Airtesting, Lab testing, 4 Stage Asbestos Clearance

Introduction

Nomenclature, appearance and regulation

A1.1 Asbestos is a term used for the fibrous forms of some naturally occurring silicate minerals that have been exploited commercially for their useful properties of flexibility, high tensile strength, incombustibility, low thermal conductivity and resistance to chemical attack. In Britain, the Control of Asbestos at Work Regulations (CAWR)10 defines and regulates asbestos as the fibrous forms of the following minerals (or any mixture containing them): chrysotile, amosite, crocidolite, fibrous actinolite, fibrous tremolite and fibrous anthophyllite. A population of airborne asbestos fibres, when viewed under a microscope, will often appear to contain many thin parallel-sided fibres and may also contain bundles of parallel fibres, split fibres, curved or wavy fibres and even matted masses. For regulatory purposes, a countable fibre is defined as any object, which is longer than 5 μm, with average width less than 3 μm and having an aspect (length/width) ratio greater than 3:1.

Outline of the method and changes from the previous MDHS

A1.2 The following method is described for the measurement of airborne fibre concentrations, and is due to replace the previously recommended guidance in MDHS39/4.5 Amendments to the EU directive are due to replace the European Reference Method, (Annex 1 of the original directive 83/477/EEC)27 with the World Health Organisation (WHO) method6 for the determination of airborne fibre number concentrations. The WHO method applies to all fibre types and is similar to MDHS59,28 although in this guidance the method is used specifically to evaluate airborne concentrations of predominantly asbestos fibres. If a mixed fibre population is suspected after the original count is completed, discriminatory counting may take place, using one or more of the techniques outlined in MDHS87.29

A1.3 The number of changes from the previous MDHS39/45 to comply with the WHO method has been minimised. However, the incorporation of the MDHS into an HSG guidance document has meant that the sampling and analytical strategy and interpretation for sampling other than for compliance monitoring has been included as chapters in the main HSG. This leaves this appendix focused on compliance sampling, analysis and the reporting of the results. This is close to the scope of the WHO method. The major changes introduced by the WHO method are as follows:

• Flow rates of up to 16 litres.min-1 may be used for short-term 10 minute sampling
• Samples should be rejected if the flow rate has varied more than +10% between flow checks
• The entire filter should be scanned at low magnification to determine whether the loading is uniform and that there is no gross aggregation of fibres or dust
• Fibres attached to >3 μm width particles are also counted
• The average fibre width is used to decide whether it is less than or greater than 3 μm
• Fields lying within 4 mm of the filter edge (or 2 μm of a cutting line) should not be counted
• Discrimination between fibre types is allowed after an original non-discriminatory fibre count has been completed (see MDHS87)
• 29subtraction of the blank count from the fibre count is allowed (but this has not been adopted)

Principle

A1.4 A sample is collected by drawing a measured volume of air through a membrane filter by means of a sampling pump. The filter (or part of the filter) is mounted on a microscope slide and rendered transparent (‘cleared’). Fibres of appropriate dimensions on a measured area of filter are counted visually using phase contrast microscopy (PCM) and the number concentration of fibres in the air calculated.

Scope and limitations

A1.5 The method measures the airborne concentration of countable fibres using PCM. Countable fibres are defined as particles with length > 5 μm, width < 3 μm and aspect ratio (length: width ratio) > 3:1. Fiber having widths less then 0.2  μm may  not be visible using this method, 30 and the PCM count represents only a proportion of the total number of fibers present. Therefore the count is only an index of the numerical concentration of fibers and not an absolute measure of the number of fibers present. As part of the UKAS accrediation requirement, laboratories will need to produce their own documented in-house method for air sampling.

Fibre discrimination

A1.6 The method does not identify the fibre type present but fibres with widths greater than about 1 μm may show optical properties that are inconsistent with asbestos and it is permissible to eliminate these fibres from the count to determine compliance with the asbestos control limit or other limits. Alternatively, other methods (eg analytical electron microscopy) can be used to determine the asbestos fibre concentration. Discrimination against non-asbestos fibres should only take place after the initial total count has been completed. Any discrimination will be dependent on the range of analytical techniques available, the skills of the microscopist and the strategy used. A hierarchy of methods is available to eliminate non-asbestos fibres such as machine-made mineral fibres (MMMF), vegetable, aramid and other fibre types. A more detailed discussion of these techniques appears in MDHS87.29 The report of the discrimination evaluation should include a statement on the type and numbers of interfering fibres which were present and the method by which the number of non-countable fibres have been eliminated from the original PCM count. To carry out fibre discrimination, the laboratory will need to be accredited for the discrimination method, as well as PCM fibre counting.

Limit of detection and limit of quantification

A1.7 Particles sampled onto a filter at best have a random distribution. This means that the precision of the count is limited by the underlying ‘Poisson’ statistics. The precision is usually expressed in terms of the confidence interval, which defines the upper and lower limits expected for a defined percentage of repeat counts. For example 95% confidence limits mean that on average 19 of the 20 values from repeat counts would be within the upper and lower limits. For low counts the lower confidence limit is 0, so a one-sided upper 95% confidence interval is used. For a count of 0 it is 95% probable that the count is <3 fibres.

