Chromatography Forum

Hello again,

I'm working with an analysis in Agilent 1200, gradient elution, Buffer/ACN , C18 column, pretty straightforward. I'm intersted in 3 relative small peaks (area about 30) , very close together (resolution about 2) eluting before the gradient starts (about 2.5 min).
The problem is that the area of the 3rd peak is different in every other injection and the %RSD may be 5 ,8, even 12%.
Regarding the other 2 peaks the %RSD is below 2%.

Any suggestions?

Thanks

Use the UV peak maximum. Increase injection volume. Are you using the right detector (MS, ELSD, CAD, FLD....)?

Not enough information to provide a specific answer (need detailed method conditions, column dimensions, sample area amt, LOD...).

Could be as simple as too low of a K prime for the samples (little or not enough retention) or perhaps poor quality integration parameters (which can easily throw off measurements of very small peaks). The most likely causes are the method itself. Here is an article with some troubleshooting info and causes when results vary. Link: https://hplctips.blogspot.com/2015/11/h ... hange.html

Hi jcwagner,

Yes, the more you tell the Forum, the better the help other members may provide.

Data that is helpful to include: the HPLC column type and dimensions, the flow rate used (so retention factors may be estimated considering the retention time(s) of the peak(s) involved--or you could note retention factor values in the post), the eluent composition, the sample/standard injection volume and the temperature at which the separation is being carried out.

In this case, it's hard to know if the sample prep steps, the autosampler (2% RSD may be high, hard for us know without further detail regarding the goal(s) of the method), the pump/mixer (if used) or as noted in the previous post, poor detector settings or poor integration settings.

Happy to help further with some additional info to consider.

The thing is that i have 3 peaks, R.T. about 2, 2.2, 2.4min , the %rsd for the first 2 peaks is 2% and for the third is 5-12%.
I don't think that column dimensions and type, flow rate, LOD and temperature have something to do with that. K is about 5.
I have run this analysis several times in the past. Always the first 2 peaks have better %RSD and for the third the RSD is from 1.5 to 3 , not 10%.
Only one of the 3 peaks is affected so the problem doesn't seem to be the "hardware".

Hi jcwagner,

I'll respectfully have to disagree, the column dimensions and flow rate allow insight into the retention factors of the peak(s) that are generated. This way we can tell if the separation is appropriate, with retention factors all > 2 for the peaks in question.

You note that the retention factor of one of these peaks is five, so I'll assume that these peaks are Really crowded together as they are all resolved by 2 apiece.

Okay, if the detector settings and integration are appropriate, then variable peak area(s) are most likely due to either autosampler trouble or pump trouble--these will readily affect peak area precision.

Hi jcwagner,

One other thing to think about is the solvent in which the standards/samples are dissolved as compared to the eluent composition at the start of the elution program. It's possible that, if the standard/sample/solvent is of stronger elution strength than the initial eluent composition, strange peak behavior could result.

3 peaks, the first 2 with %RSD <2 , the last with %RSD 10 (6 injections).
Any other ideas?

Hi jcwagner,

Other than the autosampler (variation of injection volume, perhaps a need to perform "priming injections" of standard solutions that are more concentrated than the typical standard solutions you normally use, health of the needle and injection valve) and the pump (flow rate variability can cause peak area variability--think of the pump seals, check valves and in-line/valve filters), I'm not sure, other than the software and detector settings--these last two are thought to be okay as you've used them before with this method.

Section 13. of this reference may be helpful:

http://www.waters.com/webassets/cms/lib ... a20769.pdf

Please, these injections you're making, are they of standard and sample solutions both, or standard solutions only (n = 6) from the same vial (perhaps sorption to the vial?)? Sometimes also I have found that for trace analyses, the degree of imprecision with peak height is less than that of peak area--is it possible to alter the calibration of the method you're using if this can be demonstrated?

3 peaks, the first 2 with %RSD <2 , the last with %RSD 10 (6 injections).
Any other ideas?

What compounds are they ?

Peter

If the first two peaks are fine considering RSD, this suggests that the autosampler is fine, as already has been pointed out.

For serious troubleshooting, it would be really helpful to have more information (as already requested). Column dimensions, gradient program, nature of the analytes, mobile phases, flow-rate, sample solvent,... And don't forget: a pictures is worth a thousand words. A sample chromatogram can provide very useful insights regarding peak shapes, actual resolution and other things.

Without more information, we can only guess. I'm sure we can collect a hundred possible reasons in this forum, but these probably won't help you...

Hi JCWagner,

Were you able to get the difficulty with that one analyte sorted out?