Key Takeaways
- Compact carbon capture technology: University of Houston researchers have developed a carbon capture system designed to fit inside 40-foot shipping containers, offering a smaller, more flexible alternative to conventional systems.
- Promising laboratory results: The technology achieved approximately 90% carbon dioxide capture efficiency in lab testing, demonstrating its potential for use at natural gas- and coal-fired power plants.
- Scalable solution for growing AI infrastructure: Researchers estimate about 20 modular units could capture 98% of carbon dioxide emissions from a 360-megawatt power plant supporting an AI data center.
As AI drives unprecedented demand for electricity, University of Houston researchers are developing a compact technology designed to make carbon capture more practical for power plants supporting the growing AI industry.
UH Engineering Professor Mim Rahimi and former UH Ph.D. student Ahmad Hassan, who is currently at Harvard University as a postdoctoral associate, detailed the technology, called Small Modular Electrochemical Containers (SMECs), in a study published in Environmental Science & Technology.
The SMEC is designed to fit inside standard 40-foot shipping containers, eliminating the need for some of the large, expensive equipment used in conventional carbon capture systems. Researchers demonstrated its effectiveness in laboratory testing and examined how it could be deployed at large-scale power plants.
“We’re taking a technology that traditionally requires enormous infrastructure and finding a way to fit it inside a shipping container,” Rahimi said. “That opens up entirely new possibilities for how and where carbon capture systems can be used.”
Meeting AI's Growing Energy Needs
According to the U.S. Department of Energy, data centers accounted for approximately 4.4% of the nation's electricity consumption in 2023 and could reach as much as 12% by 2028.
Meeting that demand will require substantial investments in electricity generation, including natural gas power plants capable of supplying reliable power around the clock.
However, integrating conventional carbon capture technology into these facilities presents challenges. Traditional systems often require large absorption towers, significant amounts of heat and extensive infrastructure modifications.
"With our technology, we can adjust the capacity by simply turning containers on or off."
— Mim Rahimi, assistant professor at the Cullen College of Engineering, University of Houston
“With a conventional system, you’re building a 130-foot-tall structure designed for a specific power plant. With our technology, we can adjust the capacity by simply turning containers on or off,” Rahimi said. “That flexibility is particularly important for data centers, where electricity can come from a combination of natural gas and renewable sources.”
Unlike conventional systems, the technology uses electricity to separate carbon dioxide from power plant exhaust gases before it enters the atmosphere. By combining the capture and release processes into a single-electrochemical cell, the technology eliminates the need for large absorption columns and specialized membranes.
From Laboratory to Large-Scale Deployment
At the heart of each SMEC are membraneless electrochemically mediated amine regeneration (EMAR) cells, developed in Rahimi’s laboratory and first reported in Nature Communications last year.
“The modular, symmetric architecture of membraneless EMAR provided the foundation for the SMEC design,” Hassan said. “It allowed us to combine individual cells into a compact system that can be scaled to meet different carbon capture requirements.”
Researchers evaluated the technology’s performance by conducting laboratory experiments using simulated exhaust gases from natural gas- and coal-fired power plants.
The system achieved approximately 90% carbon dioxide capture efficiency under optimized conditions and maintained stable operation through 300 cycles over 50 hours.
To explore its potential beyond the laboratory, researchers modeled how the technology could operate at a 360-megawatt natural gas power plant associated with the OpenAI’s Stargate AI data center project in Abilene, Texas.
Their analysis found that about 20 shipping container-sized units could capture 98% of the plant's carbon dioxide emissions. The proposed equipment, including supporting infrastructure, would occupy just a fraction of the data center’s 875-acre campus.
While the current research focuses on emissions from power plants, Rahimi sees opportunities to expand the technology’s applications.
“We’re working to make the technology capable of capturing carbon dioxide directly from the atmosphere rather than just from a power plant,” Rahimi said. “That could allow the system to operate anywhere.”