Date of Award

May 2022

Degree Type


Degree Name

Doctor of Philosophy



First Advisor

Michael Nosonovsky

Committee Members

Konstantin Sobolev, Benjamin Church, Mohammad Rahman, Yongjin Sung


Friction wear and lubrication, Machine learning, Surface characterization, Triboinformatics, Tribology, Wetting


Tribology is the study of surface roughness, adhesion, friction, wear, and lubrication of interacting solid surfaces in relative motion. In addition, wetting properties are very important for surface characterization. The combination of Tribology with Machine Learning (ML) and other data-centric methods is often called Triboinformatics. In this dissertation, triboinformatic methods are applied to the study of Aluminum (Al) composites, antimicrobial, and water-repellent metallic surfaces, and organic coatings.Al and its alloys are often preferred materials for aerospace and automotive applications due to their lightweight, high strength, corrosion resistance, and other desired material properties. However, Al exhibits high friction and wear rates along with a tendency to seize under dry sliding or poor lubricating conditions. Graphite and graphene particle-reinforced Al metal matrix composites (MMCs) exhibit self-lubricating properties and they can be potential alternatives for Al alloys in dry or starved lubrication conditions. In this dissertation, artificial neural network (ANN), k-nearest neighbor (KNN), support vector machine (SVM), random forest (RF), gradient boosting machine (GBM), and hybrid ensemble algorithm-based ML models have been developed to correlate the dry friction and wear of aluminum alloys, Al-graphite, and Al-graphene MMCs with material properties, the composition of alloys and MMCs, and tribological parameters. ML analysis reveals that the hardness, sliding distance, and tensile strength of the alloys influences the COF most significantly. On the other hand, the normal load, sliding speed, and hardness were the most influential parameters in predicting wear rate. The graphite content is the most significant parameter for friction and wear prediction in Al-graphite MMCs. For Al-graphene MMCs, the normal load, graphene content, and hardness are identified as the most influential parameters for COF prediction, while the graphene content, load, and hardness have the greatest influence on the wear rate. The ANN, KNN, SVM, RF, and GBM, as well as hybrid regression models (RF-GBM), with the principal component analysis (PCA) descriptors for COF and wear rate were also developed for Al-graphite MMCs in liquid-lubricated conditions. The hybrid RF-GBM models have exhibited the best predictive performance for COF and wear rate. Lubrication condition, lubricant viscosity, and applied load are identified as the most important variables for predicting wear rate and COF, and the transition from dry to lubricated friction and wear is studied. The micro- and nanoscale roughness of zinc (Zn) oxide-coated stainless steel and sonochemically treated brass (Cu Zn alloy) samples are studied using the atomic force microscopy (AFM) images to obtain the roughness parameters (standard deviation of the profile height, correlation length, the extreme point location, persistence diagrams, and barcodes). A new method of the calculation of roughness parameters involving correlation lengths, extremum point distribution, persistence diagrams, and barcodes are developed for studying the roughness patterns and anisotropic distributions inherent in coated surfaces. The analysis of the 3×3, 4×4, and 5×5 sub-matrices or patches has revealed the anisotropic nature of the roughness profile at the nanoscale. The scale dependency of the roughness features is explained by the persistence diagrams and barcodes. Solid surfaces with water-repellent, antimicrobial, and anticorrosive properties are desired for many practical applications. TiO2/ZnO phosphate and Polymethyl Hydrogen Siloxane (PMHS) based 2-layer antimicrobial and anticorrosive coatings are synthesized and applied to steel, ceramic, and concrete substrates. Surfaces with these coatings possess complex topographies and roughness patterns, which cannot be characterized completely by the traditional analysis. Correlations between surface roughness, coefficient of friction (COF), and water contact angle for these surfaces are obtained. The hydrophobic modification in anticorrosive coatings does not make the coated surfaces slippery and retained adequate friction for transportation application. The dissertation demonstrates that Triboinformatic approaches can be successfully implemented in surface science, and tribology and they can generate novel insights into structure-property relationships in various classes of materials.