A1.8 The RICE quality control program has shown that blank filter counts by PCM are low. Of the 40 blank filter samples included in the RICE inter-laboratory counting comparisons,31 39 had reference counts ranging from 0.3 f/mm2 to 2.5 f/mm2. Only 7 of the 2204 results (0.32%), had concentrations outside the acceptable limits (performance band B). These results were obtained from counts of 200 fields and showed that on average <1 fibre per 100 fields was counted. This gives an upper 95% confidence limit that <5 fibres will be counted in 100 fields and similarly <6.5 fibres in 200 fields. This relates to the ‘blank’ count in paragraph 28, so that it can be argued that 5 fibres per 100 graticule areas should be regarded as the lowest reliably detectable count above background (limit of detection). For a sample volume of 240 litres and 100 fields counted this corresponds to a calculated result of about 0.01 f/ml in the air. A sample volume of 480 litres with 200 fields counted corresponds to a calculated result of 0.003 f/ml. A further analysis of the raw data (after the removal of outliers) was also carried out using two methods. The standard method for defining the limit of detection (LOD) and limit of quantification (LOQ) is based on 3 and 10 standard deviations (5.28 and 17.6 fibres, respectively). The second method uses the underlying definitions on which the above are based, where the LOD is the 99th percentile from 0 and the LOQ is determined within a +30% accuracy (7.6 and 25.3 fibres, respectively). The results from the above analyses of the blank data are therefore consistent with the aim to count at least 20 fibres in 200 fields when measuring low concentrations and, for consistency and uniformity of reporting, this should be used to calculate the limit of quantification.

Reagents

A1.9 Acetone and glycerol triacetate (‘triacetin’) are required for filter clearance. Analytical grade reagents are not essential, but they must be clean and free from fibres. Excessive water in the acetone may reduce filter clarity. The triacetin should be free from moisture and with no evidence of hydrolysis (possibly indicated by a smell of acetic acid) or other contamination.

Apparatus

Sampling equipment

A1.10 To comply with the WHO standard method, an open-faced filter holder fitted with an electrically conducting cylindrical cowl and exposing a circular area of filter at least 20 mm in diameter, should be used for sampling. Normally the cowl should extend 1.5-3.0 times the effective filter diameter in front of the filter. Several manufacturers produce injection moulded conductive plastic sampling heads, which are pre-loaded with a suitable filter. Alternatively, metal cowls with a PTFE O-ring can be purchased. A cowled filter holder is intended to protect the filter, while still permitting a uniform deposit. The cowl is pointed downwards during sampling. Flexible tubing is required to connect the filter holder to the pump, and a cap or bung is needed for the cowl entrance to protect the filter from contamination during transport. Different filter diameters and shorter cowls can be used if they are shown to give comparable results, but must be measured to determine the effective filter area.

A1.11 The exposed area of filter must be known and the diameter of the exposed area should be measured to the nearest millimetre (mm) (ie within +5%) for each type of cowl or O-ring in use. A suitable method of measuring this is to use the filter holder and cowl to sample from a cloud of dark coloured dust. The filter is mounted on a slide in the usual way and the diameter measured using the microscope stage vernier by traversing at low magnification across the diameter of the dark area. Alternatively the diameter can be measured with vernier callipers. At least two diameters should be measured at right angles, and a minimum of three filters from similar holders or O-rings should be checked in this way. (Differences between these six measurements of more than one millimetre may indicate either a poorly fitting filter holder or an unsatisfactory clearing technique.) An uneven appearance at one edge of the deposit or signs of dust outside the exposed areas indicates that there was a leak in the sampling head.

A1.12 The membrane filters must be of mixed esters of cellulose or cellulose nitrate, of pore size 0.8 to 1.2 μm (optically clear grade). Preferably the filter should be 25 mm in diameter with a printed grid. Take care to avoid contamination when handling filters. Printed grids are on the sampling side of the filter and will be in the same plane as the particles collected and therefore provide a useful focussing aid. Any distortion of grid lines indicates poor mounting procedure.

A1.13 The pump must be capable of:

• Giving a smooth airflow
• Having flow set to within +10% for flow rates <2 litres.min-1 and within ±5% for flow rates >2 litres.min-1
• Maintaining this flow rate during the period of sampling

These values include any change of flow rate with pump orientation. For personal sampling the pump must be light and portable, and capable of being fitted to a belt or carried in a pocket. The pump’s battery must have sufficient power to operate within the specified flow limits for the duration of the measurement. If pumps for static samples are operated by mains electricity, give due regard to appropriate safety precautions. Static sampling pumps should have the facility to enable the sampling head to be positioned between 1-2 m above ground level.

Flow measurement

A1.14 The airflow must be measured by a working flow meter, sufficiently sensitive to be capable of measuring the appropriate flow rate to within the values specified in paragraph A1.13, and which has been calibrated against a primary standard. The flow meter incorporated in the pump may only be used if it meets the requirements above and it has been calibrated against a primary standard or a master flow meter with a loaded filter in line. Float type flow meters should be vertical when read. Under normal operating conditions, the measurement of temperature and pressure is not necessary, as it will only make a small difference to the total uncertainty, eg if the field temperature changes by +30°C from the calibration temperature, the correction is ~5%, which will contribute to only a small increase (0.2%) in the maximum calculated uncertainty, due to systematic errors (see Table A1.2). Pressure difference of +40 millibars from the calibration pressure will give a correction of ~2%. Only at high pressure and low temperatures, eg +40 millibars and -20°C from the calibration conditions will the readings be underestimating by more than 5%. Normally, rotameter markings are set for standard temperature and pressure (293°K and 101.3 kP (1013 mbar) and any laboratory calibration will be carried out in similar conditions. Only in exceptional circumstances and if a float type flow meter (rotameter) is used will there be a need to use the following equation to adjust the flow rate read from the calibrated rotameter:

Where: Qa = sampling flow rate in litres.min-1
Qc = calibrated flow rate the rotameter value in litres.min-1
Pa = air pressure at sampling site in kPa
Pc = air pressure at site of calibration in kPa
Ta = air temperature at sampling site in °K
Tc = air temperature at site of calibration in °K
(Note: the standard metrological unit for reporting atmospheric pressure is hectopascals (hP) where 1 hP = 0.1 kP = 1 millibar.)

