A quantitative study of the microstructure and crystallographic fiber texture in nickel electrodeposits used in radio-frequency MEMS switches, including a new transmission electron microscopy (TEM) technique for polycrystalline films

Patrick R Cantwell, Purdue University

Abstract

The microstructure of electrodeposited nickel films in radio-frequency (RF) microelectromechanical systems (MEMS) switches has been quantitatively studied to inform and validate multi-scale, multi-physics computer simulations that aim to predict the lifetime and failure mechanisms of the RF MEMS switches. The RF MEMS switches are currently under study at the Purdue University center for the Prediction of Reliability, Integrity, and Survivability of Microsystems (PRISM). An array of microstructural characterization techniques including focused ion beam (FIB) microscopy, scanning electron microscopy (SEM), X-ray diffraction (XRD), and transmission electron microscopy have be used to study the nickel film and to quantify grain size and crystallographic texture and provide information about elemental impurities and surface roughness and impurity elements. Particular emphasis has been placed on quantifying the crystallographic fiber texture of the polycrystalline nickel film as a function of film height within a single specimen using a new transmission electron microscopy (TEM) microtexture method. The TEM method employs a special type of plan view TEM sample and uses hollow cone dark field (HCDF) TEM imaging to spatially map the orientation of individual crystallites at discrete film heights. A trend of increasing 001 fiber texture with film height was discovered, which has implications for the elastic behavior of the MEMS device. The method can be applied to study fiber texture evolution as a function of height in polycrystalline films to gather data that may elucidate fundamental film growth mechanisms. The method is explained in detail. It is well-known that the elastic properties of polycrystalline thin films used in MEMS devices can deviate from bulk isotropic values and become directionally-dependent if a crystallographic texture is present. Hence, the ability to predict the actual anisotropic elastic properties of textured films is important for MEMS design and analysis. An integrated technique combining X-ray diffraction (XRD) and density functional theory (DFT) simulation is presented here for the quantification and prediction of the elastic properties of crystallographically textured polycrystalline films used in MEMS devices. The technique is rapid, efficient, and capable of analyzing individual devices in an array, making it ideal for MEMS design, analysis, and quality control. Application of the technique to the electroplated nickel bridge of an RF MEMS switch, whose critical operating parameters depend on the in-plane Young's modulus, is demonstrated. It is shown that the in-plane Young's modulus of nickel films with a perfect, single fiber texture can vary over a large range from 172 GPa to 232 GPa. Experimental results significantly outside this range cannot be explained by crystallographic texture alone. The range of Young's modulus for real films is expected to be somewhat smaller because real films rarely have a near- perfect fiber texture and sometimes have a texture that cannot be described by a single fiber of orientation. The nickel bridge of the RF MEMS switch, which has a relatively strong 001 fiber texture component as well as a weak 111 fiber texture component, exemplifies such a case. The present technique takes these texture features into account to estimate the in-plane Young's modulus of the nickel bridge in several RF MEMS switches.

Degree

Ph.D.

Advisors

Strachan, Purdue University.

Subject Area

Materials science

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server
.

Share

COinS