GCI TECH NOTES©
12, Number 6
A Gossman Consulting, Inc.
Publication August 2007
Whether you are analyzing the oxygen bomb wash from organic fuels to determine chlorine content in fuel or nitrate levels in drinking water, soil or plants, there is always the issue of the choice of analytical method. The choice of ion selective electrodes (ISE) or ion chromatography (IC) have been options for many years. Quite frankly, I have never felt very comfortable with ion selective electrodes and have always chosen to use the more expensive and time-consuming ion chromatography. That goes for recommendations to my clients using alternative fuels in cement kilns as well as our own sister company lab – ChemRight Laboratories, Inc. (CRL)
At CRL we test soil, drinking water and some plant samples for nitrates. Ion chromatography is an approved method for nitrates, as is the use of an ion selective electrode, along with a colorimetric method that uses cadmium as a reagent. Because we don’t perform tests at CRL that generate hazardous waste it was an easy decision to avoid the use of the cadmium reduction/colorimetric technique even though it has a long history of providing accurate values. As part of our operation of CRL we participate in a number of certification programs including one for soil that requires us to test a number of blind soil reference samples each quarter and submit the results for statistical analysis across a large number of soil testing labs nation wide. In the most recent quarter, we were one of only two labs that used ion chromatography for nitrates. Most of the other labs use either ion selective electrode or the cadmium reduction technique.
We graphed the results of the 2nd quarter 2007 NAPT (North America Proficiency Testing) samples. The results highlight to a startling degree some concerns about the accuracy and precision of the ISE method. The figure provides a graph of the nitrate determinations by 18 labs using ISE, 63 labs using cadmium reduction/colorimetric and the ion chromatography result provided by CRL. The statistical warning levels are calculated by NAPT and provided in their report.
The high degree of variability in the ISE data suggests that anyone using this data has real reason to be concerned about the accuracy of the results. To me ISE does not even meet the laugh test and this data strongly supports our decision to use ion chromatography to determine anions in a variety of media. The cadmium reduction method can also produce quality data. Yet, when faced with the liability for both the lab and the customer of generating hazardous waste to get data to insure adequate fertilizer or safe drinking water use of this method seems counterproductive when ion chromatography exists as a clear alternative.
The other clear aspect that this data highlights is the fallacy of relying on statistically derived control limits. Labs performing the same testing of the same parameter in the same matrix for the same customers and use should be held to the same levels of quality data, not two very different levels of quality entirely based on the statistical variability inherent in the method chosen. Clearly the use of industry-wide data quality objectives would be a better standard to hold the labs to, in order to provide the data user community a higher confidence that the data meets the needed quality for its intended use. The use of statistically derived control limits can indicate when accuracy is worse than before, but it can not confirm that accuracy is adequate. The fact that some labs passed the statistically derived limits for nitrate using ISE clearly provides a false sense of security to those data users who count on lab certification programs to confirm a consistent level of accuracy.