Moving retention time with protein method

[KarenJ]

Hi,

I have a RP HPLC method for quantitating a large protein. The separation is done with a Vydac C8, 300A, 5 um, 2.1x250 mm column. The mobile phase is a gradient of ACN/water/0.1% TFA. The gradient starts at 30% ACN and ramps to 92%ACN over 15 min. The main peak comes out at 11.3 min. on my Agilent HP1100 system. When I run this method on my Dionex Ultimate 3000, the peak comes out at 8.6 min. I have to transfer this method to another lab with an older Agilent HP1100 and the retention time on that instrument is about 14 minutes. The general shape of the chromatograms are the same on all instruments.

It was suggested to me that maybe the differences in retention time were caused by differences in instrument dwell time. I added another piece of tubing to my Agilent instrument in front of the column to try to increase the dwell time but I did not see any change relative to the solvent front.

Any ideas of what could cause this peak to come off at such different times on these different instruments?

How big of a deal is it to have different retention times on different instruments if the method is otherwise behaving well (ie. good linearity and precision) on each instrument?

Thanks,

KarenJ

[Noser222]

I guess if all the pumps are working the same then a simple check is to see if one of the systems has a gradient mixer and the other does not. Have you tried actually measuring the dwell volume on each system?

[Mark Tracy]

Do the following experiment on all three instruments. Replace the column with a restrictor loop; enough 0.003 inch tubing to make the pressure about equal to your column at normal operating conditions. Now set up a gradient of water (A) versus 0.1% TFA (B): 100% A for 5 minutes, linear ramp to 100% B over 10 minutes, hold 100% B for 10 minutes, flow rate equal to whatever you use for your method. Monitor the UV 214 signal. In a perfect system the baseline will reach 50% of maximum at 10 minutes, but in a real system, that point will be reached later. The difference between theoretical and actual is your gradient delay; this will differ between instruments.

The 1100 has a compliant pulse damper that introduces a variable delay depending on pressure. Some people rearrange the plumbing to remove this volume from the direct flow path. The UltiMate has no pulse damper, and the gradient mixer is very small. (I've met the engineers at Dionex. They were near fanatic about reducing the gradient delay on this new pump.) Also, there are various options for both systems that affect the gradient delay volume. Your autosampler loop also affects gradient delay.

Protein separations are very sensitive to the % acetonitrile, and the protein will remain stuck to the column until the right composition arrives, then it will be nearly unretained. (Try doing an isocratic separation and you will see what I mean.) That is why your separation is so sensitive to the details of the system design.

[KarenJ]

Hi,

All three pumps are gradient pumps. The Agilent pumps are binary gradient pumps with high-pressure mixing and the Dionex is a tertiary pump with low-pressure mixing.

The dwell volume of my Agilent system is 1.85 ml and the dwell volume of the Dionex system is 2.35 ml. I'm still waiting from the other lab to get the dwell volume of the other Agilent system.

If the shifting retention times were related to dwell volume, shouldn't I see longer retention times for larger dwell volumes?

Thanks,

KarenJ

[KarenJ]

Hi Mark,

Thanks for the info. about the pumps.

We measured our dwell volumes by having water as A and 0.1% acetone as B and measuring at 265 nm. We then ramp from 0 to 100% B over 15 min.

KarenJ

[tom jupille]

KarenJ, dwell volume differences are always the first place to look when you see instrument-to-instrument retention time offsets. The retention time differences should be the dwell volume differences divided by the flow rate.

As you point out, the retention time offsets are going the "wrong way" in this case, so you have to look at other possibilities:

- column chemistry. <You have probably already done this but if not, move the same physical column between instruments and see if the difference is related to the column or the instrument.

- proportioning accuracy. Look closely at the printout from your dwell volume measurement test as suggested by Mark. Is the "ramp" nice and linear or are there deviations? You can get a more accurate reading by running a "gradient step test" (a brief description is in the HPLC Troubleshooting Wizard here: http://www.lcresources.com/resources/TSWiz/hs450.htm )

- temperature. Different column compartment designs can result in different in-column temperatures. A brief discussion (and a links to a couple of articles) is located here: http://www.lcresources.com/resources/TSWiz/hs430.htm

[KarenJ]

Hi,

We have two columns that we've been using. In my lab I have been switching one column from the Dionex system to the Agilent system. The other lab gave me their column last week and I ran it on my Agilent system. The time on their column was about 0.15 min. longer than the time on my column.

If the differences in retention times is due to differences in column heating systems, gradient mixing systems, etc., what can I do to make this a transferrable method?

Thanks,

KarenJ

[tom jupille]

If the differences in retention times is due to differences in column heating systems, gradient mixing systems, etc., what can I do to make this a transferrable method?

If it's a proportioning problem, it should show up on the step test as outside the manufacturer's spec for proportioning accuracy. If one or more of the systems is out of spec, have it serviced and re-qualified. The catch is that you might be on the "ragged edge"; as Mark said in his post, proteins are a lot more sensitive to errors in composition than are small molecules.

One trick that may help a bit is to put 30% ACN in the "A" reservoir and 92% ACN in the "B" reservoir and then run your gradient 0-100%B. That spreads any proportioning errors over the entire gradient range. That often works when running proteins using narrow-range gradients (e.g., 10-30% ACN), but may not make a lot of difference in your case.

If it's a temperature control issue, you can get a hint by looking at the system back pressure. Pressure correlates with viscosity which is temperature dependent. If the shorter retention time system also has a slightly lower back pressure, that would support the temperature hypothesis. The first thing to do is to stick a thermocouple in the column oven and check that your oven temperatures are correct.

Assuming they are, the next thing is to look at the oven design. Generally, air-bath ovens give lower on-column temperatures than do heat-block ovens (metal-to-metal contact provides better heat transfer than air-to-metal). You could try putting a temperature probe in a tee downstream of the column and monitoring mobile phase temperature directly, but that's getting to be a pain in the neck (or lower ). I would try increasing the temperature setting by 5 degrees or so on the instrument with the longest retention time and see what happens. If the retention time comes down (and everything else works OK), then (if this were my problem) I'd set a fairly tight retention time window as part of system suitability and explicitly allow temperature to be adjusted as necessary to meet that spec. Obviously, you will have to set limits as to the maximum allowable temperature range, and it would be helpful to your users to include suggestions as to how much to tweak the temp for a given retention time difference.

If it's not dwell volume, not column chemistry, not proportioning accuracy, and not temperature, then I'm out of ideas.

[Bruce Hamilton]

Hi,

I have a RP HPLC method for quantitating a large protein. The separation is done with a Vydac C8, 300A, 5 um, 2.1x250 mm column. The mobile phase is a gradient of ACN/water/0.1% TFA. The gradient starts at 30% ACN and ramps to 92%ACN over 15 min. The main peak comes out at 11.3 min. on my Agilent HP1100 system. When I run this method on my Dionex Ultimate 3000, the peak comes out at 8.6 min.

As well as the above, ensure that you are making up the mobile phase exactly the same, with the same grades of reagents. I assume TFA is in both mobile phases. TFA is very volatile, so ensure that it is being accurately metered. If using volume, an air displacement pipettor will not be accurate, use a positive displacement pipettor, a volumetric pipette, or a syringe.

Those sort of differences would cause me to think either column differences ( same guards? ), different actual flow rates, or mobile phase composition.

Good luck, and please keep having fun,

Bruce Hamilton