Toward robust and user-friendly RUS analysis

Resonant Ultrasound Spectroscopy (RUS) has a proven history as an accurate technique for determining elastic constants of materials and other derived properties. The usefulness, while significant enough for extensive commercialization, is somewhat limited by several factors. Practical limitations include sample size and geometry, crystal purity, user-intensive and incomplete data collection, globally sensitive nonconvex optimization methods. We present methods and recent results addressing some of these limitations in the effort to provide a more user-friendly automated experience. We consider a modified objective function that is insensitive to missing data, alleviating the need for extended data collection, and that is less sensitive to initial estimates of elastic constants. In addition, we consider RUS model modifications that allow for elastic constant determination on samples of high aspect ratio and samples of lower crystal symmetry (e.g., monoclinic) than traditional RUS methods. We show that enforcing the Born stability criterion is helpful for improving the efficiency of the optimization process, especially for lower crystal symmetry samples.