Standard Retention Time shifting

[kpld]

Dear all,

I have been experiencing irregular shifting in my methods for a long time. I have recently replaced the pump seals, AIV catridge, ball check valve, purge valve frit filter, inline-frit filter, guard column, and capillaries for my HPLC. When running the tests, everything is fine, and the standard run was fine. But when I start running my sample in MeOH + 0.1%TFA and Water + 0.1%TFA gradient method, I get shifting, mostly elution time shifted later. I've decided to run the standard for this column again, and noticed that there was shifting in the forward direction compared to my 1st standard run when I just replace all above mentioned parts. I will outline details for the standard runs below.

Specs:
Agilent 1200 series w/ degasser and quad pump, Waters T3 250x10mm RP column + guard column, HPLC grade water and acetonitrile

Standard Method:
Sample: 6ul acetone and 0.5mg acetonaphthene in 1ml mobile phase
Injection volume: 2ul
Flow: 5ml/min
Column oven temp: 26C
Isocratic: 40% Water, 60% Acetonitrile
Equilibration time: 30 minutes

Runs:
9/5/12 - for acetonaphthene 19.35min @ 195bar
9/11/12 - 19.1min @ 210bar, 18.95min @ 208bar, 19.24min @ 204bar
9/12/12 - 19.02min @ 205.3bar, 19.1min @ 206.4bar

As you can see my pressure fluctuates slightly but is always very stable during equilibration and run, but my retention times are always 0.1min to 0.4min shifted earlier. My suspicion is the temperature in the room, because though my column is in a column oven, the cover cannot be inserted because the guard column holder is too large, thus half the column is exposed to the room temperature, which changes daily. The acetone peak elutes within the same time, though for today it was shifted about 0.05min earlier too. My boss believes it could be air trapped inside the column (from a single large air bubble entering earlier this week), but I've been running solvents at above 1500psi for hours, and see no blips in the detector. I've also backflushed the column in attempt to "purge" the bubble from the column (suggestion by boss) for an hour at 0.5ml/min MeOH.

Any suggestions will be appreciated, and I apologize for lots of text.

[tom jupille]

My first response is that the variation isn't enough to worry about.

Assuming the 110 bar on 9/11 and the 119 minutes on 9/12 are typos, your average retention time is 19.02 minutes with a standard deviation of 0.16 minutes. That's not a whole lot of variation (RSD is 0.8%). The correlation with pressure looks like it's there but fairly weak (with only four valid data points, once I eliminate the obvious typos, there's not much to go on). Ambient temperature changes could certainly do that.

If it's really a concern, see if you can find a clear peak for t0. If that is shifting, then you have a flow rate problem. If t0 remains constant, then you have a chemistry or temperature problem.

Your post didn't specify if you were running the check samples with pre-mixed mobile phase or using the pumping system to mix the mobile phase. If it was the latter, try pre-mixing (i.e., prep the mobile phase off-line and put the same phase in both the A and the B reservoirs). If the variation goes away, then the problem is probably in the gradient proportioning system. If it persists, then it's probably flow (which should show up at t0 variation) or ambient temperature.

[kpld]

Thank you for your reply, and sorry for the typos they have been corrected.

I was wondering how to calculate RSD (is it just STD divided by two)? I still feel like a 20+ second difference between runs is significant, especially if I'm collecting 1 minute fractions (thats a maximum 40% error rate) for my actual runs (these are just runs with standards).

[tom jupille]

1. RSD is the standard deviation divided by the mean. I just used Excel to calculate both.

2. Where did the "collecting fractions" come from? The importance of retention reproducibility depends on the complexity of the sample. If there are only a few peaks, then you can tolerate a lot of slop.

If you need tigher tolerances, then you do, indeed, need to find where the variation is coming from. I suggested the diagnostic tests in my previous post:

. . . see if you can find a clear peak for t0. If that is shifting, then you have a flow rate problem. If t0 remains constant, then you have a chemistry or temperature problem.

Your post didn't specify if you were running the check samples with pre-mixed mobile phase or using the pumping system to mix the mobile phase. If it was the latter, try pre-mixing (i.e., prep the mobile phase off-line and put the same phase in both the A and the B reservoirs). If the variation goes away, then the problem is probably in the gradient proportioning system. If it persists, then it's probably flow (which should show up at t0 variation) or ambient temperature.

[kpld]

Hello again,

So I called up Waters to see if they had any suggestions on what could be creating this shifting in my chromatograms, and they suggested that I equilibrate my column longer. For 5ml/min (my standard method) they calculated 96 minutes in between runs. I tried running the standard again, incorporating the equilibration time and here is what I get.



Here are the the numbers:
Run 1
Acetone = 3.114min, Acenaphthene = 18.369min, Pressure = 189.25 bar

Run 2
Acetone = 3.115min, Acenaphthene = 18.404min, Pressure = 188.9 bar

Run 3
Acetone = 3.115min, Acenaphthene = 18.399min, Pressure = 190.47 bar

Run 4 (Next Day)
Acetone = 3.120min, Acenaphthene = 18.686min, Pressure = 189.85 bar

As you can see, runs 1-3 were done on the same day, and the retention time for acenaphthene are pretty close (run 1 vs run 2 only had a 0.03min difference, acceptable for me), but when I did the same run with 96min equilibration time the next day, the peak shifted later by 0.3min, which isn't good enough.

Looking at the void volume peak of acetone(assuming its t0 for you), there is a very small shift from run 1 to run 2, and another slightly larger shift from run 3 to run 4. Does this shifting in my void volume peak means something is happening to my flow rate and magnified the longer into the run? Also for run 4, the baseline shifts upwards. Do you have any suggestions why this occurred?

I used the same method and sample as in post 1, except made the column oven temperature to 30C and sonicated my solvents for 30 minutes before I attached them to the solvent lines.

Edit: I did a run 5 96 minutes after run 4, and it had the same retention time as run 4, though baseline was shifted a little bit downwards in the chromatogram. The baseline shifting occurs at wavelength 254mm (as shown in chromatograms) but isn't present at 273mm (not shown). Here is all 5 overlaid: