Measuring heat flow through a single molecule

A new ability to measure how a minuscule amount of heat moves through a single molecule could help researchers design better materials for ultrasmall computer circuits or nanoscale thermoelectric converters (Nano Lett. 2019, DOI: 10.1021/acs.nanolett.9b02089).

Two groups working separately have measured thermal conductance at a level of a few trillionths of a watt. That achievement helped them probe exactly how heat transfer happens. “If you understand how heat is transported within a molecule, there are theories that describe how to control it,” says Hatef Sadeghi, who studies nanoscale electronics at the University of Warwick and is an author of the Nano Letters paper. But to control heat in real devices, “you need a way to measure it.”

Sadeghi was part of a team led by IBM Research–Zurich. Another group, centered at the University of Michigan, reported similar results in July (Nature 2019, DOI: 10.1038/s41586-019-1420-z).

To make their measurement, the IBM researchers suspended a silicon nitride membrane from a silicon wafer by four thin beams and placed a gold platform and a platinum sensor on top. They can heat the sensor by running current through it, and the sensor registers temperature by measuring resistance. To add their test molecules, the researchers dipped the whole setup into an oligo(phenylene ethynylene)dithiol or octanedithiol solution to deposit those organic molecules onto the gold. The silicon nitride beams have high thermal resistance, so the heat flow through them does not overwhelm the heat flowing through the single molecule, which is only 0.01% of that through the beams, says Bernd Gotsmann, a researcher in nanoscale devices at IBM.

The researchers placed the setup in a vacuum and touched it with a gold tip on a scanning tunneling microscope, which stuck to the gold on the platform. As they slowly pulled the tip away, the gold contact stretched and thinned to a thickness of a single atom and then broke, opening a gap that one of the organic test molecules could slip into. They continued to pull the tip away until the contact between the molecule and the gold broke, which produced a tiny change in heat flow. The difference in heat flow before and after the break told them how much was flowing through the molecule. Because it’s difficult to tell if a molecule was inserted within the gold gap, they repeated their measurements thousands of times over a couple months and came up with a thermal conductance of roughly 30 pW/K.

The Michigan team used a different approach but arrived at much the same measurement when they looked at one of the same molecules. “The results they get for the molecule we have in common is very similar in magnitude,” says Michigan materials scientist Pramod Reddy. Additionally, his group showed that gradually increasing the length of the carbon chain in the molecule by twos—from two up to 10 atoms—did not change the conductance. That’s different from bulk materials, in which resistance increases with the sample length.

Next, researchers will try altering the thermal conductance of various molecules— for instance, by adding side groups. Boosting resistance could lead to better thermoelectric converters, while decreasing it would reduce heat dissipation in nanoelectronic circuits.