Dear Group,
I am studying the biodegradation of 2,4,6-trinitrotoluene. To detect the various intermediates, I use an isocratic mode 50/50 ACN/phosphate buffer 50 mM, pH 3,2. The column is a Nova-Pak C18 3,9*300 mm.
Unfortunately, various isomers are not well separated. That is why I focused on a gradient mode. I can get a really nice separation between the various pics if I use the following gradient (% ACN-phosphate is changing linearly) :
time %ACN %phosphate
0 30 70
2 30 70
7,5 50 50
8,5 50 50
12,7 70 30
13,7 70 30
18 90 10
20 30 70
30 30 70

The problem is that the RT's of each compound is changing from sample to sample (for example : 2,4,6-trinitrotoluene : from 16,8 to 13,2). I never observed such thing when I was using an isocratic mode. Do you have any idea why?
Thanks for your help.

It sounds like your column is not equilibrated. It can take several column volumes to equilibrate, particullary after a complicated gradient. Typically, 3 volumes should do it, but I have one column that takes 10 volumes. If you monitor the pressure, you can get a good idea of when the column is equilibrated. Also, there is the chance that phosphate is precipitating when you have a lot of ACN, which would cause problems. Furthermore, the gradient should be run through the column a couple times at the beginning of the day before analyzing samples--again, it's equilibration.

A good rule of thum is to equlibrate for 3X the system volume + 5X the column volume. The 10 minute equil you have programmed in is not enough for a 3.9X300 mm column.

At this point, you are looking at a long run. If you get good separation, you can decrease your column size and cut back on run time. (Of course, if you are in grad school, your professor might not be sympathetic enough to buy you another column.) Remember that doubling column length only increases resolution by the 2^1/2.

I had a similar problem and the culprit is a malfunctioning multichannel gradient valve. I'm not sure what brand of instrument you are using so I can't really comment on it.

If you try all the above suggestions and it's still doing the same thing then I think you may have a hardware problem.

Does your pressure remain the same?

If you are using a quaternary pump, I think gen is on the right track. 16.8 to 13.2 is a huge difference.

You can test the gradient valve by adding a tracer to the ACN channel and testing for reproducible gradients. 0.5% acetone with the detector set to 270 nm will work. Good luck

Following up on the idea that the multichannel gradient valve (proportioning valve) could be the culprit here, please be aware that on HP/Agilent equipment (and probably others) all solvent channels must be "wet". Let one dry out and problems may arise in proportioning. Whether you'd get the symptoms you describe, I can't say. I usually keep water in the unused channels, but I have no reason to recommend that over another solvent.

on Agilent multi-channel valve and pump: after some problems a few years ago, we were told to purge ALL four at 5.0 ml/min even if only two channels are to be used. It's problem avoidance, just make and save a "Method" to do this. And on a 1050 unit, purge "ALL" doesn't really mean all, it just means all of the channels which are not 0 %.