Thermal Sensor Arrays for The Combinatorial Analysis of Thin Films

Membrane-based thermal sensor arrays were developed for the high-throughput analysis of the thermophysical properties of thin films. The continuous growth of integrated circuits and microelectromechanical systems, as well as the development of functional materials and the optimization of materials properties, have produced the need for instruments capable of fast materials screening and analysis at reduced length scales. Two instruments were developed based on a similar architecture, one to measure thermal transport properties and the other to perform calorimetry measurements. Both have the capability to accelerate the pace of materials development and understanding using combinatorial measurement methods.;The shared architecture of the instruments consists of a silicon-based micromachined array of thermal sensors. Each sensor consists of a SiN X membrane and a W heating element that also serves as a temperature gauge. The array design allows the simultaneous creation of a library of thin film samples by various deposition techniques while systematically varying a parameter of interest across the device. The membrane-based sensors have little thermal mass making them extremely sensitive to changes in thermal energy.;The nano-thermal transport array has an array of sensors optimized for sensitivity to heat loss. The heat loss is determined from the temperature response of the sensor to an applied current. An analytical model is used with a linear regression analysis to fit the thermal properties of the samples to the temperature response. The assumptions of the analytical model are validated with a finite element model. Measured thermal properties include specific heat, thermal effusivity, thermal conductivity, and emissivity. The technique is demonstrated by measuring the thermal transport properties of sputter deposited Cu multilayers with a total film thickness from 15 to 470 nm. The experimental results compare well to a theory based on electronic thermal transport.;The parallel nano-scanning calorimeter (PnSC) has an array of sensors optimized to sense changes in enthalpy. In this case heat loss sensitivity is minimized with sensor geometry and a reference measurement scheme. The minimal heat loss and small addendum result in sensitivity on the order of 10 nJ/K at heating rates on the order of 104 K/s. The sensitivity is demonstrated by measuring the characteristics of the melting transformation of a 25 nm In film. The combinatorial capabilities of the device are demonstrated by creating and analyzing a library of thin-film (290 nm) Ni-Ti-Zr samples with in-plane composition gradients. The Ni-Ti-Zr films are crystallized in-situ by local heating and the temperature dependence of the martensite transformation on Zr content is detected.;Further analysis of the Ni-Ti-Zr samples reveals that the as-deposited amorphous samples crystallize in a multi-stage process that is a function of composition. The features of the calorimetry traces are identified with the help of x-ray diffraction measurements of the crystallized samples. Crystallization at these fast heating rates results in suppression of structural relaxation, increased crystallization temperature (allowing the detection of the glass transition), and an ultra-fine nanocrystalline grain structure with non-equilibrium phases.;The characteristics of the martensite-austenite phase transformation are investigated by PnSC to determine the effects of high-temperature (900°C) heat treatments and low-temperature (450°C) thermal cycling. Heat treatments produce precipitates that vary with Zr content and alter the transformation temperature. Thermal cycling results in the accumulation of plastic deformation, which relaxes internal stresses and reduces the transformation temperature. This effect, known as thermal fatigue, is reduced in these samples due to the ultra-fine grain structure, which suppresses dislocation mobility.