This modification of the NW diameter distribution affects the luminescence properties of the ZnO NWs changing the contribution of the surface luminescence regarding the band edge emission. Shalish et al. [47] observed that the relative intensity of the UV photoluminescence peak was stronger, and the TPCA-1 visible luminescence becomes relatively weak as the size of ZnO NWs increases. They explained this size effect

in terms of bulk-related to surface-related material-volume ratio, assuming a surface layer thickness, t, wherein the surface recombination probability is 1 high throughput screening assay [47]. The intensity ratio defined by Shalish is as follows: where C is a fitting parameter learn more accounting for the efficiency of the bulk-related emission process relative to the surface and r is the wire radius. The UV-visible luminescence intensity ratios (I NBE /I DLE) calculated in our samples from the PL curves of Figure 2 are presented in Figure 8

as a function of the average wire radius (deduced from the C-TEM statistical analysis). In our case, the best fit is obtained with C = 5.8 and t = 30 nm, and Figure 8 also includes data from Shalish et al. using C = 2.3 and t = 30 nm. The trend in both is very similar with the same surface layer thickness, i.e. an intensification of the UV/visible ratio as the wire diameter increases. The ratio exhibits a clear escalation for thicker NWs (6.6 and 9 for the GNA12 irradiated NWs with fluences of 1.5 × 1016 cm−2 and 1017 cm−2, respectively). The differences of the C parameter (between our results and those of Shalish) only mean that the efficiency of the bulk-related emission process regarding the surface is higher in our case. Those discrepancies

can be explained by the fact that the compared NWs have been grown by different methods and undergone different treatments, and therefore, it is expected that they initially present different luminescence characteristics since surface state densities are notorious for their great variability. Figure 8 Experimental luminescence peak intensity I NBE / I DLE as a function of the average wire radius. Values predicted by Shalish’s data are also included. Nevertheless, if the visible emission is supposed to be mainly originated from defects related to the surface, other factors apart from the annihilation of the thinnest NWs might also be considered. Both μPL and CL data reveal an enhancement of the UV/visible ratio with the increase of the irradiation fluence. Certainly, a reduction of the point defect density in the surface would also result in the UV emission enhancement as a consequence of a net reduction of the visible emission.