A1.15 The length of the flow meter tube, the range of airflow covered and the spacing and number of markings will directly affect the accuracy of reading and the calibration. To a large extent, the accuracy of the reading of the external flow meter is part of the pump performance assessment in paragraph A1.13, if it is used to check the flow rate. The airflow and hence the float, must be sufficiently stable in the flow meter tube to enable a precise reading against the tube markings. From a practical point of view to set the flow rate to +10% at 0.5 litres.min-1 (the minimum recommended value) a minimum tube distance of 10 mm for each 1 litres.min-1 division is required. This means that the pump flow must be sufficiently stable and adjustable so the float must be able to be positioned and read to within +0.5 mm of the 0.5 litres.min-1 flow mark. Longer distances between the markings and the markings at higher flow rates, will give proportional increase in the accuracy of reading. A float type flow meter tube must be marked with an appropriate number and scale of markings to allow the flow rate to be set, within the limits defined in paragraph A1.13. If a master flow meter is used to calibrate the field flow meter, the laboratory will need to demonstrate that the two flow meters can be read and used to give sufficient accuracy, so that airflows can be set to within the ranges specified in paragraph A1.13. This is usually achieved by having larger spacing between the airflow markings than the minimum values given above.

A1.16 Bubble flow meters measure the volume of air displaced by the pump directly and have advantages in that they do not need correction for changes in air pressure and temperature and their accuracy of flow measurement is much better than a float type flowmeter, if used within the specified range of airflows. It is important to ensure that there are no leaks or significant additional constrictions in the sampling train between the sampling head and the flow meter and that the inlet of the flow meter is to atmosphere. Otherwise, any flow meter will potentially give an erroneous value.

A1.17 The primary standard or master flow meter should preferably be a flow meter whose accuracy is traceable to national standards. These should only be used for in-house calibration of the working flow meters and should be used with careful attention to the conditions of the calibration certificate. The recalibration of the master and working flow meters should be related to the amount and type of use and any evidence that is available to show their stability over time. Procedures are given in the WHO method6 for in-house calibration against a bubble flow meter, although many prefer to send their master flow meters for recalibration to an accredited calibration laboratory.

Equipment for filter clearance

A1.18 Filter clearing should be accomplished by the acetone/triacetin hot block method. A syringe is normally required to dispense the acetone and fine-tipped pipettes, or other suitable droppers, are needed to dispense triacetin.

Microscopy

A1.19 The visibility of fine fibres by PCM is dependent on the transparency of the mounted filter, the quality and cleanliness of the microscope’s optics, its correct use and maintenance, the operator’s eyesight and other factors. Differences between the smallest fibre width observable by phase contrast microscopes will contribute to differences between counters (because fibre width distributions extend below the detection limit). To maintain a uniform level of detection at the limit of visibility, the microscope and accessories should comply with the following specifications:

• A binocular stand with Köhler, or Köhler type illumination including a field iris. (The condenser (sub-stage assembly), objectives and eyepieces specified below must all be compatible with each other and with this stand.)
• A sub-stage assembly, incorporating an Abbe or an achromatic phase contrast, condenser in a centrable focusing mount, with phase annulus centring independent of the condenser centring mechanism
• A built-in mechanical stage with slide clamps and x-y displacement
• Alow powered objective (eg X 10 or X 4 magnification), which is used for carrying out checks on the evenness of the dust deposit on the filter and locating the stage micrometer and test slide ‘tramlines’
• A positive phase contrast objective (preferably par focal with the low-powered objective) of magnification X 40; the numerical aperture (NA) of this objective (which determines resolving power) must lie between 0.65 and 0.70; the phase ring absorption must lie between 65% and 85%
• An optically matched pair of binocular eyepieces, preferably of the wide field, high eye-point type, providing a total magnification of at least X 500 (one of the eyepieces must be of the focusing type and must permit insertion of a graticule);(Note: Some microscope stands may include a tube extension, which increases the total magnification. The total magnification is calculated by multiplying the objective, tube extension and eyepiece magnifications together. This total should not exceed 1000 times the NA);
• a Walton-Beckett eyepiece graticule,32 type G22, with an apparent diameter in the object plane of 100 ±2 μm (when checked against a calibrated stage micrometer) must be used to define the counting area

• various accessories including:
  • A phase telescope or Bertrand lens to ensure correct alignment of the phase ––rings
  • A green filter (optional) which assists viewing (as the optics are optimised for ––green light)
  • A calibrated stage micrometer of 2 μm divisions (eg type S12)
  • An HSE mark II or mark III test slide

A1.20 The coverslip and slide will also affect the visibility of fine fibres. Both must be of glass and of appropriate thickness. Microscope slides must be of conventional type: eg approximately 76 mm x 25 mm and preferably 0.8 mm to 1.0 mm thick. The coverslip thickness is specified/marked on the objective (eg 0.17) and the appropriate thickness must be used (usually sold as 0.16 to 0.19 mm thick, eg No. 11⁄2) and should be about 25 mm diameter or about 25 mm2. The microscope slides and coverslips should be clean and conform to relevant standards.33

Sampling

Preparation of filter holders

A1.21 If a filter holder and cowl is being reused it must be cleaned. Filters should be loaded, unloaded and analysed in an area as free from fibre contamination as practicable. Care must be taken to handle the filter at all times with clean flat-tipped tweezers and only gripping the filter at the edge, outside the exposed area. The printed grid on the filter should be placed in the filter holder, so that it faces towards the cowl. The entrance to the cowl should be closed with a protective cap or bung when sampling is not in progress. Push-fit cowls, particularly if they are reloaded, are prone to poor sealing and should always be checked for tightness. To improve the tightness of the seal, push the cowl entry down very firmly onto a hard surface with a slight rocking action (with the protective entry cap removed). Additional shrink seal bands to reduce the likelihood of leakage from push-fit cowls can be applied to the outside of the seal after loading. Screw-tightening cowls should always be checked for tightness before use: overtightening will damage the filter and cause leakage through the filter, insufficient tightening would allow leakage around the edge of the filter.

