Many reactions may be involved after UV irradiation, e.g. radicals formations and exchange with D2O.

To see the structure of the irradiated compound may be of help in the understanding what changes could be introduced by irradiation. Also to see the whole spectra could be of help.

Thanks, I have attached the structure. The only thing missing is a big solvent peak. I was looking for signs of photoisomerization but I'm pretty sure this isn't happening here.

I don't have a specific knowledge of this kind of problems, but in general, have you considered that it could just be the effect of electronic transitions? If you irradiate with UV and the electrons get promoted to some higher state where it takes them time to relax back, qualitatively the simple change in local electronic density could be enough to cause a change in the chemical shift. I don't know what the quantitative entity of such an effect would be though (I'd say it would be highly dependent on the shape of HOMO and LUMO for this specific molecule).

The structure has an extensive conjugate double bonds system and may be that Dr Sturniolo has hit the mark. To confirm this, you may run the NMR experiment after an adequate time, e.g. 24 h after the irradiation and see if you still have the same shift or not. This could also avoid the peak shift due to the different temperature of the sample after irradiation with UV. In fact also the temperature of the sample may cause shift in the registered peak and this feature is largely used for dedicated purposes.

I doubt that excited electronic states that persist for seconds. Did you measure the pH (or pD) of you sample before and after irradiation? Do you have any buffer etc in your NMR sample?

Thank you for the responses. The excited state should have relaxed by the time the NMR measurement is made (the sample is irradiated outside the spectrometer). Would temperature variations shift all of the peaks? I haven't measured the pH but I did measure the UV-Vis absorption and steady state emission, both showed a small decrease in intensity upon irradiation. My sample was just the chromophore in D2O.

This may sound obvious, but have you set the TMS signal at zero ppm in both recordings?

Another possibility is double bond isomerization (trisubstituted d.b., from Z to E). That way the appearance of the spectra (H-H coupling constants) stays the same, only shifts would change slightly.

Measure the coupling constand between the C=O carbon and the proton attached to that double bond. If it's different after the irradiation, then the configuration changed.

Well, if there were some forbidden states in the direct relaxation path excited electronic states might have much longer lifetimes. Just think of those phosphorescent paints used for glow-in-the-dark toys. But it's just a wild suggestion, I guess there are more likely causes (also, if that was the case, you would probably see two peaks for each original peak, unless the sample was fully excited and had not relaxed in the least).

The spectra have been set to zero. I will go and check the coupling constants. I thought isomerization was unlikely as I didn't expect 100% conversion to E so would have thought there would be a trace of the Z isomer.

From transient absorption data I am pretty sure that the excited state would have fully relaxed.

Ok, a few things:

1) what nucleus are we looking at here? H, D, 13C? If it's D, is the sample fully or only partially deuterated?

2) do you have a reliable assignment for the peaks? Or even just a rough guess?

3) what are the two peaks to the far right? They don't seem to shift. Also there's something at 2.85 ppm that seems to disappear altogether.

As general observations go - assuming this is not some weird instrumental artifact caused by your specific setup, it's interesting to note that the peaks shift, not split. If we had two isomers and the UV was causing a transition you'd expect the peaks of the isomer to pop up next to the original ones, and the ratio of intensities proportional to the ratio of conversion. This is not the case here. Did you observe any saturation in this peak shift with increasing UV absorption times? Did you try lowering the irradiating frequency and seeing at which energy the phenomenon disappears? I guess the formation of radicals could change the shifts, but it still feels like that should look more like the formation of multiple species (= peaks splitting) than a continuous shift. About temperature, maybe it could influence the solvation conditions of the molecule. That would change the shifts as well.

Kiri

If I look at your spectra there is something strange. The dark spectrum (green) has a lot of small peaks that may indicate degradation to some photo reaction. BUT the spectrum (red) which you label as the one recorded AFTER UV irradiation does not have these small peaks. I would expect it to be the other way. Are you sure about the labels?

Markus - Thanks for spotting that, the labels are incorrect. As you suggested, I thought the noise might indicate some kind of degradation or possibly aggregation.

Simone - The nucleus is H. I have attached the assigned spectrum, I thought the peaks on the far right might be acetone.I did not see any saturation in the shifting when increasing the irradiation time. I haven't tried lowering the power yet but will do, thank you for the idea.

If you see degradation, the formation of radicals in solution appears most probable. In this case you should consider paramagnetic shifts due to solved radicals in your solvent. You could try to irradiate your solution for twice the time and should see a higher shift, and so on... Just in case you could try to add a small amount of radical catchers to your solution and the shifts should vanish.

Also, adding radical initiators should additionally increase shift difference. AIBN would be a great choice since it would add only one proton signal at low δ, but I doubt it is water soluble. There are other water soluble initiators though.

ESR spectrum would also be of great help.

Thank you very much. I did see a higher shift with increasing irradiation time. I will try adding some radical initiators/inhibitors.

Dear Kiri, I work in the field of Electron Paramagnetic Resonance: I think Peter Kaden hits the right answer. Radical formation is very likely under UV irradiation of highly conjugated molecules, which results in paramagnetic shift in NMR spectra. Lowering of the NMR peak intensity is also a natural consequence (because of NMR line broadening).

Thanks Alfonso

I would guess your compound can form stable radical - in position which you labeled by 7: benzyl position next to imidazolon. Phynyl has hydroxyl in para position; radical can easily delocalize through the double bond to imidazole ring. Initially you form biradical - on 7 and 8. Radical in position 8 could be delocalized and shows in position 21, where it can dimerize; you can oxidize phenol to quinone - in this case radical in positon 8 would stay. Some signals would disappear in the radical.

As already suggested, some radical scavengers and initiators would be great. Try TEMPO, AIBN...

If you measure IR you might be able to see if quinone was formed. Although you should see it also in NMR - the only question is how much of the compound you have.

If you irradiate the compound for long enough, all the compound should be converted.

I thinck that the pH of the solution change during the irradiation and so the shift of HDO an of your product.

Thank you

I noticed that you have a stronger shift effect around proton 15 (and 17). Therefore, you have probably a local effect. It is difficult to know with 1D 1H NMR where your compound has been modified. You could have any modifications such as protonation, oxidation, addition, dimerization...

One of the hypothesis is modification at position 9 (protonation) which could result in a difference in the local environment.

One of the simplest method to know if you have the same compound is to reproduce the same physical conditions for both samples and measure again a 1D 1H NMR spectrum. Just sacrifice a small amount of your sample obtained after irradiation by adding it inside the tube of your original sample (without irradiation). If you obtain a unique signal pattern, it is the same compound!

If you have two patterns (so, two different compounds), I suggest to do : (1) MS spectrum to verify if your total mass have changed with irradiation. (2) 13C NMR and 2D 1H-13C NMR experiments (such as HSQC and HMBC) to better localize the modifications.

I hope it will help you.

Thanks Gildas

Kiri: all comments seem adequate to consider. Try a simple TLC after irradiation to check for molecular degradation/change of the original product. Good luck. Raul