POWER CHALLENGE — Sandia electrical engineer Bob Kaplar checks out a test circuit built under a Grand Challenge Laboratory Directed Research and Development project to evaluate the switching performance of wide bandgap and ultra-wide bandgap power semiconductor devices. (Photo by Randy Montoya)
Finding more powerful semiconductor materials
Silicon has long been the go-to material for semiconductors that power the electronic world. Now scientists are looking far beyond that omnipresent element to materials that could make everything from computers to power grids to electric cars more mighty and energy efficient.
Electrical power doesn’t travel a one-way street. In a technology called power electronics, it is converted from one form to another when you plug in a computer, drive a car, or flip on a light.
“Power electronics use semiconductor devices such as transistors, diodes, and thyristors to control the flow of electrical energy by switching electronic circuits,” says electrical engineer Bob Kaplar (1768), who is leading a Laboratory Directed Research and Development (LDRD) project exploring new, more powerful and energy-efficient semiconductor materials. “If you want to convert a DC signal into an AC signal, the actual circuit that does that is complex. But the basic idea is to turn the DC signal on and off.”
Semiconductor devices are switching systems that convert voltages and currents. The switches have been made of silicon since about the 1950s when semiconductors were first developed, replacing in many applications power conversion that uses transformers, or coils of wire around magnetic cores. Silicon is at the center of all microprocessors, computer chips, cell phones, and more. “Silicon is the core material that the device that functions as a switch is made of,” Bob says. “When semiconductor devices were invented, people started making big transistors that could handle large amounts of power.”
Because power electronics process substantial amounts of electrical energy, and energy is lost when power is converted, there has been a move over the past decade to replace silicon with other materials that would be more energy efficient. “The more you can reduce the loss, the better the energy efficiency,” Bob says. “Power electronics had been viewed as a not-so-exciting area in the past. But now there is a resurgent interest in it.”
Bandgaps and energy
The new semiconductor materials are referred to as wide bandgap. Bandgap, a fundamental materials property, is an energy range in a solid where no electron states can exist. In the electronic band structure of solids, the bandgap generally refers to the energy difference in electron volts between the top of the valence band and the bottom of the conduction band in insulators and semiconductors. If the valence band is full and the conduction band is empty, electrons cannot move in the solid. But if some electrons transfer from the valence to the conduction band, then current can flow. So bandgap is a major factor determining the electrical conductivity of a solid.
Wide bandgap refers to higher-voltage electronic bandgaps significantly larger than one electron volt (eV), typically at least three eV. The bandgap of silicon is 1.1 eV and gallium arsenide, another common semiconductor material, is 1.4 eV. Wide bandgap semiconductor materials such as silicon carbide (SiC) and gallium nitride (GaN) allow devices to operate at much higher voltages, frequencies, and temperatures than the conventional materials, so more powerful, cheaper, and more energy-efficient electrical conversion systems can be built.
Wide bandgaps have already revolutionized lighting, particularly in the area of light-emitting diodes, or LEDs, which are widely available and are replacing incandescent and fluorescent bulbs. But as transistors, or switches, in modern power electronics, they also have the potential to vastly improve the performance of electrical power grids, electric vehicles, motors for elevators and HVAC systems, and even computer power supplies. Smaller, faster switches mean less loss of power. “Faster switching also means you can make other parts of the circuit smaller, such as capacitors and inductors,” Bob says.
Wide bandgap has the potential to substantially reduce the estimated 10 percent energy loss between generating electricity and transmitting it into a home or business. “In a decade or two, the giant transformers in your neighborhood distributing power from the electric grid to homes, which now weigh 10,000 pounds, could be replaced by things the size of a suitcase that weigh 100 pounds,” says Sandia Fellow and materials scientist Jerry Simmons (1000).
