It was shown above that the pores obtained during the anodization of n-type InP, GaAs, and GaP at low current densities have definite crystallographic directions of growth, namely <111>B directions. This section describes the pore formation in these materials under high current densities.
Figure 3.16a presents the cross section micrograph of a (100)-oriented n- InP sample anodized at U = 10 V (the resulted current was >10 mA/cm2). Pores obtained under these conditions obviously have no specific crystallographic orientation, but simply follow some curved lines. In the middle of the anodized sample such kind of pores are mainly oriented perpendicular to the surface, showing only small disturbances in their direction of growth. On the other hand, near the O-ring of the electrochemical cell (the edge of the sample), or around big defects in the substrate, like doping nonuniformities, the pores are far more unstable from their perpendicular direction of growth.
Figure 3.16: a) (100)-oriented n- InP, n = 1017cm-3 , anodized at U = 10 V (j>10 mA/cm2), 5% HCl aqueous electrolyte solution. The pores prove to be unstable and are trying to grow perpendicularly to the surface of the sample forming the so-called Current-line Oriented Pores. b) (111)A oriented n- InP, n = 1018cm-3, anodized at U = 7 V (j>10 mA/cm2), 5% HCl aqueous electrolyte solution; The pores as in the case of (100)-oriented samples grow perpendicularly to the surface. The pores are curved near the O-ring of the electrochemical cell;
The lack of preferential crystallographic direction of growth hints that such kind of pores will look similarly even if the anodization will be performed on other than (100)-oriented InP substrates, e.g. on (111)-oriented samples.
An example of pores obtained at high current densities on (111)-oriented n-InP substrates is presented in Figure 3.16b. As was expected, also in this case the pores are mainly oriented perpendicularly to the surface of the anodized sample. Similarly, the most pronounced deviations from their perpendicular direction of growth are observed near the O-ring of the electrochemical cell.
Even more interesting is the fact that at the same high current density there is no difference between the pores growing from a (111)A or (111)B surface. Note that at low current densities it is not possible to obtain pores growing perpendicularly to a (111)A surface. These results suggest that at high current densities the difference between A and B planes is not anymore important from the dissolution point of view.
In our opinion, at high current densities the system is trying to minimize the ohmic losses by minimizing the path for the current  and thus "forces" the pores to grow perpendicularly to the surface of the sample or in more scientific terms perpendicularly to the equipotential lines of the electric field in the substrate. Therefore, such kind of pores, i.e. which grow perpendicularly to the equipotential lines, are called 'current-line oriented' pores or curro pores .
As it was shown above, under nearly identical conditions, except for a lower anodic current density of more than one order of magnitude, a totally different morphology was obtained for (100) and (111)-oriented n-InP samples (Figure 3.12). In this case the pores prove to be strictly oriented along specific crystallographic directions. One can conclude that a change in the current density from low to high values leads to a switch in the pore growth mechanism.
Figure 3.17: Cross section micrographs of n-InP samples anodized in turn at high and low voltages (7 and 1 V respectively): a) (100) oriented sample; b) (111) oriented sample; In both cases the upper part of the porous layer shows curro pores, whereas the lower part shows crysto pores; Taking into account that porosity of the layer with crystallographically oriented pore is lower than for the layer with Current-line Oriented Pores, it can be suggested that successive layers with different porosities can be formed by alternating the voltage from high to low values. This subject will be discussed later in Chapter 3.5 - Self-organized single-crystalline porous structures .
This conclusion is supported by the images presented in Figure 3.17 . The same (100) n-InP sample was subjected to anodization first at high and then at low current densities. As can be seen in the cross section of the anodized sample, a switch from current-line oriented to crystallographically oriented pores occurs. The upper part of the porous layer exhibits a current-line oriented morphology and corresponds to high current density (Figure 3.17a), while the lower part of the layer contains crystallographically oriented pores and corresponds to the low current densities. The same procedure was applied to a (111)B oriented sample (Figure 3.17b). However, the switch between the curro and crysto pores is not so evident due to the fact that both curro and crysto pores in this case grow perpendicularly to the surface of the sample.
