Certified Reference Material Frequently PRONOVATION Questions

Q: I have received my PRONOVATION reference Standard and certificate. What do these numbers represent?

A: "Average" is the average derived from all data points (n= number of points or tests). The normal mean square error or one sigma deviation is the "estimate" change we see from test comparisons at the 68% confidence level. The spread or two sigma deviation is the change in the estimate tested at the 95% confidence level.

Q: What numbers should I use and how should I use them?

A: The average or average is the value that you plug in the instrument and strive to maintain the most accurate measurement. You should discuss with your quality manager the accuracy you should maintain in your inspections. Most use 95% confidence. Some standards are less stringent, allowing 3 sigma or 99.7% confidence. Some are more stringent and will stick to 1 sigma or even less.

We believe that if all is well with your analyzer, you have calibrated to the mean and should immediately run the reference standard to make it accurate within the extended 95% confidence range, or even closer to the 68%1 sigma level. If your inspection exceeds or drifts beyond the 95% level, it may be time for a recalibration.

Q: The certificate says "use methods to amplify uncertainty when necessary." What does that mean?

A: While the company uses different analyzers and multiple reference comparisons to establish the most accurate values we can, our information is only a snapshot of material properties and test method capabilities. The Test Method (ASTM) cited on the certificate has developed "Precision and accuracy" tables or calculations based on inter-laboratory studies (ILS). These are more accurate examples of how your test method should behave.

Q: Does lower deviation mean better standards?

A: Not necessarily. While it may represent some part of the homogeneity of the material, there are many variables that need to be examined. Items such as multiple reference standards used, number of test points, different manufacturers of instruments, and different test methods are a few of the factors that affect the deviation results. This is why ASTM or other test methods evaluated by ILS are a more accurate methodological specific description of the test and its capabilities. Ultimately, you can only perform within the capabilities of your test instruments and methods. This means that to maintain stricter deviations, you need to check and calibrate more frequently.

Q: There is no uncertainty in my standard, or it is lower/higher than I can realistically stick to. What should I do?

A: The standard may be older, or they use a very different test method than yours (e.g. by weight, volume, titration, or mass spectrometer). Due to the method or means of determining the value, their accuracy and your accuracy may differ greatly. We recommend that you refer to ASTM or the Test Method Accuracy and Accuracy Table (Repeatability and repeatability) to help determine realistic uncertainties that need to be maintained.

Q: My reference standard is 0.500% and I want to measure the sample at 0.005%. Is that a good comparison?

A: No. The detection system of the instruments used by the company is not really linear, but quadratic. When we calibrate, we are linear only on part of the conic. This is why we strive to develop the most accurate tests using references with similar materials and concentrations.

Q: What is drift and why does my analyzer deviate from the true value?

A: The instrument has many variables that affect the test results. Changes in flow due to dust, leaks, or chemical reagent changes are only part of the story. How good maintenance, calibration and quality checks keep testing at the highest level.

Q: My 1g steel needle is certified at 70ppm±8ppm, and my 0.1g titanium is certified at 70ppm±24ppm, both with 95% confidence. Why is there a big difference?

A: The response of the instrument to a sample that weighs 10 times less is reflected in the deviation. Your instrument has an optimal detection window. The 1g steel sample provided a strong signal response, with much less variation compared to a sample with a "peak response" that was 10 times smaller. The more you deviate from the optimal detection range, the wider your relative mean square error (RSD) becomes. This is also why PRONOVATION makes similar comparisons in terms of material type, sample size and concentration.

To calculate RSD:

8ppm/70ppm*100=11.43%RSD (1g steel needle)

24ppm/70ppm*100=34.29%RSD (0.1g titanium needle)

Another example: 1g steel needle 730ppm±20ppm (20ppm/730ppm*100=2.74%RSD)

Q: The statistical K-factor is 2.1; Does that mean it is 2.1 sigma and greater than 95% confidence?

A: No, the extended uncertainty is still at 95% confidence. The K-value is the unique symmetric unbiased estimator of a statistical distribution. When everything is perfect in the world of chemical test data, k will equal 2. Factors such as rounding, instrument sensitivity capability (%RSD), data flyers (distribution symmetry), and homogeneity contribute to the calculation of mean square error and extended uncertainty.

For example:
The mean concentration is 0.00060>#br### A mean square error =±0.000147%=k=1,68% confidence =1σ
Two standard deviations =±0.000294=k=2,95% confidence =2σ
By rounding to a significant number:
Mean =0.0006>#br### A mean square error =±0.0001>#br###