Spintronic thermoelectric power generators: A step towards energy efficient electronic devices

March 21, 2014
Schematic of the spintronic thermoelectric device fabricated by the University of Utah’s researchers. This device can convert even minute heat emitted by hand-held electronic devices such as laptops, etc. into useful electricity. Credit: Gene Siegel, Shiang Teng

(Phys.org) —Imagine a computer so efficient that it can recycle its own waste heat to produce electricity. While such an idea may seem far-fetched today, significant progress has already been made to realize these devices. Researchers at the Nanostructured Materials Research Laboratory at the University of Utah have fabricated spintronics-based thin film devices which do just that, i.e. convert even minute waste heat into useful electricity.

"As enter the nano-size regime, the problem of heat generation is becoming more and more severe," says University of Utah Materials Scientist Ashutosh Tiwari, who led the research published online Friday, March 21 in the Nature publishing group's journal Scientific Reports.

"Our spintronic-based work at room temperature and don't require the continuous application of an external ," Tiwari says. "Most of the spintronic thermoelectric devices in earlier studies required the continuous application of a magnetic field to keep the device magnetized."

"Spintronics is a new branch of electronics which utilizes both the charge as well as the spin of electrons," says Tiwari.

"Tiwari conducted the research with graduate students Gene Siegel, Megan Campbell Prestgard, and Shiang Teng. The study was funded by the U.S. National Science Foundation's Condensed Matter Physics Program, Sensors and Sensing Systems Program, and the University of Utah's Materials Research Science and Engineering Center."

"The most important and fascinating aspect of our study is that these devices are not made of traditional thermoelectric materials which, when heated, generate a voltage simply because of the movement of charge carriers. This, known as the Seebeck effect, has a fundamental limitation," says Tiwari. "Specifically, for achieving heat-to-electricity conversion efficiencies which are acceptable for practical applications, the electrical conductivity of the thermoelectric material should be maximized while its thermal conductivity should be simultaneously minimized. These two requirements are contradictory."

He adds, "Our spintronic-based devices function on an altogether different concept known as spin-caloritronics. Here, thermal and electrical transport occurs in different parts of the device and hence these devices are not plagued by the problems encountered by their traditional counterparts."


For making spintronic thermoelectric devices, Utah researchers deposited thin films of a material know as bismuth-doped YIG (Bi-YIG) using a 25 nanosecond pulsed laser. Over the Bi-YIG film, a 10 nm thick layer of platinum was deposited using electron beam evaporation. The prepared bi-layer structure was kept in a magnetic field for a few minutes to magnetically polarize the Bi-YIG film. After this, the was removed and a temperature gradient was applied across the bilayer. This temperature difference leads to a current of low-lying excitations of localized spins, known as magnons, in the Bi-YIG. "When this magnon current enters into the platinum layer, it is converted into a charge voltage through a process known as the inverse spin Hall effect," explains Gene Siegel, first author of the paper.

The researchers' trick was to generate a very large roughness on the surface of the Bi-YIG films by using very high energy density laser pulses. Rough surfaces resulted in very large stray fields, which gave rise to large magnetic coercivity in the films. "Because of the large coercivity, once these devices are magnetized they remain magnetized and don't require any external field for operation," says Siegel.

"Our experimental findings are in excellent agreement with the predictions of the magnon transport theory," says Tiwari.

"Tiwari's group's research opens the doors for the development of spin-driven thermoelectrics which can turn into electricity, and make efficient electronic devices," says Ajay Nahata, Director of the University of Utah's NSF MRSEC on Next Generation Materials for Plasmonics & Organic Spintronics.

Explore further: Using heat to make magnets

More information: Siegel, G., Prestgard, M.C., Teng, S. & Tiwari, A. Robust longitudinal spin-Seebeck effect in Bi-YIG thin films. Sci. Rep. 4, 4429; DOI: 10.1038/srep04429 (2014).

Related Stories

Using heat to make magnets

October 17, 2013

EPFL scientists have provided the first evidence ever that it is possible to generate a magnetic field by using heat instead of electricity. The phenomenon is referred to as the Magnetic Seebeck effect or 'thermomagnetism'.

Controlling magnetism with an electric field

February 18, 2014

There is a big effort in industry to produce electrical devices with more and faster memory and logic. Magnetic memory elements, such as in a hard drive, and in the future in what is called MRAM (magnetic random access memory), ...

Recommended for you

Creating antimatter via lasers?

September 27, 2016

Dramatic advances in laser technologies are enabling novel studies to explore laser-matter interactions at ultrahigh intensity. By focusing high-power laser pulses, electric fields (of orders of magnitude greater than found ...


Adjust slider to filter visible comments by rank

Display comments: newest first

1 / 5 (1) Mar 21, 2014
These Guys should Set up their Labs Around the Edges of Volcanoes!
not rated yet Mar 21, 2014
But if you have a device, like a laptop motherboard with the chips and all, that you want to keep below 50 degrees C and you're in a room that is 25 C, the amounts of energy you can possibly recover between that is miniscule because the absolute maximum theoretical efficiency is just 7.7%

With a non-ideal device the recovery efficiency is just pointlessly low. Imagine paying even a hundred dollars more for a laptop that runs 20 minutes longer on a battery. Is it really worth it?
5 / 5 (1) Mar 21, 2014
50 degrees C and you're in a room that is 25 C

The inside of the processor actually runs a lot higher than 50, but they design them with heat sinks built into them. Internal temperatures over 100 C are common.

paying even a hundred dollars more for a laptop that runs 20 minutes longer

Yeah, there's a big difference between a novelty that works in a lab and something you can actually sell to people. I notice they are using platinum, which isn't cheap.

Recovery of waste heat is an important goal though. It could greatly reduce the cost of just about everything we use. It might not work well in a laptop (yet), but think about the factories that make the parts of that laptop (or parts for your car, or the bread on your sandwich). The patent for a good thermoelectric converter could be worth a lot of money.
not rated yet Mar 21, 2014
The inside of the processor actually runs a lot higher than 50, but they design them with heat sinks built into them. Internal temperatures over 100 C are common.

That's the reason why they have to be cooler on the outside, because the thermal resistance from the actual transistor junctions to the surface of the chip casing hinders the heat flow. The limiting factor is how you bond the actual chip to the chip casing, and the chip casing to a heat spreader plate if any, which is usually done with thermally conductive glue or resin, and the actual heatsink is then attached on top of that by more paste or glue.

The part where your thermoelectric element comes in is in between the heat spreader and the heatsink, and doing that adds another gap which has to be filled with thermal paste, which increases the total thermal resistance and narrows the temperature difference across the TEG by a significant amount.

So getting 25 C of temperature difference is asking a lot.
4 / 5 (2) Mar 21, 2014
The patent for a good thermoelectric converter could be worth a lot of money.

It has to be extremely cheap if you want to utilize it with low temperature differentials because the energy thus recovered is not worth a whole lot of money.

Some applications would be interesting, like adding generators to central heating systems that run on gas or wood.
not rated yet Mar 23, 2014
yes, so ideally, you'd want the thermoelectric element to be inside the chip, theoretically. If you could 'soak up' the waste heat and recycle it in an efficient way inside the chip, that would change everything (in computer chip design).

Most of my comments are actually in regard to thermoelectric applictions on larger scales, and at higher temperatures, where the majority of our waste happens.

Please sign in to add a comment. Registration is free, and takes less than a minute. Read more

Click here to reset your password.
Sign in to get notified via email when new comments are made.