Peak symmetry vs spectrum shape

[greg07]

Hello,

I have little experience with 3D view and spectral analysis. I would like to ask for help in deciding whether the asymmetry/symmetry of the peak affects the shape of the spectrum. I inject salicylic acid (225 to 280 nm) and for the Waters instrument I get a symmetrical peak (1.2) at 270nm, the MAX at 238nm and the expected shape of the spectrum. After transferring the phase, column and sample to the Shimadzu (LC20) apparatus, the peak has an asymmetry of 1.7 (little back tailing), the MAX at 228nm and the wrong shape of the spectrum on the shorter wavelength side - starts well from the left, but then a downward jump occurs and futher good again. It seems that all elements of the method for Waters and Shimadzu are identical. Could this difference in asymmetry cause a bad spectrum?

[img][IMG]http://img4.imagetitan.com/img4/small/24/24_viber_image_2021-09-06_21-44-02x.jpg[/img][/img]

Best Regards
Greg

[lmh]

I am guessing wildly here: I think you have a more worrying thing going on. You shouldn't really get a change in peak shape on moving between two instruments, if your method is robust, and the column/solvents are the same.
My worry is that your salicylate peak may be being affected either by equilibration of the column, or by injection solvent. These are both things that are very dependent on system. For example, if your equilibration time has been optimised on the Waters system so that it is just long enough (but only just), it might not be adequate on the Shimadzu system, and then you are injecting a sample onto an inadequately-equilibrated column, and getting a bad peak shape. Equilibration time mess-ups happen when the delay volume between pump and column is different on the new system, or the time it takes to make an injection is different (as this contributes to the equilibration time). Or, another possibility is that the two systems differ in their approach to injection; some loop injectors, doing partial-loop injections, will fill the remainder of the loop with various things including needle wash solutions, which affect the strength of the injection solvent, and can cause peak shape problems, particularly to early peaks.
And the knock-on problem is that if your peak is eluting early, and you've now got some system change meaning that it's eluting as a badly-shaped peak at a slightly different time, then it may be eluting with different background UV absorbance.
And, of course, the spectrum you see depends on the spectrum on which the detector was zeroed, which might not be the same for the two systems.

[Consumer Products Guy]

I'm not sure exactly what you're doing, or if my comments are even relevant.

We always assayed salicylic acid at a higher wavelength (say your 270nm) using acid in the mobile phase and isocratic conditions, on RP18 column, and never had any issues.

[Multidimensional]

If I understand the information supplied, you have run the same sample on two completely different HPLC systems (using the same HPLC column and mobile phase) and observed differences in the peak spectral data. That would be 100% normal. One can not expect two different HPLC systems to provide the same results. Not only are all of the HPLC modules different, but the settings used are also different(esp the detailed detector settings for each DAD. The flow cell type and path-length, the flow path design (tubing layout, ID's etc) and perhaps even the injection solvent used are also different. Some or all of these things may contribute to differences in results (again, normal). If the goal is to transfer a method to a different HPLC system (goal of obtaining similar results), then the HPLC systems used will require very extensive detailed set-up and testing, to account for all of these differences. Moving column and liquid alone are insufficient to achieve this.

BTW: We usually perform quant work for that compound at around 300 nm (depending on exactly where the peak is for the choosen method as it will vary depending on the solution it is dissolved in AND the mobile phase run it). *There are multiple peaks which is useful for qualitative ID during analysis.

Detector settings used (DAD have too many to list here) can change the output, peak shape and spectra seen. We often say that detectors simply provide output, but the operator's training and experience are needed to setup the detector and method to provide useful output.

Differences in injector design, injection solution/volume etc alone can change the peak shape.
Differences in overall flow path, the same.
Differences in flow cell geometry also can account for differences in signal response and peak shape. These are all fundamental aspect of HPLC.

[vmu]

Reduce the sample amount either by reducing the concentration or by reducing the injection volume and see what will happen. The absorbance on the left side of the spectrum under question is too high. The signal overload disturbs the spectrum shape in the low-wavelength range.

[Multidimensional]

Good spotting 'vmu'. The abs scale is over 2 in that image !!!!! Yes, that is certainly yet another problem to add to the list. Way too high a concentration for evaluation, but this too can be grouped with all the other training issues which need to be addressed before using HPLC.

[lmh]

yes, well spotted vmu!. And the bit where it's 2 is at completely the wrong wavelength isn't it? You've got a hump around the right wavelength and a truly enormous OD at much too short a wavelength. Could it be that something else has coeluted? If you overshoot on the detector, you flatten the places that should have high OD, not raise unduly the places that shouldn't.