Though thin film thermoelectrics have been of research interest for the last decade, scientists have struggled to develop them into solid-state devices in practical geometries due to the difficulty of maintaining a temperature gradient across films. Still, interest remains high as researchers pursue efficient, unobtrusive power sources for wearable technology and small medical devices. The thermoelectrics research group at Northwestern University presents a solution to this problem – turn the planar 2D energy harvester into a helical 3D coil.
According to Stephen Kang, a recent graduate of the group, thin films are unsuitable for miniature devices like wearable technology simply because they are too thin to maintain a temperature difference across their thickness.
“It’s maybe analogous [to] a situation of wearing thin clothes; the temperature of your skin will be almost the same as the atmosphere if your clothes are too thin. No temperature difference means no electrical voltage.”
The 3D array their group developed overcomes this issue by instead allowing the out-of-plane heat to travel through the in-plane direction of the film. The group’s experimental and computational finite element analysis studies show that the architecture provides stability even with a thin material.
Applicability over scalability
In principle, the array that the group developed can be scaled as large as processing capabilities will allow. The team claims, however, that the architecture is ideal for powering miniature devices.
“The idea of an energy harvester is that, for devices that have a very small operational power demand, you integrate the harvester with the device so that you can power the device from energy obtained from the device environment,” says Kang. “Then you don’t have to worry about changing the battery or plugging in a power cord.”
The group’s goal is to therefore focus on improving their efficiency, measured by the power supplied by a given area of an array, rather than scaling their technique for larger devices.
The efficiency of the 3D array depends on the material used. In their model system, the group chose silicon because of its well understood properties, rather than its potential as a thermoelectric – its figure-of-merit is low compared with commonly studied and more competitive thermoelectric materials like Bi2Te3 and Cu2Se. In bulk, however, it is difficult to modulate the properties of thin films of Bi2Te3 and Cu2Se as the group did with silicon. Differences in mechanical properties could also hinder the geometry’s performance. According to Kang, this opens doors for organic or polymer materials as thermoelectrics.
“Our device design is a place where the mechanical properties of organic or polymer materials could offer a great advantage,” Kang said. “Although their inherent thermoelectric properties are inferior to inorganic materials (i.e. lower figure-of-merit), their mechanical properties will allow designs that can potentially get more power out of the harvester.”
More details can be found in the latest Science Advances.