Deep Level Passivation

One of the results of this variety in hydrogen-defect reaction pathways is that it largely clouds one of the hopes of the hydrogenation experiments, namely that the susceptibility of deactivation could provide information on the defect microstructure and the nature of the bonding with hydrogen. 

Considerably more work is needed to clarify the passivation mechanism for deep levels. No information is yet available on the susceptibility of these levels to passivation by other species such as the alkali metal ions Na+, Li+, K+ or species like F. Such experiments may shed more light on the deactivation mechanism. DeLeo et al. (1984) have predicted that alkali metal impurities will not passivate vacancy dangling bonds. This is experimentally testable in a relatively straightforward fashion—both Li and F can be introduced into Si at high concentration by a number of methods. 

Chantre, A., Pearton, S.J., Kimerling, L.C., Cummings, K.D., and Dautremont-Smith W.C. (1987). Appl. Phys. Lett. 50, 513. 

Corbett, J.W., Peak, D., Pearton, S.J., and Sganga, A. (1986). Hydrogen in Disordered and Amorphous Solids (G. Bambakadis and R.C. Bowman, Jr. eds.) Plenum Press, New York, New York, p. 61. 

Pearton, S.J. (1985). Proc. 13th Inti. Conf. Defects in Semicond., ed. L.C. Kimerling and J.M. Parsey, Jr. Metall. Soc. AIME, PA. 14a, 737. 

---

In order to remove tlie unwanted electrical activity associated witli deep-level impurities or defects, one can eitlier physically displace tlie defect away from tlie active region of tlie device (gettering) or force it to react witli anotlier impurity to remove (or at least change) its energy eigenvalues and tlierefore its electrical activity passivation). 

In this chapter we will list the deep-level centers passivated by atomic hydrogen in the major elemental semiconductor, namely Si, and discuss their thermal stability and the possible passivation mechanisms. As is the case with any aspect of hydrogen in semiconductors, much more work has been performed in Si than any of the other materials. 

Figure 1 shows a deep level transient spectroscopy (DLTS) (Lang, 1974) spectrum from a Au-diffused, n-type Si sample before and after hydrogenation of 300°C for 2h (Pearton and Tavendale, 1982a). The well-known Au acceptor level (Ec - 0.54 eV) was passivated to depths > 10 pm under these conditions and was only partially regenerated by a subsequent... 

Most of the other metal-related deep levels in Si are also passivated by reaction with hydrogen (Pearton, 1985). Silver, for example, gives rise in general to a donor level at Ee + 0.54 eV and an acceptor level at Ec - 0.54 e V (Chen and Milnes, 1980 Milnes, 1973). These levels are very similar to those shown by Au, Co and Rh and raise the question of whether Au might actually be introduced into all of the reported samples or a contaminant, or whether as discussed by several authors there is a similar core to these impurity centers giving rise to similar electronic properties (Mesli et al., 1987 Lang et al., 1980). This problem has not been adequately decided at this time. It has been... 

The passivation of deep level defects and shallow impurities in semiconductors by hydrogen has been studied extensively in recent years (Pearton et al., 1987, 1989 Haller, 1989). For Si in most cases, complexing with hydrogen eliminates the electrical activity of a defect.Once passivated, the... 

Hydrogen Passivation of Deep Level Impurities and Defects 372... 

In addition, a number of other deep level impurities have been hydrogen passivated. They include nickel, cadmium, tellurium, zirconium, titanium, chromium, and cobalt (Pearton et al., 1987). Most of these studies have been qualitative, and important work remains to be done if the hydrogenation of these and most probably additional impurities, such as gold, palladium, platinum and iron, is to be fully understood. 

In GaAs, there is a huge number of unidentified deep levels that have been evidenced by DLTS. Some of them depend upon the growth technique. Several works, including early ones, reported the passivation of some of these deep level defects by hydrogenation. 

In MBE grown GaAs three dominant electron traps are usually observed Ml at c - 0.17 eV, M3 at c - 0.28 eV and M4 at c - 0.45 eV. Exposure of MBE grown material to a hydrogen plasma for 30 minutes at 250°C completely passivates these three deep levels as shown in Fig. 10 (Dautremont-Smith et al., 1986). After five minute anneals at 400°C or 500°C, the passivation remains complete while the shallow donors are fully reactivated. A five minute annealing at 600°C partially restores the electrical activity of M3. Therefore the thermal stability of the neutralization of deep levels in MBE material is much higher than in other materials and is compatible with most technological treatments. 

As far as the passivation of deep level defects by hydrogen is concerned, their understanding is rather poor, partly because the microscopic structure of these deep level centers is largely unknown. The thermal stability of the passivation of these deep centers has the advantage of being usually compatible with the temperature used in the process of III-V devices. This point might already create an interest in the field of applications. 

We proceed now to discuss the analogous hydrogen-related complexes with shallow-level impurities. We will find that the way hydrogen incorporates itself and the character of the complex after it is formed is independent of whether the passivated impurity introduced a deep level or a shallow level. Therefore, many of the basic arguments regarding the complexes will prevail as we consider the shallow-level defects. 

---