[STM]

Dear all,

Please kindly enlighten me on what it means and how to confirm purity of a substamce by HPLC. I read in some articles that involve synthesis and all they say is that e.g. the purity was confirmed by HPLC to be 97%.

For instance, if a compound has fluorescence property, does it mean i inject it into an HPLC under its favourable conditions (eluent, flow rate, ex/em wavelength etc) and look for how neat its chromatogram is?

I really wish to get a clear insight on what is meant by confiming purity by HPLC and how it is done. Other useful text, articles, weblinks helpful in this regard will be highly appreciated.

Thank you.

[Furanone]

Based on running chromatograms using a known 100% pure substance and comparing the peak areas for equal injections (or more likely average of say three injections each) to the unknown substance to be tested for purity, your unknown would give an average 97% of that same peak area of the 100% pure substance (assuming all other things held constant ie. detector response, sample injection volume, etc.). The 97% pure sample could also show some other small peaks (as the 3% impurities) but not necessarily, if the impurity is solvent.

[Babcock_Hall]

Just to amplify what Furanone said, if you saw two peaks in a ratio of 97:3, it would not necessarily mean that the product distribution is 97:3.  It depends on whether or not the detector is equally sensitive to both compounds.

[STM]

Thank you for your response @ Furanone and Babcock_Hall.

@Furanone: What will you advice in a situation where the standard or 100% pure substance doesnt exist? In the articles that I saw, the compounds were not commercially available. Hence, they had to be synthesized. The authors usually after reporting the NMR will then say the purity was confirmed by HPLC to be ..% pure. This is what I dont get clearly how they arrived at the % purity that is been quoted.

[Arkcon]

They have simply said, "It is 97% pure by HPLC."  That means the typical method for analysis doesn't show more than 3% impurities.  Now, by some other analytical method, it might be 90% pure, but they didn't clam that, so they don't have to test that.  Heck, just adding a mass spec detector could show their 97% pure peak has a significant mass impurity that's co-eluting.  Or running the HPLC on a chiral column could show its made of equal proportions of various diasteriomers, but that's up to them to decide if its worth checking.

Its possible to scan a peak under a PDA, and determine, within the limits of separation, if there is something under the peak that's distorting the spectrum.  Even something that doesn't absorb at the target wavelength is to distort the spectrum across the peak.  But I don't know if there's any evidence, in a research paper, that people are doing that.  But when developing a method meant to analyze a pharmaceutical, as an example, the regulatory agencies will want to see that.

[Furanone]

Thank you for your response @ Furanone and Babcock_Hall.

@Furanone: What will you advice in a situation where the standard or 100% pure substance doesnt exist? In the articles that I saw, the compounds were not commercially available. Hence, they had to be synthesized. The authors usually after reporting the NMR will then say the purity was confirmed by HPLC to be ..% pure. This is what I dont get clearly how they arrived at the % purity that is been quoted.

I am guessing if all you can get commercially is the 97% then there is a good reason for this, and that is usually to purify that extra 3% is not economically feasible (Law of Diminishing Returns = lots of extra fractionation steps for very little extra compound). This is why 95% ethanol is easy and cheap to get (distillation limit before azeotrope), but anhydrous ethanol costs you your first born. Another thing to keep in mind is that it is usually listed as >97% or 97%+ which means it could be slightly higher than the reported 97%.

Can you not use this as your standard and just take into account in your calculations that the standard peak is 97% of what a pure 100% compound peak would be to correct for your assay peaks?