Modifications to the membrane filter method with phase-contrast microscopy (PCM) for fibre-counting to support a new occupational exposure limit value

The occupational exposure limit value (OELV) currently in force for the concentration of asbestos fibres in air in most of the world is 0.1 f cm-3. This value has been considered in some methods as the limit of detection (LOD) for the membrane filter method with phase-contrast microscopy (PCM). There is a directive within the European Union (EU Directive 2023/2668) to lower the OELV in member States to 0.01 f cm-3 from 21 December 2025. In the EU, OELVs are established with specific requirements for the expanded uncertainty around the target value, and to meet those requirements the OELV should not be a simple LOD. A value of approximately one-half of the OELV, if determined accurately, can give reasonable assurance of compliance with the OELV. There are several options available which can allow modification of the membrane filter method with PCM to measure 0.005 f.cm-3 as a time-weighted average (TWA) over a period of hours. As there will be fewer fibres on a sample, greater precision in counting is required. To achieve greater precision, (i) the filter background must be below 2.5 f.mm-2, which is achievable where filters are manufactured in a clean environment, (ii) microscopists must receive training and evaluation, e.g. through participation in a proficiency test program, and (iii) the sample size for counting will need to be increased by increasing the volume of air sampled and counting additional graticule areas. These modifications to the methods, taken together, should allow accurate measurements at 0.005 f cm-3, thus demonstrating compliance with an OELV of 0.01 f.cm-3 for a 4-h or 8-h TWA sample. Although these changes bring some additional cost in sampling and analysis, the PCM method could be retained. Continuing with the PCM methods has the advantages of low cost, capability for on-site analysis, and traceability to risk assessments based on prior PCM data.