Increasing retention times at high temperature

[JayVee]

Hello,

The pharmaceutical lab I work in uses both Waters and Agilent HPLC (recent additions to the instrument pool). One of the separation methods gives good plate count on Waters (c.a. 4000), and much poorer on Agilent (c.a. 2000). The method temperature is 50 Celsius. The columns used for the comparison were the same type and same supplier (I can't guarantee they were the very same column though).

I'm quite confident the plate count issue comes from the temperature inhomogeneity of the Agilent oven (the oven heats up one side of the column more than the other, which causes band broadening).

When I keep the column against the fins of the heater to get a more even column temperature, the plate counts rises to 2900.
When I keep the column against the fins of the heater and keep some insulating material between the column and the compartment door to prevent heat loss, the plate count reaches 3700, nearly as good as the Waters system.

However, the retention time of the peaks also increase with repeat injections when I insulate the column. The plate count/resolution remain good, but the peaks take longer to elute.

Any suggestions as to what the cause might be?

Notes. The method is isocratic, 0.6 mL/min, 70% ACN, 30% amonium acetate buffer. The analyte is derivative of Rapamycin. The column (YMC pro C18, 3um, 150mm) is on right heater (6uL volume) to preheat the mobile phase properly. Pressure was identical between the repeat injections. Pressure with the insulation was a bit lower than without insulation, which I attribute to the lower viscosity when the column is hotter.

Thanks

[syx]

we worked with azithromycin and found the same condition. this substance need high temperature to retain in the column. BP mentions the temperature is 70 C.

[unmgvar]

the main issue here is the difference in the way heating is transfered in both ovens.
the Waters oven is a air circulated type, the agilent oven is a contact type.

in the agilent oven you must make sure that the column is touching the metal or that something else is providing a transfer bridge then more heat is transfered.

another important factor regardless of oven type is the warming of the mobile phase.
when it hits the column it must be at oven temp. for that it is necessary that to have enough tubing (not too much either) in the oven to warm the mobile. in an agilent oven case, also find a way to make the tubing touch the plate of the oven in order to really warm. the pre-heater is not always sufficient in itself.

viscosity is lower because the mobile phase is hotter because heat is better tranfered.

[JayVee]

I would expect the retention times to get shorter with increasing temperature.
When I insulate the column in the Agilent HPLC, with good contact with the heater, is the column actually reaching 50 C, and getting damaged? Is this why the retention times increase with the insulation, instead of getting even shorter ?

In that case, would conditionning the column before using it for testing be a good idea ?

[tom jupille]

Retention times *usually* decrease with increasing temperature, but "usually" does not mean "always". Conformational changes and pKa changes have both been known to make retention increase at higher temps. If you're getting about the same retention and the same plate count at the same back pressure, I wouldn't worry about the fact that the change was the "wrong way".

[Kostas Petritis]

I agree with Tom and this is especially true for biopolymers. I have seen in mixtures of peptides that although most of them are eluting faster with increasing temperatures, some of them are eluting later (generally this happen with longer peptides).

[bily200677]

My idea is same to Tom's. SO i think you should draw the figure to reflect the relation of tempreature and the retention time to this compound. I believe "Actions speak louder than words"
Good luck !

[mardexis]

Is this a 3 mm ID column?

For the Agilent system there is no need to insulate the column or worry about placement relative to the metal fins. Temperature control is provided by heating the mobile phase within the metal fins probably much more effectively than the Waters air design. As long as the cover is on the instrument, heat supplied to column by the heated mobile phase should provide constant control at or near the set temperature. At lower flow rates (such as a 2.1 mm ID) you might have to worry a bit more about a radial thermal gradient eroding performance but the extra column broadening will be your main concern.

Hope this helps.

[JayVee]

Is this a 3 mm ID column?

It is a 3mm ID column indeed. The flow is 0.6 mL/min.

Without the insulation, peaks are significantly broader than with the insulation. Without either contact between the fins and the column, or insulation, they are so much broader than the system does not pass the system suitability - the resolution between the peak of interest and the next peak is too low.

Mardexis, can I ask you the name of the company you work for in Redwood city ? I work in Clonmel, Ireland, and we could very well be colleagues.

[mardexis]

Hmmm.... As long as liquid is actually flowing through the 6 uL heater and the cover is on there shouldn't be a problem. Oh! Unless the left and right sides are set to different temperatures. There is a setting to make them always the same and that box should be checked. If they are set at different temps and the column is in contact with metal at a different temp (ie, the other fin) then there will be a definite problem.

I work at Codexis but Abbott Vascular is just a couple of hundred meters away.

[HW Mueller]

If syx´s contribution is considered, and if I understand JayVee correctly then there is no problem here.

[JayVee]

If syx´s contribution is considered, and if I understand JayVee correctly then there is no problem here.

I think the method might have been badly designed from the start. The method is such that a slight increase in temperature increase the retention times, so it's not robust. The processing method forces the baseline between the peak of interest and the next peak, which are just resolved, so if there is a small rentention time shift, the baseline drop happens on one of the peaks.

And also, the "buffer" isn't actually a buffer, it's a solution of ammonium acetate, without acetic acid .

It tried my experiment again with a different column, different mobile phase and different sample solution, and the shift in retention happened again (higher temperature = longer retention).

The analyte has a conjugated Nitrogen (N-CO-CO). Do you think its protonation might change at higher temperature? Maybe a buffer would help to stop it from happening ?

Lastly, it's probably not column damage, because the retention times return to their previous value when the column return to its previous temperature (i.e. when I de-insulate it).

[/img]

[HW Mueller]

Tom already answered your question regarding the rt shift. Going into it more deeply is difficult.
You are right on buffring of a "pure" NH4OAc solution, but adding HOAc does not necessarily produce a good buffer (this is explained in more detail in another chain).

[danko]

Hi JayVee,

You mentioned that your mobile phase comprises 30% amonium acetate buffer, but you didn’t mention the concentration of amonium acetate. So, I can’t be too certain, but if the concentration is relatively high and if your analyte has a hydrophobic site, then you are experiencing a HIC interaction.
The higher temperature facilitates the breakage of the structured water, surrounding the hydrophobic site of the molecule thus exposing it to the stationary phase which in this case is also hydrophobic. This promotes stronger attraction of the analyte to the stationary phase and thus longer retention.

Best Regards