Divalent dissolution

This fast removal of Si-F species can be ascribed to the weakening of the Si backbonds induced by the strong polarizing effect of F [Ubl], The weak back-bonds are then attacked by HF or H20. This reaction scheme for the dissolution process is supported by quantum-chemical calculations [Trl]. The observed dissolution valence of two for Jelectron injection current and Si-F bond density [Be22] are experimental findings that are in support of the divalent dissolution mechanism, as shown in Fig. 4.3 [Lei, Ge7, Ho6]. 

Fig. 4.3 Reaction scheme proposed for the anodic, divalent dissolution of silicon electrodes in HF.

Divalent dissolution is initiated by a hole from the bulk approaching the silicon-electrolyte interface which allows for nucleophilic attack of the Si atom (step 1 in Fig. 4.3). This is the rate-limiting step of the reaction and thereby the origin of pore formation, as discussed in Chapter 6. The active species in the electrolyte is HF, its dimer (HF)2, or bifluoride (HF2), which dissociates into HF monomers and l ions near the surface [Okl]. The F ions in the solution seem to be inactive in the dissolution kinetics [Se2], Because holes are only available at a certain anodic bias, the Si dissolution rate becomes virtually zero at OCP and the surface remains Si-H covered in this case, which produces a hydrophobic silicon surface. 

For the electrochemical dissolution of Si in electrolytes composed of anhydrous HF and an organic solvent a reaction is proposed that is similar to the divalent dissolution in aqueous HF. However, molecular hydrogen is not observed and four charge carriers are consumed per dissolved silicon atom, as in the tetravalent case [Pr7, Ril]. 

The photocurrent doubling discussed above can be understood as a consequence of the divalent dissolution reaction as shown in Fig. 4.3. Dissolution for current densities below JPS is initiated by a hole in step 1 and proceeds under injection of an electron in step 2. For the case of an n-type electrode, one photon is required to generate one hole, but the electron injected in the dissolution process doubles the current without consumption of another photon. Hence the resulting current density is twice as large as observed at a reference photodiode. Because step 2 of the reaction depicted in Fig. 4.3 is independent of type of doping it can be concluded that electron injection also takes place at p-type electrodes. There is, however, no simple way to detect these injected electrons because the electrode is under depletion in this regime, as discussed in Section 3.2. 

The origin of the electron injection peak at the end of the dissolution of an oxide film is not understood in detail. Silicon interface atoms with three Si-O bonds and a single Si-Si bond are proposed to be responsible for the effect [Mai]. On the other hand, during the dissolution process silicon interface atoms with one Si-O bond and three Si-Si bonds lead to a configuration identical to the one for which electron injection is observed during divalent dissolution (Fig. 4.3, step 2). In any case, the injected charge exceeds by a factor of 3 to 5 the charge expected... 

For homogeneously doped silicon samples free of metals the identification of cathodic and anodic sites is difficult. In the frame of the quantum size formation model for micro PS, as discussed in Section 7.1, it can be speculated that hole injection by an oxidizing species, according to Eq. (2.2), predominantly occurs into the bulk silicon, because a quantum-confined feature shows an increased VB energy. As a result, hole injection is expected to occur predominantly at the bulk-porous interface and into the bulk Si. The divalent dissolution reaction according to Eq. (4.4) then consumes these holes under formation of micro PS. In this model the limited thickness of stain films can be explained by a reduced rate of hole injection caused by a diffusional limitation for the oxidizing species with increasing film thickness. 

FIGURE 5.62. Mechanism of the anodic dissolution of silicon in concentrated hydrofluoric acid solutions (divalent dissolution). After Memming and Schwandt. ... 

The divalent dissolution of Si was then described by a two-dimensional reaction scheme at a kink site on a silicon surface (not shown) assuming the surface to be covered by fluorine instead of hydroxyl groups [8]. It was further postulated that the unstable silicon difluoride changes into a stable tetravalent form by a disproportionation reaction [9] ... 

On the other hand, the decrease of tj) from 4 down to 2 is much more difficult to interpret. As already mentioned above, gravimetric experiments have shown that the decrease is due to a change in the dissolution valency of Si from IV to II. This analytical result is supported by the fact that H2 is formed as soon as the dissolution valency changes (Fig. 8.9), because H2 formation under anodic bias is only possible if Si is dissolved in the divalent state (see Eq. 8.3). Very recently, a model has been presented in which it is assumed that Si(I) (which is just a mobile surface radical) catalyses the divalent dissolution [23]. 

The mesa-type appearance of remnants on Si surfaces after photocorrosion in fluoride-containing solution in the divalent dissolution region (see Section 2.4.1), measured by contact-mode AFM (CM-AFM) (Figure 2.38), can be reproduced by combinations of the three low-index (1x1) H-terminated surfaces (111), (113), and (110) as shown on the left-hand side of Figure 2.38. 

Figure 2.J8 CM-AFM images of mesa-type stmctures obtained during divalent dissolution of Si (region I in Figure 2.39) together with a crystallographic...

Figure 2.40 SRPES data for anodic photocorrosion of n-Si(lll) in dilute NH4F solution in the divalent dissolution regime (cf Figure 2.39, region i) (a) sample emersion slightly anodic from open circuit potential (b) sample emersion near first photocurrent maximum.

Jakubowicz J, Jungblut H, Lewerenz HJ (2003) Initial surface topography changes during divalent dissolution of silicon electrodes. Electrochim Acta 49 137-146 Jakubowicz J, Smardz K, Smardz L (2007) Characterization of porous sihcon prepared by powder technology. Physica E 38 139-143... 

As already shown by Uhlir [10] and Turner [9], the dissolution mechanism for Si in concentrated HF is quite different from that of Ge in aqueous solutions. The fact that Si was dissolved in the divalent state is little surprising because, in general, the stability of a divalent state of an element decreases in the IVth group in the direction Pb-Sn-Ge-Si-C. Therefore, a divalent silicon ion formed by reaction (8.2) is expected to be unstable. Originally the divalent dissolution of Si was then described by a two-dimensional reaction scheme at a kink site on a silicon surface (not shown) assuming the surface to be covered by fluorine instead of... 

---