Peak splitting after 200 inj.

[Rdsxbaseball]

Hi,
I am experiencing a peak splitting problem with one of my assays. We are using a Gemini-nx column with corresponding guard. The mobile phase is a sodium phosphate/meoh gradient. Injection volume is 30ul and the mp ph=7. The column was working fine for a couple of months and getting around 900 inj. Now it barely is getting 200inj. Any ideas on what could be causing this?

[tom jupille]

How often are you changing the guard cartridge? And is the 200 injections on only one column or on several (you may just have gotten one bad column)?

[Rdsxbaseball]

Tom,
At first I thought I just hit a bad column but going through about 5 columns, the exact something is happening. I've tried another instrument with the same setup and no change. Peaks still split. Before the peak splitting showed up, the columns were getting ~900 injections with a guard change at ~300-400 injections. Could it be a solvent issue? Changing columns to another brand, there is no splitting but the lifetime is only ~400 injections.

[tom jupille]

Well, that eliminates the "one bad column" hypothesis.

Some additional things to look at:

If you have more than one peak in your chromatogram, did they all split the same way? If they did, that suggests that the problem is an inlet flow profile anomaly -- in this case most likely a headspace at the column inlet. That really shouldn't be happening at pH 7, especially with a "hybrid" support like the Gemini. Which brings up the question "were all of the failed columns from the same lot?". Can you contact Phenomenex and try to find out if there was some systematic change at there end between old "good" columns and newer "bad" columns?

If different peaks look "different", then that suggests that you're looking at a chemical problem. You can get peak shape problems if the diluent is stronger than the initial mobile phase, but those usually show up right away, not after several hundred injections.

Is there something in your sample that might be changing the column chemistry (extreme pH? surfactants? lipids/fats/waxes?)?

[Rdsxbaseball]

According to phenomenex, no change in manufacturing. We have two analytes of interest and the only peak that splits is the first analyte. After looking all over the net, it's leading me to believe that the inlet frit of the column is becoming clogged. but I'm not totally convinced by this because the second analyte peak doesn't split. I've never seen anything like this before, so that's why I'm looking for help. What is the main cause of not all peaks splitting in hplc? that is the question.

[HPLCaddict]

Clogging of the inlet frit usually leads to "shouldering" peaks, only in extreme cases you will see split peaks. But as Tom already pointed out, this will be the case for ALL peaks. Since only one of your peaks shows splittting and the other one is fine, this points to a "chemical" problem, not a "physical" one.
Splitting of early eluting peaks usually is caused by a solvent/mobile phase mismatch. In you're case, as you're seeing this phenomenon not from the beginning but only after ~200 injections, this explanation unfortunately is not really straightforward . It seems some sort of column changing over the course of your analyses is also involved...
As a quick experiment, I'd try to inject smaller amounts of your sample on a "bad" column and/or, if there is significant difference between sample solvent and mobile phase, prepare a sample in the initial mobile phase and inject that one. With either approach, you should see some improvement concerning peak splitting if solvent/mobile phase mismatch is the root cause.

[juddc]

Have you tried turning the column around and running an injection or three? I wouldn't throw out the solvent - MP mismatch theory yet. You CAN see such a situation at ~200 injections if they've occurred during different runs and an error was made in sample preparation. Depending on column size a 30-ul injection volume is a pretty good slug. I concur with HPLC addict's suggestion of a smaller injection volume.

Alternately, could there be something stuck on a frit elsewhere in the system perhaps.

Pictures might help - any possibility that we might see a chromatogram?

Good luck!

CJ

[Rdsxbaseball]

So today I back flushed a column that I saw peak splitting in at 200 inj. I then took the column and injected the same sample twice, 15ul first, then 30ul right after. The 15ul injection did not show any split peaks, whereas the 30 ul injection did (the first analyte). So, I guess there was some column overloading going on. I am going to work on validating the 15ul injection volume and see if we can get the columns to last past 900 injections (previous amount of injections I saw with injecting 30ul). I will keep everyone posted and big thanks to everyone steering me in the right direction.