Chromatography Forum

Hi

I am doing LC-MS/MS and have been having problems with carryover, I have been analysing nicotine and diclofenac (+their respective metabs) and have been getting carryover with them both and also their metabs, sometimes up to 10%. I have completely changed the autosampler and still have the problems (not the rotor valve hopefully). I have tried alternative needle flush solutions with multiple flushes and this has no real effect. I have also tried to wash everything off the column with a long wash (95% organic) however this also has no effect. I have run out of ideas can anyone help?

Mike_D

If you do repeated blank injections, do the peak areas/heights decrease with successive injections, or stay roughly constant.

If the former, it looks like true carryover (you may have to tear apart the injector after all ). If the latter, consider the possibility of environmental contamination (i.e. sample "dust" floating around in your lab ).

I have had problems with carryover, as Tom suggested try shooting several blanks after a real sample and see if response falls off.

If so, could try another wash solvent. I like on the people at Waters suggested, 1:1:1:1 water:acetonitrile:methanol:isopropanol containing 1% formic acid.

I once had a problem that I had somehow contaminated my mobile phases. If that is the case, the longer you hold at the initial gradient conditions, the bigger the response when you make an injection..

see the following LCGC article..

http://www.lcgcmag.com/lcgc/data/articl ... rticle.pdf

http://www.forumsci.co.il/HPLC/Carryover_problems.pdf

good luck, aggravating!

I'm just a novice LC-MS user, but could it be interference coming from your source? Do you still get the carry over when analysing by UV-DAD (presuming you are?).

Hi (thanks for all the responses)

To answer the first reply I have done repeated blank injections and the carryover trails off over 3 injections, so it must be my injector.

I have given 1:1:1 water:methanol:acetonitrile a go previously and this did not reduce the carryover, I will give the suggested flush a go.

Im not running UV-DAD, normally I work with 14-C so have got a radio detector and a UV however this work does not use a radiolabel so I am working a bit blind (for me anyway).

Many thanks

Mike_D

Another thing to check is to run a non-injection gradient after a high standard. If you still see it, then it's coming from your column or some other permanently swept part of your flow path. Alternately, since you are doing LC-MS/MS, you could run some flow-injections to either rule out your column or autosampler. If your LC method is isocratic, then never mind.