast, it is more surprising to see the instrument give a result of 1.4 fibres/ml for a concentration measured by the conventional method of 8.2 fibres/ml (Table 4, line 1, experiment 5). In this case, users do not know that they are outside the domain of use of the Fibrecheck FC-2. Examination of the data of Table 4 also shows that there appears to be a limit in certain cases below which the instrument gives results far removed from the real concentration (Table 4, lines 2, 6 and 9, experiment 5).

Other tests have been published showing a positive correlation (r = 0.68) between the measurements of the Fibrecheck FC-2 and the membrane filter method (Independent Environmental Services, 1994). These involved field data, the special feature being that these data concern low level pollution situations. Ninety four per cent of the data was lower than the detection limit of phase contrast optical microscopy (0.01 fibres/ml), only two concentrations out of 255 being >0.1 fibres/ml. This constitutes a considerable test limitation, a fact pointed out by the authors. To correctly evaluate the capacity of the instrument to measure a fibre number concentration, surely it is necessary to have fibres to measure.

CONCLUSION

It emerged from this study that the FM-7400 is the more versatile instrument as regards the nature of the fibres. When set to the sensitivity at which it was calibrated, the ratio of the concentration measured by the instrument to the concentration measured by the conventional method varied from 0.5 to 1 for the different types of fibres studied (chrysotile, glass wool and ceramic fibres). The weak point of the instrument is the fact that it can change sensitivity for no apparent reason. In our experience, each time it is used it must be verified that the instrument is correctly set to the sensitivity level at which it has been calibrated, to avoid any confusion. In addition, the calibration certificate supplied by the manufacturer must indicate the sensitivity used.

The Fibrecheck FC-2 is a much less versatile instrument. On the basis of a calibration allowing correct detection of asbestos fibres, it greatly overestimated the concentration of man-made mineral fibres. In the state where it is normally calibrated, a fine aerosol of chrysotile was not detected. For man-made mineral fibres its response is highly dependent on the nature of the fibres. It appears to lack sensitivity at low concentrations and it is not always obvious to users that they are outside the advised domain of use.

These instruments must be calibrated and verified with the types of fibres to be studied. Unfortunately, this operation is not always accessible to the user and requires the services of a specialized laboratory, the manufacturer not always being in a position to carry this out. As things stand at present, the evaluation of fibre concentration in atmospheres where several varieties of fibres may coexist causes problems, whatever the instrument envisaged.

Acknowledgements—The authors would like to thank Nathalie Carabin, Simone Lima and Marie-Cécile Starck for their participation in this study.

*

Author to whom correspondence should be addressed. Tel: +33-3-85-50-20-23; fax: +33-3-83-50-20-60; e-mail: edmond.kauffer@inrs.fr

Fig. 1. Length and diameter distribution of the chrysotile fibres generated.

Fig. 1. Length and diameter distribution of the chrysotile fibres generated.

Fig. 2. Length and diameter distribution of the glass wool fibres generated.

Fig. 2. Length and diameter distribution of the glass wool fibres generated.

Fig. 3. Length and diameter distribution of the rock wool fibres generated.

Fig. 3. Length and diameter distribution of the rock wool fibres generated.

Fig. 4. Length and diameter distribution of the ceramic fibres generated.

Fig. 4. Length and diameter distribution of the ceramic fibres generated.

Fig. 5. Concentrations measured by the two direct-reading instruments, for all the fibres tested, compared with the concentrations measured by the conventional method. The results obtained at different relative humidities are distinguished. The slope of the straight line is equal to one.

Fig. 5. Concentrations measured by the two direct-reading instruments, for all the fibres tested, compared with the concentrations measured by the conventional method. The results obtained at different relative humidities are distinguished. The slope of the straight line is equal to one.

Fig. 6. Ratio of the concentration measured by counter 1 to the mean geometric concentration measured by counters 1 and 2 for the five experiments. The vertical bar represents the confidence interval.

Fig. 6. Ratio of the concentration measured by counter 1 to the mean geometric concentration measured by counters 1 and 2 for the five experiments. The vertical bar represents the confidence interval.

Fig. 7. Concentration measured by the FM-7400 per humidity level at the reference concentration for the five experiments. The vertical bar represents the confidence interval. The letter H or S indicates that the instrument was set to high or standard sensitivity. The scale on the vertical axis is logarithmic.

Fig. 7. Concentration measured by the FM-7400 per humidity level at the reference concentration for the five experiments. The vertical bar represents the confidence interval. The letter H or S indicates that the instrument was set to high or standard sensitivity. The scale on the vertical axis is logarithmic.

Fig. 8. Concentration measured by the Fibrecheck FC-2 per humidity level at the reference concentration for the five experiments. The vertical bar represents the confidence interval. The letter M or I indicates that the instrument was in its modified or initial state. The scale on the vertical axis is logarithmic.

Fig. 8. Concentration measured by the Fibrecheck FC-2 per humidity level at the reference concentration for the five experiments. The vertical bar represents the confidence interval. The letter M or I indicates that the instrument was in its modified or initial state. The scale on the vertical axis is logarithmic.

Table 1.

Historical overview of the instruments over the period April 1998–April 2000

FM-7400 Fibrecheck FC-2 
April 1998. Calibration by MIE (high sensitivity)  
 December 1998. Calibration of the instrument by the manufacturer with chrysotile at INRS 
July 1999. Cleaning carried out by the seller  
September 1999. Calibration by MIE (high sensitivity)  
 November 1999. Calibration of the instrument at the manufacturer’s premises with caffeine fibres 
December 1999. Flow rate problem; repair carried out by the seller  
February 2000. Short circuit on the laser power supply; repair carried out by the seller  
FM-7400 Fibrecheck FC-2 
April 1998. Calibration by MIE (high sensitivity)  
 December 1998. Calibration of the instrument by the manufacturer with chrysotile at INRS 
July 1999. Cleaning carried out by the seller  
September 1999. Calibration by MIE (high sensitivity)  
 November 1999. Calibration of the instrument at the manufacturer’s premises with caffeine fibres 
December 1999. Flow rate problem; repair carried out by the seller  
February 2000. Short circuit on the laser power supply; repair carried out by the seller  
Table 2.

Structure of experimental design

Line no. Relative air humidity level Fibre no. concentration level 
 1 
 2 
 3 
 4 
 5 
 6 
 7 
 8 
 9 
10 
11 
12 
13 
14 
15 
16 
Line no. Relative air humidity level Fibre no. concentration level 
 1 
 2 
 3 
 4 
 5 
 6 
 7 
 8 
 9 
10 
11 
12 
13 
14 
15 
16 

Three levels of relative air humidity were studied (1 = 20%; 2 = 45%; 3 = 70%). Four fibre number concentrations were retained. The concentration range targeted, but not always achieved, extended from 0.1 fibres/ml for level 1 to 4 fibres/ml for level 4.

Table 3.

Characteristics of the aerosols generated

Type of fibre Mean geometric diameter (µm) Geometric standard deviation Geometric mean length (µm) Geometric standard deviation Coefficient of correlation 
Chrysotile 0.19 1.42 2.38 1.71 0.11 
Glass wool 0.70 1.94 4.66 1.95 0.37 
Rock wool 0.34 1.72 5.23 2.00 0.32 
Ceramic 1.28 1.50 6.17 1.75 0.72 
Type of fibre Mean geometric diameter (µm) Geometric standard deviation Geometric mean length (µm) Geometric standard deviation Coefficient of correlation 
Chrysotile 0.19 1.42 2.38 1.71 0.11 
Glass wool 0.70 1.94 4.66 1.95 0.37 
Rock wool 0.34 1.72 5.23 2.00 0.32 
Ceramic 1.28 1.50 6.17 1.75 0.72 
Table 4.

