Require suggestion regarding solvent selection

[shehzad]

I am using Acclaim surfactant column with ELSD detector. Mobile phase using is acetonitrile and 0.1M amonium accetate ( 85:15 gradient). Standards (i.e. C12-13 7PO SO4) is diluted with ACN and got a good straight line. But my samples contain NaCl and is giving phase separation when diluted with ACN and has given unexpected very low peak area.
Is it necessary to use solvent that is well soluble in mobile phase? If yes than please suggest any other solvent for my case.
Can this insolubility of sample in solvent i.e. ACN be the reason of low peak area?
My sample is 2-3 month old. Can this be the reason of low detection?

[Vlad Orlovsky]

Acclaim column has basic groups on the surface, if you mobile phase is not strong enough to elute cloride ion you will be accumulating it on the column, it will eventually come out, but you can incorporate a wash after each run

[shehzad]

Acclaim column has basic groups on the surface, if you mobile phase is not strong enough to elute cloride ion you will be accumulating it on the column, it will eventually come out, but you can incorporate a wash after each run

Can this be a reason of low area under peak of anionic surfactant?

[James_Ball]

But my samples contain NaCl and is giving phase separation when diluted with ACN and has given unexpected very low peak area.

Try diluting with a mixture of ACN/H2O, maybe even using your mobile phase to dilute the samples. The added water may help keep everything in solution. Also if you can make your standards and samples in the mobile phase you will get much better chromatography of early eluting compounds.

[shehzad]

Actually this time i run after dilution with mobile phase. I am amazed to see that by taking solvent as mobile phase (i.e. 85% ACN and 15% 0.1M ammonium acetate) peaks are coming at retention time of 10 min, where as previously by taking only ACN as solvent it was coming at 30 min retention time. What could be the possible reason? Does it happen normally?

Second thing now the straight line is not passing through origin. Any suggession plz

[tom jupille]

What could be the possible reason? Does it happen normally?

Vlad answered that question a couple of posts back. Part of your retention is ion-exchange. The ammonium acetate provides competing ions to elute your compounds.

Second thing now the straight line is not passing through origin. Any suggession plz

The ELSD is notorious for being non-linear. Try a quadratic fit and see if that doesn't work better.

[shehzad]

I got a perfect straight line but this straight line doesn't pass through origin. Can i still use it?

[tom jupille]

I got a perfect straight line

There is no such thing as a perfect straight line in nature. A straight line may look like a reasonable fit, but a quadratic may be better. You have to look at the residuals and look at the p-value to determine that. Check out these articles by coleman & Vanetta:
http://www.americanlaboratory.com/913-T ... agnostics/
http://www.americanlaboratory.com/913-T ... continued/
http://www.americanlaboratory.com/913-T ... concluded/

but this straight line doesn't pass through origin. Can i still use it?

That depends on lots of things:
- how far off are you? If the origin is within the prediction interval, then technically, your line *does* pass through the origin.
- how likely are you to get samples that are down near LOQ? If all of your samples are in the middle of the linear range, and you always run bracketing standards, then whether the calibration passes through the origin is relatively unimportant.
- if you analyzing over a wide range of concentrations (e.g. several orders of magnitude), then consider doing a weighted least-squares fit (or using a log-log fit) to compensate for the heteroscedasticity of chromatographic data.
http://en.wikipedia.org/wiki/Heteroscedasticity

[shehzad]

Thank you very much for knowledge sharing and help.
I prepared standards by myself after dilution and couldn't find pure standard.
My samples concentration is in between my tested standard concentration i.e 30ppm to 1000ppm.

[shas904]

Straight line doesn't pass through origin and intersect x-axis at 225 ppm. R square value is 0.9986. I am using ELSD detector Base line was at zero and blank run was also at zero. I am using bracketing standards i.e. my sample concentration is in between my standard concentration. I prepared my standards by myself after dilution of sample and couldn't find pure standards. Retention time is 8 min. What can be possible reason? and is this acceptable?



- how far off are you? If the origin is within the prediction interval, then technically, your line *does* pass through the origin.
- how likely are you to get samples that are down near LOQ? If all of your samples are in the middle of the linear range, and you always run bracketing standards, then whether the calibration passes through the origin is relatively unimportant.
- if you analyzing over a wide range of concentrations (e.g. several orders of magnitude), then consider doing a weighted least-squares fit (or using a log-log fit) to compensate for the heteroscedasticity of chromatographic data.
http://en.wikipedia.org/wiki/Heteroscedasticity[/quote]

[tom jupille]

If you ran standards down to 30 ppm, I'm not sure how you could get a y-intercept of 225 ppm if the response is truly linear.. I'll repeat what I said earlier: ELSD response is notoriously non-linear.

R^2 values by themselves tell you little about the linearity. You have to look at the distribution of the residuals and then run a p-test for the linear fit versus a higher-order (quadratic) fit. Read those sections from the Coleman and Vanetta series. They explain it a lot better than I can.