UOP-389 Calibration Standards

I am currently reviewing the UOP-389 methodology for trace metals in organics by ICP-OES, and the method seems to only use one blank and one standard at 10 mg/kg while promising detection limits between 0.01 and 0.11 mg/kg. The analytes are segregated by minimizing potential interferences between analytes of interest. I am worried that only 1 point of nonzero quantification may not be sufficient to reliably produce results with such low MDLs. 

Has anyone had experience with this method? Would I be wrong to assume that I should make more standards with this methodology to have a more robust calibration? There isn't much literature around this method's use, but it seems to be accepted in my field (renewable fuels) as reliable. Other labs that perform this methodology also seem to minimize the nonzero calibration points, only using 1 or 2 at most.

Any insight into this would be appreciated!

  • I can see your concern for using only a 1 point calibration curve. The working range of a curve should be established by the MDL on the low end, and the Linear dynamic range (LDR) on the high end. Each wavelength for a given element will have it's own characteristic MDL and LDR. For the wavelength you are using, you should measure the MDL on your instrument at your analytical conditions. Once you have determined the MDL, you could make up a low end calibration standard at the MDL level and include it on your curve. The other option would be to calibrate using the 1 point method, and then prepare initial verification control standards at known concentrations in order to check the validity of the curve. I would recommend using weighted fit for your curve. 

  • HI, Tina

    What is MDL?

    How to use weighted fit, and what it means?

    Thank you

  • MDL is the method detection limit.

    On the ICPOES software standards tab, there is an option to check "weighted fit" for the calibration curve. Statistically for a general calibration curve without weighted fit, more weight is given to the highest calibration standard on a calibration curve. But with weighted fit, it applies more weight to the lower calibration point(s). This means you have good accuracy on the lower end of the curve and improved accuracy on low concentrations. 

  • Thank you for quick response. I understand now. I am doing 8900, always ignore weighted fit, and low standard always has high %bias to their preparation targets. I will try weighted fit  next time. 

    Thank you

Was this helpful?