Sample concentrations?

[Sante]

Hey everyone.

I'm trying to develop a Reversed-phase HPLC method for two quite similar pharmaceuticals A and B. They are separating fine, but I have encountered a problem with concentrations...

Concentration of A is few mg/ml and B few mcg/ml. Because of this the spike of B is naturally very small. I have tried increasing injection volume, but this inflicts problems with A spike going out of shape.

The only solution I can figure is to dilute the sample for the analysis of A and use raw sample with large injection volume for B... or just trying more different mobile phases.

Would anyone know any other ways to get through this? I don't like the idea of showing anyone a method where I'm measuring a little spike next to a few thousand mAu monster

I have a system with variable wavelenght UV-detector.

[David Blais]

Sante,

If I were in your position, I would just run what you have proposed using the two solutions to analyze for each compound. Without knowing what your compounds are, I would hesitate to suggest changing the mobile phases, you probably will not see a great change (if any) in peak response, and you will more than likely sacrifice your resolution, specificity, etc.

One other thought - have you checked the lamdba max of these compounds? You said they were similar, but perhaps the wavelength at which they have a maximum absorbance is different. Then you could program your method to switch wavelengths during the run so that the second, smaller (in concentration) compound was detected at a wavelength where it showed greater absorption. I realize this is a stretch, but it's worth a shot.

[HW Mueller]

If they absorb at the same wavelength one can run the high concentration analyte at a wavelength which is off the max. One losses a bit of robustness, though, when using the shoulder of the spectrum rather than its max.

[Dan]

Sante,

We have a similar situation. Our pharmaceutical product has three actives and the difference from lowest strength active to highest is 2.5 orders of magnitude.

We didn't want the additional work of extra dilutions.

We chose to use two different wavelenghts for the UV detection. One wavelentgh is optimized for the lambda max of two of the analytes (the lower concentration ones). The other wavelength is not a lambda max for the third (highest concentration) analyte, it is a wavelength on the shoulder of its spectrum.

As HW Mueller pointed out, there can be a problem with robustness when using a wavelength on the shoulder. In our case, the detection of the analyte becomes non-linear above the 130% level. However, this is still OK for our use as our samples are never that high in concentration.

So, we make one sample preparation. The options then are:
1) Inject once and use two different detection wavelengths.
(Use either a dual channel detector or two detectors in series).
2) Inject twice: once at each wavelength.
3) Switch detection wavelengths during the run.


Regards,
Dan

[Sante]

Thanks guys!

I feel a bit embarassed, should have thought that myself.

I'm not quite sure if using wawelength off lambda max will work, since the difference in peak response is huge, but these suggestions are definately worth a try...

In general, I feel that there's great atmosphere of helping less experienced people on this forum. This kind of knowledge sharing is rare in times like these, when only thing that counts is $.
Keep up the good work!
I hope someday I have enough knowledge to help others.

[HW Mueller]

If your peaks overlap wavelength switching won´t work, if there is no overlap you can choose even a wavelength where A doesn´t absorb at all........ if you don´t want to see it anymore.