Imperfections such as grain boundaries in a material affect how efficiently it can transport heat at the nanoscale. A team of researchers led by Kenneth Goodson at Stanford University has now developed the first experimental technique to visualize local variations in heat flow near individual grain boundaries in polycrystalline diamond and has observed a nearly two-fold reduction in local thermal conductivity in the immediate vicinity of the disordered boundaries. This finding will be important for developing materials that are more efficient at carrying away excess heat in electronic devices.
Unwanted heat is a major problem in electronics, and the issue gets even worse as devices become smaller. It is in fact a major stumbling block to miniaturizing components further. To address this challenge, researchers have suggested that synthetic polycrystalline diamond could be ideal as a next-generation heat-sink material. However, understanding how grain boundaries affect heat flow in this material, is crucial to determining how good it might be for thermal management applications.
“Until now, researchers have studied the relationship between grain structure and heat transport using indirect techniques that average over the effects of thousands of grain boundaries. While useful, these measurements provide little direct information about the nature of heat flow in the immediate proximity of an individual boundary, explains team member and study lead author Aditya Sood. “Our approach, which is based on an ultrafast optical pump-probe technique called time-domain thermoreflectance (TDTR), allows us to measure thermal transport down to the level of single grains.”
Constructing 2D maps of thermal conductivity
“The technique involves using short pulses of light, each lasting about a trillionth of a second, to heat up the surface of a material,” he explains. “As heat leaves the surface, we use a second set of pulses to probe the rate at which temperature decays on the nanosecond timescale. This rate of heat flow can be quantitatively linked to the local thermal conductivity of the underlying material.”
By focusing the laser light to a small spot and raster scanning the sample using a high-precision stage, the researchers say they can construct 2D maps of thermal conductivity. In these experiments, one of the main challenges was to be able to precisely locate the grain boundaries in the material and to make sure that the thermal transport measurements were made at the same location on the sample. Working closely with electron microscopy expert Mark Goorsky of the University of California Los Angeles (UCLA), the researchers did this using a technique called electron backscatter diffraction (EBSD), which produces mosaic images of the local crystalline orientation of each grain.
“A unique feature of our work is that we use both techniques in a correlative manner, which allows us to construct a direct visual link between the local thermal conductivity and the underlying grain structure,” Sood tells nanotechweb.org. “This correspondence has never been visualized before with such clarity.”
Two-fold reduction in the local thermal conductivity
The researchers observed a nearly two-fold reduction in the local thermal conductivity of their sample – boron-doped polycrystalline diamond. “This dramatic suppression of heat flow occurs because of the strong scattering of thermal energy carriers (phonons) at the disordered grain boundaries,” explains Sood. “Interestingly, we detect this lowering in thermal conductivity up to a few microns away from the grain boundary. We believe that this non-local effect stems partly from the transport of heat by phonons with long mean-free-paths – that is, atomic vibrations that transmit thermal energy over large distances without scattering.”
The result could be important for thermal management of electronics devices using materials such as high-conductivity synthetic polycrystalline diamond. “Although the specific details are certainly material-dependent, we have shown that heterogeneities in the underlying microstructure of a material need to be considered when evaluating its performance as a heat-sinking substrate.”
“I believe that this is going to become increasingly important as device dimensions start to become comparable to the intrinsic defect length scales in polycrystalline substrates.”
“In addition to grain boundaries, the new correlative microscopy approach might also come in handy for studying the interactions of vibrational heat carriers with other types of crystalline defects,” says Sood. These could include dislocations, precipitates and pores. “An important field that could benefit from this is thermoelectrics, in which multi-scale defects are intentionally incorporated into materials to impede thermal conduction. Understanding the underlying physics here will require accurate experimental measurements of heat flow near individual defects, something that a technique like ours might provide.”
The team is now busy working on extending its measurements from 2D to 3D. “Many technologically important materials grow with an anisotropic microstructure, such that heat flows very differently within the plane as compared to across the thickness,” explains Sood. “Constructing a 3D microscopic picture of heat transport would allow thermal engineers to design better materials for heat management and routing.”
The Stanford-led project was funded by the Defense Advanced Research Projects Agency (DARPA). As well as the collaboration with Mark Goorsky’s group at UCLA, the Stanford researchers also worked closely with Samuel Graham and colleagues at Georgia Tech. The research is detailed in Nano Letters DOI: 10.1021/acs.nanolett.8b00534.