Efficient High Current Process Transfer and Device Matching Strategies for sub-90nm Manufacturing

Abstract

As devices scale down to 90nm and beyond, process transfer between different high current (HC) implanter platforms requires precise matching of multiple characteristics of implant process environment. Matching traditional recipe parameters such as dose, energy and tilt/twist no longer secures achieving target device characteristics with tolerances needed for high volume manufacturing (Δ<1% for Idsat, Vt). Thus, beam characteristics, previously considered to be second order, such as beam angular properties (divergence and steering), beam current density, and energy contamination (EC) levels must be taken into account to achieve accurate device matching between different high current implanter platforms. While device sensitivity to beam angular properties increases as devices shrink, at reduced ion energies traditional implanter matching procedures relying on SIMS and Rs measurements on bare wafers lose sensitivity to variations of beam angular characteristics. Thus, to achieve efficient qualification of new high current implant tools new approaches delivering precise process and device matching using minimum number of device split lot runs are needed. In this paper, using 90nm CMOS device data, we demonstrate that in order to transfer existing HC processes to new high current tool successfully, key parameters determining ion beam characteristics need to be quantified and matched between the tools with considerations for unique responses of 3D PR patterned device structures to differences of beam angular and EC components of implant process environment.