It was concluded above that, according to the current burst model, the dissolution during the formation of crystallographically oriented pores is dominated by direct dissolution and therefore they expose very strong anisotropic features. On the other hand, the strong anisotropic features should partially disappear if the dissolution via oxide formation is dominating the whole dissolution process. Thus, looking at the main characteristics of the curro pores, i.e. they have no preferential crystallographic direction for dissolution, it can be concluded that in this case the main dissolution mechanism is the dissolution via oxide formation. Due to the fact that anodic oxides are normally amorphous, the dissolution at the pore tips tends to be more isotropic, i.e. from the electrochemical point of view there is less difference between different crystallographic planes.
As concluded in the last section, crystallographically oriented pores growing along <111>B directions are common to GaAs, InP and GaP. Thus, it is interesting to know if the Current-line Oriented Pores are also characteristic to GaP and GaAs or they are inherent only to n-InP.
Figure 3.18: (100)-oriented n- GaP, n = 1017 cm-3, anodized at j = 60 mA/cm2, 5% H2SO4 electrolyte solution; a) Overview of the porous layer (cross section); b) A higher magnification of the pores; At such high current densities also in GaP the pores do not grow anymore along <111B> crystallographic directions, but start to grow perpendicularly to the surface resulting in a current-line oriented porous layer. The current-line pores in GaP prove to be much more unstable, i.e. showing branching characteristics, and their diameters are much bigger as compared with those obtained in n- InP. The observations of Current-line Oriented Pores in GaP underlines once again the similarities from the electrochemical point of view between InP and GaP, in spite of the fact that one is an indirect ( GaP) and another is a direct (InP) semiconductor.
Figure 3.18 presents the anodization results obtained at high current density for (100)-oriented n-GaP samples, n=1017 cm-3,j=60 mA/cm2. As in the case of n-InP, the pores do not show any preferential crystallographic directions of growth and simply grow perpendicularly to the initial surface of the sample. Thus, the pores obtained in n-GaP at high current densities can be also catalogued as current-line oriented ones.
Nevertheless, there are some major differences between the Current-line Oriented Pores in n-InP and n-GaP, shown in Figure 3.18. First of all the Current-line Oriented Pores in n-GaP are much more unstable (Figure 3.18b). Secondly, the pore diameters are much bigger in GaP, in the range of 3-4 Ám, whereas for the same doping level the diameters of the Current-line Oriented Pores in n-InP are in the submicrometer range. Last but not least, the ratio pore-diameter/pore-wall-width in GaP is much bigger than in InP.
These differences can be explained assuming stronger passivation efficiency for the pore walls and probably also weaker oxidation and less qualitative oxide in InP as compared to GaP. According to the current burst model, less oxidation in InP means smaller pores. Additionally, in order to break the more effective passivation and consequently to dissolve InP, a stronger electric field at the pore tips is necessary, i.e. a smaller radius of curvature, which is directly related to smaller pore diameters. On the other hand, in GaP the passivation at high current densities is not so efficient, therefore the electric field inside the space charge region can be lower, i.e. the radius of curvature can be bigger and consequently the pore diameters are bigger. Also, low pore wall passivation leads to a small space charge region in the semiconductor and consequently thinner pore walls.
In order to increase the pore wall passivation in GaP a small amount of HCl was added to the aqueous H2SO4 solution (5 ml of 35 % HCl was added to 200 ml of 5% H2SO4). The pores obtained with this mixture are presented in Figure 3.19. It is obvious that the curro pores in this case look more similar to those observed in InP, i.e. the pores grow much more stable and pore diameters are smaller. Thus, it can be concluded that HCl is a better passivation and oxidation agent for GaP as compared to H2SO4.
External macroscopic voltage oscillations represent important characteristics for both materials, GaP and InP, while anodizing them at high current densities. These oscillations will be discussed in more details in one of the next chapters (see section 3.4.2 - Observation of Voltage Oscillations in pore formation regime ). They provide important information about the pore formation mechanism.
Figure 3.19: (100)-oriented n- GaP, n = 10e17 cm-3, anodized at j = 60 mA/cm2, 5% H2SO4 + HCl electrolyte solution. a) Overview of the porous layer. b) A higher magnification of the pores. This kind of pores are much more similar to the curro pores observed in InP;
So far, Current-line Oriented Pores have been observed only in n-type InP and GaP, but not in GaAs. In GaAs at high current densities the so-called tetrahedron-like pores and domains of crystallographically oriented pores are formed. This type of pores in GaAs will be discussed in the following two sections.