Sampling period, flow rate and volume

A1.22 The sampling procedures and strategy should be designed where possible to give sample densities within the range for optimum accuracy (100-650 f/mm2) or to ensure that the minimum limit of quantification is based on at least 20 fibres. The recommended flow rates and sampling times for various sampling strategies are given in the main guidance.

Pump preparation

A1.23 Pumps with poor flow control may change their flow rate during the initial warm-up period. To stabilise the flow rate, some pumps may need to be run for 10-15 minutes before resetting the flow rate, unless there is evidence (eg sampling data and/or manufacturer’s instructions) to show that this is unnecessary with the type of pump in use. A separate filter and filter-holder should be dedicated to this, and may be used for several pumps before being discarded. Pumps should be capable of maintaining flow for the intended sampling period (eg up to 4 hours). Particular care should be taken with short period samples because flow instability at the start may have a significant effect on the apparent volume collected.

Sampling

A1.24 For personal sampling, the filter holder should point downwards and be fixed to the worker’s clothing (eg upper lapel, hood or shoulder), as close to the mouth and nose as practicable, and preferably within 200 mm. Due regard must be given to localised concentrations: in such cases, the sampling head should be positioned on the side expected to give the higher result. If a respirator is worn, the sampling head should be positioned away from the clean exhaust air. Static samples are taken using a downward pointing filter holder positioned some 1-2 m above floor level and away from any walls or large obstructions. Each filter holder should be uniquely identified and the person or position it is used to sample recorded, along with the date and other relevant site information (eg the type of activity taking place and any environmental factors that may affect the results).

Taking the sample
Time and flow-rate recording

A1.25 At the start of the sampling period, the protective cap must be removed from the filter holder, the pump started and the time noted. The flow rate should be measured and recorded at the start and the end of the sampling period. At the end of the sampling period the time should be recorded, the pump stopped and the protective cap replaced on the filter holder. The sampling period must be measured to within ±2.5%. The average flow should be calculated. Periodic checking and adjustment of the flow rate can be made and this should be recorded and used in the final calculations of the average flow. The flow variation between the start and the end of the sampling period should be maintained within ±10% or the sample rejected.

Filter handling and transportation

A1.26 The preferred procedure is for the filter to be transported in the capped filter holder, but if for some reason this is not possible, the filter may be removed in a clean area and carefully placed (with exposed face upwards) in a clean tin or similar container with a close-fitting lid. The filter should be handled with tweezers, which are used to grip the unexposed edges. Sprays (eg cytology fixative) must not be used to ‘fix’ the dust to the filter. If a tin or container is used for transport, unless it can be guaranteed that it will be carefully handled and remain upright, adhesive tape should be used to secure the clean unexposed edge of the filter to the container. The filter can be cut free for mounting and analysis using a surgical scalpel with a rolling action. Care must be taken not to contaminate the filter at any stage or to dislodge any deposit. The filter holder and cowl or the container must be thoroughly cleaned and dried before reuse.

Blanks

A1.27 There are three types of blanks.

• Sampling media blanks are generated when filters are extracted from a box of unused filters. They are mounted and counted before sampling to check that the batch of filters is satisfactory; the initial procedure is to select at least 4 blank filters from each manufacturers’ batch (or a minimum of 1% from larger batches) before the filters are used. Individual blank filter counts should not normally exceed 3 fibres per mm2 (21⁄2 fibres per 100 fields) and if laboratory records show that the proportion is consistently higher, the causes (including the source of supply) should be investigated.
• Field blanks are generated when filters are taken from satisfactory batches to the sampling area and subjected to the same treatment as filters used for sampling, (the cap is removed and replaced after a few seconds). The filters in capped, cowled heads are taken to the sampling area without having air drawn through them and without them being attached to the worker (the cap is removed and replaced after a few seconds). A field blank should normally be nominated for each job and sent with the field samples to the laboratory.
• Laboratory blanks are generated when filters, extracted from satisfactory filter batches, are mounted and counted to check for laboratory contamination, if a field blank has indicated a need for investigation. A laboratory blank may be evaluated with each batch of routine samples, or afterwards, if contamination due to laboratory sources is suspected.

A1.28 The type and number of blanks that are available for analysis and are analysed, will depend on a number of factors. Sampling media blanks are analysed before sampling to check they are suitable for use. The sampling agency is responsible for initiating field blanks and the laboratory will normally know from the sampling information supplied that they are included. The laboratory should always ensure that at least one field or one laboratory blank is prepared for each batch or group of samples, so that it can, if necessary, check whether the source of any fibre contamination was due to the filter preparation. Normally, if low counts are obtained from some of the field samples it will not be necessary to analyse the field or laboratory blanks. If elevated counts are obtained on all the field samples, at least one field blank (or if not available, one laboratory blank) must be counted for each batch to exclude the possibility of contamination. Laboratories should investigate the source of any blank contamination and monitor the batch-to-batch consistency of membrane filters. Due to the low numbers of background fibres present, blank counts must not be subtracted from sample counts.

