In contrast to GaP, InP is a direct band semiconductor. In this connection it is interesting to study the similarities and difference between luminescence properties of porous InP and GaP.
The most extensively studied optical property of porous InP is the photoluminescence (PL). These investigations were motivated by several reasons.
Figure 4.10: a) SEM and b) panchromatic CL images in cross-section taken from a porous InP structure with spatially modulated degree of porosity. The insert shows the interface between two neighboring porous layers. In contrast to n-GaP, here the bulk material shows much stronger luminescence than the porous layers. Also the porous layers with crystallographically oriented pores have a higher CL intensity than the porous layers with Current-line Oriented Pores (the white thin lines on the SEM and CL images are porous layers with crystallographically oriented pores).
The samples used for these experiments were (100)-oriented S-doped n-InP with a free carrier concentration n=10e18 cm-3. The CL was excited with a continuous 15 keV, 0.25 nA electron beam at normal incidence and was collected from 125 Ám2 regions to reduce electron beam induced effects.
Figure 4.10a shows the SEM image in cross-section taken from a sample subjected to anodization in 5 % HCl electrolyte. In order to modulate the degree of porosity as a function of depth, the anodic etching current was periodically switched on and off, i.e. the anodic current changed from 80 mA to zero and vice versa with a period of 0.2 min. The etching takes place only during the first 0.1 min of the cycle, while in the second 0.1 min no dissolution occurs (no current flows). However, the process interruption for 0.1 min proves to be very important because in the meanwhile the pore tips become passivated. When the current is switched on again, a new nucleation phase is required, i.e. a new nucleation layer (NL) emerges similar to the NL related to the first stage of the dissolution starting at the initial surface of the sample. Applying this procedure, a structure was obtained with a spatially modulated porosity consisting of 14 alternating layers of high and low porosities (Figure 4.10a). Although neighboring layers exhibit quite different morphologies and degrees of porosities, the interfaces prove to be rather sharp, see the insert in Figure 4.10a.
The porosity relief caused by periodical switching of the dissolution current on and off was found to give rise to spatial modulation of CL intensity (Figure 4.10b). Note that the higher the degree of porosity the lower the CL efficiency. The spectra presented in Figure 4.11 also show that the CL intensity from the porous regions with Current-line Oriented Pores is nearly three orders of magnitude lower than that from the bulk n-InP and two orders lower that the intensity from regions exhibiting crystallographically oriented pores. This decrease in CL intensity with porosity can be understood if it is supposed that the non-radiative recombination centers (mainly at the surface) in porous n-InP are not passivated by the species in solution during the electrochemical etching and they also cannot be passivated in air, e.g. by oxide formation. Therefore, the rate of non-radiative recombination of free carriers increases with porosity (surface exposed to the air), i.e. the number of surface defects increases with the degree of porosity.
Figure 4.11: CL spectra of bulk InP (squares) and porous layers produced at low current densities ( crystallographically oriented pores, spheres) and high current densities (Current-line Oriented Pores, triangles) in 5 % HCl electrolytes, T = 80 K.
Both the degree of porosity and the transverse dimensions of pores and skeleton walls can be effectively modified by changing the electrolyte composition. In particular, in 5 % HCl electrolyte the transverse sizes of pores and InP wall thickness is equal to 200 and 120 nm respectively, leading to a degree of porosity of about 55 %. The SEM top view of sample taken after removal of the NL shows the spatial distribution of pores to be uniform (Figure 4.12).
Figure 4.12: SEM micrograph of the top surface of an InP sample after anodization in 5 % HCl electrolyte with subsequent removal of the NL. The inserts show cross-sectional SEM images taken from samples anodized at (a) 5% and (b) 10% electrolytes respectively. The pore walls and the pore diameter can be reduced if the concentration of the electrolyte is increased.
According to the inserts a) and b) presented in Figure 4.12, doubling the electrolyte concentration results in a nearly 3-fold reduction of the skeleton wall thickness. This in turn further reduces the intensity of luminescence. As seen from Figure 4.13, the intensity of the near-band-edge CL decreases 100 and 4000 times after sample anodization in 5 % and 10 % HCl electrolytes respectively.
Please note that at 10 % HCl the pore walls are in the range of 50 nm. In this range the quantum size effects can already be expected. Indeed, a small upward shift of the frequency (?5 meV) of the CL band maximum in the sample anodized at 10 % electrolyte in comparison with that of the as-grown specimen is found. One should be careful when intending to increase the concentration of the electrolyte in order to decrease the thickness of the pore walls. This is caused by the fact that by increasing the concentration the number of undissociated HCl molecules increases as well, which can etch chemically (and uncontrollably) the pore walls and thus destroy the porous structure.
Figure 4.13: CL spectra of bulk InP (curve 1) and porous layers produced by anodization in 5 % (curve 2) and 10 % (curve 3) HCl electrolytes. T = 80 K. The CL decreases as the porosity is increased. A frequency shift at high degree of porosity is observed (curve 3) which can be attributed to quantum size effects.