This work centers on assessing the environmental and economic impact of materials and processes as early in their development as possible. We leverage information along the trajectory from lab bench synthesis to prototype processing to final scaled manufacturing.

We are working to understand at an early stage the economic and environmental impact of materials.

Predictive Synthesis


(Above): Machine-learned latent synthesis vectors from [1].

Advances in computational materials design have enabled rapid screening for desirable properties of both real and virtual compounds. However, the pace of commercially-realized advanced materials may now be limited by trial-and-error synthesis techniques. The goal of this research project is to advance computational learning around materials synthesis approaches by creating a predictive synthesis system for advanced materials design and processing — to do for materials synthesis what modern computational methods have done for materials properties.

[pullquote align=”full” cite=”” link=”” color=”#16a085″ class=”” size=”16″]We are working to build a curated database of solid state synthesis methods [/pullquote]

We are working to build a curated database of solid state materials and their published synthesis methods compiled from thousands of materials synthesis journal articles, as well as algorithms developed through machine learning approaches, capable of predicting synthesis routes for novel materials based on chemical formulae and other known physical input data. More information can be found at

[1] E. Kim, K. Huang, S. Jegelka, E. Olivetti, Virtual screening of inorganic materials synthesis parameters with deep learning. npj Comput. Mater. 3, 53 (2017).

Predictive Assessment: Environmental Impact of Electronics

Evaluation of the environmental impact of manufactured products can be costly, time consuming, and uncertain, particularly for products that evolve on rapid time scales. This is particularly seen in the semiconductor industry, in which process flows for wafer manufacturing are complex and the product profiles change every two years or less.

Our group is developing a methodology to perform quantitative evaluation of the carbon footprint at reduced cost. The methodology being investigated is underspecification, which involves using a high level assessment of the product/component in question based on engineering models, publicly available studies and industry expert input to screen for high impact elements with high degree of uncertainty of the footprint. Data are refined around these primary drivers of impact until uncertainty is reduced to a desired level.

We are applying this methodology to characterize integrated circuits (ICs), the manufacturing of which is resource- and energy-intensive and involves hundreds of different process steps. The IC impact data is screened to identify the elements with the most leverage to reduce the uncertainty using Monte Carlo simulations. The goal of this research is the development of a streamlined, modifiable carbon footprint tool that has uncertainty incorporated directly into it and includes upstream raw materials up to the impact of the wafer fabrication facility.