Concentrations measured by the conventional method and by the different instruments for the different types of fibres studied and the different lines of the experimental design

Line  (1) Chrysotile  (2) Glass wool fibre  (3) Rock wool fibre  (4) Ceramic fibre  (5) Glass wool fibre 
Plan HU MO FM FC  MO FM FC  MO FM FC  MO FM FC  MO FM FC 
70 2.683 2.47 1.49  12.74 5.06   2.820 1.940   4.675 2.5 2.63  8.227 3.13 1.399 
70 0.085 0.0689 0.093  0.122 0.13 2.12  0.092 0.058 0.255  0.117 0.079 0.129  0.127 0.0529 0.007 
45 0.333 0.388 0.266  0.645 0.66   0.387 0.251 1.697  0.808 0.666 1.825  0.525 0.263 0.12 
70 0.494 0.611 0.305  1.475 1.4   0.763 0.476 3.358  1.43 1.17 3.038  0.898 0.542 0.276 
70 0.355 0.304 0.131  0.658 0.72   0.277 0.242 1.896  0.682 0.645 1.969  0.437 0.249 0.134 
45 0.078 0.063 0.001  0.147 0.135 2.56  0.153 0.053 0.778  0.215 0.0997 0.29  0.082 0.0492 0.006 
45 0.349 0.321 0.001  0.707 0.675   0.502 0.242 2.496  0.925 0.539 1.529  0.372 0.254 0.199 
45 2.368 2.72 1.64  7.363 5.35   8.802 2.070   4.235 2.27   2.61 1.91 1.29 
45 0.087 0.0703 0.000  0.138 0.141 2.063  0.227 0.051 1.254  0.098 0.0955 0.51  0.075 0.0503 0.009 
10 20 0.707 0.611 0.316  2.788 1.39   5.452 0.487   1.385 0.979 3.349  1.132 0.508 0.684 
11 20 0.345 0.376 0.281  0.972 0.679   3.207 0.242   0.738 0.432 1.474  0.658 0.25 0.357 
12 45 0.683 0.616 0.537  2.122 1.33   4.018 0.495   1.205 1.05 2.521  0.777 0.497 0.412 
13 20 2.476 2.41 1.43  14.10 5.09   12.368 2.030   4.398 2.04      
14 45 0.576 0.63 0.45  1.682 1.36   2.312 0.484   1.105 0.953 2.507  0.538 0.495 0.311 
15 45 2.368 2.61 1.85  9.413 4.97   20.223 1.950   4.407 2.68   2.927 1.89 1.706 
16 20 0.059 0.0679 0.001  0.222 0.134 3.612  0.620 0.056 3.558  0.108 0.117 0.452  0.095 0.0505 0.053 
Line  (1) Chrysotile  (2) Glass wool fibre  (3) Rock wool fibre  (4) Ceramic fibre  (5) Glass wool fibre 
Plan HU MO FM FC  MO FM FC  MO FM FC  MO FM FC  MO FM FC 
70 2.683 2.47 1.49  12.74 5.06   2.820 1.940   4.675 2.5 2.63  8.227 3.13 1.399 
70 0.085 0.0689 0.093  0.122 0.13 2.12  0.092 0.058 0.255  0.117 0.079 0.129  0.127 0.0529 0.007 
45 0.333 0.388 0.266  0.645 0.66   0.387 0.251 1.697  0.808 0.666 1.825  0.525 0.263 0.12 
70 0.494 0.611 0.305  1.475 1.4   0.763 0.476 3.358  1.43 1.17 3.038  0.898 0.542 0.276 
70 0.355 0.304 0.131  0.658 0.72   0.277 0.242 1.896  0.682 0.645 1.969  0.437 0.249 0.134 
45 0.078 0.063 0.001  0.147 0.135 2.56  0.153 0.053 0.778  0.215 0.0997 0.29  0.082 0.0492 0.006 
45 0.349 0.321 0.001  0.707 0.675   0.502 0.242 2.496  0.925 0.539 1.529  0.372 0.254 0.199 
45 2.368 2.72 1.64  7.363 5.35   8.802 2.070   4.235 2.27   2.61 1.91 1.29 
45 0.087 0.0703 0.000  0.138 0.141 2.063  0.227 0.051 1.254  0.098 0.0955 0.51  0.075 0.0503 0.009 
10 20 0.707 0.611 0.316  2.788 1.39   5.452 0.487   1.385 0.979 3.349  1.132 0.508 0.684 
11 20 0.345 0.376 0.281  0.972 0.679   3.207 0.242   0.738 0.432 1.474  0.658 0.25 0.357 
12 45 0.683 0.616 0.537  2.122 1.33   4.018 0.495   1.205 1.05 2.521  0.777 0.497 0.412 
13 20 2.476 2.41 1.43  14.10 5.09   12.368 2.030   4.398 2.04      
14 45 0.576 0.63 0.45  1.682 1.36   2.312 0.484   1.105 0.953 2.507  0.538 0.495 0.311 
15 45 2.368 2.61 1.85  9.413 4.97   20.223 1.950   4.407 2.68   2.927 1.89 1.706 
16 20 0.059 0.0679 0.001  0.222 0.134 3.612  0.620 0.056 3.558  0.108 0.117 0.452  0.095 0.0505 0.053 

