2
$\begingroup$

I help with the supply chain of a small supplement company, sometimes the HPLC lab reports we get back have a weight % over 100%. What does this mean, how does this kind of thing happen? Obviously the sample can't be over 100% pure.

enter image description here

$\endgroup$
  • $\begingroup$ Rounding error maybe? $\endgroup$ – Todd Minehardt Feb 9 '17 at 1:32
  • $\begingroup$ So like a rounding error from the weight calculation algorithm they're using? I've seen it occasionally with different contract labs. $\endgroup$ – Veta Feb 9 '17 at 1:36
  • $\begingroup$ How many different parameters are being measured and how much over 100% is the total? Or is it just one compound coming back as >100% ? $\endgroup$ – airhuff Feb 9 '17 at 1:37
  • 1
    $\begingroup$ You could ask the lab for specs like uncertainty of the reported values. $\endgroup$ – airhuff Feb 9 '17 at 2:23
  • 1
    $\begingroup$ Not given the additional information you have provided. I would strongly suggest "within analytical uncertainty of the method", but you would need to verify with the lab what the accuracy of that measurement is. $\endgroup$ – airhuff Feb 9 '17 at 2:51
3
$\begingroup$

By far the most likely reason for reported purities of over 100% when using HPLC is the "analytical uncertainty of the method", but let me illustrate exactly how this is likely to arrive.

A standard procedure for estimating purity is:

  1. Run an exhaustive series of known standards with (assumed) exactly known purities, or weight percents, on the HPLC.

  2. Plot this calibration data with peak area on one axis, and the known weight % on the other axis.

  3. Fit a line to this data. The slope of this line is called the "response factor" or sometimes just the "response".

  4. Run some unknown samples and measure the peak area for the unknowns.

  5. Using the response factor, calculate a weight percent for the unknown sample.

If there are analytical uncertainties in the calibration standards, then the response factor may be randomly off by small amounts from the "true" value of the response factor. Similarly, if there are uncertainties in the peak area of the unknown sample, even using the "true" response factor may give a value that is slightly off.

Statistically, these errors should be random, so that sometimes when the "true" purity is 100.000 % the measured value will be 100.6%, and sometimes it will be 99.4%.

Chemically, what is the source of the errors? Well, maybe the autosampler that injected the unknown sample was supposed to inject 2.0000 μL of sample for analysis, but due to instrument imperfections actually injected 2.012 μL instead. Maybe there was a tiny power surge in the AC lines going to the UV detector that was measuring the peak right when the compound of interest eluted, leading to an ever-so-slightly greater signal at that moment...

The list of possible causes is nearly endless, but none of these things are really cause for concern, unless you really need to know the purity to three or more decimal places. From the report you provided, you can safely assume that the uncertainty in their analytical method is at least 0.6%, but not much else.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.