New Temperature Sensor May Simplify Thermal Monitoring for Circuits
Researchers at the University of Tokyo recently developed a new type of thermal sensor on a lightweight rolled-plastic base.
As devices simultaneously shrink and increase power, heat is becoming an issue in applications where it would not have been a consideration a few years ago. Simply cutting a few vent holes into the case or dropping in a fan is no longer adequate as a cooling method. These quick fixes may have worked at one time, but today, engineering the thermal properties of a design is as important as the EMI and signal integrity.
Heat map of an enclosure. Image (modified) used courtesy of MentorMAD/Wikimedia Commons (CC BY-SA 4.0)
Designers must now measure and monitor heat in a way that will fit in a design, fall within a budget, and accurately portray the real-world operational thermal environment. Researchers at the University of Tokyo (UT) may deliver a solution that fits all of those parameters.
Common Practices for Thermal Characterization
The most expensive chips today have thermal sensors built in to protect against overheating and thermal runaway. Mission-critical PC boards may also include these sensors in key places. However, cost and space limitations often prohibit monitoring more than a few critical areas; engineers must instead do their best to address thermal problems before production begins.
During development, designers concerned about heat typically add sensors to suspect areas of a prototype and characterize the thermal properties during shakedown runs. Thermal simulation software also aids in characterizing the product.
Limitations of Thermal Monitoring—And a Solution
While these approaches are reasonable and mostly effective, they run into limits with compact equipment or equipment operating in difficult-to-simulate settings. Thermal sensors, consisting of wire and small semiconductor components, are not small enough for the current crop of ultra-miniaturized semiconductors and compact designs. The sensors may only provide a thermal view of discrete points rather than the entire system in operation.
To address this limitation, a team at UT has developed a flexible film thermal sensor that may help designers accurately characterize and monitor circuit components at a lower cost and with less interference with the physical layout of the product. The team produced the sensors by sputtering material deposition on a PET film and etching the sensor.
The University of Tokyo used a sputtering and etching process to develop a flexible thermal sensor. Image used courtesy of The University of Tokyo
The thin film can go where many other sensors cannot. Further, it can be put in place during manufacturing for lifetime monitoring without a significant impact on the mechanical arrangements of the product.
How the Thermal Sensor Works
Most thermal sensors rely on the thermoelectric Seebeck effect (SE), the heating of two dissimilar materials (usually metals or semiconductors) resulting in current flow. Alessandro Volta himself discovered the roots of the thermoelectric effect all the way back in 1794. The phenomenon is named after Thomas Seebeck, who independently rediscovered it in 1821.
When heat is applied to the joined end of two dissimilar materials, the heat differential between the hot, joined end and the cooler non-joined ends of the circuit excites electrons enough to cause some of them to move from one material to the other through the joint. This flow of electrons is proportional to the heat differential and can be measured.
Typical thermocouple. The temperature differential between Tsense and Tref excites electrons enough to cause them to move through the junction of the two wires. Image used courtesy of Wikimedia Commons (CC0 1.0)
The new sensors from UT use a less commonly known but related thermoelectric/thermomagnetic effect called the anomalous Nernst effect (ANE). Like The Seebeck effect, ANE converts heat into electricity. However, ANE relies on magnetic materials and operates in a plane perpendicular to the heat. This, in concert with UT’s method for depositing magnetic materials based on iron and gallium on plastic film, yields a flat-area sensor.
The etched circuit uses an alternating arrangement to cancel out the Seebeck effect while providing an accurate reading of the Nernst effect. Image used courtesy of the University of Tokyo
One challenge in using the ANE is that the thermoelectric SE is stronger and obscures the ANE area reading. The UT process neutralizes the SE by alternating patterns. This allows the etched circuit to deliver a more accurate thermal picture of the area.
Potential Applications of the New Thermal Sensor
Prior to this research, thermoelectric sensors were large, awkwardly sized, fragile, and difficult to integrate into more than point-source applications. This research opens the possibility of flexible, form-fitting thermal sensors that can fit in just about any application.
Because this technology is unobtrusive, it can accurately characterize thermal properties during development. Designers can also install this sensor in many devices permanently to actively manage heat over the life of the device. Together, robust thermal design and accurate thermal monitoring can extend a device's lifetime. The inventors of the new sensor hope that in addition to electronics applications, the medtech industry may pick up their technology to generate heat maps of the human body for diagnostic purposes.