HU, relative air humidity in per cent in the dust generation chamber; MO, concentration measured by the conventional method; FM, concentration measured by the FM-7400; FC, concentration measured by the Fibrecheck FC-2.

FM-7400. Set to high sensitivity for experiments 1, 4 and 5, to standard sensitivity for experiments 2 and 3. Returned to seller for maintenance between lines 12 and 13 of experiment 2.

Fibrecheck FC-2. Modified state for experiments 1, 2 and 3, initial state for experiments 4 and 5.

For experiment 5, four values considered abnormal measured by the conventional method were deleted, as was one line of the experimental protocol.

The values in bold italic were not taken into account in the analysis.

Table 5.

Values of the parameters and their confidence intervals measuring counter effect and homogeneity in the chamber for the five experiments (the numbers of measurements taken are also indicated)

 Counter effect  Homogeneity of the chamber 
 Value Lower limit Upper limit No. of measurements  Value Lower limit Upper limit No. of measurements 
(1) Chrysotile 1.01 0.98 1.04 96  13.5 10.3 17.5 48 
(2) Glass wool fibre 1.16 1.13 1.19 96  6.01 3.80 8.76 48 
(3) Rock wool fibre 1.03 1.00 1.06 96  6.20 3.75 9.03 48 
(4) Ceramic fibres 0.91 0.89 0.94 96  13.2 10.1 17.2 48 
(5) Glass wool fibre 1.08 1.05 1.11 86  18.0 14.0 23.2 43 
 Counter effect  Homogeneity of the chamber 
 Value Lower limit Upper limit No. of measurements  Value Lower limit Upper limit No. of measurements 
(1) Chrysotile 1.01 0.98 1.04 96  13.5 10.3 17.5 48 
(2) Glass wool fibre 1.16 1.13 1.19 96  6.01 3.80 8.76 48 
(3) Rock wool fibre 1.03 1.00 1.06 96  6.20 3.75 9.03 48 
(4) Ceramic fibres 0.91 0.89 0.94 96  13.2 10.1 17.2 48 
(5) Glass wool fibre 1.08 1.05 1.11 86  18.0 14.0 23.2 43 
Table 6.

