IN THIS ISSUE Monthly Meme! Latest News ‣ New Companies Planning to Join MDS-Rely ‣ New Pitt Undergraduate Engineering Data Analytics Certificate New Center Project ‣ " Comparative analysis of Machine Learning techniques in predicting structure property relationships for composite dielectric materials" Software Packages ‣ XCTimage ‣ BayesO ‣ netSEM Upcoming Events ‣ Thursday, November 17th 4-5PM - Technical Seminar presented by Ina Martin ‣ Biweekly Project Meetings Available To All Members Hellos and Goodbyes ‣ Welcome to our new Business Development Director , David Ruvolo! ‣ Thank you to Maureen Barcic! ‣ Check out our new Instagram page! Job Openings and Opportunities ‣ Open Position at NSF Meme of the Month! LATEST NEWS New Companies Planning to Join MDS-Rely Thank you to the following companies who have formally indicated their interest and plans to join the Center! New Pitt Undergraduate Engineering Data Analytics Certificate Center co-director Paul Leu introduced a new undergraduate certificate in engineering data analytics that has officially launched in Fall 2022. This new program was inspired by Applied Data Science minor at Case Western Reserve University created by Directors Roger French and Laura Bruckman. NEW CENTER PROJECT Comparative analysis of machine learning techniques in predicting structure property relationships for composite dielectric materials Alp Sehirlioglu, Ina Martin, and Laura Bruckman, Case Western Reserve University Dielectric coating applications are myriad, ranging from heatsinks, to waveguides, to electrostatic clutches. A data science approach will allow creation of a generalizable model for relating composite materials structural and optical properties to the capacitance that is critical to function at both local and macroscopic scales. The deposition conditions used to form the dielectric films, and the nature of the underlying substrate, will dictate the structure and composition of the film, and in turn, these will affect their stability during performance, e.g. wear cycles in a dielectric clutch. Creating models to understand these structure-property relationships as they pertain to real-world applications is a way of enabling progress in this specific technology, while also feeding broader applications of these materials systems. The objective of this project is to identify predictive abilities and deep learning capabilities of different approaches of varying complexity to aid choice of machine learning tools in addressing data science problems while building workflows with greater reliability along the supply chain. We will be starting biweekly meetings for this project soon that all members can attend. SOFTWARE PACKAGES XCTimage XCTimage is a Python package that provides a suite of tools and functions for the analysis of X-Ray Computed Tomography images. It contains image pre-processing and feature extraction techniques in 2 and 3 dimensional data. The package provides efficient, user-friendly feature extraction to serve as a foundation for robust, scalable machine learning pipelines. It has a broader goal of becoming the foundation for analysis of XCT imaging across domains and materials. BayesO BayesO (pronounced “bayes-o”) is a simple, but essential Bayesian optimization package, written in Python. It is designed to run advanced Bayesian optimization with implementation- specific and application-specific modifications as well as to run Bayesian optimization in various applications simply. This package contains codes for several surrogate models such as Gaussian process regression and random forest regression, so that sequential model-based optimization can be implemented. netSEM Network structural equation modeling ( netSEM ) is a data-driven modeling technique that has been developed at the SDLE Research Center and with a public version available as a R software package CRAN. Adapted from structural equation modeling (SEM) which is used to explore linear relationship between observed and latent variables, netSEM also captures the non-linear relationships between variables and explores the degradation pathway in a <Stressor|Mechanism|Response> (<S|M|R>) framework, providing a predictive <Stressor|Response> model (data-driven model) as well as an inferential and mechanistic <Stressor|Mechanism|Response> model that allows for rank ordering of degradation mechanisms. netSEM selects the best relationship between variables based on statistical significance such as adjusted R2 and also rank order models using Akaike Information Criterion (AIC), and Bayesian Information Criterion (BIC). UPCOMING EVENTS "Stabilizing Transparent Conductive Oxides as a Route to Long-Lived Thin Film Photovoltaics" Ina Martin | Operations Director, Materials for Opto/electronics Research and Education (MORE) Center | Case Western Reserve University Date: Thursday, November 17th 4:00-5:00PM Location: * virtual (Zoom) Zoom Link Here! Abstract: My research encompasses the chemistry and physics of materials deposition, characterization, and degradation processes necessary to understand and improve the materials and devices that benefit society. Here I will demonstrate how surface and bulk measurements can be combined to understand the degradation of