Retention time shift

[StephenO]

If you have a validated HPLC method should I see shifts in Retention time (by 1 min) from Sample Set to Sample set?

There are some differences between the sample sets in that the Reagents for the mobile phase are from different suppliers but the same grade and also they are run on different HPLCs (both Waters Alliance).

[tom jupille]

Details, details, details! We need details!

Isocratic or gradient?
Are the instrument-to-instrument differences consistent or random?
Do the same instrument differences show a trend or are they random?
Neutrals or ionizable? if the latter, approximate pKa? buffer used? pH?
Temperature?

[mattmullaney]

Hi StephenO,

Mr. Jupille is on-the-money with his assessment, more details into a problem are always better to have. I guess it's kind of like when I was a TA back in grad school...it can be hard to remember that the students don't know what the teacher does...assuming they know little is better than not (though the quicker or more-prepared students may get a bit flummoxed by this approach).

One quick thought, since it came up recently at my job, are you Really Certain that the salts from different vendors are exactly the same? By this I mean all of them are anhydrous or the same hydrate salt...ionic strength of buffer can make a Huge Difference in analyte retention time. When you note that the mobile phase reagents are the same...is this traceable to the Certificates of Analysis for these materials?

Just had another thought...was the temperature of the water the same when buffers were prepared for both instruments, and/or the pH adjustments made? Again, this can cause huge (unexpected?) shifts in retention time. Our lab is kind of warm (26-27 deg Celsius), so we take care to standardize our pH meters and measure buffer pH using a constant temp bath to hold the solutions we are using at the time.

[StephenO]

Hi Tom See below

Isocratic or gradient? ISOCRATIC
Are the instrument-to-instrument differences consistent or random? RANDOM
Do the same instrument differences show a trend or are they random? RANDOM

Neutrals or ionizable? if the latter, approximate pKa? buffer used? pH?
Mobile phase 30:70 (0.01M Ammonium Aceate : Methanol) pH to 5.2 with Glacial Acetic Acid

Temperature?
Column Temp 30 degrees

[mattmullaney]

Hi Again, StephenO,

Isocratic, changes in RT are random amongst two HPLCs of the same type and from the same manufacturer (Waters Alliance), pH around 6 or so considering buffer + MeOH proportion in eluent,
30 deg Celsius.

Thinking about it a bit, the things that can affect retention time are, assuming the same column volume for two columns used remains the same, are:

Differences in the stationary phase in the two columns (not likely here).

Differences in the eluent (possible here, if each of two preps are used for each Alliance used even if the Same Eluent Additives are Used).

Differences in eluent temperature settings (probably not likely, but may be, or something like one HPLC could be under an AC/heating unit and the other not, but random retention time differences are hard to explain this way.).

Differences in eluent flow rate (probably not likely...as random differences in retention time are observed on both HPLCs...that said, are the backpressure values steady with pressure ripple less than 2% of mean value on each HPLC, are pressures within, say 100 or so PSI on both HPLCs, are IDs of capillary tubing the same on both HPLCs...are there leaks or air bubbles...random air bubbles may be most likely here, are you pre-mixing the ammonium acetate and the MeOH or using the pump and GPV to mix the two together on the HPLCs?).

Differences in the packing density of the stationary phase within the two columns (probably not likely).

My new set of thoughts:

Please check your pressure ripple values for both HPLCs when they are running...random air bubbles on both HPLCs is a possible explanation for what you observe.

Please ensure the ammonium acetate salts that you are using really are both similar through looking at their Certificates of Analysis...though this doesn't explain random retention time variation.

Please check into the pH adjustment of the ammonium acetate solution part of the eluent.

If it's easy for you, please check into the retention time of a non-retained substance (say uracil or sodium nitrate) on both systems with the two columns.

If your analyte(s) are neutral, there may be less of a chance for retention time shifting...unless the sample matrix itself is problematic. When I ran DNPH derivatives of carbonyl compounds on air samples from automotive exhaust, I found that I needed to increase the vendor-recommended ammonium acetate concentration from 4 mM to 12 mM to have stable retention times for all 15 analytes...standards were never a problem with retention time variablity...).

If your analyte(s) are polar, or even ionic, please check into their pKa values (as Tom noted above). In my opinion, the greatest likelihood of the randomly shifting retention times could be that the buffer may not be concentrated enough, or maybe that the eluent pH that you are working at is too close to the pKa of the analyte(s) you are trying to look at.

(That said, the other things I list may also contibute to the shifting retention times...though they are seemingly less likely at least to me.)

Best Wishes--and if Tom has further comments I'm sure he'll send them along.

[HPLCaddict]

Did I understand correctly that even on one instrument you get erratic retention times? Using the same mobile phase or does the retention time change only after a different batch of mobile phase is applied?
If you don't get steady retention times when using only one instrument this points either at system malfunction or some inherent flaw in the method itself. Before checking possible differences between those two machines, I'd rather try to set up the method smoothly on one HPLC first.

[mattmullaney]

HPLCaddict has a point that I missed...he is a sharp one. This method is validated. If the retention times are randomly shifting in a one-minute range on one HPLC alone, well, that would seem to point to an eluent flow rate problem, specifically...transient bubbles.

And, of course, it's the best idea to ensure that the difficulties are completely isolated from how the HPLC instrument is performing. If you can get stable retention times on one Alliance, that is an important piece of information to have.

[StephenO]

In a regulated environment could you argue this shift between retention times (1 minute) in sample sets or would it be expected to have the same component retention everytime you run a sample set, given that the below parameters could be different between sample sets.

Different Analyst preparation.
Different time of the year.
Different reagent supplier.
Different HPLC Used.

Thanks for all your guidence.

[mattmullaney]

Hi StephenO,

Wow...what is the retention factor of the analyte? If the value of k for the analyte is relatively large, say 5 or greater, then perhaps there is an argument to make if we're talking about multiple analysts and instruments.

If we're talking about a one minute shift in retention time for one analyte on one instrument during the course of one sample set run...now, this is a different matter altogether.

I'll respectfully ask the same question as HPLCaddict...are the differences in the retention times for the analyte observed on one instrument, on every instrument for each sample set run (retention times are always shifting on every single instrument tried), or between two instruments?