Experiment 1: chrysotile

 Level FM-7400  Fibrecheck FC-2 
  Value Lower limit Upper limit No. of measurements   Value Lower limit Upper limit No. of measurements 
Instrument effect  1.00 0.90 1.09 16  0.76 0.64 0.89 12 
Instrument reproducibility  7.83 2.30 19.6   16.0 3.13 40.6  
Instrument effect per humidity level 1.04 0.85 1.24  0.76 0.54 1.02 
 0.99 0.87 1.13  0.93 0.72 1.21 
 0.97 0.80 1.16  0.63 0.49 0.80 
Humidity effect 2/1 0.96 0.76 1.20 8/4  1.26 0.87 1.78 5/3 
 3/1 0.94 0.71 1.22 4/4  0.85 0.56 1.24 4/3 
Concentration effect 2/1 0.89 0.69 1.13 4/4  2.30 1.16 4.08 4/4 
 3/1 1.04 0.79 1.33 4/4  1.02 0.69 1.48 3/4 
 4/1 1.05 0.81 1.35 4/4  1.03 0.70 1.49 1/4 
 Level FM-7400  Fibrecheck FC-2 
  Value Lower limit Upper limit No. of measurements   Value Lower limit Upper limit No. of measurements 
Instrument effect  1.00 0.90 1.09 16  0.76 0.64 0.89 12 
Instrument reproducibility  7.83 2.30 19.6   16.0 3.13 40.6  
Instrument effect per humidity level 1.04 0.85 1.24  0.76 0.54 1.02 
 0.99 0.87 1.13  0.93 0.72 1.21 
 0.97 0.80 1.16  0.63 0.49 0.80 
Humidity effect 2/1 0.96 0.76 1.20 8/4  1.26 0.87 1.78 5/3 
 3/1 0.94 0.71 1.22 4/4  0.85 0.56 1.24 4/3 
Concentration effect 2/1 0.89 0.69 1.13 4/4  2.30 1.16 4.08 4/4 
 3/1 1.04 0.79 1.33 4/4  1.02 0.69 1.48 3/4 
 4/1 1.05 0.81 1.35 4/4  1.03 0.70 1.49 1/4 

Values of instrument effects, instrument effects per humidity level, humidity and concentration effects as well as instrument reproducibility (the different levels for humidity and concentration are given in Table 2, the explanation of the different parameters measured are given in Data processing).

For the humidity and concentration effects, the values in bold italic correspond to a significant effect.

Table 7.

Experiment 2: glass wool fibres

 Level FM-7400  Fibrecheck FC-2 
  Value Lower limit Upper limit No. of measurements  Value Lower limit Upper limit No. of measurements 
Instrument effect  0.71 0.64 0.78 16  16.92 14.50 19.47 
Instrument reproducibility  17.0 8.92 29.3   9.67 2.38 31.5  
Instrument effect per humidity level 0.53 0.44 0.65     
 0.82 0.71 0.93     
 0.83 0.67 0.99     
Humidity effect 2/1 1.55 1.21 1.93 8/4     2/1 
 3/1 1.56 1.16 2.04 4/4     1/1 
Concentration effect 2/1 0.97 0.73 1.26 4/4     0/0 
 3/1 0.76 0.58 0.99 4/4     0/0 
 4/1 0.53 0.40 0.69 4/4     0/4 
 Level FM-7400  Fibrecheck FC-2 
  Value Lower limit Upper limit No. of measurements  Value Lower limit Upper limit No. of measurements 
Instrument effect  0.71 0.64 0.78 16  16.92 14.50 19.47 
Instrument reproducibility  17.0 8.92 29.3   9.67 2.38 31.5  
Instrument effect per humidity level 0.53 0.44 0.65     
 0.82 0.71 0.93     
 0.83 0.67 0.99     
Humidity effect 2/1 1.55 1.21 1.93 8/4     2/1 
 3/1 1.56 1.16 2.04 4/4     1/1 
Concentration effect 2/1 0.97 0.73 1.26 4/4     0/0 
 3/1 0.76 0.58 0.99 4/4     0/0 
 4/1 0.53 0.40 0.69 4/4     0/4 

Values of instrument effects, instrument effects per humidity level, humidity and concentration effects as well as instrument reproducibility (the different levels for humidity and concentration are given in Table 2, the explanation of the different parameters measured are given in Data processing).

For the humidity and concentration effects, the values in bold italic correspond to a significant effect.

Table 8.

Experiment 3: rock wool fibres

 Level FM-7400  Fibrecheck FC-2 
  Value Lower limit Upper limit No. of measurements  Value Lower limit Upper limit No. of measurements 
Instrument effect   0.26  0.20  0.34 16   5.10  3.84  6.83 
Instrument reproducibility  55.6 33.6 98.8   35.9 16.7 78.9  
Instrument effect per humidity level  0.10  0.06  0.17   6.17  2.75 11.84 
  0.25  0.17  0.36   5.12  3.52  7.29 
  0.73  0.42  1.22   4.52  2.84  6.72 
Humidity effect 2/1  2.63  1.29  4.77 8/4   0.95  0.37  1.97 4/1 
 3/1  7.53  3.38 15.07 4/4   0.84  0.31  1.76 4/3 
Concentration effect 2/1  0.73  0.31  1.43 4/4     1/0 
 3/1  0.56  0.24  1.12 4/4     3/0 
 4/1  0.64  0.27  1.27 4/4     4/0 
 Level FM-7400  Fibrecheck FC-2 
  Value Lower limit Upper limit No. of measurements  Value Lower limit Upper limit No. of measurements 
Instrument effect   0.26  0.20  0.34 16   5.10  3.84  6.83 
Instrument reproducibility  55.6 33.6 98.8   35.9 16.7 78.9  
Instrument effect per humidity level  0.10  0.06  0.17   6.17  2.75 11.84 
  0.25  0.17  0.36   5.12  3.52  7.29 
  0.73  0.42  1.22   4.52  2.84  6.72 
Humidity effect 2/1  2.63  1.29  4.77 8/4   0.95  0.37  1.97 4/1 
 3/1  7.53  3.38 15.07 4/4   0.84  0.31  1.76 4/3 
Concentration effect 2/1  0.73  0.31  1.43 4/4     1/0 
 3/1  0.56  0.24  1.12 4/4     3/0 
 4/1  0.64  0.27  1.27 4/4     4/0 

Values of instrument effects, instrument effects per humidity level, humidity and concentration effects as well as instrument reproducibility (the different levels for humidity and concentration are given in Table 2, the explanation of the different parameters measured are given in Data processing).

For the humidity and concentration effects, the values in bold italic correspond to a significant effect.

Table 9.

Experiment 4: ceramic fibres

 Level FM-7400  Fibrecheck FC-2 
  Value Lower limit Upper limit No. of measurements  Value Lower limit Upper limit No. of measurements 
Instrument effect   0.71 0.62 0.81 16  2.30  1.59 3.28 12 
Instrument reproducibility  19.4 4.64 38.8   56.5 27.8 117  
Instrument effect per humidity level  0.68 0.53 0.88  4  2.83  1.39 5.23  3 
  0.71 0.59 0.85  8  2.36  1.42 3.72  6 
  0.74 0.56 0.97  4  2.00  0.96 3.61  3 
Humidity effect 2/1  1.06 0.76 1.41  8/4  0.93  0.39 1.93  6/3 
 3/1  1.11 0.74 1.60  4/4  0.79  0.28 1.78  3/3 
Concentration effect 2/1  1.05 0.72 1.48  4/4  1.17  0.48 2.49  4/0 
 3/1  1.14 0.78 1.58  4/4  1.11  0.45 2.31  4/0 
 4/1  0.75 0.52 1.05  4/4      4/0 
 Level FM-7400  Fibrecheck FC-2 
  Value Lower limit Upper limit No. of measurements  Value Lower limit Upper limit No. of measurements 
Instrument effect   0.71 0.62 0.81 16  2.30  1.59 3.28 12 
Instrument reproducibility  19.4 4.64 38.8   56.5 27.8 117  
Instrument effect per humidity level  0.68 0.53 0.88  4  2.83  1.39 5.23  3 
  0.71 0.59 0.85  8  2.36  1.42 3.72  6 
  0.74 0.56 0.97  4  2.00  0.96 3.61  3 
Humidity effect 2/1  1.06 0.76 1.41  8/4  0.93  0.39 1.93  6/3 
 3/1  1.11 0.74 1.60  4/4  0.79  0.28 1.78  3/3 
Concentration effect 2/1  1.05 0.72 1.48  4/4  1.17  0.48 2.49  4/0 
 3/1  1.14 0.78 1.58  4/4  1.11  0.45 2.31  4/0 
 4/1  0.75 0.52 1.05  4/4      4/0 

Values of instrument effects, instrument effects per humidity level, humidity and concentration effects as well as instrument reproducibility (the different levels for humidity and concentration are given in Table 2, the explanation of the different parameters measured are given in Data processing).