metal oxides, specifically aluminum doped ZnO (AZO) used as a transparent electrode in solar cells. Further, I will discuss how surface modification can be used to suppress bulk degradation of these materials. Damp heat (DH) degradation of polycrystalline metal oxide thin films degrades the material through the diffusion and reaction of moisture through the grain boundaries. In a transparent conductive oxide (TCO), the resistivity of the film increases with degradation. Surface modifiers can be used to “cap” reactive moieties, to improve the bulk stability of the material. I will present results on how 3- aminopropyltriethoxysilane (APTES) modification of thick (~0.8 μ m) AZO films significantly impedes the electrical degradation of the material caused by DH exposure, without significantly affecting the initial optical, electrical, or structural properties of the AZO films. Upon 1000 h of DH exposure, resistivity of both systems increased and can be attributed only to decreased mobility, as carrier concentration was consistent. APTES modification slowed the increase in AZO resistivity over 1000 h of DH exposure; however, the protective nature of APTES modification became critical after 1500 h. At this extended exposure time, macroscopic degradation was observed only for bare AZO including pitting and delamination and was accompanied by an increase in resistivity and decrease in carrier concentration. X- ray photoelectron spectroscopy (XPS) data show that the APTES layer stabilizes the oxygen binding environment of the AZO surface, thereby improving the stability of the material. Further, when applied to Cu(In,Ga)Se 2 (CIGS) solar cells, the application of the nm-scale modifier not only mitigates AZO degradation in damp-heat exposure, it arrests the degradation of the full CIGS device. Biweekly Project Meetings Available To All Members If you are interested in attending any biweekly project meeting, please visit our Members portal, which can be accessed via the link at the top right of our Center website . You can also go to a specific project and raise a request to get access to the Zoom links to attend any of these meetings. You can also access prior recordings and presentations of any biweekly meetings. 1. netSEM Modeling for Service Life Prediction of Polymers Prof. Laura Bruckman November 15, November 29... Tuesdays 1 - 1:30 PM 2. Achieving Reliable Laser Powder Bed Fusion based Additive Manufacturing via Machine Learning of in-situ Optical Profilometry Monitoring Data Prof. Xiayun Zhao December 3... Fridays 4:15 - 4:45 PM *no biweekly meeting on November 18 3. Image Machine Learning of Printed Metal Films for EMI Shielding Profs. Leu/French/Iyengar November 23, December 7...Wednesdays 1:30 - 2 PM HELLOS AND GOODBYES Welcome to our new Business Development Director, Dave Ruvolo! Welcome Dave Ruvolo. Dave is the new Centers Administrator for the Swanson School of Engineering and will be working as an industry liaison and helping manage and run MDS-Rely amongst several Centers at Swanson including the Center for Advanced Manufacturing , MOST- AM , and AMPED . He’s been with Pitt since 2013 working in a variety of roles in both academic and administrative departments. Please join us in welcoming Dave to the team! Thank you to Maureen Barcic! Thanks to Maureen Barcic who has served as our MDS-Rely Center Business Development Coordinator since March of this year and is retiring. Maureen has really gone above and beyond her responsibilities in reaching out to new companies and helping organize our Spring and Fall meetings in 2022. Prior to joining our Center, Maureen worked as part of the Pitt Co-op office beginning in July of 1987 and became the director in October of 1989. Thank you, Maureen, and we wish you the best in retirement! Check out our new Instagram page! See more funny jokes/memes as well as pictures from some of our recent activities and events! JOB OPENINGS AND OPPORTUNITIES Open Position at NSF The National Science Foundation (NSF) is currently looking for an Engineering/Science Analyst to join their team. The analyst will employ data science, data analytics, visualization, and an array of multimedia communication tools to justify and support recommendations and decisions in support of promoting the progress of science. Find out more and apply here! Submit News Fill out a news form here! Submit Job Openings *For MDS-Rely members only Fill out a job opening form here! Interested in partnering with Case Western or Pitt Professors? Please contact Dr. Roger French or Dr. Paul Leu for more information! CONNECT WITH US! Copyright © 2022 Materials Data Science Rely, All rights reserved. Our mailing address is: Case Western Reserve University White Building, Room 538 10900 Euclid Avenue Cleveland, OH 44106 Want to change how you receive these emails? You can update your preferences or unsubscribe from this list Learn more here! Check out our calendar with upcoming events here! @mdsrely This email was sent to <<Email Address>> why did I get this? unsubscribe from this list update subscription preferences The Center for Materials Data Science for Reliability and Degradation (MDS-Rely) · 10900 Euclid Ave · Cleveland, Ohio 44106 · USA Subscribe Past Issues RSS Translate