When absorption and excitation spectra do not match, there is a severe problem. There can be a number of reasons. First: sample impurity. That has to be checked first. Absorption spectroscopy alone often is not sensive enough, depending on relative fluorescence quantum yields of sample and impurity and/or differences in absorbance at the excitation wavelength. Keep in mind that spectral calibration is not a simple issue. Also photoproducts can lead to difficulties. Check as a function of irradiation time.

ideally excitation spectrum should be a surrogate of the absorption spectrum i.e. they should be identical/superimposable provided the emission to which the excitation belongs originates from the same ground state species to which the absorption belongs. However, it may so happen that emission actually comprises of overlapping emissions from different species which originate from different ground states, in that case excitation may be different from absorption. Before estimating Stokes shift, I suggest you do a scanning emission using different excitation wavelengths in order to differentiate the different ground state species if at all present. Or you can do a scanning excitation spectrum for different emission wavelengths to confirm the same. I feel your compound has different ground state species which should be assigned first. After which Stokes shift can be estimated from the absorption and its corresponding emission.

Hi, Rupashree! Thanks a lot for your kind answer! At first I have the same idea as yours. It is strange that when I used the two different excitation wavelengths including 260 nm and 300 nm to excite the compound, the maximum emission intensity always appears at 350 nm. It seems that the emission under those two excitation wavelengths comes from the same electronic transition, which indicates that there are two excited-states for it. But the excited-state which needs 260 nm light does not like emitting too much photos at recorded wavelength (400 nm).

I would look at Stokes shift (SS) as an excitation-wavelength-dependent parameter. That is, there is one value of SS corresponding to 260 nm and another one for 300 nm. From the point of view of the intended application, whatever it is, you'll hardly be able to freely choose the excitation wavelength and adjust it to the maximum of efficiency, but rather deal with available UV LEDs, excimer lasers or discharge lamps. However, if you want to compare several different compounds, measuring SS from the maximum of the excitation band makes more sense if you compare identical transitions, be it excitation via charge transfer bands or transition to the same excited state. In any case, you should not be surprised that absorption spectra and the spectra of emission do not fully coinside. First of all, most of the fluoresence spectrometers are quite unreliable when you measure around 250 nm, so that the correction of the spectrum does not bring much confidence. Second, you normally measure excitation of luminescence in reflection, not in transmission, which may also cause certain disagreement. Is you sample powder-like? If it scatters light strongly enough, and the refence sample is a simple blank substrate, then you strictly measure not the absorption but absorption + scattering, which again may influence the result. Furthermore, do you correct absorption for luminescence? If you don't, you have a following situation: while measuring absorption near the excitation peak, the detector of your absorption spectrometer collects not only the transmitted light you intend to measure (at the excitation wavelength) but also excited light (luminescence emitted by your sample). The absorption peak apears weaker than it actually is. Finally, additional absorption maybe due to some impurities in your sample, which absorb exciting radiation but do not emit any light. Who knows, this list of what could go wrong is not exhaustive. You should carefully check if you appropriately perform your measurements because most of the "wonders" are just the result of inappropriate measurement conditions or mistakes in the analysis of the spectra.

Hi Roman! Thanks a lot for such detailed deduction!

Actually, my compound dissolved in solvent. If this situation that both transmitted light and excited light were measured by UV-photometer results in the distortion of adsorption spectrum, will the UV photometer equipped with PDA detector gives a correct absorbance curve?

You see, if you would provide more details in your question, the discussion could be better focused on important issues. Once your detector measures spectrally resolved intensity, the correction for luminescence should be quite straightforward and is probably done automatically. I have never worked with such absorption spectrometers but maybe you'll find some information in the manual of your equipment. After all, if you have a diluted sample, the contribution from luminescence could appear negligible. Scattering could also be ruled out. What is left are possible non-luminescent impurities (assuming that the cuvettes you use do not significantly absorb light) and insufficient sensitivity of the fluorescence spectrometer in the UV-range. If the agreement between the spectra of absorption and excitation of photoluminescence cannot be sufficiently improved, I would rather use the peaks in the excitation spectra, not in absorption.

Hi Roman! I used very diluted sample (~1e-5 M). I worked it out just now. Since my compound is relatively big. This 260 nm is mainly absorbed by one part of the molecule whose radiative rate is very very low. The 300 nm is mainly absorbed by another part which has a large radiative rate. When I excite the molecule with 260 nm and 300 nm, emission peak appearing at 350 nm both results from the part 2 of the molecule. This is related to impurity theory you suggested. I will also choose maximum wavelength in excitation spectrum for stokes shift calculation since they are both caused by the same electronic transition of part 2.

I agree with the messages of Roman.

Before you start your investigation, you must always have a correct elemental amalysis of your compound (+ or - 0,1 % of the calculated value) and correct NMR and MS values (Structure proof).

Without these values, you cannot work in a scientific manner.

Very important:

You must use spectroscopical pure solvent and it must have A CUT OFF wavelenght allowing the absorption/Emission a.s.o...

A list of solvents with related cut off wavelenght can be found in internet or companies.

When absorption and excitation spectra do not match, there is a severe problem. There can be a number of reasons. First: sample impurity. That has to be checked first. Absorption spectroscopy alone often is not sensive enough, depending on relative fluorescence quantum yields of sample and impurity and/or differences in absorbance at the excitation wavelength. Keep in mind that spectral calibration is not a simple issue. Also photoproducts can lead to difficulties. Check as a function of irradiation time.

You should first check the absorption of your solvent versus air.

If some 260 nm is absorbed by your solvent deviations appear in the excitation spectrum.

The absorption over the whole range should be less than 0.1 or your excitation spectrum will deviate.

I encountered similar problems during my PhD and wrote a technical note about it:

http://dare.uva.nl/en/record/129375

See also last reference note to test anthracene in cyclohexane.

Hi, Klaas and Rene! Thanks a lot for your kind suggestions! I think my problem may come from my solute itself. This molecule contains two main groups. They can absorb 260nm and 300nm light, respectively. But only the second group can emit 350nm light. So the adsorption spectrum is different from the excitation spectrum. How do you think?

Without knowing the structure of the molecule you're working with I can't be certain, but if your 260 nm absorber has a fluorescence emission spectrum that overlaps the absorption spectrum of your 300 nm absorber then intramolecular FRET just became an option.

In this case, since you're really dealing with two fluorescent species with different quantum yields, it is highly likely that the excitation and absorbance spectra would not match. This is because the absorbance comes from two species, but the emission only comes from one species.

Hopefully this helps.

Thanks a lot for your kind answer! It really helps me!

Hi all, thanks for having a discussion on this topic. I am also facing similar type of problem with sensor molecule. Can anyone provide me any reference on sensor molecule having different UV-VIS absorbance and photoluminescence excitation value..