Inject Samples and Inject Standards in Empower 3 sample set

If running a standard at the start of your run, does the function HAVE to be set to Inject Standards for any system suitability calculations to display? To confirm, I run 6 injections of a standard and the tailing and %RSD of peak needs to be a certain value or range. No amount of response factor is required, which would require the injection to be of a "Standard" type as far as I know. The reason I don't want to Inject Standard is its skewing the amounts and the calibration curve when reviewing a sample set and cycling through the injections ie the points from the lower strength standard are obscuring the calibration curve for later sample amounts. 

I know I could, based on a previous question just clear calibration after checking the system suit standards to start the curve again but I want to know is there any issue simply attaching the function of Inject Samples to a standard?

Best Answer

  • MJS
    MJS
    Answer ✓
    For the system suitability calculations, that is solely dependent upon the processing method, not injection type.  Empower doesn't care if it is a standard or unknown to calculate resolution, plate counts, etc. 

    Using a report method table's summary function to calculate a %RSD is solely dependent upon what is filtered in the table, not injection type.

Answers

  • DavidHPLC
    edited June 2019
    AFAIK, you'll get system suitability parameters for all injections (either Standard or Sample) but for %RSD calculation you'll need to set injections to Standard type.

  • Hi DavidHPLC, would that be the case even if I was using the report method summary functions to calculate RSD?
  • Yes, it would work in the report method as well.

    I believe the only area where this would cause you problems is if you had the processing method setup to shutdown the instrument if the system suitability injections exceed a threshold (say RSD greater than 2%). In this case, if you don't setup standards as such, Empower won't have any injections to compare to.

    We have a lot of older test methods where I work that still run full calibration curves. We very commonly make replicate injections of standard injected as samples to satisfy % RSD requirements. Only the cal curve injections are set to inject as standards. We've never had any problems with reporting, calculation various peak attributes (tailing, resolution, plates, k', RSD, etc.).

  • shaunwat said:

    We have a lot of older test methods where I work that still run full calibration curves. 

    What's your new approach to calibration curves?
  • We try to run single point methods as often as possible. In some instances we observe bias on the low end of the method range that requires us to revert to performing the analysis utilizing a five or six point calibration curve.
  • DavidHPLC
    edited June 2019
    Single point as in single concentration of the standard, right? Shouldn't be bias on high/low end  covered within the validation process (50-150% I think)?
  • Yes, you are correct on both.

    We try to run a using a standard concentration at our specification level. During our accuracy evaluation we will spike and calculate recovery using the single standard and a full calibration curve. As long as accuracy requirements are satisfied and calculated recoveries using both single point and full curve agree within 2-3%, then we state the method is acceptable to use as a single point method.

    The flaw to this is sometimes we have to over a wide range. For example, I recently developed a monomer screening method for one of our products (medical device) that had to cover a range of 0.1 ppm - 10,000 ppm. Typically a single point prepared in the middle of the range will be very inaccurate at 0.1 ppm; if you perform the single point at 0.1 or 10,000 ppm, then the other extreme will be inaccurate.

    In the case of the above, not only was a 7 point curve required, I also needed to force a zero intercept because the 10,000 ppm point biased the 0.1 ppm point. Granted, this could be an argument that the method needs to be split into two ranges.

    Our end application was to characterize our cure process. Because we didn't know crosslinking rates of individual monomers, we couldn't really predict which cure samples would need to be tested with the lower or higher range method. In addition, we were testing hundreds of samples...

  • Aha. Thank you for the explanation.