I have been runing a method that has two peaks, isomers, on a C1 column. It is an isocratic method so to avoid errors attributed to the proportioning valves the mobile phase has been premixed. The RT initial for the fist peak is around 19 min where later in the run it will be at 9 min. The mobile phase is 62% buffer and 38% MeOH.
Also, after the column is cleaned and the machine is ran again the RT is correct. But, several hours latter the RT time will be shorteneed

What is your sample like is a clean sample? It sounds like something in your sample is sticking on the column.

The retention sites on your column are retaining something in your sample strongly. This keeps the retention sites from the peak of interest, hence decreasing retention times.

The first thing I would try is to clean up your sample prep, assuming you can. If not then you may have to clean the column between injections. This will increase your run times considerably.

The sample is a bulk drug substance, therefore it is as clean as I can get it. With 38% Meoh going through it I had hoped it would be able to keep it clean. We had thought that tolune, the sample contains a small amount of it, might be hanging up in the column. I had injected a sample with a large amount of tolune and then watched the preceding chromatograms. If anything the RT was a little latter. The run time is already 45 mins and the samples do not have that great of stability, so I would hate to make it even longer by having a clean out stage.
But the big question is what could be retaining in the column that would block the retention sites?
Also, I would take a while to determine if the cleaning were doing anything to it since it takes around 10 hours for it to start shifitng in.

Does the retention time decrease gradually (e.g., you lose a few seconds each run) or does it happen abruptly (i.e., retention stays constant for a long period of time and then suddenly shifts)?

It keeps shifting in gradually. It will go in a little and then come back at then go in a little more. I would say no more than a 30s shift of main components on a single injection.

I had a similar problem, and I checked my favorite reference book, "Practical HPLC Method Development," 2nd ed., Snyder et al. (p. 217) They state that this can be caused by poorly buffered mobile phase - either the wrong buffer, too low buffer concentration, or a pH out of the buffer's range. I increased my buffer concentration, and the retention time stabilized. Their suggestion is to start no lower than 20 mM. Hope this helps.

Thank you for the information. I belive it is the proporting valve opening. I ran the samples on another machine and the retention time appears to be conisistant. Our theory is one of the organic resevoirs leaked some into the pump. Since we were runining isocratic. We are going to do some more work on it. We are also, having what appears to be stability problems. The area counts for these runs are decreasing rapidly but no impurity is increasing.

Have you tried using a "premixed" mobile phase instead of on-line mixing?

As I had mentioned in the first paragraph, I did use a premix mobile phase to avoid the on-line mixing problem. But, there appears to be a leak on the proporting valve of one of the organics in the other resevoir.

My apologies; I didn't read all the way back to the beginning of the thread. :(

However, you can still use a premix to confirm your proportioning valve hypothesis: just put the premixed mobile phase in *both* reservoirs and see if the problem goes away.

A premix was placed on another instrument and the other reservoirs had water, since it was easier to make than more buffer, plus an excess of water would not effect the RT as much as an organtic. Thank your for your assistance.