43651-AC5
Probing a New Form of Surface Chemistry: Interactions with Point Defects
In the same way that gases react with surfaces from above, point defects within a solid can react from below. Such defects include vacancies and interstitial atoms. Little attention has been paid to this form of surface chemistry, even though the rates govern many aspects of important phenomena such as solid-state diffusion, photostimulated power production and catalysis in semiconductors, solid electrolyte sensors, and metal oxide catalysis. To better understand such chemistry, we employ measurements of solid-state diffusion in semiconductors, together with electron microscopy of extended defects.
Our basic idea is that saturating dangling bonds at a semiconductor surface (or interface with another solid) controllably interferes with both the annihilation and generation of defects. A surface having many dangling bonds can annihilate interstitial atoms by simple addition of the interstitials to dangling bonds. However, if the same surface becomes chemically saturated, annihilation requires the insertion of interstitials into existing bonds. Such insertion should have a higher activation barrier and a correspondingly reduced probability of occurrence. By analogy, surfaces with many dangling bonds should be capable of generating self-interstitials more rapidly than saturated surfaces. The atoms on atomically clean surfaces have lower coordination than atoms on saturated surfaces, so that interstitial creation from a clean surface should require less bond breaking and exhibit a lower activation energy. Higher generation rates should ensue. In addition, electrically charged dangling bonds at a surface or solid interface can also interact electrostatically with charged defects in the semiconductor bulk, either attracting or repelling them depending upon the exact charge states.
This past year we showed that such principles can be employed in the technologically relevant application of integrated circuit manufacture. The continual downscaling of silicon devices for integrated circuits requires the formation of pn junctions that are progressively shallower, incorporate increasing levels of electrically active dopant, and sustain minimal implantation damage. In the case of boron implanted into silicon, we demonstrated experimentally the existence dopant pile-up within the first few nanometers of the surface, which several years ago our group predicted would be the “smoking gun” to demonstrate electrostatic interaction between charged surface and bulk defects. Moreover, we tentatively demonstrated that the charge state of the surface dangling bonds can be altered by photostimulation.
We also demonstrated with isotopic self-diffusion measurements that low-energy ion bombardment of a solid-solid interface between silicon and an overlayer of silicon dioxide can controllably vary the rate at which bulk defects in the silicon are annihilated at the interface. As the degree of interface bond rupture due to noble gas ions increases, the probability of annihilation first increases by about a factor of ten, and then (surprisingly) decreases again.
We have performed isotopic self-diffusion measurements to examine the role of surface effects in controlling defect populations in metal oxides. For the case of titanium dioxide, results suggest that the effects mimic those in silicon. In particular, an atomically clean surface greatly accelerates the formation of oxygen interstitial atoms at the surface, which then sink into the bulk and enhance the diffusion of oxygen there.
