Factors controlling the resistance of Ohmic contacts to n-InGaAs

The authors report specific contact resistance (rho(c)) values for Mo-, Ti-, TiW-, Pd-, and Pt-based Ohmic contacts to n(+)-In0.86Ga0.14As that are deposited with either collimated sputter or electron-beam deposition. Palladium-based contacts with 5 nm of electron-beam evaporated Pd have a specific contact resistance of 7.6 +/- 0.5 x 10(-9) Omega cm(2), while identical collimated sputter deposited contacts have a specific contact resistance of 4.1 +/- 0.4 x 10(-9) Omega cm(2). Contacts with a very thin evaporated Pd layer (2 nm) or a sputtered Pd/Ti 1/1 nm x 4 multilayer have stable rho(c) values of 5.4 +/- 0.5 x 10(-9) and 5.0 +/- 0.5 x 10(-9) Omega cm(2), respectively, after 2 h at 270 degrees C. The rho(c) of sputter deposited TiW-based contacts is an order of magnitude lower than for identical evaporated contacts (1.6 +/- 0.3 x 10(-8) vs 2.6 +/- 0.3 x 10(-7) Omega cm(2)) and is stable during annealing, while sputter deposition of Pt-based contacts yields r(c) values that are half an order of magnitude lower than similar evaporated contacts (4.5 +/- 1.0 x 10(-9) vs 1.7 +/- 0.2 x 10(-8) Omega cm(2)). However, the rho(c) values of the sputtered Pt contacts approach those of evaporated contacts after annealing. Surface preparations consisting of UV-ozone treatments followed by immersion in buffered oxide etch or ammonium hydroxide yield rho(c) values that are half an order of magnitude lower than values yielded by similar preparations in which HCl is used for oxide removal. The authors also discuss how to avoid over- and underestimation artifacts that may be encountered in measurements of very low specific contact resistances of Ohmic contacts to semiconductors that have low sheet resistances. (C) 2012 American Vacuum Society. [http://dx.doi.org/10.1116/1.4705730]