Chromatography Forum

My C18 column has pH limits of 2-8. My mobile phase is 10% acetic acid, 1% phosphoric acid, 5% acetonitrile in water. The aqueous portion of the mobile phase has a pH of about 1.63. Solvent B is 100% AN.

I am using a gradient method on a new C18 column that goes 100% A to 40% A. My first sample run was really good, but subsequent runs had an unresolved peak coming off the column near the dead volume and nothing else. After flushing the column with high percentage AN, I was able to reproduce the first chromatogram, but the same thing happened again with subsequent runs.

I am trying to figure out what is happening. I'm trying to decide if there is an issue with a) the autosampler, b) highly retained sample components that affect subsequent runs, c) loss of bonded phase due to low pH, or d) something else. Total number of runs at low pH has been 10 or fewer. I don't know how to diagnose loss of bonded phase, and if it would happen so fast. Any input is welcome!

If flushing with ACN restored performance, that suggests that loss of bonded phase is not the problem (that would be a "one-way" trip).

Yes, it does seem like loss of bonded phase isn't the primary problem. But I am also interested in finding out more about how that would manifest itself.

It sounds like your equilibration time may not be adequate

A strange unretained or nearly unretained peak is not a sign of a C18 function coming off. However, if the manufacturer recommends the use of a pH larger than 2, I would heed that recommendation and adjust the pH accordingly. The recommendation is likely to be based on data, and it is in agreement with my expectation for the stability limit of a monofunctionally bonded C18.

Please note that there is not a permanent stability of the packing at pH 2.1 nor an instantaneous instability at pH 1.9! You may get away with running the column at pH 1.6 for a while. But don't cry, if you will not be able to reproduce next week what you did today...

the peak could be coming from the previous injection.
best thing to do is to extend the first run you do. since the peak in question comes out at the beginning of the next injection, 10 minutes more in the run time should do it.