Bracketing Standards?!

[seamoro]

We are currently running a 6 point calibration curve, we then run 20 samples and have a bracketing standard after the 20 samples. We include Internal standard in in the calibration, samples and bracketing standard. When we include Internal standard in this manner we have no problem meeting our acceptance criteria of bracketing standard -/+20% response. But when we take the internal standard out when get bad results for a bracketing standard e.g -/+60% response. So what can we do? we would like to remove Internal standard as it interferes with highly positive samples.

[Don_Hilton]

I assume the interrerence with positive compunds is partial coelution of the analyte with the internal standard? And you, do not have this coelution in the calibration curve? Without seeing the chromatograms, two further questions come to mind. Does this positive sample have such a huge peak for the analyte that it is outside the range of the calibration curve? Or, is there a matrix interference that is beign interpreted as the tail of the analyte peak - or is changing the chromatography?

It sounds like there is a robustness issue and a bit mroe examination of the method might be requried.

[rb6banjo]

Do you have the flexibility to choose a different internal standard? I would choose a different internal standard - one that elutes away from everything of interest.

[seamoro]

The analysis is total recoverable hydrocarbons so for a positive sample it is not individual peaks but a large broad band of hydrocarbons. so it is impossible not have the internal standard affected by co-elution. In relation to choosing another Internal Standard....yes we thought of that but it is not possible with TRH analysis for the same reason i stated before. In fact really the use of internal standard is completely wrong but we must use it as its the only way of getting our bracket standards to pass. At this moment in time we are adding internal standard to our cali, samples and bracketing standards but when we have a positive we process our cali without internal standard and then use this to quant or sample (we also subtract the the internal standard from the sample). It is not an ideal situation!

[rb6banjo]

What detector do you use for this analysis? Can you show us a chromatogram with your internal standard labeled?

[CE Instruments]

Why have you chosen to bracket the standards ? This is very unusual in GC as in general the detectors do not show significant variation over a say a 20 sample run. Bracketing normally comes in with LC and especially where Pharma is concerned. Here it tends to be because you do get a reduction in recovery with some detectors and they need RSDs in the +/- 1% level not your 20% I am puzzled that your standards could vary 20% over such a short analysis run. Bracketing assumes any change in response occurs as a linear change and gives averaged results; are you certain this is correct for what you need ?
Internal standard is recommended to help where there are issues with either sample preparation or with sample introduction. You MUST get an accurate clean peak for your internal standard otherwise your results will be thrown out by variation caused by inaccurate integration of that peak ! It does not seem appropriate for an analysis that quantifies any positive result to give a total Carbon answer that must mean summing all peaks.

A more normal analysis would be external standard and simple area %, running a post run standard as a check to be certain drift has not occurred. What data system are you using , why does it not support bracketed standards without internal standard ?