The biggest source of untapped energy on the planet is waste heat.
Laptops, PCs and tablets constantly give off heat when they operate, as do much larger industrial machines. Of all the energy humans produce, about 70% is heat that is never recaptured. What if this wasted energy could be put to use?
Engineers at the University of Virginia, George Washington University and the National Institute of Standards and Technology have formed a UVA-led team to address this challenge. Their research will innovate the design and manufacture of coin-sized thermoelectric generators and devices that can convert waste heat into electricity.
Their goal is to design and document a single-step fabrication technique of thermoelectric devices with 40% increased efficiency and 20% cost reduction over conventional manufacturing practices. A $500,000 grant from the U.S. Department of Energy's Advanced Manufacturing Office supports the research.
The team employs a multi-pronged and well-coordinated approach to achieve this goal. Materials science is one prong. The field of materials science provides a steady supply of new and improved energy-conversion materials that could be used to develop thermoelectric generators.
Team member Saniya LeBlanc, associate professor of mechanical and aerospace engineering at George Washington University, develops energy technologies with novel materials and advanced manufacturing techniques.
"Although many improvements have been made in thermoelectric materials' performance, the same improvement in device performance has not been realized," LeBlanc said. "There's a gap between the materials and the devices that we need to bridge to make thermoelectric technology commercially viable. Additive manufacturing is that bridge. AM ties materials and device innovations together. It's an innovative approach to manufacturing that allows us to achieve unprecedented improvements in both efficiency and cost."
How the device is made is as important as the material it's made from. Ji Ma, UVA assistant professor of materials science and engineering and a principal investigator on the project, will use additive manufacturing to improve the geometric structure of thermoelectric materials and devices.
"Additive manufacturing allows us to rapidly fabricate complex geometric structures such as hollow cylinders, pyramids and combs," Ma said. "Additive manufacturing also saves raw materials and reduces reliance on mechanical tools that need to be constantly calibrated and maintained."
Improvements to thermoelectric generators and coolers, as well as flexible thermoelectric devices fabricated by additive manufacturing--picture wrapping a heat exchanger around a pipe--are an urgent need. Ma will help the team link laser processing to the resulting structure and properties of thermoelectric parts. "The payoff of our research will be realized in techniques that optimize production," Ma said.
The third prong of the team's effort is finding the optimal combination of materials, geometries and fabrication techniques to produce a device that can generate as much electricity from heat as possible, whether from its ambient environment or its own operation. Prasanna Balachandran, UVA assistant professor of materials science and engineering and mechanical and aerospace engineering, employs advanced computer modeling techniques to identify the most promising materials.
Optimal additive manufacturing processing parameters for thermoelectric materials are unknown. Establishing the process parameters for a new material is traditionally long and arduous, relying mostly on sizeable test matrices and trial-and-error experiments. From Balachandran's modeling, LeBlanc and Ma can conduct more meaningful experiments to understand the process-property-performance relationships that determine thermoelectric device performance.
Finding the best 3-D blueprint for a device involves printing a lot of samples. Ma estimates that manufacturers could spend six to 12 months to develop and prove a new process. With LeBlanc's materials and Balachandran's models, the team can optimize material properties and speed production simultaneously.
"In place of a regular test grid, working methodically through each combination square-by-square, we can jump to the square where we expect to find high-value data about the properties we most care about, such as fewer defects and desired microstructure," Balachandran said.
Balachandran's machine learning data set resides at the center of a virtuous production cycle of process parameter control, high-throughput sample fabrication and high-throughput characterization that yields optimized parameters to additively manufacture the device.
The fourth prong of the team's approach is to quickly characterize the printed materials' internal structure to look for defects as well as desired properties. Martin L. Green, who leads the materials for energy and sustainable development group at the National Institute of Standards and Technology, will transfer his experience with high-throughput characterization to the team's task.
Altogether, the team shares substantial expertise in additive manufacturing of thermoelectric materials, and are the first to demonstrate successful, layer-by-layer laser additive manufacturing of bismuth telluride, the working material for most thermoelectric generators. The interplay of machine learning and high-throughput characterization can work for other material classes or systems that play a predominant role in energy conversion and storage.
"The research we conduct with this grant will have a long tail," Ma said.