S3-1 [Invited] Ext. Abs. the 5th International Workshop on Junction Technology 2005 Non-contact Measurement of Sheet Resistance and Leakage Current: Applications for USJ-SDE/Halo Junctions V.N. Faifer, M.I. Current, T. Nguyen, T.M. H. Wong, V.V. Souchkov Frontier Semiconductor, 1631 N. l' Street, San Jose, CA 95112 USA Tel: 1-408-452-8898, net: fsmI00@frontiersemi.com 1. Shallow Junction Scaling and Leakage With gate lengths of advanced CMOS logic devices scaling to 50 mn and less, the proportional scaling of source/drain extensions over the years 2003 to 2007 estimates junction depths of 20 mn and less and sheet resistances in the range from 200 to 900 Ohm/square [1]. The continued increase in chip transistor count above 10 9 per device has driven a strong effort to understand and constrain the sources of leakage currents and other non- functional power drains. The increased impact of short channel effects for deeply scaled CMOS transistors indicate that tight process controls (of the order of 1% or less for dose and energy) are required for source/drain extension implants in order to maintain reasonable (5% range) controls on threshold voltage and other key transistor characteristics [2, 3]. At the same time, the predominant metrology for implant process control used over the last two decades, sheet resistance measurements with 4-point probes, encounters severe challenges due to physical and electrical punch through for junctions of 50 rn and less [4]. Even with the use of "soft", non- penetrating probes, junction leakage currents continue to cause large errors in direct contact measurements due to mixing of current flows thorough the junction and into the sub-junction layers. Optical interference and reflection methods give structural information on junction depth and damage density but have limited or no sensitivity to electrical activation of shallow junction dopants and do not provide high-precision data on implant dose for implants which produce amorphous layers, such as ultra-shallow source/drain extension implants [5]. The sum of these factors point to the need for a new metrology for process control of doping process (implant and annealing) which can provide sub-1% control on implant dose for junctions of less than 20 nm and provide practical insight into conditions which lead to reduced leakage current for these ultra-shallow junctions. An additional goal is to provide a metrology that can accurately monitor implant and annealing conditions over the full range of doping applications; for ion energies from sub-keV to MeV and doses ranging from 1011 to >1015 ions/cm2. 2. Principles of the Measurement The basis of the measurement is to use photo- excitation of carriers in the junction and wafer substrate and to monitor, in a spatially resolved manner, the generation and drift of carriers with two electrodes, a transparent electrode at the center of the probe and second electrode some small distance away (Fig. 1) [6]. Modulated light beam V i Outside -. electrode Transparent electrode - Surface junction -_ Illuminated region - Figure 1. Sketch of the photo-excitation and drift of carriers with a modulated light source and two capacitor electrodes, one transparent for transmission of LED light, for monitoring the junction photo-voltage (JPV) in a spatially resolved manner. The analysis consists of measuring and modeling of the junction photo-voltage, (JPV), signals, V, and V2, captured by the transparent and non-transparent electrodes when modulated light flux, 0 (t)= <Po(x,y)(J-cos(2rft)), produces electron-hole pairs in the semiconductor material containing a surface junction, where (Do(x,y) is the light flux distribution at the surface of semiconductor with lateral coordinates x, y and f is a light modulating frequency [7]. The surface voltage distribution v(x,y,t)= vo(x,y)*cos(2rft+p7(x,y)) depends mainly on the drift of carriers along the p-n junction. The surface voltage signals, VI and V2, can be determined by the integral over the area of these electrodes, V (t) = Const ffv (x, y, t)dxdy s The dynamics of photo-induced carrier creation, recombination, drift and diffusion is described by a set of continuity and Poisson's equations yielding solutions for the carrier motion and recombination and corresponding spatial distribution of surface voltages in the p-n junction [8]. Operating in the regime of low excitation levels, where vo(x,y) <<kT/q, where k is the Boltzmann constant, T is Kelvin temperature, q is a charge of electron, the surface voltage is proportional to the absorbed light flux. Also in the low light excitation regime, the variation of the surface charge region width, W, induced by illumination is small, which allows the use of one-dimensional Poisson equations at each point.