by
lmh » Mon Apr 20, 2015 1:24 pm
yes, the response to a deuterated standard does depend on the summed concentration of the deuterated standard itself and the unlabelled compound (and also anything else that coelutes).
You can see this because if you create a standard curve using variable amounts of unlabelled compound, and constant amount of deuterated, you will see that the response for the deuterated will drop off. Some software will automatically show you a calibration curve for the internal standard (it will be an external calibration because the internal standard doesn't have its own internal standard...!!), and this will make it very clear. It will be a single-point calibration, and unless the range of your unlabelled calibration curve is very narrow, the single-point calibration will look awful, with very bad replicate points.
But it doesn't matter. In fact this is why we added the internal standard. Already with external standard calibration it didn't matter provided you fitted the calibration curve properly. The assumption with external calibration and a nice curvy curve-fit is that if ionisation efficiency is lower at 10uM than 7uM, this is already accounted for in the curve's shape, and we assume that the effect is always the same, whether it's a standard at 10uM or a sample at 10uM. Internal calibration goes a step further and acknowledges that 10uM in a sample might mean 10uM + 5uM rubbish, and be more severely affected than 10uM pure sample. So we look at it relative to 2uM (for example) internal standard. This means that if there is extra junk in the sample that means we measure the 10uM sample with only half the efficiency we managed in the standard, we also see only half what we'd expect of the internal standard - the ratio remains the same, and the result gives the correct answer, even though all the peak areas involved are less than in the standards.
This is also why internal standard calibrations often give much better linearity than external standard. The external standard curves off because of the varying ionisation efficiency. The internal doesn't because when the external calibration peak area of unlabelled analyte has dropped to only 70% of what you expected, the internal standard peak area has also dropped to only 70%. There is a direct link between the scatter of those internal standard points in my previous paragraph, and the improvement in linearity!
