relative response factor

[martinamarie]

Hi, I have an impurity, that is analysed at a different wavelength to my active. Can I still calculate the RRF using the slope of the active and the impurity at each wavelength.


Thanks

Tina

[moraespack]

Tina, if you only want to estimate the amount of the impurity, I think you can choose the best wavelength between both compounds.
If you already know what impurity is , another alternative is to look for a standard that is most similar as possible to your impurity and quantify it. Someone would criticize me, but if your impurity is about 0.1 % for example and you are only estimating, if you have an error of 30% in the amount of impurity, ís not a big deal if the result is 0.10 or 0.13% when you are estimating.
I hope it helps,

Marcelo

[martinamarie]

thank you, I will try

tina

[krickos]

Tina, if you only want to estimate the amount of the impurity, I think you can choose the best wavelength between both compounds.
If you already know what impurity is , another alternative is to look for a standard that is most similar as possible to your impurity and quantify it. Someone would criticize me, but if your impurity is about 0.1 % for example and you are only estimating, if you have an error of 30% in the amount of impurity, ís not a big deal if the result is 0.10 or 0.13% when you are estimating.
I hope it helps,

Marcelo

In short, I would not recommend that approach (error of 30%) when doing pharmaceutical impurities (I think martinamarie does that) with regard to the ICH Q3 A+B guidelines trigger points for identification and qualification (clinical trials) of impurities.

[HW Mueller]

This has been discussed many times. If one has no data about an unknown, a guess could be off0%, 30% or a 1000% . . . ., how would you know?
I don´t understand the original question as I don´t see any problem with quantification with analytes being measured at different wavelength.

[lmh]

if an impurity is analysed at a different wavelength to the active compound, then it is a chemically different group that you are looking at (or at least a modified group), and there's no reason to expect the extinction coefficient of group A at wavelength 1 to be anything like that of group B at wavelength 2.

It would be wrong to use the response factor (related to extinction coefficient) of one group at one wavelength for the other group at the other wavelength and expect any sort of accurate estimate of purity (not even in the gross ball-park range, let alone tens of percent). I'm very much afraid that HWMueller is so right!