(Note: the WHO method calls for subtraction of the blank count, but this guidance has not adopted this procedure except as an additional stage, if contamination on the blanks has been found. Evidence shows that the blank count should normally be low and will make little difference to compliance measurements.)

Filter clearing and mounting

A1.29 If additional analytical work to discriminate between fibre types is required (see MDHS87)29 then samples and blank filters may be cut in half with a scalpel using a rolling action, with the filter carefully held at the edge. Half of the filter can then be mounted, and the other half suitably stored and kept for subsequent investigation if necessary. All samples must be uniquely labelled.

A1.30 The acetone-triacetin mounting method must be used. The principle is that condensing acetone vapour collapses the filter pores, adhering the filter to the glass slide and turning it into a transparent plastic film with any asbestos fibres contained close to the upper surface. Triacetin is used to provide the interface between the collapsed filter and the coverslip. The mounted slide will keep for years without noticeable deterioration, although small-scale fibre movement may occur. Slides should be stored carefully and not subjected to extremes of temperature. They should be preserved with all relevant records for at least six months so that the result can be checked if necessary.

A1.31 The filter to be mounted is placed centrally on a clean microscope slide, sample side upwards, and preferably with grid lines parallel to the slide edges. It is important that the filter is free from excessive moisture as water interferes with the clearing process. If samples have been exposed to high humidity, it may be necessary to dry the filters before mounting. This can be done by placing the filters and containers in a warm air cabinet (without a fan), or slide warmer before mounting, making sure the lid of the container is at least partly removed to allow water vapour to escape. A ring of metal or inert plastic placed around the filter helps to localise the spread of acetone and improves the efficiency of clearing. The minimum volume of acetone to completely clear the filter should be used (~0.25 ml). The slide (which must be clean) is placed under the outlet orifice of the hot block. The acetone is injected slowly into the hot block so that the vapour emerges in a steady stream over the filter. The filter should clear instantly. This small amount of acetone minimises fire and health risks. However, all sources of ignition should be remote, and the acetone storage bottle should be stoppered when acetone is not being extracted. Acetone vapour is highly flammable and slightly toxic and the appropriate safety precautions should be taken before this procedure is used. (The procedure may be conducted in a fume-cupboard to minimise inhalation of acetone vapour.) The slide may be placed on a hot plate at 50-60°C for a few seconds to evaporate any excess acetone before applying triacetin and the coverslip. When the acetone has evaporated, a micropipette or other suitable dropper is used to place a drop or two of triacetin on the filter (~120 μl). This must be just enough to cover the filter when the coverslip is in place, without overflow around the edges. The clean coverslip is lowered gently at an angle onto the filter so that all the air is expelled. It should not be pressed onto the filter, or moved, once it has been lowered into place. If necessary any excess triacetin can be carefully removed from the slide using a corner of a tissue, or similar, to absorb the excess fluid by capillary action. The coverslip should not be touched or wiped in any way. At this stage the filter will still appear grainy under the microscope. If counting is to take place immediately, it should be placed on the hot plate (eg for up to 15 minutes at 50-60°C) to produce a more transparent mount. If left overnight at room temperature, the filter will ‘clear’ without any heating. The slide should be kept clean in a horizontal position (coverslip side up), until it is cleared and counted.

Evaluation

A1.32 The microscope must be adjusted and used according to the manufacturer’s instructions, and the analyst must check its performance at the beginning of each counting session (or more frequently if any adjustments have been made). A typical sequence for checking that the microscope is correctly adjusted is:

• Place, centre and focus the working stage micrometer, preferably using bright-field illumination. If necessary use the low-powered objective to help locate the 0-100 μm scale, then return to the X 40 objective
• Adjust field iris and condenser height to obtain Köhler or Köhler type illumination
• Check (and adjust if necessary) that the inter-ocular distance is correct for the user, the image has sharp focus in both oculars and that the Walton-Beckett graticule is also in sharp focus
• Measure and record the diameter of the Walton-Beckett graticule against the stage micrometer (it should be in the range of 98-102 μm). The measured diameter must be used in calculations
• Remove the stage micrometer and replace with the HSE mark II/mark III test slide
• Centre and focus the test slide using phase contrast microscopy, (if necessary use dark field illumination and a low-powered objective to help locate the two sets of parallel grooves (tramlines) in which the test grating is located, before inserting the X 40 phase objective)
• Check using a Bertrand lens or phase telescope that the phase rings are concentric and centred: adjust if necessary
• Check and readjust the field iris and condenser height at the working magnification to obtain Köhler or Köhler type illumination
• Record which of the seven bands is just visible (lines only partly seen) by traversing from the most visible to the least visible
• The ridges of block 5 of an HSE mark II test slide must be visible, while only parts of block 6 ridges may be visible and none of block 7 ridges should be visible at the working magnification. Mark III test slides issued with a red certificate require that block 4 must be visible while only parts of the block 5 ridges may be visible and none of the block 6 ridges should be visible
• The focus and condenser focus will need readjustment before each filter is counted

Other sequences can be used provided all the necessary adjustments and checks are made.

