Absorption and luminescence spectroscopy in rubrene

The optical absorption and photoluminescence spectra of organic molecular crystals depend on the optical properties of the molecules, on the way the molecules are arranged in the crystal matrix, and how the molecules interact with each other. The rubrene single crystal has a large optical anisotropy that has a strong influence on the absorption and luminescence spectra that are observed under different experimental conditions. Although transport properties of rubrene single crystals have been extensively studied, considerably fewer studies have explored their optical properties.

Rubrene , being an anisotropic molecular crystal, absorbs light very differently depending on its polarization inside the crystal, and similarly the spontaneous emission from its electronic excited states is very strongly polarized.

These features are very important because they mean that the absorption and fluorescence spectra that you are going to measure will change a lot depending on light polarization, and also on the angle of incidence of the light that you use for excitation, and how you collect the light that is emitted. Not only this, but there are some variants of rubrene crystals that sometimes grow in such a way that their fluorescence spectrum is characterized by a more or less strong absorption emission band (near 650 nm) that is not observed in intrinsic rubrene.

[Image: Luminescence in a rubrene crystal']

Rubrene crystal excited in the middle, emitting fluorescence all over the place.

In addition to all of this, rubrene crystals tend to grow with wide flat surfaces perpendicular to the c-axis, which is the axis that has the largest absorption, and the axis along which spontaneously emitted fluorescence is polarized. If such a crystal is put in a spectrophotometer to measure absorption in the same way one would do when measuring the absorption of some molecule in solution, then the resulting absorption coefficient will turn out to be weak, and to have the first strong absorption band corresponding to a higher vibrational level, because the direct transition to the lowest vibrational state is forbidden for light polarized perpendicular to the c-axis, as would almost always turn out to be the case if one illuminates the natural large surface of a rubrene platelet.

[Image: Single crystal Rubrene's absorption and fluorescence spectra']

These are the absorption and luminescence spectra of rubrene. Note that the absorption spectra are as shown in the figure, while the a- and b-polarized luminescence is so much weaker than the c-polarized one that the curves had to be multiplied by 10. From Irkhin et al, 2018

Several of the artifacts that can happen while measuring absorption and fluorescence spectra are discussed and demonstrated in our paper from 2012. The figure above summarizes the final results for the intrinsic absorption coefficient and luminescence of rubrene single crystals.

There are many notable observations here. One example is the shoulder in the b-polarized emission spectrum near 2.2 eV. This is what the spectrum should look like when 2.2 eV light is not also absorbed. But then things are complicated. A measurement that filters the purely b-polarized luminescence well, which implies using polarizers but also implies detecting only the light emitted perpendicular to the flat faced of a rubrene platelet, should deliver a spectrum that looks like what is given in the figure above. But if the crystal is not micrometer-thin and is illuminated with a-polarized light that reaches deeper into the crystal, then the higher-energy light that is emitted is attenuated as it travels towards the surface, and the 2.2 eV band is suppressed. However, if that measurement is performed by collecting the light through a high numerical aperture objective, then the 2.2 eV band re-emerges, because of the simple fact that such an objective collects light that is emitted in a cone from the surface, and therefore contains a small contribution of c-polarized emission, which comparatively enormous in rubrene, gets added to the spectrum, and then one again sees a shoulder near 2.2 eV, because of the intrinsic c-polarized emission that is not blocked anymore.

[Image: oscillating spectrum when going through a surface defect']

Effect of a surface defect when measuring the fluorescence spectrum of a rubrene crystal. The movie shows what happens when traveling with the detection spot over a surface defect while measuring the luminescence spectrum on the natural large facet of rubrene crystals.

Another example is what happens when there is a scratch or a small defect on a rubrene crystal excited from the facet perpendicular ot the c-axis. Any excitation through that fact will create excitons deeper inside the crystals that will emit most of their fluorescence polarized perpendicular to the surface. This light normally never reaches the detector. What we detect instead is the small amount of b-polarized light that is emitted because the selection rules are relaxed when going from a higher vibrational state to the ground state. But as soon as the surface is not perfectly flat anymore, any c-polarized light that would otherwise undergo total internal reflection at the surface will be scattered by the defect. The defect glows bright orange and the 2.2 eV shoulder of the rubrene spectrum becomes very, very prominent. The movie shown to the right is a simulation based on the actual parameters of the emission spectra, but what actually happens in the experiment is very very similar to this simulation.

And so on... a lot of things need to be considered and done right in order to obtain accurate fluorescence spectra of rubrene single crystals. It also turns out that observing the luminescence spectrum of rubrene is a very good way to observe the effect of even small quantities of impurities or crystal defects. This despite the fact that defect concentration can correspond to less than one molecule out of thousands of intrinsic rubrene molecules. But to understand how this is at all possible (one naturally asks, why wouldn't the luminescence from those thousands of intrinsic molecules completely overwhelm what defects can do?), one needs to talk about the excitons that are created by light absorption, and that do not really immediately go back to the ground state emitting photons. Instead, they prefer to split into two triplet excitons, that do not emit light individually but can meet later on to then merge again into a singlet excitons and only then emit the luminescence that can be detected. This process is already dominant in rubrene when using conventional laser beams of just a few mW (see here) to excite the luminescence, and since the triplet excitons can travel far and wide inside the crystals (see here), the fluorescence that they emit can be sensitive to them encountering a defect.

Photoluminescence and absorption spectra of single crystal rubrene