Recombination velocity

In the depletion region for a band bending U - Ujb> 100 mV, where a reasonably low surface recombination velocity is found, the PMC signal can consequently be approached by... 

If minority carrier current (1BC, dotted line, symbols in Fig. 3.2) is detected at the collector, it can be concluded that the emitter is no sink for minorities. The absolute value of IEB depends not only on the charge state of the emitter-base junction and surface recombination velocity, but as well on bulk diffusion length and on sample thickness. However, the latter two parameters are constants for a given sample. 

This is the regime of cathodic currents. The silicon atoms of the electrode do not participate in the chemical reaction in this regime. An n-type electrode is under forward bias and the current is caused by majority carriers (electrons). The fact that photogenerated minority carriers (holes) are detectable at the collector indicates that the front is under flat band or accumulation. A decrease of IBC with cathodization time is observed. As Fig. 3.2 shows, the minority carrier current at the collector after switching to a cathodic potential is identical to that at VQcp in the first moment, but then it decreases within seconds to lower values, as indicated by arrows in Fig. 3.2. This can be interpreted as an increase of the surface recombination velocity with time under cathodic potential. It can be speculated that protons, which rapidly diffuse into the bulk of the electrode, are responsible for the change of the electronic properties of the surface layer [A17]. However, any other effect sufficient to produce a surface recombination velocity in excess of 100 cm s 1 would produce similar results. 

A current of photogenerated holes observed in regime 4 at the collector at low VeB of about 1 V for an illuminated n-type substrate indicates that no significant SCR is present at the front side the n-type electrode is in inversion. If the bias is increased this current disappears indicating an SCR or an increased surface recombination velocity at the emitter. In regime 4a all photogenerated holes are consumed by Feb- Breakdown of the junction in regime 4 does not lead to pore formation. 

For not too low doped samples (D W), however, the contribution of 1SCR is usually negligible. If the surface recombination velocity at the illuminated front is low, IBPC then only depends on sample thickness D, illumination intensity eP, and minority charge carrier diffusion length ID. 

Double-sided electrolytic contacts are favorable for this method of diffusion length measurement because they are transparent and the required SCRs are easily induced by application of a reverse bias. Therefore homogeneously doped wafers need no additional preparation, such as evaporation of metal contacts or diffusion doping, to produce a p-n junction. Furthermore, a record low value of surface recombination velocity has been measured for silicon surfaces in contact with an HF electrolyte at OCP [Yal], Note that this OCP value cannot be further decreased by a forward bias at the frontside, because any potential other than OCP has been found to increase the surface recombination velocity, as shown in Fig. 3.2. Note that contaminations in the HF electrolyte, such as Cu, may significantly increase the surface recombination velocity. This effect has been used to detect trace levels (20 ppt) of Cu in HF [Re5j. 

These devices showed EL enhancements to ammonia, methylamine, di-methylamine, trimethylamine, and sulfur dioxide that increased in magnitude with concentration until saturation was reached [14]. The LEDs with larger active layers produced the greatest change in EL intensity with exposure to sulfur dioxide and the amines. Intensity changes were attributed principally to surface recombination velocity effects, as the significant forward biases employed should eliminate the depletion width. 

Electron-hole recombination velocities at semiconductor interfaces vary from 102 cm/sec for Ge3 to 106 cm/sec for GaAs.4 Our first purpose is to explain this variation in chemical terms. In physical terms, the velocities are determined by the surface (or grain boundary) density of trapped electrons and holes and by the cross section of their recombination reaction. The surface density of the carriers depends on the density of surface donor and acceptor states and the (potential dependent) population of these. If the states are outside the band gap of the semiconductor, or are not populated because of their location or because they are inaccessible by either thermal or tunneling processes, they do not contribute to the recombination process. Thus, chemical processes that substantially reduce the number of states within the band gap, or shift these, so that they are less populated or make these inaccessible, reduce recombination velocities. Processes which increase the surface state density or their population or make these states accessible, increase the recombination velocity. 

In summary, a simple chemical picture of surface recombination is presented. The surface or grain boundary recombination velocity decreases when the appropriate surface species is reacted with a strongly chemisorbed species. It increases when a species is weakly chemisorbed. We shall now illustrate this concept for six extensively studied semiconductors, Ge, Si, GaP, GaAs, InP and InSb. 

