I am having a bit of a problem using our LC-MS (Finnigan LCQ- ion trap) to quantify analytes. For one clean peak that elutes from the LC-column, I get numerous m/z signals, for instance at the correct mass, at double the mass, mass+Na, double mass+Na, etc. Whilst using negative ionisation I can avoid the Na-adducts, but I still have problem with multimeres. I have tried increasing sheath- and auxiliary-gas flows (better ionisation, smaller droplets) but the problem is still (to some extent) there. Has anyone experienced the same problems (and maybe solved them)? If not, how do you go about quantifying in this case?

Hi,

This problem is usually caused by a analyte concentration that is much to high. Try a lower concentration, it usually helps.

The "annoymous" post is correct and I would try it but you will probably still see some multimeres. First, I would ask if you are certain these are multimeres, i.e. do the peak sizes vary as you vary the injection volume or concentration? Since they are double the expected m/z value its a good chance that they are dimers but before doing quantitation I would want to prove this to myself.

If you have a diode array detector, it is easy to use it to see if there are any gross impurities in the "single, clean peak".

Next I would try to vary the cone voltage, that is, create a different ionization environment. I would try both raising and lowering it but perhaps raising it slightly would be a better chance of breaking up the dimers. Of course, don't raise it too high or you will begin to see fragmentation through the process of "up front CAD (collisionally assisted dissociation). One of the great advantages of LC/MS over classical GC/MS is the ability to vary the cone voltage and thereby "fine tune" your spectrum.

Usually, but not always, Na adducts are a result of using glassware somewhere in the sample prep process. You may want to try to modify your method to use polymer-ware in place of glass. Eliminating Na adducts may be better than using negative ion mode because many compounds show inferior ionization efficiency in one mode vs. the other. For quantitation, in fact, you need to establish whether negative or positive ionization gives you the best quantitative range.

Finally, if all else fails, use the capability of your data system to sum the results from individual m/z values into a single intensity value and use this for quantitation. Obviously, before doing this though you must prove to yourself that the peak at each m/z value represents the analyte and only the analyte of interest.

Another thought occurred after I sent the last message. I assumed you are doing APCI since you were able to switch ionization modes and see a significant signal in both (+) and (-) modes. However, if you are doing ESI and you are able to see both positive and negative ions with good signal strength, this may indicate you have not optimized the mobile phase pH for your compound.

Clearly, this is a fairly complex molecule, such as a peptide, that has both acidic and basic groups on it. For ESI, if you are running positive ion, you want the mobile phase pH to be adjusted down so that all basic groups on the molecule are protonated and all acidic groups are neutral. This can be done in the mobile phase itself, or if this ruins the chromatography, then adjust the pH post-column but before the ESI interface by "teeing" in another line.

If it is a complex molecule you may now begin to see multiple-charge ions, creating peaks at 1/2 , 1/4 and so on of the expected m/z. For this, and a variety of other reasons, it is good practice to begin to refer to m/z (mass-to-charge ratio), not "mass" when speaking about mass spectra.

If you do see multiple charge ions, there are software products available which can deconvolute these peaks into a single peak at the expected m/z value for an M+H ion. Ask your LC/MS manufacturer about this type of software.

Fred Klink
STM

Thank you very much for the advice. I have increased the MeOH content of the mobile phase (now 80%, was 20) and have decreased the flow to 100 microliters/min (was 1 ml/min). Results were rather dramatic, though not yet optimal.
Tuning of cone voltage is done rather well by the LC-MS software. DAD shows peaks to be "clean" - at least as far as UV-impurities are concerned. Although the possibility of using the software to calculate peak areas of numerous m/z is possible, I would like to optimize toward receiving one signal. I will keep you posted as to progress.