[mir]

What physical property decides what response factor a compound got in a certain detector like:
1. Flame ionization detector
2. Electron Ionization-MS

?

You have a reaction with several biproducts that shows up in the chromatogram (along with your internal standard):
A --> C + D + E + F + noise

It should be possible to predict the response factor? Lets say, you could use a descriptor of some sort, like "average oxidation state of carbon" or "HOMO-LUMO Gap" to predict this?

The last detector I mentioned, EI-MS, gives Total Ion Count as response. This must surely be linked to how fast the analyte is decomposing after the electron-impacts. And today there is a lot of kinetical data around radical reactions, maybe they might help if you know your product? Lets say, you got a compound with a lot of oxygen, logically this compound might decompose really quick, to release carbon monoxide and CO2.

And logically the substrate (A) shouldn't be so much different from the product (if its a reaction with no huge changes in structure)?

Do you have any thoughts around this subject (response factors)?

[Custos]

I think it would be extremely difficult to determine response factors a priori from a samples chemical structure. There some "rule of thumb" factors such as halogenated compounds give higher responses in FID, but to quantitate that would be very difficult. The practical and much more accurate method is to use an internal standard of know concentration calibrated against the analyte of choice.

[mir]

But can you see, the tremendous potential inside this? Often in environmental chemistry, you know the structures, but not the concentrations. Most of the compounds isn't commercial available.

I find it very hard to believe nobody has done research at this? Or done some thought around it...?

Of course there is a problem with this: The response from every instrument are biased by noise from either the instrument or by the treatment of the sample. So, I guess a way of investigating responsefactors, must be done by a chemometric treatment to eliminate the noise.

There some "rule of thumb" factors such as halogenated compounds give higher responses in FID

You see: We can draw conclusions a priori.

[Custos]

You see: We can draw conclusions a priori.

Qualitative, not quantitative.

For environmental analysis the only reliably accurate method is to but or make the test substrate, and construct a calibration curve with a known internal standard. Your suggestion is intellectually attractive but I suspect the variables that go into making a response factor are too complex and interrelated to ever be calculated with great precision. Look at the example of logP calculations based on structure. These can be pretty good usually, but it's also not uncommon to be out by up to a whole logP unit from the measured value. For most purposes that may not be a problem but for applications where accuracy is paramount (environmental analysis) it's just not good enough.