Focus: Temperature Difference Leads to Magnetism
Computer simulations suggest that creating hot and cold regions within a specific arrangement of semiconductors generates internal electric currents and magnetic fields. If borne out by experiments, the new effect, reported 8 July in Physical Review B, could lead to improvements in electronic devices that heat up in use. Experts are intrigued by the effect but remain cautious about its practical importance until they see experimental data.
Temperature differences, or “gradients,” can have important effects on the flow of current in a semiconductor because the electrons or holes tend to drift from hotter regions toward colder ones. Such thermal effects can also interact with electric and magnetic fields, as in the so-called thermoelectromagnetic effects. For instance, in the Nernst effect, when a semiconductor is exposed to a temperature gradient and a magnetic field at right angles to each other, a small electric field is produced in the third direction. Semiconductors in electronic devices often develop temperature gradients, so researchers need to understand their effects.
Junqiao Wu of the University of California, Berkeley, and his colleagues noticed that in the known thermoelectromagnetic effects, the magnetic field is never induced but always one of the “inputs,” or applied fields. They wondered whether a magnetic field would result if a semiconductor were subjected to an electric field and a temperature gradient.
The team ran computer simulations of a two-micron-wide sample consisting of an n-type semiconductor (electrons carry current) on top of a p-type semiconductor (positively-charged holes carry current). Near the interface, such a structure–which is common in electronics–generates a so-called depletion region, where electrons diffuse down into the p-type material and holes diffuse up into the n-type material. The fixed charges left behind create an electric field pointing down. Next, the team’s simulation assumed that the left edge was 10 millikelvin cooler than room temperature, and the right edge was 10 millikelvin hotter.
In the simulations, a current vortex developed in each material. In the n-type semiconductor, which was on top, electrons moved to the right at the top edge and to the left just above the interface, with the holes executing nearly a mirror image of this motion below the interface, in the p-type material. These vortices generated a magnetic field that pointed outward, toward the viewer.
The vortices are results of a complex simulation, and It’s difficult to explain them in physical terms. But part of the story is that half of each loop of current goes through the depletion region, the area within perhaps 100 nanometers of the p/n interface, where there are fewer charge carriers than in the rest of the material. This lack of charge carriers turns out to allow the temperature gradient to have a stronger effect on the mobile charges that remain there–pushing them from right to left–than it has on charges outside the depletion region. Away from this zone, near the upper and lower edges of the structure, there is much higher conductivity, which allows the charges to more easily flow left-to-right, against their usual thermal diffusion direction.
In addition, the vertical electric field effectively acts at the center of charge of the electron or hole “cloud,” whereas the temperature gradient acts at the center of mass, says Wu. He says that perpendicular forces acting at different places generate a torque on the charges, which partly explains the rotation, an effect described theoretically by others in 2005 .
The magnitude of the effect can be large, and the eddy currents could soak up energy if the structure were part of a circuit, the researchers say. In fact, Wu says, similar arrangements of semiconductors in commercial devices that heat up may be running with slightly reduced efficiency because of this effect. The solution would be to minimize the vortices by aligning the direction of the temperature gradient with that of the electric field in future designs. The team believes that a better understanding of the relationships among temperature, currents, and electromagnetic fields may help engineers improve electronic designs in other ways as well.
Gang Chen of the Massachusetts Institute of Technology in Cambridge says Wu’s paper “certainly points to an interesting phenomenon that should be further explored” with experiments. Charles Marcus of Harvard University agrees, saying whether “it has technological relevance [would] depend on the magnitude.” Wu says his team is currently designing experimental tests.
–Saswato R. Das
Saswato R. Das is a freelance science writer in New York City.
- K. Mohseni, A. Shakouri, R. J. Ram, and M. C. Abraham, “Electron Vortices in Semiconductor Devices,” Phys. Fluids 17, 100602 (2005)