A1.33 The slide with the mounted filter is placed on the microscope stage. The sample should be examined with a low power objective to check the uniformity of the deposit and that there is no gross aggregation of fibres or dust on the mounted filter. The filter should be discarded if badly non-uniform or overloaded with particles so that it is difficult to count. Fibres on the filter must be counted using at least X 500 magnification (Note: if higher magnifications are used they should not exceed 1000 times the numerical aperture of the objective lens). The fine focus must be adjusted upwards and downwards by several micrometers at each new area to ensure that all fibres are seen. The counting should proceed according to the following rules:

• graticule areas for counting must be chosen at random to avoid bias and to be representative of the exposed filter area. Fields lying within 4 mm of the filter edge (or 2 mm of a cutting line) should not be counted. Fields should be rejected if:
  • A filter grid line obstructs all or part of the field of view
  • More than one eighth of the graticule field area is occupied by an ––agglomerate of fibre and/or particles, by discrete particles or by air bubbles
  • The microscopist judges that fibres are so obscured that they cannot be ––counted reliably

If the number of rejected fields exceeds 10% of the number accepted, or the microscopist judges the sample to be uncountable or biased, this should be noted in the final report.

• A countable fibre is defined as any object which is longer than 5 μm, with an average width less than 3 μm and having an aspect (length/width) ratio greater than 3:1 (fibres attached to particles are assessed as if the particle does not exist and are counted if the visible part of the fibre meets the above definition);
• The following recording rules apply:
  • A countable fibre with both ends within the graticule area is recorded as one ––fibre
  • A countable fibre with only one end in the graticule area is recorded as half a ––fibre
  • A countable fibre passing through the graticule area, and having no ends ––within that area, is not counted
• A split fibre is taken to be one countable fibre if it meets the definition above, otherwise it should be ignored; a split fibre is defined as an agglomerate of fibres which at one or more points on its length appears to be solid and undivided, but at other points appears to divide into separate strands; the width is measured across the undivided part, not across the split part
• Loose agglomerates of fibres are counted individually if they can be distinguished sufficiently to determine that they meet the definition above
• Fibres in a bundle and tight agglomerates of fibres, where no individual fibres meeting the definition of a countable fibre can be distinguished, are taken to be one countable fibre if the bundle or agglomerate as a whole meets the definition above
• If the width of the fibre varies along its length, a representative average width should be considered.

A1.34 The number of graticule areas counted depends on the sampling situation as follows:

• For evaluations related to personal sampling in connection with compliance sampling and the assessment of respirator protection, at least 100 fibres must be counted or 100 graticule areas must be inspected, whichever is reached first; at least 20 graticule areas must be inspected even if these contain more than 100 fibres
• For evaluation of other samples (eg clearance indicator, background, reassurance and leak sampling), 200 graticule areas must be inspected on samples of the minimum of 480 litre volume. If the collected air volume (v) is more than 480 litres, the number (n) of graticule areas inspected may be reduced proportionately according to the formula n = 96 000/V. For example, if 960 litres is collected, only one hundred graticule areas need be examined. It may not be necessary to examine n graticule areas if a clear decision can be reached at an earlier stage: for example, if 30 fibres in 200 fields would give a calculated concentration of 0.015 f/ml, then it may be possible to report an enclosure as unsatisfactory as soon as a count of 30 fibres is reached (even if only a few graticule areas have been examined). Where 2 or more samples are being pooled to obtain 480 litres (or a larger volume), V is the total volume of the pooled samples and n is the same number of graticule areas inspected on each of the pooled filters

Calculation of results

A1.35 The airborne concentration is given by the formula:
C = 1000 N D2 / V n d2 fibres per millilitre (f/ml)
where N is the number of fibres counted;
n is the number of graticule areas examined;
D (mm) is the diameter of the exposed filter area;
d (μm) is the diameter of the Walton-Beckett graticule;
V (litres) is the volume of air sampled through the filter.
When pooling 2 or more samples, V is the total volume sampled, N is the total number of fibres and n is the number of graticule areas examined on each filter (which is the same for each filter and not the total number examined). The concentration result must be calculated correct to sufficient decimal places to determine whether it is above or below the specified limit (eg for 0.2 f/ml calculate to 2 decimal places, for 0.01 f/ml calculate to 3 decimal places).

A1.36 The following points should be noted in relation to measurements.

• Sampling for comparison with control limits may be carried out for a set time or for the duration of a specific activity. Ideally, a time-weighted average should be calculated based on a continuous 4-hour sampling period. However, in many circumstances (eg asbestos removal and maintenance work), continuous 4-hour or 8-hour sampling is not possible, as the duration of the work is too short and activity-related sampling is carried out). A time-weighted average for a fixed time period can be calculated if additional sampling of other activities (eg tea and lunch breaks) are included or an assumption that no exposure is taking place is made.
• Even when continuous sampling is possible, high particulate concentrations may require that shorter sampling periods are used to yield countable filters and samples must be pooled together, to calculate a time-weighted average.

Recording and reporting results

A1.37 All relevant sampling and analytical information should be recorded. The sampling records should include relevant site information and contain sufficient information: to establish the traceability of any calibrations, to calculate the results and to assure the quality of the sampling. The analytical records should contain sufficient information: to establish the traceability of the calibrations, to calculate the results reported and to assure the quality of the analysis. The report should include sufficient information on the sampling and analysis so the results are traceable and the purpose and outcome of the sampling are clear. Normally, results are also covered by a laboratory’s UKAS accreditation (see CAWR 2002)10 and additional information (see Lab 30)20 may be required to be recorded and reported. As sampling and analysis may be carried out by different individuals and bodies, the analytical report should either append or contain the appropriate sampling information. All test reports should conform to ISO/IEC/17025 requirements.20

A1.38 Any report should include the following information:

