Mass Spec impurities from DMSO solvent

[tadah1]

Hi,

Our lab has recently acquired a MSQ mass spec (single quad) from Dionex (now ThermoFisher) as part of a preparative HPLC-CAD-MS system. Due to solubility issues most compounds are dissolved in DMSO and injected straight onto the column without any at-column dilution step. We scan a range from 100-1000 Da in the TIC and are having major problems with DMSO and DMSO adducts in that mass range. Specifically, with the m/z=120, and m/z= 157 peaks corresponding to [DMSO+acn+H]+ and [2*DMSO+H]+. The intensity of these peaks drops of as the DMSO elutes off the column, but the baseline stays significantly elevated due to these peaks even at 15 minutes after the DMSO elution. As a consequence, the analyte peaks appear very weak in the TIC. This is only happening with the MSQ mass spec from ThermoFisher. The mass specs from other manufacturers show the DMSO peaks but at much lower intensities. I think this is effect is due to the geometry of the ion source chamber (since the probe is fixed at a certain angle to the cone) but I am not sure. We use generic gradient runs from 1-99% acetonitrile on C18 columns at flow rates between 35-50 mL/min. Mobile phase = water, acetonitrile (TFA). And no, we cannot get away from DMSO, DMF or NMP as our sample solvents. Any suggestions or insight is appreciated. Thank you.

[Arne]

Hi,

We have used DMSO before when doing high throughput screening of drugs and encountered the same ions from DMSO. We now use 5% DMSO in one of our mobile phases for peptide analysis to improve signal.

Do you have a targeted assay? You could monitor just your target compounds rather than scanning. Is it a problem to use XICs instead of the TIC chromatogram? Don't you use that anyway for quantitation? Could you scan 160 to 1000 instead?

I advise against using TFA in your solvents when doing MS detection. That depresses overall signal. Use formic acid instead.

[JMB]

Your extreme HPLC gradient, together with the very wide scan range, suggests that you are trying to isolate low MW impurities from a high MW final product. Is this correct ?

Are you using LC-MS to guide fraction collection because of the lack of a suitable UV wavelength or UV detector ?

If you MUST scan over the m/z 120 and 157 signals, then you might try to decluster these ions. I am not familiar with your MS instrument and your source geometry is fixed, but do you have the option of varying cone or repeller voltages or temp. of interface or drying gasses etc ?

If you cannot decluster then, per Arne's suggestion, set up an MS scan program that monitors the appropriate [M+H]+ ions in consecutive time segments.

Any further info. you could provide, within the limits of confidentiality, such as compound classes, approx. MW etc would be helpful in considering other approaches.