Multi-response optimization of ultrathin poly-SiGe films characteristics for Nano-ElectroMechanical Systems (NEMS) using the grey-Taguchi technique

Poly-SiGe can be used for monolithically integrating Micro/Nano-ElectroMechanical Systems (M/NEMS) with its driving circuitry in a MEMS-last approach. For these applications, it is important to have poly-SiGe films with a low tensile stress, a low resistivity and a high deposition rate. This paper presents a systematic procedure for the simultaneous optimization of these 3 properties for CVD deposited ultrathin (100+/-5nm) poly-SiGe films by using the grey-Taguchi approach. Seven process variables were identified as important parameters for controlling the deposition process and the resulting film properties, namely the deposition temperature, the silane, germane, diborane and hydrogen flow rate, the chamber pressure and the shower head-heater spacing. By using 4 different levels for each process variable, 32 unique experiments were defined based on an L32 orthogonal array. The optimal combination of process parameters was determined by applying the grey relational analysis (GRA) for multiple performance characteristics. The analysis of variance (ANOVA) showed that the deposition temperature has the highest influence on the multi-performance characteristics (contributing ~41%). The projected optimized process resulted in a 100nm-thick poly-SiGe film with a tensile stress of 43MPa, a very low resistivity of 1.39m@?-cm, a deposition rate of 0.34nm/s, a germanium concentration of 87%, a cauliflower surface morphology with a root-mean-square roughness of 4.2nm, an elastic modulus of 101+/-0.81GPa and a strain gradient of -2.0x10^-^2/@mm. The optimized film is expected to be desirable as structural layer for M/NEMS applications such as nanoswitches, nanoresonators, biosensors etc.