Annealing effect on PL efficiency

[sttan]

I've some ZnO films grown on sapphire substrate. The ZnO film, however, is a highly-oriented polycrystalline film. The annealing was carried out at 800'C at different ambient, (N2, O2, etc.)

After annealing, as expected, the crystallinity improved (narrowing FWHM of XRD). Resistivity increased (due to the well-developed grain boundaries after annealing and annihilation of some shallow defects). However, the PL results puzzled me. I integrated the PL spectrum (radiative emission efficiency) of the as-grown and annealed samples. I found that the efficiency degraded after the annealing. This is abnormal phenomenon. Generally, at this annealing temperature, the sample has better crystallinity and PL efficiency. Of course, i know that the effect of annealing depends much on the ZnO films such as stoichiometry, etc. But it is generally accepted that 800'C annealing will improved stuctureal and optical properties of ZnO. I'm wondering why my batch of samples show different characteristic.

Here is my speculation, the degrade of PL efficiency is obviously due to the increase of non-radiative recombination centers (defects). Since the FWHM of the XRD narrowed after the annealing, most of the non-radiative defects are generated at or near the grain boundaries. Hence, i get better crystalliinity of the films after annealing, while degrade in PL efficiency due to the creation of non-radiative defects. Am I right to say so? Is there anyway to prove my speculation?

 

[Gokul43201]

Here is my speculation, the degrade of PL efficiency is obviously due to the increase of non-radiative recombination centers (defects). Since the FWHM of the XRD narrowed after the annealing, most of the non-radiative defects are generated at or near the grain boundaries. Hence, i get better crystalliinity of the films after annealing, while degrade in PL efficiency due to the creation of non-radiative defects. Am I right to say so? Is there anyway to prove my speculation?

It would seem to me that a reduction in the FWHM can come about as a result of strain relaxation after annealing. With lower strain gradients, you have sharper parameters and hence narrower peaks. If you do an analysis of the peak widths* (before and after anneal), you might be able to isolate the effects of the broadening.

* See my very brief discussion of XRD peak broadening in the recent XRD thread by anil.bose

 

[sttan]

Hi Gokul, thanks for your reply.

I do understand that the narrowing of XRD FWHM could be due to the increase of grain size and/or strain relaxation after annealing. However, for polycrystalline films, the grain size effect could be dominant over the strain relaxation effect. In fact, I have carried out a study at my previous batch samples (with various XRD peaks, typical polycryrstalline films), while the annealing was done at lower temperature of 500'C. Williamson-Hall graph was plotted to estimate the strain relaxation after annealing. It was found that the strain relaxation is very small and hence its effect on FWHM narrowing could be negligible compare to grain size effect. Also, the grain size of my films was estimated to be in the range of 50 ~ 60 nm using Scherrer's formula, which is considered accurate way of grain size estimation for polycrystalline films.

Back to my question again. If the narrowing of FWHM is mainly due to the ripening of the grains. Is my speculation correct? Is my statement convinced enough?