Accuracy for Method Validation of fatty acids

[kristind858]

Hello,
Does anyone know what the acceptance criteria is for accuracy of fatty acids such DHA and EPA? I am using a combination of the AOCS Ce 1i-07 and AOCS Ce 1b-89.

I am spiking my placebo samples (soybean oil) with a known amount of standard (Eicosapentanoic methyl ester and/or Docosapentanoic methyl ester), then performing the same sample prep steps as I would a regular sample (base then acid catalysis to methylate triglycerides and any free fatty acids).

My calculated DHA and EPA amount is more than the known amount I spiked into the sample. I am getting around 22-24% difference from the true value for each of the 3 concentrations I created. Plotting Peak Area vs. Amount of Standard (mg/g), my R^2 = 0.9999. I have tried spiking the placebo sample after it has been weighed and also spiking the test tube before sample is weighed and drying with nitrogen (proved a little bit better but not by much. Percent difference from true value= 16.9%) . There is some systematic error occuring. I have created newly prepared standards and ISTD solutions as well.

[kristind858]

I think my issue is with the weighing of my standard. I am trying to do quantitative transfer by weighing the ampule, emptying the contents, washing with solvent, then weighing the empty ampule. I am using isooctane with BHT which may be leaving residual BHT and possibly not all of the isooctance is evaporating quick enough. Plus any broken off glass will lead to a larger standard weight. Hexane would be a better choice but this is not the solvent used in the method. Also by opening the ampule, it creates a greater surface area for wter to adsorb. I think the best way to weigh such a small sample would be direct weighing into the volumetric flask. I will try this

[cityanalyst -2]

I joined this forum today. I don't know weather you would read this??

DHA and EPA are PU fatty acids. They are highly unstable. I do not have your methods in my hand now.You should get theoretical Response(TRF) about 0.96. But in my experience you always get less response than TRF. That mean you get high recovery.Try to use theoretical value for calculation.