Upper end of concentration range

[Peter Apps]

Hi WK

Is it the mls added / 100 mls or the v/v from the density that you use on your x-axis ?

Peter

[WK]

Hi Peter,
From the %v/v determined by tables I convert to %w/w both the ethanol and propan-1-ol for use in the calibration tab in the method.
I then convert samples back to %v/v that have been run in the sequence using factors and the sample density.
WK

[Peter Apps]

Maybe the inaccuracy is in the tables. What if you plot peak areas vs the volume % based on your own volumetric measurements ?.

Peter

[WK]

Hi Peter / Yama001
I ran some upper levels last night with the suspected outlier removed and the calibration is much
improved so I will repeat this over the weekend. The control samples run later were all satisfactory.
Thanks for all your help.
Its only by these discussions that I learn !
Regards
WK

[aldehyde]

Hi WK, sounds like you've already tried a lot. Have you attempted the larger volume injection yet? I am interested to see if that helps.. I have seen that the reproducibility of injections at 0.1 or 0.2 uL is a bit worse than at 1 uL. I don't know if it is an effect of the sampler or the syringe. Some people get along just fine with small volume injections (certainly helps keep things clean!) but it all depends on what level of precision you need.

I would investigate the effect of your injection volume if only to rule it out.

[WK]

Hi aldehyde,
I did that initially after Peter suggested it - but it didnt improve my problem which I am certain was the outlier.
It seems I get sufficiently good work with 0.1ul 200:1.
This weekend I have prepared some more levels/controls to spread the calibrations better in each of the 3 ranges.
I will report when I have processed the data.
Best Regards
WK

[mckrause]

Flatten out your concentration bias by calculating in log-log space. It will help your top end significantly (as well as your bottom end)