gradient method validation

[Ericz]

The LC method is a gradient followed by an isocratic run. The compound was eluted after the isocratic step. However, the results are reproducible, i.e, it's the same retention from run to run. Is this still a valid gradient method?
Thanks in advance!

[tom jupille]

If you can demonstrate that you get consistent results, then it's valid (by definition), and I'm not sure whether the isocratic/gradient distinction is important.

That said, can you clarify what you mean by "after the isocratic step"? Remember that your gradient system has a delay volume between the pump and the column, and the column has a dead time. So if, for example, you have:

• - a gradient delay volume (properly called "dwell volume") of 1.5 mL
  - and you're using a 150x4.6 mm column (which has a dead volume of about 1.5 mL)
  - at 1 mL/min
  - and you run a 15-min gradient,

the end of the gradient doesn't get to the end of the column until 18 minutes.

[DR]

This is one of those instances where a picture would be worth a couple of hundred words...

[Ericz]

Sorry I can't have a plot here for the method!
The gradient method is a short one used for LC/MS. It's basically isocratic (0.5min), gradient (1.0min B-organic 5 to 100%), isocratic(1min),gradient(0.5min to 5%B) (**) back to initial and then isocratic for equilibration (1min). The compound eluted somewhere around the second "gradient" step(**). The column is 30X2mm with flow rate of 300µL/min. LC is a Shimadzu 10A VP with only a pre-mixer. Formic acid of 0.1% is used in both organic and aqueous phases. The results are indeed repeatable.

More comments will be highly appreciated

[Uwe Neue]

You got to give some of the information that Tom has talked about, otherwise one does not know what is going on.
Your LC system
Your flow rate
the column dimensions
the gradient run time before the isocratic step
the length of the isocratic step
the elution time of your compound

and best all other gradient details.

[tom jupille]

Eric & Uwe : looks like crossed messages.

Eric: the missing piece of information is the dwell volume in your system.

[Uwe Neue]

Tom: Eric provided the relevant information by editing a previous message. This is why Q and A look inverted.

Eric: Tom is correct. The factor that determines where the peak elutes is the composition of the gradient at the column outlet. In order to calcualte this from the information provided, one would need the gradient delay volume of your instrument. From the information that you gave, your peak elutes somewhere between 2.5 and 3.5 minutes. You actual gradient is executed at the gradient mixer between 0.5 and 1.5 minutes. In order to see if this peak actually elutes in your gradient, you need to know, at what time your gradient gets to the column outlet. If you had a gradient delay volume of 0.9 mL (including the column), you gradient would get to the column outlet at 0.5 + 0.9/0.3 = 3.5 minutes.

This is why you need to figure out what the gradient delay volume of your system is.

[Ericz]

The column is about 90µL (30x2mm). The premixer is about 100µL and other tubing may be around 10µL.

Why the compound elution is far away from the estimated time? Is the mothed still valid?

Thanks for all of you!!!

[Uwe Neue]

I am not familar with your system. Is this a high-pressure gradient mixing system (=two pumps) or a low-pressure gradient mixing system (=the gradient is mixed before the pump). If the later is the case, you would need to add the components of the pump to the gradient delay volume.

[Ericz]

It's a high pressure system with two pumps and a pre-mixer.

Thanks!

[Uwe Neue]

The only conclusion that I can draw from all the information that you gave us is that your peak elutes only when you run the reverse gradient back to starting conditions. This is weird...

What is your column?

(unless there is some volume in your system that you do not know about)

[Ericz]

Thanks! This what I don't understand either.

The column was small (30x2 mm). It would be good if the compound eluted before back gredient started (RIGHT?), but it didn't. Actually it happened to a few compounds as well (Sorry I can't say anything about the compounds' themselves). I did have another method that makes more sense, i.e, the compound eluted before the back gradient began. My first thought was that the pump didn't deliver the right amount of mobile phase even if it showed the percentage was changed as set up. However I'm still wondering whether there is any explanation or there is actually a problem in the method itself.

Has anybody come cross the similar case?

There is another topic about Fast LC. Is it related to my case?

[bhaskarac]

Here I have a other question.

If our method is a gradient and Solvent-B is a mixture of two solvents (i.e. Methanol and acetonitrile), in Robustness study of analytical method validations, how can we distinguish different organic ratio.
comments polease

Thanks in advance,
Bhaskar

[putnam]

Ha!

Gradient delay volume has been the bane of my existence since working for my current employer. All methods here were developed on HPLCs from one vendor ... all identical and with, from my experience, rather large mixing chambers. The methods are all gradients and the gradient begins at t=0 minutes. To make things worse, many of the gradients are quite drastic, say, from 0% B to 100% B in 20 minutes.

Problems arose when methods were transferred to Europe to a laboratory that had HPLCs from a different vendor with, from previous experience, I know has a significantly smaller mixing volume. They failed to achieve adequate resolution of the two peaks in the method. I knew this was due to difficulties with the gradient, but trying to explain this to the biologists who developed the method was not so easy. They couldn't understand why V(d) was different from t(d). Anyway, after calculating V(d) of the different systems, I instituted an isocratic hold in the method that allowed the vendor laboratory to achieve the same results that we do on our instrumentation.

The question is ... do I need to re-validate the method? I don't think so, because I can show that the analyte is now in the same "environment" (in terms of gradient composition) at different reference times throughout the method on each HPLC system, whereas before this was not the case. In fact, I believe, without the isocratic hold - re-validation would be necessary as the methods clearly, because of differences in V(d) have different retention characteristics.

Do I have this right? If so, what is the more elegant explanation?

[DR]

...
Do I have this right? If so, what is the more elegant explanation?

Right - I'd say "An initial isocratic hold may be added to simulate the #mL dwell volume of the instrument that this method was developed on." - put that right in the system suitability part of the method.

You're lucky that it wasn't the other way around. I've been handed gradient methods developed on systems with <.5mL Vd and been expected to get results on a system with huge Vd (people complain when you make the run time longer).