Home AI News Accelerating Biomass Conversion with AI-Based Catalyst Development

Accelerating Biomass Conversion with AI-Based Catalyst Development

Accelerating Biomass Conversion with AI-Based Catalyst Development

The Significance of Biomass as a Renewable Energy Source

Biomass is a type of organic matter, including plants, wood, and agricultural waste, that can be used as a renewable source of energy. Unlike fossil fuels, biomass comes from living organisms and can be replenished relatively quickly. It has the potential to be converted into different forms of energy, such as heat, electricity, and biofuels, and can play a role in reducing greenhouse gas emissions and promoting sustainable development.

The Abundance of Biomass in Rural Areas

Rural areas with farms, prairies, and ponds are rich sources of biomass, including crops like corn, soybeans, and sugar cane, as well as switchgrass and algae. These materials can be transformed into liquid fuels and chemicals with various applications, including renewable jet fuel for air travel in the United States.

The Role of AI in Developing Low-Cost Catalysts for Biomass Conversion

One of the challenges in converting biomass into valuable products like biofuel is finding affordable and effective catalysts. However, researchers at the U.S. Department of Energy’s Argonne National Laboratory have developed an AI-based model to speed up the development of a low-cost catalyst based on molybdenum carbide.

When biomass is subjected to high temperatures, it produces pyrolysis oil with a high oxygen content. A molybdenum carbide catalyst is used to remove this oxygen content, but the catalyst’s surface attracts oxygen atoms, which reduces its effectiveness. To solve this problem, researchers propose adding a small amount of a new element, such as nickel or zinc, to the molybdenum carbide catalyst. This reduces the bonding strength of oxygen atoms on the catalyst surface, preventing its degradation.

Using Supercomputing and Deep Learning to Enhance Catalyst Development

The research team utilized the Theta supercomputer at Argonne to simulate the behavior of surface atoms binding with oxygen and those nearby. They created a database of 20,000 structures for oxygen binding energies to doped molybdenum carbide and used this data to develop a deep-learning model. This model allowed them to analyze tens of thousands of structures within milliseconds, providing faster and cost-effective results compared to traditional computational methods.

The Chemical Catalysis for Bioenergy Consortium received the research team’s findings and will use them to conduct experiments and evaluate a shortlisted group of catalysts. The team plans to expand their computational approach in the future by examining over a million structures and exploring different binding atoms, such as hydrogen. They also aim to apply the same technique to catalysts used in other decarbonization technologies, like producing clean hydrogen fuel from water.

Source link


Please enter your comment!
Please enter your name here