HPLC-UV, low signal bias

[Keaka]

I've got a RP-HPLC/UV method for acetate determination running calibration standards and then followed by sample with my compound of interest along with the acetate counter ions that I am interested in.

Now, my samples are consistently 5 to 15% below theoretical acetate content. If you consider typical assay variation of +/- 15%, especially for GLP work..that actually aligns with the theoretical value pretty good.

The calibration standards are tight.. withiin +/- 5% deviation.

Could it be my compound is quenching fluorescence? How would I find out?

[HW Mueller]

What fluoresces?

[Keaka]

I'm trying to quantitate acetate counter ion.

4 Acetate standards

Samples contain my peptide + acetate counter ion

[danko]

But the question was: What fluoresces?

For the record – acetate does not fluoresce.

Which wavelength have you chosen, by the way?

Best Regards

[Keaka]

Sorry, completely wrong/bad choice of words for my original question there, was thinking of something else.

I'm not talking about fluorescence of acetate, but the detector signal. My calibration is consistently very linear and the detected signal from my peptide samples are somewhere in the middle of the curve.

But as I said the calculated results always have a low bias. As far as counter ions go its possible something else snuck in there at very low percentages but I'd like to eliminate the method as a source of the low bias.

[peptidemetdev]

Versus theoretical, our counter-ion results are frequently biased low. However, we see the same thing with GC-FID results from an outside testing agency. I haven't ever bothered looking into how the theoretical value is calculated (I'd presume it has to do with how many basic groups there are, as well as the concentration of acetic acid in the final solution before lyophilization).

The best way to determine if the bias is due to your methodology or due to your sample would be to find a secondary standard and see how much acetate you actually recover. I see this a lot in reports we get for similar sorts of tests from agencies we use. They use one reagent from one vendor to prepare their curve, and then inject another reagent from another vendor to test the curve (looking for something like 95-105% recovery of the secondary standard). We don't bother to do that here, but I've always thought it was a good practice.

[HW Mueller]

Keaka, do you mean that your samples may be saturating the detector absorbance (blackout....)? In that case the samples would have to have a higher baseline absorption than your standard, since that was linear.
If you havn´t done it you will have to do a careful recovery study with different spike amounts.