pH and UV response

[lcollard]

Hello all,

Is it possible that pH influence the UV response of a molecule ? ;

I've try to analyse a compound and an impurity in HPLC (230nm) with acidic conditions (H3PO4 additive) and then with basic conditions (NH4+HCO3-).

It appear that for the same sample i can see 0,1% Area of the impurity with acidic conditions and 1,5% Area of the impurity in basic condition.
It is very strange...I don't think that theses additives "eat" my UV response.
Could it also be possible that my compound deteriorates during the analysis in this impurity due to my eluent pH ? (= more concentration and then more response)

Thanks all for your help

Laurent

[Noser222]

Though not the only possible cause, pH can definitely affect the UV absorbance.

[zokitano]

For example take chromate/dichromate or phenolphtaleine (indicator) and record their UV spectra in solution with low and high pH. You'll get the difference.

pH change can dramatically alter the UV spectrum of compounds.

Regards

[tom jupille]

pH change can dramatically alter the UV spectrum of compounds.

Which is why area% (by itself) is a terrible way to measure purity!

[mbicking]

Which is why area% (by itself) is a terrible way to measure purity!

Well, said. I have seen this approach for years, and never understood why it still persists. Assuming that everything has the same response factor as the active is convenient, but often/usually wrong.

Is there no effort to correct this?

[alicee!]

Which is why area% (by itself) is a terrible way to measure purity!

Well, said. I have seen this approach for years, and never understood why it still persists. Assuming that everything has the same response factor as the active is convenient, but often/usually wrong.

Is there no effort to correct this?

agreed.. but i havent been able to find another way.. i mean.. i dont have quantified purified impurities to run along.. to be able to compare area values and quantitate the impurities in the sample.. how do u do it? i have been quite looking for the answer to this!! thankss..

[mbicking]

The area % method is very common, although it often produces wrong results. However, it is accepted by many regulations, so you can use it.

But you must be very clear what conditions are used for your analysis. In your case, specify the pH.

Maybe some others with regulatory experience have a better suggestion.

[HW Mueller]

It´s accepted?? There was a discussion not too long ago that appeared to aslo suggest that, but then there was a citation of an article in which some nebulous response factors were mentioned . . . . ?

Incidentally, it has been discussed by several participants how this is done correctly: You have to isolate impurities and characterice them (in the physical organic chemistry lab where I did my Ph.D. we had to send out samples for elemental analysis, besides NMR, UV, IR, MS. . . )

[mbicking]

By "accepted" I do not mean to imply "acceptable." But these analyses usually are satisfying regulatory requirements for impurities in API's. Enough said?

The ones I have reviewed closely (I don't work directly in the industry) will usually be described as a method for "related substances." This allows you to make the "assumption" that since they are all related, they must all have the same response factor (yeah, right, I know). If you have an actual standard, you can make an adjustment in the calculations, but the small peaks just assume the same RF, since the amounts are too small to collect and analyze. I think these days you get the MW from MS and let it go at that.

As Tom said, from a scientific standpoint this is a terrible approach. But it is easier to understand, I guess, and certainly cheaper than doing all the other work.

It would be interesting to see if RF's for a large number of compounds were more consistent at, say, 205 or 210 nm, since just about everything organic absorbs in this range. Then the peak area % method might be somewhat less inaccurate.

[juddc]

It would be interesting to see if RF's for a large number of compounds were more consistent at, say, 205 or 210 nm, since just about everything organic absorbs in this range. Then the peak area % method might be somewhat less inaccurate.

I would expect that one would not see any greater consistency there, either. Considering that the number of unconjugated double bonds and Mol Wt of coupounds can vary just as greatly as anything else, I'd not expect any more consistency

[HW Mueller]

Even though almost everything absorbs near 205nm, there are still huge differences among different substances. Also, as Tom pointed out most recently, there is a great possibility that one has a steep slope there.

mbicking, I really meant to ask whether this "accepted" is due to a misunderstanding of regulations or whether they really "allow" this. If they really do, I just want to restate that I disagree emphatically. One should remember that sometimes slight variations in a molecule decide between activity or inactivity, etc., in vivo. Thus, one can imagine that a "trace impurity" (or one that is not seen at all) might be in reality a hefty % of the total and be the "carrier" of activity, rather than the active.

Again, it is surprising that some people are extremely frugal here in the face of enormous expenses that follow in form of the animal and hospital trials.

[mbicking]

HW M:
Well, again, I do not spend all of my time in pharma labs, so I may not be the best source of info, but the methods I have seen are making the assumption that the response factors are all the same, unless there is a standard available, which allows you to input a different RF for that peak. These are registered API's, so I know the information has been through regulatory review. So, the general approach is "accepted."


Anyone who knows anything about absorbance spectrophotometry should recognize the dangers of this approach, but it is certainly the easiest way to get values. Perhaps someone in the industry could shed some light on the history and justification, but I believe this approach has been around for many years. The "related substances" approach is probably the argument that will be used; at least is has some scientific validity.

[HW Mueller]

Even the assumption that a synthesis yields only related substances is quite bold.