After the last newsletter I was reminded about the change to the filter specifications in Method 25 to correspond with the filter requirements in Method 5.
Filters. Glass fiber filters, without organic binder, exhibiting at least 99.95 percent efficiency (<0.05 percent penetration) on 0.3 micron dioctyl phthalate smoke particles. The filter efficiency test shall be conducted in accordance with ASTM Method D2986-71, 78, or 95a (incorporated by reference—see §60.17). Test data from the supplier's quality control program are sufficient for this purpose.
These specific ASTM methods are no longer applicable as they have been withdrawn by ASTM. I believe the test standards for HEPA filters are now the applicable methodologies since HEPA filters can be effective at the 0.3 micron level. The Institute of Environmental Sciences and Technology (IEST) is the more probable source for the certification procedure as they seem to provide the HEPA filter test methods used by the American Society of Heating, Refrigerating and Air-Conditioning Engineers (ASHRAE) for their filter class determination and the certification of clean rooms or other uses of HEPA or better filtration.
This actually leads into the biggest question posed, which was why not use Method 25A if Method 25 has any positive bias? This is a good question but very difficult to answer because Method 25A has multiple bias potentials which can result in a net positive or negative bias depending on the factors involved.
An FIA is generally not used in conjunction with a Method 5 style filter, which means some sources where Method 25 would show a small loading the FIA of Method 25A will show a much larger loading. This type of positive bias is what led us to discover the significant difference in the different filters used for Method 25 over the last three decades. The FIA in use was in series with an FTIR, which was seeing a problem due to small particulates. The addition of a Method 5 filter in line stopped that issue but also dropped the FIA concentration significantly. The Method 25 samples taken during this time were repeated using a different filter and showed a significant difference in reported concentrations between the filters used. Of course, this bias can be limited by the addition of a Method 5 filter to the equipment requirements of Method 25A as well. Unless there is an across the board change it will be impossible to know if and when this effect occurs since we know the difference in particle size may not be that significant from Method 5 specifications. I do not have the data available for the Method 25A comparisons before and after filtration, but I do have the data from the Method 25 comparisons between a Method 5 equivalent filter and an non-equivalent filter, which is impossible to distinguish by any simple means. These are the same sources with similar flows, conditions, etc., but with the differing filters. The average reduction of the reported concentration in ppmC over three runs were 94% for one, 95% for another, and 97% for the third. If the Method 25A responses were similar the positive bias would be huge in comparison. Of course, these are inlets to control devices so the Method 25A results would look more realistic in comparison to the Method 25 in these cases. The potential for other sources, including outlets, which would have similar impacts is really unknown as this does not seem to have been a subject of much available study.
The second and more common issue with the FIA is the difference in response between different compounds. This has been documented since the 1960s with the work of Dietz and continuing through to the present time. The FID, which is commonly called a “carbon counter” works well on many of the hydrocarbons by having a generally consistent response in relation to the number of carbons in the atom. This is especially true for individual paraffin compounds, but not as much for alcohols or aldehydes, for example. I say individual since nearly every reference to a relative response is for the specific compound, which leaves a question as to the effect of a mixture of multiple compounds. I am going to reference three more recent publications which deal with FIA response relative to propane or methane as a standard since Method 25 also uses propane but analyzes everything converted to methane for a consistent response factor.
The two documents which reference propane as the standard are a US Army research contract report from the early 2000s, http://www.afsinc.org/files/1410-240%20real%20time%20emission%20measurement%20public.pdf, and an undated technical information paper by J.U.M. Engineering, http://www.jum-aerosol.com/images/E-Fakt-02.pdf . I am enclosing the entire JUM document but only the chart from the Army report along with this newsletter for reference to other compounds of interest and their relative responses. The one which references methane as a standard is a technical document from a manufacturing company named Inficon, http://www.raecorents.com/products/gasmonitoring/Photovac-Inficon-MicroFID-II/FID-response-factors-diaf56a1-2012.pdf . I am only including the chart for methanol from this document, since the propane response is given at a different concentration from all of the other comparisons. The differential between methane and methanol is also more significant to the point being made.
The JUM document indicates a concentration of 100 ppm methane would be measured as 42 ppm propane while 100 ppm of methanol would be measured as 23 ppm propane. The second FIA reported similarly with the 100 ppm methane measuring as 42 ppm propane and 100 ppm methanol as 23 ppm propane The Army research indicates the first FIA they used would report the same 100 ppm methane as 34 ppm propane, but the methanol would be similar with the 100 ppm being reported as 24 ppm propane. The Inficon information, which is based on a methane calibration shows a similar negative bias for methanol as the 100 ppm methanol sample would be reported as 24 ppm methane using their relative response for 100 ppm methanol. The 150 ppm propane would be reported as 326 ppm based on the methane calibration and the response factor listed for that level. The fact there are such differing relative response factors for some of the concentrations of different compounds is somewhat confusing, but it may be related to their specific instrument. The difference between the reported and actual concentrations indicated by these comparisons can create a significant negative bias if the impact of the relative response is not corrected. Again this information is all based on the comparison of individual compounds and does not include any impact that might result from or be magnified by the presence of multiple compounds, differing levels of moisture, or temperature in the sample stream.
On an associated note, I have been informed that my patent for the on-site Method 25 analyzer has been approved. Now starts the long journey into getting the production prototype and drawings completed in preparations for commercial production at some point in the future.
Triangle Environmental Services, Inc.
P.O. Box 13294 122 US Hwy 70 E
Research Triangle Park, NC 27709 Hillsborough, NC 27278
(919) 361-2890 (800) 367-4862 Fax: (919) 361-3474