• The name or letterhead of the body carrying out the work
• The full postal address of the body and other electronic contacts
• The UKAS accreditation mark and number (and any appropriate disclaimer)
• The printed name(s) of the person(s) who carried out the work
• The printed name and signature of the person who authorised the release of the report (this may be the same person who carried out the work)
• The date the report was authorised for release
• A suitable report identifier or number

The sampling report should also include:

• The location of the sampling (eg name and address)
• The date of sampling
• The type of sampling being carried out
• The sampling information for each sample, including:
  • A unique identifier (eg sample number)
  • The type of sample (eg personal or static and compliance, background, ––clearance etc)
  • The position of the sample: (eg the name of the or location)
  • The sampling time started and ended for each period
  • The calculated volume of air sampled
  • Reference to any specific activities or events taking place during the sampling ––(eg during demolition, immediately after demolition etc)

The analytical report should also include:

• The method of analysis used and for each sample
• The sample number
• The volume of each sample (if not given elsewhere)
• The fibre concentration
• The limit of quantification

(The reported concentration should not imply greater accuracy than can be justified by the limit of quantification, eg a 480 litre volume sample with 200 fields counted will be reported as <0.01 f/ml or rounded to two decimal places if >0.01 f/ml.)

A1.39 As well as containing the information reported, the sampling and analysis records may also include:

• The sampling strategy, including any variations from standard procedures (eg for very dusty conditions, sampling times may need to be very short to prevent overloading and a stopwatch may give a more accurate measure of the sampling period)
• Any relevant environmental conditions which may significantly influence the results (eg fog or rainfall if sampling outside, large temperature and pressure differences between the calibration and field conditions for float type flow meters)
• The type of filters in use and batch number
• The type and identifier for the flow measurement device
• The type and identifier for the air sampling pump
• The identifier for the timing device
• The measured flow rate at the start and end of each sampling period and any checks in between
• The name of the analyst(s) carrying out the fibre counting
• The identifier for the stage micrometer
• The identifier for the test slide
• The measured diameter for the Walton-Beckett graticule
• The block number where the gratings are still visible on the HSE test slide
• Number of graticule areas examined for each sample
• Number of fibres counted for each sample
• The measured diameter and calculated area of exposed filter
• The upper and lower 95% confidence limits of the fibre count
• The overall uncertainty
• Any additional information for the discrimination counting (see MDHS87)29 strategy

Human factors

A1.40 Take care to ensure that the ergonomic and working practices and the working environment in a laboratory have no adverse influence on the accuracy of the counts. There should be sufficient legroom and clearance so that adjustable seating can be adjusted to allow the microscopist to sit in a well-supported, relaxed and comfortable manner. To avoid eye fatigue, the light intensity of the microscope and surroundings should be comfortable to view. Also, the ambient light should not be brighter than the microscope and should not reflect off the coverslip or optics or cause any other source of glare. Any peripheral view beyond the microscope should be, if possible, an unobstructed distant view in unchanging light. Alternatively, a matt background shield can be used. Poor posture may lead to neck and back strain and poor lighting or incorrect microscope adjustment will lead to eyestrain, all of which will adversely influence counts produced by the microscopist. Adjustments of the inter-ocular distance and for the different focal lengths of each eye are quick but important and should be carried out by the microscopist at the beginning of each counting session. The eyes should not be too close to the oculars and if high eye point wide-field binoculars are used those not wearing glasses should make use of the eye shields. The microscope image should be sufficiently vibration-free that particles in the field of view are both steady and clear.

A1.41 Limits must be placed on the amount of fibre counting undertaken by analysts in specified periods because fatigue can adversely affect the quality of counts. The number of graticule areas examined in any 8-hour period by one counter should not normally exceed 2400, the equivalent of 12 samples if 200 graticule areas are examined on each. Counters are recommended to take a break at least after every third or fourth slide counted in succession, and if long shifts are worked, additional quality assurance (QA) measures may be necessary. The length and frequency of the fibre counting sessions will depend on the microscopist, the type of samples and the laboratory conditions. The number of samples evaluated in a day also differs from microscopist to microscopist: typically, counters may take 10-25 minutes to evaluate a sample with a sparse dust deposit, but longer for greater numbers of fields and more difficult samples.

Accuracy

A1.42 It is not possible to know the ‘true’ airborne fibre concentration of a given dust cloud and the absolute accuracy of the method cannot be assessed. However, some information is available about relative bias associated with sample evaluation. Microscopists generally undercount dense deposits. When sampling fibres in atmospheres relatively free from interfering particulates, the density range for optimum accuracy should be in the range 100-650 fibres/mm2;34 for densities above 650 fibres/mm2 the results may be underestimates (but no attempt should be made to correct them) and above 1000 fibres/mm2, fibre levels are subject to increased undercounting and are normally too dense to count. In mixed dust situations, the presence of other fibres and particles may interfere with the accuracy of results. Chance superimposition of non-fibrous particles may cause fibres not to be counted fully, by a proportion, which depends on the mean size and concentration of the non-fibrous particles but not on the fibre concentration.35 In practice, the effects of chance superimposition on counts are small compared with subjective effects and will not be important for the counting rules defined in this method. An important factor is that the counting procedure can result in systematic differences in counts produced by different microscopists within and, more particularly, between laboratories. These intra- and inter-laboratory differences must be controlled by proper training and periodic quality checks.

