Retention time increase with pressure increase

[skunked_once]

I am analyzing some natural compounds extracted from sunflower florets using a C-18 column with C-18 guard column. The mobile phase (isocratic) is 75% water (pump A) and 25% acetonitrile (pump B) at a flow rate of 1 mL/min at 40 C. At the beginning of the runs, the pressure is about 1300 psi. After several runs, the pressure creeps up as high as 3900 psi. The last peak in my sample elutes at 52 min at 1300 psi but shifts to 57 minutes at 3900 psi. All of the peaks in the sample shift to longer retention times as the pressure increases. Your first guess (and mine) is that the flow rate was decreasing at the higher pressure but this is not the case. I measured 59 mL of mobile phase in 60 minutes at both 1300 and 3900 psi. If I reverse flush my guard column, the normal pressure and retention times are restored so I know that the guard column is collecting sample crud. Any ideas?

[tom jupille]

As a wild guess, the compressibility compensation on one or both of the pumps is incorrectly set so that you're pumping the correct volume, but the ratio is off ?????

[Petrus Hemstrom]

Two guesses:
1. The crud and the sample interact (ionic interactions?) perhaps not so likely with the long retention time.
2. Heating and or pressure effects from the increased backpressure. (there are some papers on this from David McCalley and also from Tanaka on UPLC-HPLC selectivity differences) Cant remember what was the conclusion.

Add a restriction capillary to simmulate the increased backpressure but without the crud will tell you if it is the crud or the pressure that is the root of your problem.

[skunked_once]

Tom and Peter,

Thanks for your replies. When I run the next sample set next week I am going to premix the mobile phase and monitor the results. Will report the results.

[mbreslav]

Premix A and B phases. You can also run each pump separately with the premixed mobile phase for troubleshooting.

[skunked_once]

When I run the next sample set next week I am going to premix the mobile phase and monitor the results. Will report the results.

So, I premixed the acetonitrile/water (25:75) and ran a set of my floret extracts. The first run was at 1983 psi and I verified the flow rate at exactly 1 mL/min by solvent collection. The internal standard eluted at 26.1 min. and my last peak eluted at 49.8 min. Pressure gradually increase to 2212 psi at run seven but the flow rate was still exactly 1 mL/min. The internal standard eluted at 26.5 min. and the last peak eluted at 52.0 min. Since the flow rate did not change, the issue seems to be a guard column crud buildup problems causing the retention time increase.

[mbreslav]

Sorry, overlooked your previous post. <..> If this is a buildup, this should be caused by the hydrophobic material that holds your analytes. (You may want to consider SPE or extraction sample preparation procedure). I am curious if your peak shapes are changing with the retention time change.

[skunked_once]

If a picture is worth a thousand words, a chromatogram is worth a few. Here are the two runs about which I am posting.


[url][/url]


There doesn't seem to be any change in peak shape, just a retention time increase.
The chromatogram in black is the first run and the one in blue is run #7.

[mbreslav]

Let's put a new guard column insert, and see what happens.

[skunked_once]

Let's put a new guard column insert, and see what happens.

I've replaced the guard column in earlier sample sets and the same thing always happens. Retention times are normal with a fresh guard column and normal pressure but as pressure increases with more sample injections, the retention time start to increase.

[tom jupille]

Looking at those chromatograms, the percentage change in retention time is different for different peaks. That eliminates flow as a culprit (confirming what you found when you measured the flow) and points the finger at chemistry. You're seeing the same problem with premix vs on-line mixing of the mobile phase, so that pretty much eliminates proportioning problems. To borrow from Sherlock Holmes ". . . when you have eliminated the impossible, whatever remains, however improbable, must be the truth.". That leaves CRUD building up on the guard cartridge. I can think of three possible solutions:
- change the guard frequently
- change the sample prep to get rid of more CRUD
- use a switching valve to backwash the guard after each run.

[mbreslav]

You can also develop procedure where you inject your standards mixture as a system suit before, in between samples, and at the end of your samples sequence.

[skunked_once]

Thank you all for your comments. I have completed the analysis of these samples for the time being but will work with your suggestions if/when I run these samples again. Results of success and/or failures will be reported at that time.