The subject of Automatic Particle Counters (APCs)

APC’s have occupied the attention of users, designers and experts over the past 35 years or so, in an attempt to improve accuracy and traceability of the measurements from these instruments. At the last meeting of the ISO group responsible for drafting standards for the fluid power industry (ISO/TC131/SC6 Contamination Control), the project leaders gave an update of the project and explained how potential confusion can be averted. Mike Day of CMS Consultants Ltd was at that meeting as part of the UK team and has prepared this article.

APCs count  nd size the very small particles that can be present in a wide range of fluids in various industries e.g. fluid power, lubrication and fuel to name most, so that their numbers can be controlled. The size parameter is the µm. APCs were first introduced in the early 1970’s and have since proved to be indispensible for the control of particulate contamination in both service and associated process fluids. The principle is seen in Figure 1; the passage of a particle through the sensor reduces the amount of light received by the detector and this produces an output voltage pulse, the value of which is proportional to the size of the particle and is obtained by calibration.

Traceability

As decisions are often based on the value of the data, it is essential that it is both correct and valid otherwise incorrect decisions will be made. This will waste valuable time as well as costing money. Important aspects in this is calibration and traceability of the measurement and the development of these are detailed below:

In the early models (1970s), the various channels of the were adjusted to ‘trigger’ when a voltage pulse was fed into the instrument which represented the passage of a spherical particle sized d µm. This was based upon the ratio of the projected area of the particle to the area of the detector, thus V = Ap*Vb/Ad,, where: A Is the area of particle, A the area of detector and Vb is the base voltage of the instrument. This was a theoretical calibration with little traceability.

Also in the 1970’s researchers at Oklahoma State University in the USA developed a calibration method for APCs based upon A.C. Fine Test Dust (ACFTD), which was then being used to test filters and it was a logical development to use this as a calibrant. Identical samples of a suspension of ACFTD were prepared and the number of particles were sized and counted using an optical microscope with the longest dimension as the sizing parameter. This was adopted as ISO 4402 [1] in 1977. Whilst this method offered standardisation, there was virtually no traceability to national standards of measurement and batch to batch control of the ACFTD was not good enough and lacked the consistency that was necessary for a reference material. Perhaps of greater importance was that the size vs. number relationship was shown to be incorrect in the ‘80s.

Also in the late 1970s, two calibration methods were developed using latex spheres. One used mono-sized spheres (ARP 1192) which had traceability to national standards and the other used distributed latex spheres (CETOP RP94H) which was a secondary calibration and loosely traceable to a mono-sized calibration standard. These two methods were dismissed by ISO

/TC131/SC6 as the carrier fluid was water with a different Refractive Index to the fluids used by the hydraulics industry and both gave variable results.

In 1997, ISO were informed that ACFTD would no longer be available and had to develop  an  alternative.  The  alternative chosen was SAE 5-80, a test dust of identical material to ACFTD, but with significantly less numbers  of  particles  below  5µm, which lessened coincidence errors in the APC. This was renamed ISO Medium Test Dust (ISOMTD) and formed part of a family of test dusts defined in ISO 12103 [2]. The particle size distribution (PSD) of calibration samples (SRM 2806) was certified by the National Institute is Science & Technology (NIST) in the USA  using a Scanning Electron Microscope  (SEM)  with  Image Analysis software, and traceability was obtained. The PSD of the dust was obtained by sizing in the particles in the sample on the basis of the diameter of a  sphere  whose  area  was  the  same  as the particle, hence the term equivalent spherical diameter’.

Unfortunately, this new measurement gave slightly different counts to the previous ACFTD calibration and, to lessen the effects on industry, ISO selected the measurement size that would give notionally the same particle count had the APC had been calibrated using ACFTD to ISO 4402. This is seen in Table 1 As the equivalent sizes were often decimal sizes, the nearest integer size was selected to simplify communication, as seen in Table 1. To distinguish between the ‘old’ and ‘new’ calibration data, the new size was labelled µm(c).

This new method of calibration has been published as ISO 11171 [3] and gives a more accurate and traceable calibration. This will enhance both the repeatability of measurement with the same instrument and the reproducibility of measurement between different instruments and laboratories. The standard also included procedures for determining the performance and accuracy of the APC.

