Manufacturing Bits: May 29

Utilizing Heat For Energy
One of the big problems in electronics in general, and semiconductors particular, is heat. And it’s not just about leakage current anymore. Heat is a problem at every level, from circuit design to the materials being used inside the chips, as well as warpage between die caused by heat after they are packaged together. Heat can prematurely age chips as well as destroy them, and it can affect the performance of devices throughout their lifetime.

While this has spawned strategies ranging from dark silicon to multiple power domains and even microfluidics, one alternative may be to actually utilize that heat for additional power generation. Researchers at MIT say they have found a way to double the amount of electricity generated by thermoelectric materials.

“If everything works out to our wildest dreams, then suddenly, a lot of things that right now are too inefficient to do will become more efficient,” according to Brian Skinner, a postdoc in MIT’s Research Laboratory of Electronics. “You might see in people’s cars little thermoelectric recoverers that take that waste heat your car engine is putting off, and use it to recharge the battery. Or these devices may be put around power plants so that heat that was formerly wasted by your nuclear reactor or coal power plant now gets recovered and put into the electric grid.”

The idea behind thermoelectric materials is that when one side of a material is heated, excited electrons move toward the colder side of that material, where they aggregate. That, in turn, creates a measurable voltage. But in most materials, electrons are confined to specific bands, and it has proven extremely difficult to induce electrons to cross the gaps between those bands.

Skinner and Liang Fu, associate professor of physics at MIT, have taken a different approach. Rather than focusing on semiconductors or insulators, they have focused on topological semimetals, which currently are being studied for quantum computing. What’s intriguing about these man-made materials is the band gap is zero.

When thermoelectric materials are heated on one side the electrons leave behind holes, according to the researchers. Normally, positively charged particles that accumulate on the cold side would cancel any beneficial effects. But by applying a magnetic field, electrons and holes can be reversed. So electrons would move toward the cold side, and holes move toward the hot side. That can significantly increase the amount of electricity generated by heat.

It’s still early in this research, but the implications of this development is intriguing for lowering power across a wide range of devices.

Time Crystals
Quantum computing has set off a new wave of research, including whether structures can exist in time that do not repeat in space. This is something of a mind-bender, because it brings the fourth dimension into the center of research rather than treating it like an unalterable part of an equation.

These so-called time crystals have been discussed and theorized for years. What makes them particularly interesting is they could increase the longevity of structures or phenomena that exist only for extremely short periods of time.

“Nature has given us a system that wants to be coherent over time,” says Vladimir Eltsov, senior scientist and leader of the ROTA research group at Finland’s Aalto University. “The system spontaneously begins to evolve in time coherently, over long periods of time, even infinitely long.”

How this actually works with real-world applications isn’t clear yet. But slowing or eliminating decay of devices could have a huge impact on electronics of all sorts.

The researchers at Aalto demonstrated a time quasi-crystal and observed its transition to a superfluid time crystal. They said that by understanding how and when time crystals materialize, they may be able to develop coherency in other devices, regardless of environmental factors.

Observing Strain Across Six Atoms
Adding strain into materials is a well-known approach to improving performance because it allows for improved electron mobility. But how strain will work in increasingly thin materials at advanced process nodes isn’t entirely clear.

Researchers at the University of Connecticut’s Institute of Materials Science have found that a six-atom-thick bilayer of tungsten diselenide showed a 100X increase in photoluminescence when subjected to strain. In the past, that luminescence was significantly less.

“Experiments involving strain are often criticized since the strain experienced by these atomically thin materials is difficult to determine and often speculated as being incorrect,” says Michael Pettes, assistant professor of mechanical engineering at the school. “Our study provides a new methodology for conducting strain-dependent measurements of ultrathin materials, and this is important because strain is predicted to offer orders of magnitude changes in the properties of these materials across many different scientific fields.”

This has potential implications for materials that are even thinner, such as graphene, which is just one atom thick. Graphene is a tough, single-cell material that is extremely conductive, but researchers have found it difficult to work with because it does not inherently have a band gap. Researchers at the Catalan Institute of Nanoscience and Nanotechnology reportedly have devised a way to grow graphene with a band gap of 1 electrovolt (eV).

A band gap is critical to controlling the flow of electrons, and in particular to turn off that flow. Silicon’s band gap is 1.14eV, while III-V materials such as aluminum nitride and gallium nitride are even higher, at 6.0 and 3.4eV at 302K.

Pettes and Wei Wu, a graduate student in Pettes’ lab, measured the impact of strain engineering on tungsten diselenide by encapsulating it in a layer of acrylic gas and then heating it in an argon gas chamber. Then they subjected it to a bending device and tracked the effects with a Horiba Multiline Raman Spectrometer at Harvard University.

“Our new method allowed us to apply around two times more strain to the 2-D material than any previous study has reported,” says Pettes. “This is the first time that extrinsic control over an indirect-to-direct electron band gap transition has been conclusively reported. Our findings should allow computational scientists using artificial intelligence to design new materials with extremely strain-resistant or strain-sensitive structures. That is extremely important for the next generation of high-performance flexible nanoelectronics and optoelectronic devices.”

Michael Pettes, left, and Wei Wu check a specially engineered device they created to exert strain on a semiconductor material only six atoms thick. Source: Peter Morenus/UConn