The existence of today s silicon based microelectronics technology is evidence for the low surface recombination velocity of oxidized Si. The velocity is less than 103 cm/sec.5,6,7... 

We explain the low surface recombination velocity by the sweeping of surface states from the region between the edges of the conduction and valence bands upon oxidation of the surface. The standard free energies of formation of crystalline and fused quartz from bulk Si are -192 and -191 Kcal/mole respectively. The standard free energy change for a Si surface is likely to be... 

We note in this context that in Si based MIS (metal-insulator-semiconductor) solar cells one of the roles of the 20-60A thick Si02 layer may well be reduction of the recombination velocity at the Si surface. Chambouleyron and Soucedo noted a decrease in the recombination velocity at the conductive Sn02/Si interface8 relative to that at the Si surface and Michel and Lasnier find a recombination velocity of less than 2xl04 cm/sec at the conductive indium tin oxide/Si interface.9 In both cases heating the metal oxides present on the elemental Si produces an intermediate Si02 layer. 

Exposure of silicon to atomic hydrogen increases the surface recombination velocity.111213 The free energy of formation of SiH4, the most stable of the hydrides of silicon, is only — lOKcal/mole. Since four electron pairs are shared in the formation of the molecule, the free energy of formations per Si-H bond is only -2.5 Kcal or about O.leV. Because of the weak chemisorption, heating of the silicon to temperatures above 500 C is adequate to release the hydrogen. Our model explains the increase in surface recombination velocity by the weak chemisorption of hydrogen, which may increase the density of surface states within the band gap (see Fig. 2b). 

Casey and Buehler have shown that the surface recombination velocity of n-InP ( 5xl017 carriers/cm3) is low, 103cm/sec.17 Suzuki and Ogawa have recently reported a sequence of surface treatments that cause substantial changes in the surface recombination velocity of InP.18 They found that in freshly vacuum cleaved (110) faces v, is much greater than at air exposed faces and that the quantum efficiency of band gap luminescence increases by an order of magnitude when the freshly cleaved face is exposed to air. This suggests that the surface recombination velocity is reduced when 02 is chemisorbed. 

The changes are explained as follows The density of surface states within the band gap on freshly cleaved InP is high. As a result, the surface recombination velocity is high and the luminescence efficiency is low. Chemisorption of oxygen splits the surface states, as large band gap, colorless InP04 is formed.19... 

Reduction in the surface recombination velocity of GaP, from 1.7 x 10s cm/sec to 5xl03 cm/sec, is observed upon exposure to a CF4 plasma in which fluorine is known to be present.20 Again, the product of chemisorption of fluorine on the surface is likely to be a large band gap material such as GaF3, which straddles the edges of the conduction and valence bands of GaP. 

The presence of arsenic at the interface implies that surface states within the band gap will be introduced (see Fig. 1). We associate the high surface recombination velocity with the presence of arsenic. The formation of elemental As on the GaAs surface explains the difference in behavior of InP and of GaAs. In InP the thermodynamically stable phase that results from oxidation of the surface is colorless InP04 which straddles the band gap. In GaAs it is Ga203 and small band gap As. 

Woodall et al.36 have analyzed the relationship between surface recombination velocity and the steady state band gap luminescence in GaAs. They calculate for 534nm excitation that a decrease in vs from 106cm/sec to 104cm/sec will triple the quantum efficiency at a 2.5Mm deep p-n junction if the hole diffusion length, Lp, is 3jim, and the electron diffusion length, L is 4/im. 

Ions that are not chemisorbed do not affect the performance of semiconductor liquid junction solar cells.32 Weakly chemisorbed ions produce inadequate splitting of surface states between the edges of the conduction and valence band and increase rather than decrease the density of the surface states in the band gap and thus the recombination velocity. Bi3+ is an example of such an ion. As seen in Figure 5, it decreases the efficiency of the n—GaAs 0.8M K2Se-0.1M K2Se2-lM KOH c cell.30 Since the chemisorption of Bi3+ is weak, the deterioration in performance is temporary. The ion is desorbed in 10 min. and the cell recovers. 

---