Researchers at the University of Michigan have taken an important step towards thermoelectric materials which are efficient enough to be used for large-scale generation of electricity from waste heat in places like power plants.
Thermoelectrics are materials which generate electricity directly from a temperature gradient. This makes them ideal for harvesting energy which would otherwise be wasted as heat, in cars, buildings, and even power plants.
However, currently available thermoelectric materials are not efficient enough to make this commercially viable in most cases. The standard measure of the effectiveness of a thermoelectric system is called it's "figure of merit" - the best-performing materials we have now have a figure of merit between 2 and 2.5, and they will need to reach 4 in order to compete commercially with existing technologies.
Now, a group of researchers led by Prof. Pierre Ferdinand Poudeu at the University of Michigan has found a way of manipulating the nanostructure of a thermoelectric material to drastically improve it's performance. Prof. Poudeu explains how the discovery came about:
Professor Pierre Ferdinand P. Poudeu
"In the field of thermoelectrics, the main goal is to reduce the thermal conductivity of the semiconductor, whilst maintaining a good electrical conductivity. This is most often done by manipulating the nanostructure of the semiconducting matrix.
"The main purpose of the nanostructures is to scatter phonons (vibrations in the lattice) - in other words to reduce the conduction of heat through the material. The problem is that those nanostructure also scatter electrons, which reduces the electrical conductivity as well.
"So our idea was to engineer the nanostructures in such a way that they don't scatter electrons. We wanted to keep using nanostructures, that scatter phonons without interfering with electrons."
The material the team chose to work on was an alloy of titanium, zirconium, nickel and tin, which is actually not a very effective thermoelectric material - it has a figure of merit of around 1.
The researchers investigated the crystal structure of the alloy, along with some other related materials. They noticed that there was a lot of empty space within the lattice in their chosen alloy, but other materials had a structure in which these vacancies were occupied.
By introducing additional nickel atoms into the alloy matrix, they found they could create nanoscale domains in the material with a different, denser lattice structure, but which was very similar in other respects. Podeu commented:
"The benefit of this approach is that we were able to keep a very high coherency between the two phases, because they are made of the same materials arranged into a different structure.
"With that kind of lattice coherency, the electrons can flow from one phase to the next without scattering, but phonons are still scattered, because these different domains have different masses."
Hi-resolution TEM image of one of the quantum dot domains (paler area) within the semiconductor matrix. This domain is about 30nm across. Image credit: Image credit: Pierre Ferdinand P. Poudeu
So that approach was successful in it's goal of reducing conductivity of heat without affecting electrical conductivity - it boosted the figure of merit of the alloy up to 1.5 - 2. But there was another result from the work which was more surprising:
"We were also able to increase the inherent ability of the matrix to convert heat into electricity - something we call the thermopower of the matrix. So we had created a double benefit - reducing the thermal conductivity, and increasing the thermopower - all whilst maintaining the same electrical conductivity. This has not been achieved before in thermoelectric research."
Poudeu's team has been working on this family of thermoelectric semiconductors for several years, as part of the University of Michigan's extensive program of nanotechnology research which spans teams in the departments of physics, materials science and engineering, as well as medicine and biological sciences.
These particular fascinating results were first published in 2011, but it is only recently that the researchers have developed a fuller understanding of why they were able to increase the thermopower of the material in the way that they did.
Rather than simply altering the structure of the alloy, the nanoscale domains which were added to the lattice behave like quantum dots - structures so small that the effects of quantum mechanics can directly affect their electrical properties, instead of being averaged out over the bulk material.
These quantum dots were acting as traps for low-energy electrons, whilst allowing higher energy electrons to pass through unaffected. This had the effect of increasing the thermoelectric conversion efficiency of the alloy - effectively creating a brand new material with a distinct electronic structure. Prof. Poudeu is enthusiastic about the implications of their discovery:
"Now that we have this understanding, it will be possible to apply this concept to systems like lead telluride, or bismuth telluride, which are the best-performing thermoelectric materials right now. It is just a matter of finding the right elements that can be included into the bulk matrix to yield a high enough coherence between the phases to achieve the effect. We are hoping that other research groups will join in with this effort to produce even better thermoelectric materials."
Considering how drastically Poudeu and his team were able to boost the performance of the alloy they worked on, it seems very likely that the goal of commercially viable thermoelectric materials is well within grasp.