Ghost peak in HPLC

[Technician]

Hi,
I am new in chromatography. I have to fit a problem with a HP 1050 system. The user has a ghost peak between 1.24 and 1.28 minutes with blank injection. Eluents are water (+0.05%TFA) and acetonitril (+0.05TFA), flow of 5ml/min. The column is a Chromolith Speed ROD RP-18e 50-4.6mm. Gradient changes from 90% - 10%Water during the method. We flushed the system with acetonitril, changed the precolumn an column, I cleaned the stator and rotor of the Rheodyne switch valve with acetonitril in ultrashall (same with the needle seat) and I wiped the needle witch acetonitril. During and after these actions I made blank injections, nearly always the same result. Of course I read the other topics describing similar problems. The last thing I know to do is changing the millipore water and acetonitril. Have you any other advices to give me? Thanks in advance!

[Technician]

I forgot: we use a 1050 VWD with 214nm wavelength.

[tom jupille]

First of all, try running a "dummy" gradient (do not activate the injection valve, just run the gradient). If the ghost peak remains, then it is not coming from the injector.

Next, run a series of dummy gradients with differing equilibration times between the gradients. If the height of the ghost peak increases with increasing equilibration times, then it is coming from contamination in the "A" solvent of the mobile phase. Sources could be the water (as you have speculated), any buffers added to the aqueous phase, a pH electrrode used during mobile phase preparation, residual detergent on glassware, etc.

[Consumer Products Guy]

Adding to what Tom said (great idea to run gradient with "no injection" to immobilize the autosampler, just leave cell for vial number blank): why even worry about a ghost peak at that retention time? If an analyte elutes that early, you need to change the chromatography, you must retain it better (longer). If the analyte is well-resolved from 1.28 minutes, why pursue?

[Technician]

Many thanks for your answers. I will first try the "dummy" gradient without injection.

If an analyte elutes that early, you need to change the chromatography, you must retain it better (longer). If the analyte is well-resolved from 1.28 minutes, why pursue?

Unfortunately I'm not sure what you're meaning. The method the laboratory uses since several months takes 2.2min. The ghost peak appears always at about 1.26min (1.24 - 1.28 ). In their chromatogram it disturbs as they want to see only the expected substance.

[tom jupille]

For "Consumer Products Guy": I just sat in on a Bioanalytical LC-MS class taught by John Dolan. One interesting aspect is that apparently "ballistic" gradients run on very small columns turn out to be chromatographically equivalent to the longer gradients a lot of us are familiar with, so that short retention times are not necessarily indicative of problems as I would have suspected.

Also, thanks for the tip on leaving the vial number blank in order to force a dummy injection!

[DR]

In the event that the ghost peak is proportional th the equilibration time and the water is the prime suspect, you can clean it up by prefiltering it through a Varian (via 3M) Empore Extraction disk, or, as long as you're using a high pressure mixing system, you can stick a small C18 column between the A pump outlet and the mixing chamber (as long as the run isn't long enough to allow breakthrough fo the junk from the water).
You also have to worry about breakthrough with the extraction disk - just move it to another flask every other liter and give it a little MeOH flush and water chaser, then resume filtering the water for the A phase.

[Technician]

After changing the eluents, cleaning the bottles, and not having better results, the user changed the precolumn and column (we did this already before), but this time also the mounting plate of the precolumn. And in fact, the contamination has been in this mounting plate. Now we have good result without ghost peaks. Thanks for your help!

[tom jupille]

Wow! That's one I wouldn't have suspected.

Just goes to show that you can't afford to overlook anything!

And thanks for posting the outcome so we can all learn from it.