Page 1 of 1

Qualitative method for unknown components

Posted: Thu Nov 01, 2007 9:00 am
by chromdo
Hello!

I am developing a method to detect unknown components in pharmaceutical preparations. Herefore I am using an HPLC coupled to PDA and a spectral database, with the identification based on Rt (10% window) and UV-spectra. The aim is to develop a cheap, robuste and reproducible method. After a lot of reading and experimenting I developed the following m and s phase:

Acetonitrile - 50%
(high optical transmission down to 200 nm)
Phosphate buffer + TEA pH 3,2 - 50%
(high optical transmission down to 200 nm, TEA to prevent tailing, buffer for reproducible UV-spectra and retention times)
Column: Waters symmetry C18, 5um (because of the high reproducibility)

I chose for a isocratic system because of the robustness and the opportunity to recycle the mobile phase. The ratio of ACN-buffer I determined with a sample which contains a critical pair and a substance with a long retention time, however this critical pair (paracetamole-caffein) could not be sufficiently seperated whilst the run time is still acceptable.
A disadvantage I experience now is that a lot of components of interest (75%) elute in the first 5 minutes and the later in the last 20 minutes. Which is explainable, because of the isocratic elution and a lot of protonated bases at this pH. Most of the time this not a problem because a lot of samples only contain one or two substances, but for some flu-medicines and painkillers this results in overlaying peaks to a degree that the correct UV-spectra can not be obtained. In my head I played with a couple of solutions: :idea:

- Gradient elution - but the HPLCs here are not very reproducible in their gradients and the mobile fase can not be recirculated (the people here really want their method to be cheap)

- Another mobile phase (less ACN) when the components seem to overlay, so only when the chromatogram is not sufficient - but that means making two databases.

- Adding a ion-pairing agent. Because most of the quick eluting drugs are basic drugs with the pKa above 4. When letting them interact with the ion-pairing agent more retention and therefore probably better seperation can occur. However, when adding for example disodiumhexanesulfonate, does this substance not react strongly with the TEA? Can both of these substances still work sufficiently in each others presence?

- leaving it this way - this is what my colleague suggest, however I don't really want to settle with this.

The method only has to be qualitative, so a perfect resolution is not necessary, only the correct UV spectra must be obtained. I did a lot of reading, but I don't have a lot of practical experience. Whilst reading the forum I saw that you obviously have a lot. I was hoping that you could point me in a good direction.


San :arrow:

Re: Qualitative method for unknown components

Posted: Thu Nov 01, 2007 9:29 am
by chromdo
FYI:

I am making my own database, so it is no problem to change the conditions.

Posted: Thu Nov 01, 2007 4:37 pm
by DR
Developing a spectral DB is a nice idea, in theory. The problems you will run into are:

Each and every API of interest to you will probably require a different LC method to get it away from its impurities.

Many related substances and degradation products have UV spectra that are annoyingly similar to that of the API.

UV is not a good fingerprinting tool, you will need other means for definitive IDs of new impurities.

PDA is a useful tool for telling you if your API and RS are all separated from each other, but it is rarely, if ever useful for making an identification.

So, in practice... you might better spend your time on a MS or NMR database if you have the luxory of access to either type of instrument at the end of an LC.