Table 10.

Experiment 5: glass wool fibres

 Level FM-7400  Fibrecheck FC-2 
  Value Lower limit Upper limit No. of measurements  Value Lower limit Upper limit No. of measurements 
Instrument effect  0.53 0.46 0.61 15  0.43 0.34 0.52 11 
Instrument reproducibility  8.74 2.36 23.3   18.0 2.98 53.4  
Instrument effect per humidity level 0.46 0.34 0.60  0.64 0.42 0.92 
 0.67 0.57 0.79  0.43 0.31 0.55 
 0.49 0.39 0.61  0.29 0.18 0.43 
Humidity effect 2/1 1.50 1.07 2.06 8/3  0.71 0.41 1.08 6/2 
 3/1 1.10 0.75 1.54 4/3  0.49 0.24 0.78 2/2 
Concentration effect 2/1 1.04 0.74 1.41 3/4  0.88 0.38 1.46 1/4 
 3/1 1.23 0.87 1.65 3/4  1.32 0.84 1.95 1/4 
 4/1 1.02 0.70 1.44 3/4  1.44 0.79 2.46 1/2 
 Level FM-7400  Fibrecheck FC-2 
  Value Lower limit Upper limit No. of measurements  Value Lower limit Upper limit No. of measurements 
Instrument effect  0.53 0.46 0.61 15  0.43 0.34 0.52 11 
Instrument reproducibility  8.74 2.36 23.3   18.0 2.98 53.4  
Instrument effect per humidity level 0.46 0.34 0.60  0.64 0.42 0.92 
 0.67 0.57 0.79  0.43 0.31 0.55 
 0.49 0.39 0.61  0.29 0.18 0.43 
Humidity effect 2/1 1.50 1.07 2.06 8/3  0.71 0.41 1.08 6/2 
 3/1 1.10 0.75 1.54 4/3  0.49 0.24 0.78 2/2 
Concentration effect 2/1 1.04 0.74 1.41 3/4  0.88 0.38 1.46 1/4 
 3/1 1.23 0.87 1.65 3/4  1.32 0.84 1.95 1/4 
 4/1 1.02 0.70 1.44 3/4  1.44 0.79 2.46 1/2 

Values of instrument effects, instrument effects per humidity level, humidity and concentration effects as well as instrument reproducibility (the different levels for humidity and concentration are given in Table 2, the explanation of the different parameters measured are given in Data processing).

For the humidity and concentration effects, the values in bold italic correspond to a significant effect.

