Improvement of the dose uniformity in plasma immersed ion implantation by introducing a vertical biased ring

Investigation of the ion dose non-uniformity caused by sheath-lens focusing effect on silicon wafers

“…The dependence of r imp on plasma parameters for a dc bias was already reported for disk and square electrodes of about 10 mm, a size much smaller than the plasma chamber (25 cm in diameter) [17]. In present work, and PIII in general, the sheath size is just a few times smaller or even comparable with the reactor size or only a few times smaller, and it is therefore important to verify the influence of plasma density and implantation parameters on discrete and modal focusing effects.…”

“…The same solution but for a larger scale was suggested to avoid the ion dose non-uniformity observed in PIII of silicon wafers [21]. However simulations showed that such a guard ring cannot prevent the back side ions to reach the probe (wafer) surface the modal focusing effect and that the discrete focusing effect is still taking place on the guard ring [17]. In order to prove this for PIII experiments, we used the stage setup presented in Fig.…”

“…As an alternative solution to the guard ring a vertical ring was proposed but not verified in real PIII situation [17]. In order to do so we used the arrangement presented in Fig.…”

“…Thus, it was demonstrated that the modal [14] and discrete [15] focusing effects are responsible for the formation of a passive surface (no ion impact) and the presence of an intense focusing ring that significantly degraded the ion flux uniformity [16]. A solution to improve the ion flux uniformity by introducing a vertical ring (VR) was actually proposed [17,18] but its effectiveness was not yet validated on real PIII conditions.…”

Controlling the ion flux on substrates of different geometry by sheath-lens focusing effect

“…A "guard-ring" (often used for planar Langmuir probes) was also used to surround the wafer stage, shifting (for certain processing parameters) the discrete focusing ring outside the wafer surface. Since the guard-ring cannot prevent modal focusing, we proposed instead a biased verticalring [32]. Another solution is illustrated in Fig.…”