Machine Learning and High-Throughput Robust Design of P3HT-CNT Composite Thin Films for High Electrical Conductivity

Abstract

<p>Combining high-throughput experiments with machine learning allows quick optimization of parameter spaces towards achieving target properties. In this study, we demonstrate that machine learning, combined with multi-labeled datasets, can additionally be used for scientific understanding and hypothesis testing. We introduce an automated flow system with high-throughput drop-casting for thin film preparation, followed by fast characterization of optical and electrical properties, with the capability to complete one cycle of learning of fully labeled ~160 samples in a single day. We combine regio-regular poly-3-hexylthiophene with various carbon nanotubes to achieve electrical conductivities as high as 1200 S/cm. Interestingly, a non-intuitive local optimum emerges when 10% of double-walled carbon nanotubes are added with long single wall carbon nanotubes, where the conductivity is seen to be as high as 700 S/cm, which we subsequently explain with high fidelity optical characterization. Employing dataset resampling strategies and graph-based regressions allows us to account for experimental cost and uncertainty estimation of correlated multi-outputs, and supports the proving of the hypothesis linking charge delocalization to electrical conductivity. We therefore present a robust machine-learning driven high-throughput experimental scheme that can be applied to optimize and understand properties of composites, or hybrid organic-inorganic materials.</p>

A Machine Learning and Computer Vision Approach to Rapidly Optimize Multiscale Droplet Generation

“…17−19 Thus, using discrete inkjet-deposited droplets of varying semiconductor compositions for high-throughput experimental characterization elicits an accelerated search of this vast composition space for an optimum composition. However, studies such as Bash et al 15 rely on a domain expert to finetune experimental conditions that establish flow instability control prior to semiconductor characterization and, hence, require an understanding of system physics prior to optimization. Rather, in this study, the nonlinear physical relationships between control parameters are iteratively learned by the probabilistic ML surrogate model as data is collected.…”

“…Depositing millimeter-scale droplets onto a substrate is a process useful for high-throughput characterization of material, most notably for finding optimized semiconductor materials that maximize efficiency or stability. , Semiconductors such as perovskites have vast and complex composition spaces which make it challenging to discover optimum compositions. − Thus, using discrete inkjet-deposited droplets of varying semiconductor compositions for high-throughput experimental characterization elicits an accelerated search of this vast composition space for an optimum composition. However, studies such as Bash et al rely on a domain expert to fine-tune experimental conditions that establish flow instability control prior to semiconductor characterization and, hence, require an understanding of system physics prior to optimization. Rather, in this study, the nonlinear physical relationships between control parameters are iteratively learned by the probabilistic ML surrogate model as data is collected.…”

Online Preconditioning of Experimental Inkjet Hardware by Bayesian Optimization in Loop

“…Hence, searching this vast composition space via conventional synthesis methods is highly time and cost inefficient. Therefore, we look towards a method of efficient high-throughput experimentation in which we reduce our samples to the most fundamental form -droplets -which are still characterizable to determine the performance properties of the composition [1]. Utilizing conventional inkjet deposition methods elicits this high-throughput experimental exploration of semiconductor composition spaces.…”

“…However, inkjet deposition methods are not inherently optimized to generate semiconductor droplets that can be reliably characterized. For droplets of semiconductor material to be reliably characterized they must have (1) high geometric uniformity and (2) high yield. Geometric uniformity ensures all droplets are characterized under the same conditions to reduce variability.…”