Integrated Experimental and Computational Design of Alloys for Wear and Corrosion Resistance Across the Temperature Spectrum

Abstract

Metals and alloys with exceptional resistance to wear and corrosion, capable of withstanding a wide range of temperatures and harsh corrosive environments, are critical for diverse applications, including aerospace, automotive, energy production, and medical devices. However, achieving this dual functionality is challenging due to an inherent trade-off: strengthening phases that enhance wear resistance often increase corrosion rates through micro-galvanic coupling. Simultaneously optimizing these conflicting properties represents a significant materials design challenge, further complicated by the vast compositional space of emerging high-entropy alloys (HEAs), also known as multi-principal element alloys (MPEAs) and compositionally complex alloys (CCMs). This Ph.D. thesis seeks to address these challenges by designing robust metals and alloys with outstanding wear and corrosion resistance, spanning conditions from ambient aqueous environments to extreme high-temperature oxidative settings. The research progresses through computational, experimental, and machine learning approaches, addressing key challenges in alloy design. Initially, the effects of alloying composition on the mechanical and corrosion properties of equiatomic and non-equiatomic NiCrFeCo alloys were systematically studied using density functional theory (DFT) calculations. Specifically, the influence of chromium (Cr) concentration was evaluated for compositions containing approximately 16, 25, and 34 at.% Cr. The results revealed that Cr plays a critical role in enhancing both mechanical and corrosion properties. Alloys with higher Cr content exhibited increased Young's modulus and Poisson's ratio, as well as superior corrosion resistance. Relaxing the equiatomic constraint further enabled simultaneous improvements in these properties, leading to the identification of an optimal non-equiatomic composition, Ni22Cr34Fe22Co22, which achieves a balance between mechanical strength and corrosion resistance. Subsequently, the focus shifted to designing microstructures in NiCrFeNb-based superalloys to enhance high-temperature tribo-oxidation behavior. The carefully engineered microstructures from additive manufacturing and post-printing heat treatment promoted the formation of layered spinel-based oxides (AB2O4, where A and B include Ni, Fe, and Cr) on the tribo-oxidized surfaces. These oxides enabled sustained self-lubrication, resulting in remarkably low coefficients of friction, ranging from 0.13 to 0.34, at temperatures between 600°C and 900°C. The superior lubricity was attributed to the structural incommensurability between AO4 and BO6 layers within the oxide, as well as shear-induced phase transformations during wear, offering a novel mechanism distinct from conventional solid lubricants. This work highlights the critical role of bulk microstructure design in achieving improved tribo-oxidation performance in Ni-based alloys at high temperature oxidative environment. Finally, a study on corrosion-resistant NiCrFeCo-based HEA design was conducted integrating computational, experimental, and machine learning approaches. Designing complex alloys with exceptional corrosion resistance across a wide range of corrosive environments, from acidic solutions to chloride-containing conditions, has traditionally relied on trial-and-error methods. This conventional approach is time-consuming and often inefficient, limiting the ability to explore the vast compositional space of modern alloys, such as HEAs. To overcome these challenges, this work adopts a machine learning-driven approach to predict and optimize corrosion behavior in HEAs. A deep neural network (DNN) was developed, integrating physical principles such as the Pilling-Bedworth Ratio to enhance the prediction of corrosion rates. The model was iteratively refined by synthesizing and experimentally testing alloy compositions with high prediction uncertainty, ensuring improved accuracy and reliability. Experimental variability was accounted for using Gaussian noise and dropout techniques, while the emphasis on physical descriptors over simple compositional data enabled more universal and generalizable predictions. The framework was validated by successfully synthesizing two novel HEAs with outstanding corrosion resistance across diverse environments. Future research will focus on expanding this approach to achieve simultaneous wear and corrosion resistance in HEAs, addressing the dual demands of modern materials in harsh environments.