Chromatography Forum

Would it be ok to calculate the linearity slope of impurity by varying injection volume in HPLC? is that correct? not too much impurity available. Thanks

It is quite correct I would say. I’ve had this discussion several times with different people. And the dominating argument against the notion has been “injection volume variation”. And it’s legitimate enough as a separate weighing, dissolving, potentially dilution as well.
My point is: You know the variation of the injection system you use and if you don’t know it you better determine it. Now most users determine the injection variation at a single level, say 10 µL. But if you find diverse volumes useful as you du in your case, then you can qualify the system you use according to your needs.
What I’ve done previously – because I find it very useful in deed, to inject varying volumes as a mean of loading varying amounts – is to inject the different volumes at the relevant levels (in replicates of course) and evaluate the potential variation. What I´ve found at several occasions is, that the procedure is much more accurate and precise than weighing, dissolving and diluting separate samples.
For the documentation sake, you could inject standards, or control samples at the relevant levels in your analytical sequence and thereby document the feasibility of what you’ve done.
Best Regards
P.S. From some of the discussions I remember, people forgetting that there is a variation as well in case of injecting the same volume multiple times. I’ve found at several occasions that the variation is virtually the same irrespective of the procedure – whether its multiple levels (normalized) or multiple injections at the same level.

Best Regards

Thank you Danko. Excuse my novice questions: Both API and impurities would have to be at the same concnetrations to obtain the slope and then calculate the RRF using this approach, say at +/- 50% of specification level for the impurity? or API linearity at +/-50% of nominal concnetration and impuritiy at +/-50% of specification level? sorry I am new at this and gets a little confusing.

Shadow, we use a calibration line of impurity and API in the same concentration range to calculate rrf.

Thank you aceto_81! would this calibration curve go down to LOQ or +/- 50% of reporting threshold for the impurity? Thank so much

We try to make a calibration curve going from LOQ to our impurity limits, but if we find that there is to much variation in the calibration lines, we may adjust to use higher concentrations.

Thank you very much , this is very helpfull. One more question, are RRF the same regardless of chromatographic conditions? different methods? Thanks a lot

Nope, RRFs are dependend on the method, a different pH can give different responses.

HTH

Ace

thanks aceto_81

Would it be ok to calculate the linearity slope of impurity by varying injection volume in HPLC? is that correct? not too much impurity available. Thanks

Hi everyone,
I don't think it is a good idea to variate the injection volume. This is useful to qualify your injector and it is a method used to check the linearity of the injector.
As far as I know injecting 10 microliter of a sample at a concentration of 1mg/ml is different than injecting 100 microliter of a sample at a concentration of 0.1 mg/ml. Accuracy and/or precision might be affected.
Moreover the EP and USP agree that any reduction in injection volume can be made, as long as system suitability passes. But no increase is admitted. So to validate a method one should keep the same injection volume.

Would it be ok to calculate the linearity slope of impurity by varying injection volume in HPLC? is that correct? not too much impurity available. Thanks

Hi everyone,
I don't think it is a good idea to variate the injection volume. This is useful to qualify your injector and it is a method used to check the linearity of the injector.
As far as I know injecting 10 microliter of a sample at a concentration of 1mg/ml is different than injecting 100 microliter of a sample at a concentration of 0.1 mg/ml. Accuracy and/or precision might be affected.
Moreover the EP and USP agree that any reduction in injection volume can be made, as long as system suitability passes. But no increase is admitted. So to validate a method one should keep the same injection volume.

Doing work with EPA methods it is not allow to vary the injection volume, but for research work or specialty work it can be allowed. One thing to remember is that the solvent for the calibration should be matched with that of the initial mobile phase at time of injection. If you are injecting a calibration in pure Methanol into a mobile phase that is 10% Methanol and 90% Water then as you vary the injection volume you are having slight changes in Methanol concentration hitting the column, the larger the injection the larger the 100% methanol slug passing through the column and that can cause variations in response and retention times. Just something to be careful of when trying this.

Nour wrote:

I don't think it is a good idea to variate the injection volume. This is useful to qualify your injector and it is a method used to check the linearity of the injector.

Is that so? And why not? What will happen?

And if it's only useful for checking the linearity of the injector, why bother anyway.
Checking something you're not allowed to utilize – great way of keeping yourself busy?
------------------

To James Ball.

Matching the mobile phase strength is always a very good idea. Using weaker sample solvent than the mobile phase, when possible, is even a better yet idea.

Best Regards

If in your Performance qualification the injector linearity is qualified, then you can use the different injection volumes to find your relative response factors.
But keep the solvent strengths in mind: if you inject 100µl of sample in methanol on your column and you get ugly distorted peaks, then you can't use it!


Ace

I'd see it in a pragmatic way: If you're using different injection volumes you're checking both the linearity of the sampler and the method. Of course you should check the data accordingly, i.e. verify linearity, and not use it only for calculation of relative response factors. If the data is fine, bingo.

yes HPLCaddict, that would be the idea. Thank you!