References

AFNOR. (
1991
) Détermination de la concentration en nombre de fibres par microscopie optique en contraste de phase. Méthode du filtre à membrane. Standard X43-269. Paris-La-Défense: Association française de normalisation.
Baron PA. (
1993
) Measurement of asbestos and other fibers. In Willeke K, Baron PA, editors. Aerosol measurement—principles techniques and applications. New York: Van Nostrand Reinhold.
Baron PA. (
1994
) Direct-reading intruments for aerosols. A review.
Analyst
 ;
119
:
35–
40.
Besson P, Lalanne FX, Wang Y, Guyot F. (
1999
) Multi-parameter observation of environmental asbestos pollution at the Institut de Physique du Globe de Paris (Jussieu Campus, France).
Ann Occup Hyg
 ;
43
:
527–
41.
Cluff DL, Patitsas AJ. (
1992
) Size characterisation of asbestos fibers by means of electrostatic alignment and light-scattering techniques.
Aerosol Sci Technol
 ;
17
:
186–
98.
Droz PO. (
1982
) Some field experience with a fibrous aerosol monitor. In Proceedings of the fourth international colloquium on dust measuring technique and strategy, Edinburgh, 20–23 September. pp.
199–
208.
Du Toit RJJ. (
1982
) Vickers M.88 Versus Manual Count. In Proceedings of the fourth international colloquium on dust measuring technique and strategy, Edinburgh, 20–23 September. pp.
133–
42.
Görner P, Fabriès JF. (
1990
) Techniques de mesure automatique des aérosols atmosphériques.
Cah Notes Doc
 ;
140
:
595–
626.
Iles PJ, Shenton-Taylor T. (
1986
) Comparison of a fibrous aerosol monitor (FAM) with the membrane filter method for measuring asbestos concentrations.
Ann Occup Hyg
 ;
30
:
77–
87.
Independent Environmental Services. (
1994
) Table of results from the “fibrecheck” field-trials carried out to compare its performance to that of the accepted method for determining airborne asbestos fibre concentration. Cramlington, UK: Independent Environmental Services.
Jones AD, Gale RW. (
1982
) Industrial trials with the Vickers M88 rapid asbestos fibre counter.
Ann Occup Hyg
 ;
25
:
39–
51.
Kauffer E, Vigneron JC, Villa M, Martin P, Grzebyk M. (
2000
) Mesurage de la concentration en nombre de fibres dans l’air. Etude comparée de la méthode du filtre à membrane et des appareils à lecture directe. Résultats préliminaires.
Cah Notes Doc Hyg Sécur Travail
 ;
178
:
55–
63. (Also published in
Gefahrstoffe Reinhaltung Luft
 ,
2001
;
61
, after translation into German.)
Kerker M, Matijevic E. (
1961
) Scattering of electromagnetic waves from concentric infinite cylinders.
J Opt Soc Am
 ;
51
:
506
.
Lilienfeld P. (
1987
) Light scattering from oscillating fibers at normal incidence.
J Aerosol Sci
 ;
18
:
389–
400.
Lilienfeld P, Elterman PB, Baron P. (
1979
) Development of a prototype fibrous aerosol monitor.
Am Ind Hyg Assoc J
 ;
40
:
270–
83.
Mie G. (
1908
) A contribution to the optics of turbid media, especially colloidal metallic suspensions.
Ann Phys
 ;
25
:
377–
445.
Phanprasit W, Rose VE, Oestenstad RK. (
1988
) Comparison of the fibrous aerosol monitor and the optical fiber count technique for asbestos measurement.
Appl Ind Hyg
 ;
3
:
28–
33.
Prodi V, De Zaiacomo T, Hochrainer D, Spurny K. (
1982
) Fibre collection and measurement with the inertial spectro-meter.
J Aerosol Sci
 ;
13
:
49–
58.
Rihn B, Kauffer E, Martin P, Coulais C, Villa M, Bottin MC, Vigneron JC, Edorh A, Martinet N. (
1996
) Short-term crocidolite inhalation studies in mice: validation of an inhalation chamber.
Toxicology
 ;
109
:
147–
56.
Rood AP, Walker EJ, Moore D. (
1992
) Construction of a portable fiber monitor measuring the differential light scattering from aligned fibers.
Aerosol Sci Technol
 ;
17
:
1–
8.
Schneider T, Holst E. (
1983
) Man-made mineral fibre size distribution utilizing unbiased and fibre length biased counting methods and the bivariate log-normal distribution.
J Aerosol Sci
 ;
14
:
139
.
Spiegelhalter DJ, Thomas A, Best NG, Gilks WR. (
1996
) BUGS: Bayesian inference using Gibbs sampling, version 0.5 (version ii). Cambridge: MRC Biostatistics Unit.
Timbrell V. (
1975
) Alignment of respirable asbestos fibres by magnetic fields.
Ann Occup Hyg
 ;
18
:
299–
311.
WHO. (
1985
) Reference methods for measuring airborne man-made mineral fibres, Environmental Health no. 4. Copenhagen: World Health Organization.