Accurate Sheet Resistance Measurement on Ultra-Shallow Profiles
- PDF / 138,490 Bytes
- 6 Pages / 612 x 792 pts (letter) Page_size
- 42 Downloads / 207 Views
0912-C05-07
Accurate Sheet Resistance Measurement on Ultra-Shallow Profiles Trudo Clarysse1, Alain Moussa1, Frederik Leys1, Roger Loo1, Wilfried Vandervorst1,2, Mark C. Benjamin3, Robert J. Hillard3, Vladimir N. Faifer4, Michael I. Current4, Rong Lin5, and Dirch H. Petersen5 1 IMEC, Kapeldreef 75, Leuven, B-3001, Belgium 2 INSYS, Electrical Engineering department, Katholieke Universiteit Leuven, Kasteelpark Arenberg 10, Leuven, B-3001, Belgium 3 Solid State Measurements, Inc., 110 Technology Drive, Pittsburgh, PA, 15275 4 Frontier Semiconductor, Inc., 1631 North First Street, San Jose, CA, 95112 5 Capres A/S, Scion DTU, Building 373, Kgs Lyngby, DK-2800, Denmark
ABSTRACT The accurate and reliable characterization of the sheet resistance of ultra-shallow (USJ) profiles is a key issue in the development of future CMOS technologies. Typically, conventional means, such as in-line four point probe measurements, have a limited accuracy due to the substrate contribution resulting from too much probe penetration, especially in the presence of highly doped underlying layers (such as well/halo-profiles). In this work, a series of advanced Boron doped layers (132 nm down to 2 nm) have been grown with Chemical Vapor Deposition (CVD) on medium and lowly doped substrates and have been characterized with a large variety of state-of-the art non-penetrating/non-contact sheet resistance tools. INTRODUCTION The sheet resistance of ultra-shallow source/drain and extension implants is one of the crucial parameters for optimal CMOS transistor performance [1]. Earlier work has illustrated that one needs to be very careful with the interpretation of conventional four point probe measurements for sub-20 nm layers, especially in the presence of medium (well) and highly doped (halo) underlying layers, as probe penetration and junction leakage can cause very serious distortions (lowering) of the observed values due to substrate shorting [2]. It has been shown before that reducing the penetration of the probes to virtually zero tends to increase the measured sheet resistance values considerably on lowly doped substrates [3] bringing them closer to their expected values. In this work we study into more detail the behavior of three recently developed zero-penetration and/or non-contact tools both on medium and lowly doped substrates and compare them with conventional means. STRUCTURES For this purpose two batches of low temperature Chemical Vapor Deposited (CVD) [4] Boron doped layers have been grown with a dopant level of about 2x1019 at/cm3 and with thicknesses ranging from 132 down to 2 nm (table I). The Reduced Pressure Chemical Vapor Deposition (RP-CVD) system used in this work is a standard ASM Epsilon 2000 production epi reactor. This tool is a horizontal, cold wall, single wafer, load locked reactor with a lamp heated graphite susceptor in a quartz tube. Epitaxial layers were deposited on blanket 200 mm (001) Si
wafers. Before deposition, the wafers were cleaned in a diluted NH4OH/O3 solution. The native oxide was removed by an in-sit
Data Loading...
