Fluorescence fundamentals (?)

[Mattias]

I have work a quite long time with LC, but have just recently started to use a fluorescence detector. I have probably forgotten some of the fundamentals.

I have an impurity peak that is able to fluorescence. I started be recording the DAD-spectrum of the impurity in UV. Then I set the excitation wavelength to the found UV-max (220, 253, 270 nm), one at the time. Then I made a scan of the emission when the impurity was in the detector. There is a scan function for the excitation wavelength too, but then I had two unknowns (excitation and emission wavelength).

I get emission spectra that look OK, but they all look the same!

Questions:
1: Is it correct to use the UV-max as the excitation wavelength?

2. Why do I get the same emission bands for all excitation wavelengths?

3. I think the noise is quite high (about 0.5 mV at the lowest gain). It is worse S/N then the UV signal. (The lamp is brand new). My idea was to get fantastic QL . Can acetonitrile be a problem? Any good ideas to reduce the noise?

Thanks!

[danko]

Hi Mathias,

What are the emission maxima for each of these 3 excitation wavelengths, you fond?
If the compound in question is a large molecule, it just might be light scattering you’re witnessing.
By the way, if it shows that the acetonitrile was causing the S/N problem, then it must be badly polluted, because it normally doesn’t fluoresce or scatter.

Best Regards

[Mark Tracy]

Yes, using the UV maxima is the best place to start for excitation. I'm a bit surprised that the emission spectra are the same, they should have the same maxima, but not necessarily the same intensities. You should always avoid emission wavelengths that are integral multiples of the excitation wavelength, because the n-order effects in grating monochromators; all you will see is the scatter. After you have a nice emission setting, go back and optimize the excitation, usually it is slightly different than the UV maximum. Also, the greatest intensity is not necessarily the best S/N. Finally, the best combination of wavelengths is hardware-dependent.

Your operator's manual should have a S/N test you can do. The test uses the Raman scattering of water, so you don't need any special preparation for the test other than good H2O.

[Mattias]

Danko: the emission maxima are 292, 373 and 575 nm. The molecule is quite small, but contains three phenyl groups.

Mark: The intensities are not the same, but the autoscaling of the spectra makes them look the same. The emission peaks are very broad, the peak at 292 nm is almost 100 nm wide (maybe this is normal?)

I did the S/N test in the manual (Dionex detector RF2000), and it passed. But isn't the noise limit set very high in that test?

[HW Mueller]

Mattias, probably you have noticed by now that a review of fluorescence would be a good thing.
Just a few more, hopefully helpful, points.

The fluorscence yield can be very strongly dependent on polarity of the medium (solvent), especially of nonrigid molecules. The stuff may fluoresce very strongly in hexane, little or not at all in ACN. In other words, the absorbance might be a more sensitive detector than fluorescence even of a known fluorescing substance, hence a lower S/N.

To fluoresce the molecules have to absorb light, the more absorption the more fluorescence (if it fluoresces at all).

Scattering and fluorescence are easily distinguished from each other: The scattering wavelength moves with the excitation wavelength, fluorescence does not as you saw (are they called Jablonski diagrams that show how the fluorescence occurs from the lowest vibrational state of the electronically excited state, therefore always giving the same spectrum shape?).

The higher order fluorescence, mentioned by Mark, can be a real nuisance, watch for it, don´t be fooled by it.

[AdrianF]

Something to bear in mind, flourescence of some compounds is quenched by dissolved oxygen. It is vital to throughly and continuosly degas to get consistent results.

Reagents and water should also be of the highest purity available to avoid background fluorescence.

[Mark Tracy]

The emission spectrum can be very broad or very sharp. There are some commercial fluorescent dyes that emit nearly white light; they are the 'brighteners' in laundry detergent. FMOC derivatives, on the other hand, have a very sharp emission band. Don't worry about it.

[danko]

Another thing one should take in account is that the emitted light can be absorbed by the very same solute (fluorescence energy transfer). In your case, if you excite your molecule at 220 nm, it will emit light at max 292 nm. But it also will absorb energy/light in this region (max 270 nm). So my advice would be: Don’t use 220 nm as an excitation wavelength, although you might find the strongest absorbance at this wavelength.
Generally, fluorescence is a very useful technique, but one has to be more watchful compared to the pure absorbance measurements. As mentioned above, there can be quite a few factors to consider, such as range, linearity, temperature dependency, solvent dependency (also nature and concentration of other solutes) and more. It has many advantages too, but you should need them badly in order to go for fluorescence.

Best Regards

[grreddy]

is there any articles on fluorance detector

g.r.reddy.

[HW Mueller]

Any book on fluorescence will do, also instruction manuals for fluorescence detectors/spectrometers can be very helpful.
Selfabsorbance of fluorescence is not a problem if one uses an emmission line at which the molecule doesn´t absorb (I can not remember any substance we ran where the emission max was on top of an appreciable absorption tail).

[Mark Tracy]

Check out Molecular Probes: http://probes.invitrogen.com/

[Mattias]

Mark> Thanks for the tip, I bought the Amplex Red kit from Molecular Probes and it works beautifully. The samples are stable for days!

The only problem I had was that the reaction was inhibited by povidone. It was solved by extracting the samples in toluene.