Precision

A1.43 Counting precision depends on the number of fibres counted and on the uniformity of the fibre distribution on the filter. The latter may be described reasonably by the Poisson distribution. Theoretically the process of counting randomly distributed (Poisson) fibres gives a coefficient of variation (CV) = 1/ N1/2, where N is the number of fibres counted. Therefore the CV is 0.1 for 100 fibres and 0.32 for 10 fibres counted. In practice, however, the actual CV is greater than these theoretical numbers due to an additional component associated with subjective differences between repetitive counts by one microscopist and between replicate counts by different microcopists: this CV is given approximately by the formula (N + 0.04N2)1/2/N, where N is the mean number of fibres per evaluation;36 . If n fibres are found in a single evaluation, the mean of many repeated determinations on equal areas is expected to lie within the confidence limits M95 and M05 on 90% of occasions₃₆ where:
0.91 M₉₅₂ - (2n + 2.25)M₉₅ + n₂ = 0,
0.84 M₀₅₂ - (2n + 4)M₀₅ + n₂ = 0.

These equations have been used to calculate the upper and lower confidence limits shown in Table A1.1. It can be seen from this that counting more than 100 fibres gives only a small increase in precision. Also, the method loses precision as fewer fibres are counted; this loss of precision increases as counts drop below about 10 fibres. Inter-laboratory CVs can be twice the intra-laboratory coefficients, or even greater if quality control is poor.

Uncertainty budget

A1.44 The UKAS Lab 30 document,20 which gives guidance on ISO/IEC 17025,37 requires that the variation associated with each part of the measurement is used to calculate the overall uncertainty (also referred to in UKAS documents as total uncertainty and expanded uncertainty). Various levels of complexity38, 39 are recommended, but the overall uncertainty for air monitoring of fibres is a function of the systematic, subjective and random errors associated with the air sampling and fibre counting. It is normal practice to represent the overall uncertainty in terms of the 95% confidence interval (this is equivalent to the standard uncertainty times, a coverage factor of 2 for a normal distribution). In a first approximation, the uncertainty due to the systematic errors associated with the analysis are small, when compared to the random errors due to the placement of the sampler and the random distribution of fibres on the filter. As fibre counting is carried out manually, there is also a substantial subjective error that will vary with a whole range of factors for the same counter and between counters. Due to the large random and subjective variability associated with manual fibre counting and that the underlying random Poisson distribution is not symmetrical, the overall uncertainty can only be derived from repeated blind measurements of the same sample. The uncertainty from between (inter-) laboratory counts is best represented by the RICE performance limits.

A1.45 Examples of the factors that contribute to the systematic variability within a laboratory are given in Table A1.2. Both the maximum allowed variability as defined in the method and the typical values measured are evaluated. In some cases a maximum variation has not been specified in the method and a value of 1% has been adopted. In practice,38 if the variation is <20% of the largest variable, it is usually regarded as negligible. It can be seen from Table A1.1 that the 95% confidence interval from fibre counting, for a count of 20 fibres is between 48 and 187%. The systematic uncertainty from sampling and calibration is calculated (Table A1.2) and combined with the fibre counting variables (Table A1.3) using the root sum square method that treats all the variables as independent. As the maximum allowed systematic uncertainty in Table A1.2 is between +26% of the mean count, this gives a calculated 95% confidence limit for the overall uncertainty of between 55-191% of the average value. The more likely value of the systematic error calculated in Table A1.2 is ~11%. Even assuming the systematic error is the maximum allowed the overall effect of the systematic errors is less than 20% of the fibre counting error for counts of up to 200 fibres (see last two columns of Table A1.3). This means all the systematic errors from calibrations, timing and flow measurement could be regarded as negligible, compared to the random and subjective errors from fibre counting and the 95% confidence limits derived from the equations in paragraph A1.41 and Table A1.1, and these adequately describe the overall uncertainty.

Quality control

A1.44 Employers must ensure that the laboratories which they use for the sampling and analysis of airborne asbestos meet the necessary standards set out in ISO/IEC 17025.37 Employers can satisfy this responsibility by using laboratories that hold UKAS accreditation for asbestos sampling and asbestos fibre counting. UKAS publishes guidance,20 which discusses accreditation for asbestos sampling and analysis. Laboratories are responsible for ensuring the work is carried out by competent trained staff. The training may include formal training and qualification from organisations participating in the Faculty of Occupational Hygiene within BOHS S301 module for asbestos and other fibres. New staff will need at least a qualification in the BOHS proficiency modules P403 asbestos fibre counting and P404 air sampling and clearance testing of asbestos. Suitable in-house training and performance monitoring of laboratory staff must be carried out to ensure the quality of the results.

A1.45 An essential part of quality assurance is participation in internal and external quality control schemes. This is particularly appropriate for this method because of the large differences in results within and between laboratories obtained with all manual fibre counting methods. Laboratories using this method therefore must participate in the Regular Interlaboratory Counting Exchanges (RICE) scheme. This provides a measure of the laboratory’s performance in relation to other counting laboratories. Participation in RICE must be supplemented by checks on internal consistency, which should aim to measure and control the individual counter’s performance relative to other counters in the laboratory. The internal quality control scheme should incorporate the use of both reference samples (ie those which have a well-defined result established as a mean of a number of determinations), routine samples (ie those which have been analysed in the course of normal work) and blanks. Participation and assessment of individual performance should be carried out at least once a month. Systematic records of quality control results must be maintained and the assessment of performance must be to a defined set of criteria.

A1.46 If it is suspected that the HSE mark II/mark III test slide has deteriorated in quality due to damage or other factors then it should be re-evaluated. HSL should be contacted for advice (see paragraph A1.47).

Advice

A1.47 Advice on this method may be obtained from the Health and Safety Laboratory (HSL) (see ‘Further information’ for their address). Suggestions for improvement should be sent to the same address.