And if electric vehicles could tap the potential for wide bandgap power electronics to withstand higher temperatures, they might not need a liquid cooling system, reducing the system’s complexity and improving vehicle range because the car would weigh less. “There are non-civilian applications as well,” Bob says. “The military wants small power converters on unmanned aerial vehicles, and the Navy is interested in electric ships. You want as much power as you can get in a confined space. These advantages are pretty universal.”
Leapfrog to the generation after next
Sandia is researching SiC and GaN, but it’s also working to leapfrog over these next-generation materials to the generation-after-next, ultra-wide bandgap materials such as aluminum nitride (AlN), which has a bandgap of 6.2 eV. The Ultra-Wide Bandgap Power Electronics Grand Challenge LDRD project that Bob is leading is at the end of its first year. Grand Challenges are three-year LDRD projects that focus on bold, high-risk ideas with potential for significant national impact.
“Potential benefits like shrinking system size and high-temperature operation become even greater with ultra-wide bandgap materials,” Bob says. “We’re also interested in other harsh environments. There the challenges become greater.”
AlN and GaN are compatible enough to be mixed. That allows researchers to take small steps toward developing AlN by gradually increasing the amount of AlN versus GaN to study behavior and the effect of lattice mismatch between the semiconductor and the material it’s grown on, Bob says.
Estimates predict SiC could perform 100 times better than silicon for power switching, GaN could be 1,000 times better than silicon, and AlN could be 10,000 times better than silicon. However, their potential can’t be tapped until researchers better understand how the materials work, develop mature techniques to process them, and address reliability concerns, particularly for high-consequence uses.
Lots of energy in a small package
The Ultra-Wide Bandgap Grand Challenge is the flagship Laboratory Directed Research and Development project for Sandia’s Power on Demand Research Challenge aimed at developing electrical power systems with the smallest size and weight, while handling the largest possible amount of energy. The research challenge tackles underlying fundamental science questions, engineering applications, and technical challenges for devices, materials growth, and power systems.
The Grand Challenge covers three areas: materials growth; device design, fabrication, and testing (including demonstration of efficient switching); and defects and radiation resistance. It’s exploring ways to grow ultra-wide bandgap materials with fewer defects and different device designs to exploit the properties of materials other than silicon.
Although some devices using SiC and GaN are on the market, thorny problems remain, Bob says. Common performance issues include defects, incompatibility with the microelectronics substrates on which the materials are grown, and the impact of integrating a device into a larger system. Sandia researchers can evaluate those problems impartially, building on expertise from decades of nuclear weapons work.
The downside to wide and ultra-wide bandgap technology is that it is not as mature as the silicon industry, which has a huge manufacturing infrastructure. “It’s easy to control the properties of silicon and related materials,” Bob says. “People know how to process those really, really well.”
Researchers are not at the point of making a power converter out of ultra-wide bandgap materials but envision such a device several years down the road. “At the circuit level we’re characterizing the devices, measuring how much voltage we can put across before it goes into breakdown and how fast the switching transient is when we turn it on and off,” Bob says. “We have voltage targets for the devices we’re building. We can see all the pieces fitting together and moving toward the devices and circuit demonstrations.”
About 50 scientists are working on the Ultra-Wide Bandgap Power Electronics Grand Challenge, and Bob hopes the research will continue when the challenge ends. “This is forefront materials science,” he says. “This is a brand new field to go beyond the wide bandgap materials, and not many people are working on it. The impact potential is huge.”
Different methods exist to convert electrical energy from one form to another. For example, direct current (DC, or constant) voltage can be converted to a lower DC voltage by connecting two resistors in a series arrangement known as a voltage divider. But that method of conversion is inefficient and wastes much of the power as heat. An alternative approach is to switch the DC voltage on and off, and take the average of the resulting on-off-on signal. That switching approach is known as power electronics and is much more efficient, so that little power is wasted as heat. The switching approach also allows for a wider variety of types of power conversion, for example, from a lower DC voltage to a higher DC voltage, from DC to alternating current (AC), from AC to DC or from one AC frequency to another.