The latest situation

The number of people using the NIST SRM 2806 calibration samples has increased subtantially the last 20 years and there is  a limit to the numbers of suspensions that can be produced in    any one batch. Thus, re-certification is necessary when a new batch is produced. As the first two batches (SRM 2806  and 2806a) were produced from the  same  material,  recertification was not necessary, and the original certification applied.

The next batch (SRM 2806b) was certified using an improved SEM with automated analysis to improve both the speed of analysis and its accuracy. A 10% increase in particle size was observed and this caused a shift in the calibration curve, for example, 10 µm(c) particles became 11 µm(b) particles. This 10% shift in particle size has a significant impact on observed particle counts, and reported filter removal ratings. To compensate for this a size correction factor of 0.898 has to be applied to correct a 2806(b) calibration to a 2806(c) and µm(c),

i.e. µm(c) sizes =µm(b) x 0.898. Unfortunately, the latest version of ISO 11171: 2016) – which is still current – was somewhat ambiguous and it gives two options to obtain the calibration curve which will report different results on the same sample. This is in the process of being revised.

Note that it is inevitable that there will be differences in the sizing of these particles during certification because the angular shaped particles will adopt different orientations on the analysis membrane and thus present a different area to the detector. This is despite the extraordinary lengths that NIST go to in order to control variability.

ISO TC131/SC6 recently took a decision to ‘standardise’ on µm(c) as the measurement standard in ISO 11171 and all subsequent editions will feature this. SC6 have also instructed NIST to develop a ‘Consensus Standard’ which will simply give the numbers of particles per mL at the µm(c) size in the calibration suspension, thus greatly simplifying the issue.

What to do about your current calibration.

The actions necessary will depend on the  calibration  method used. If the calibration uses µm(c) then there are no actions necessary as this is the standardised unit now and in the future; both historical data can be compared and current specifications  can be used. If µm(b) calibration is used, it:

The actions necessary will depend on the  calibration  method used. If the calibration uses µm(c) then there are no actions necessary as this is the standardised unit now and in the future; both historical data can be compared and current specifications  can be used. If µm(b) calibration is used, it:

  • will give increased counts compared to µm(c) data
  • may lead to failure to meet µm(c) specifications may result in extra time and money being expended to determine if the increased counts are due to real issues
  • will give discontinuities with historical data
  • may cause investment in testing to validate revised specifications for entire size range
  • necessitate a repeat of the above factors when both subsequent batches of SRM 2806(x) and ISO 11171 revisions are released So, it is recommended that users of SRM 2806(b) calibration data convert the µm(b) sizes to µm(c) using the conversion factor in ISO 11171:2016, i.e. 0.898, to give the equivalent decimal µm(c) sizes. The calibration threshold values can be plotted against the decimal µm(c) values to obtain the µm(c) calibration curve for that calibration and the integer µm(c) sizes obtained by interpolation. Alternatively, a curve fitting routine can be used for more accurate interpolation. The ISO committee recommend the constrained cubic spline interpolation method developed by C.J.C. Kruger [4]. This curve fitting routine can be used to obtain the revised µm(c) calibration data and also to obtain the particle count data on the basis of µm(c) calibration  to see if the SRM 2806(b) data exceed earlier specifications.


Future calibrations

As stated above, NIST are working on the so-called ‘Consensus Standard’ so that future calibration samples will be certified to    the µm(c) base and calibration. Thus, it will be a simple matter     of setting the APC to record the certified numbers at the sizes required that are given on the NIST certification sheet. This will greatly simplify the process of APC  calibration,  overcome possible confusion caused by a recertification  process,  improve the supply of calibration samples and avoid running out  of samples. It will also greatly reduce the amount of time spent by members of the ISO Working Group that has to be involved with the validation of any new batch of suspension material.

References

  1. ISO 4402: ‘Hydraulic fluid power – Calibration of automatic count instruments for particles suspended in liquids – Method using classified AC Fine Test Dust’, International Standards Organisation, Geneva, Switzerland, 1991.
  2. ISO 12103-1: ‘Road vehicles – Part 1: Arizona test dust’, International Standards Org., Geneva, Switzerland, 2016.
  3. ISO 1 171: ‘Hydraulic fluid power – Calibration of automatic particle counters for liquids’, International Standards Org.,

Geneva, Switzerland, 2016

  1. C.J.C Kruger: ‘Constrained Cubic Spline Interpolation for Chemical Engineering Applications’, Korf Technology Ltd, Ontario, Canada, 2003, www.korf.co.uk/spline.pdf.