Dear all members,

I am a newcomer for LCMSMS. Recently I analyse a drug using LCMSMS system. Initially I have no problem in doing it at 10ppb level. When I try to dilute my standard solution further and then analyse, I find that the peak shape becomes irregular (not Gaussian any more). The chromatographic peak becomes split when the conc. goes down to 0.5 ppb. Is it normal in LCMSMS analysis? How can I remedy it so that a sharp and nice peak shape can be obtained at low conc.? The drug that I am analysing is a nitroimidazole. Thanks all in advance.

Is your analyte prepared in mobile phase? If you have a UV detector, what do the peaks there look like. It also could be that something in the source is interfering with ionization. Are the non-gaussian peaks always the same shape, or do they vary from injection to injection? If they are always the same shape, it is an LC problem. If they vary from injection to injection, it is problably an MS problem. Try stronger ionization conditions (vary temp, etc). Also, make sure that your source is clean. A lot of problems can be caused by a dirty spray chamber. You don't say what kind of instrument you are using, but the sample introduction needle or capillary must be free of blockage, in addition to the surfaces of the source being clean.

Hope this helps.

Michele,

Thanks for your advice. Actually we are using ESI+ mode of API4000. The problem did not appear when conc. of standards were high enough, say 5-10 ppb, but did happen when it was down to trace level such as 0.5ppb. The "pattern" of the peaks were not consistent, probably what you say was the dirtiness of ion source. I'd try to clean it and see if it would improve.
Anyway, thanks a lot.

At low signal levels, LC/MS 'peaks' can become very noisy. Unlike the UV detector you have probably been using, there may be a significant (on the microsecond time scale) time difference between scans, causing this jaggedness. On our Finnegan Deca system, this display problem is controlled by changing the 'smoothing' factor, which instructs the computer to blend a number of data points together into a smoother curve. I typically use 7-11 points, but have experimented with higher numbers.

Perhaps your system has a similar parameter. You are pushing the edge of the envelope! Can you get the system to generate more data points/sec by perhaps reducing scan ranges, number of events/segment, or other means? This might also help.

The Sciex instruments do have a smoothing function, however, this can affect the accuracy of quantitation. Rather than apply a mathematical function after acquisition, I would try acquiring more data points over the width of the peak. Also, the Sciex 4000 is capable of easily seeing most compounds well below 0.5 ppb. I have used the 3000 and have seen peaks as low as 0.02 ppt. I hadn't, however, realized how small of a compound you are working with, Umberto. For analytes with molecular weights that low, APCI is often a better choice.