External Standard Method

[tien]

Hi all,
In USP <621> CHROMATOGRAPHY:
External Standard Method: The concentration of the component(s) quantified is determined by comparing the response(s) obtained with the sample solution to the response(s) obtained with a standard solution.

Is there any specific statement in USP or other official documents require the concentration of the standard solution to be similar to that of the sample solution?

Thank you all in advance.

[Blazer]

Not that I'm aware of. It's common practice since it can allow you to eyeball the approximate concentration of your sample. This can give you a leg up on beginning to troubleshoot an issue if the peaks aren't similar. That said, as long as your results are demonstrated to be accurate, it doesn't really matter what your standard concentration is.

[Kreall]

Generally if you run only one concentration of standard than you should use approximately the same concentration as your sample. The pickle is to know you b value from standard curve.

If the b value in equation y= ax + b is significant different from 0, than running different concentration from your sample will make you results underestimated or overestimated. In this case making the same concetration as your sample is right way.

If you know that you b value from standards curve is not significant different from 0, that propably using different concetration from your sample should make your result accurate.

Kreall

[danko]

I will always prefer to generate a standard curve, covering the region I find relevant.
Two rationales for that. 1: If you only have one level standardization the other/lower end of the curve will be assumed to be zero and that is very, very seldom the case. So, that was the first source of determination error. 2: How do you know what concentration the sample would have, especially if you test stability samples or recovery samples etc.? So, that was the other source which might result in necessity for re-analysis if the std. and sample concentrations are too far from eachother.
Actually, there’s a third rationale for generating std. curve: You test the linearity in every analytical series, which is a typical SST parameter.

Best Regards