Chromatography Forum

Hello,

I'm looking for some help from somebody far smarter than I am.

I am doing suppressed anion chromatography on a Metrohm 940. On the anion side, no matter what calibration range I do (small range or large range with orders of magnitude) my y-intercept is always negative.

This has been the case for 3 separately prepared 6-point calibrations with good precision and 96-103% accuracy.

My anion eluent is 3.2mmol sodium carbonate and 1mmol sodium bicarbonate. Suppressor is regenerated by 150mmol sulfuric acid and 100mmol oxalic acid.

My most recent calibration curve had the following y-intercepts for a calibration range of 0.5 - 6ppm.
Linear model: -0.0164440
Quadratic model: -0.0201090
Neither are forced through zero and have no weighting applied. I tried all weighting options and the y-intercept remains negative. Forcing through zero compromises result accuracy.

When analyzing cations (un-suppressed), this negative y-intercept is not observed and so I don't think it's a product of poorly made standards.

Could anybody offer some help as to why this is happening and some guidance on whether this is acceptable or not?

Thank you.

Dear Pear

Suppressed anion calibration curves are typically slightly non-linear. The typically recommended qaudratic fit is not perfect but better than linear.
The negative offset is quite typically and comes from the unlinearity.
What we recommend is to use calibrations only within the applied concentration range.

I hope that this helps.

Dear Pear

Suppressed anion calibration curves are typically slightly non-linear. The typically recommended qaudratic fit is not perfect but better than linear.
The negative offset is quite typically and comes from the unlinearity.
What we recommend is to use calibrations only within the applied concentration range.

I hope that this helps.

Hi Markus,

Thank you so much for your reply. I tried a Quadratic fit, but the y-intercept was still negative; even with all possible calibration weighting options and across the calibration range, accuracy was almost identical as linear.

It has to do with how the detector works and the trace conductivity of the eluent. You have a slightly positive baseline, but the conductivity of the of the peaks is not completely additive with that of the baseline so it shifts to a negative intercept.

Think about it this way, if you had a perfectly linear curve with perfect zero intercept, then you shift the baseline up slightly (which raises the zero point of the vertical axis) then you perfect zero intercept becomes slightly negative, by the amount you shift the baseline positive.