The Michigan Engineer News Center

Thruster on a chip

The principle for each is simple: rather than blasting hot gas out of the back of a spacecraft, ion thrusters and nanoFETs use electrical energy to shoot out streams of particles.| Medium Read

Getting payloads into space takes muscle and money – a lot of both. Each space shuttle guzzled 1,400,000 pounds of chemical fuel to reach Earth orbit. But once in space, out of gravity’s grasp, the rules change. Out there it doesn’t take as much force to move a spacecraft around. Nevertheless, once in space many satellites and probes use chemical thrusters, which consume a lot of fuel to accelerate a craft very quickly to its maximum velocity – Voyager I burned all of its fuel within a total of 15 minutes – one burn on its way to Jupiter and one course correction – before coasting the rest of the way to Saturn. Deep-space probes and satellites with long lifespans need additional fuel to adjust position and orientation in space; for those kinds of craft, chemical thrusters aren’t adequate. But there are alternatives. One is the ion thruster. Another, still under development, might be the nano-particle field extraction thruster (nanoFET).

The principle for each is simple: rather than blasting hot gas out of the back of a spacecraft, ion thrusters and nanoFETs use electrical energy to shoot out streams of particles. Ion thrusters fire ionized atoms of a gas such as xenon. NanoFETs emit particles that are a thousand times larger than atoms but still a thousand times smaller than the width of a human hair.

Ion engines and nanoFETs provide very low thrust – it takes a long time for a stream of minute, low-power  particles to accelerate a craft to a high velocity. In the late 1990s, NASA’s Deep Space 1 (DS1) became the first interplanetary probe to use ion propulsion as its primary driving force. Although the engine created less than 100 millinewtons of thrust – about the force needed to hold up a single AAA battery against gravity – it ran for more than 600 days and accelerated the craft to a speed of more than 4 kilometers per second (about 9,000 miles per hour). The only propellant its ion engine needed was xenon gas, which is four-times heavier than air – just 15 percent of the total mass of DS1 was xenon propellant. In comparison, the fuel carried by a chemically-powered spacecraft can account for up to 80 percent of the vehicle’s mass.

NanoFETs will outperform ion thrusters – for several good reasons. Nanoparticles have so much more mass than xenon atoms that, at similar velocities, the nanoparticles pack more punch. Unlike atoms, nanoparticles can be charged very easily – and the greater the charge, the greater the acceleration and the resultant thrust. NanoFETs might outlive ion engines, which suffer from erosion of their electrodes over time. And whereas chemical rockets have demonstrated fuel efficiencies up to 35 percent, and ion thrusters have demonstrated fuel efficiencies more than 90 percent, Professor Brian Gilchrist thinks nanoFETs will do better.

Gilchrist, a professor in the departments of Electrical Engineering and Computer Science, and Atmospheric, Oceanic and Space Sciences, said that the original concept for nanoFET “was motivated by a DARPA request to develop an electric propulsion system for use in space. We weren’t selected, but it got us thinking about the challenges of that project and how it could be done better. The entire nanoFET concept played to one of Michigan’s real strengths – we knew it would be a multi-disciplinary project and we could bring together researchers from aerospace engineering, electrical engineering, materials science, plasma science… we had everything in one place to solve some tough problems.” Gilchrist soon found himself part of team that included Professor Alec Gallimore, director of Michigan Engineering’s Plasmadynamics and Electric Propulsion Laboratory (PEPL).

Gallimore, an Arthur F. Thurnau Professor of Aerospace Engineering and Michigan Engineering’s Associate Dean for Research and Graduate Education, explained that, in figuring out an approach to the development of nanoFETs, PEPL researchers took a look at today’s routine use of nanoscale engineering to create labs-on-a-chip – silicon wafers jammed with vast numbers of transistors and microfluidic channels that transport liquids. This led to the concept of a thruster-on-a-chip, two centimeters square, wafer-thin, perforated with a million micro channels through which an electric field would fire off trillions of nanoparticles. “If our engineers could pull off Brian’s basic idea, we’d have a nanoFET that could deliver up to ten times as much thrust as an ion engine of similar size,” Gallimore said. “It’s a beautiful idea but very tough to execute.”

The first step was to prove the concept. Under Gallimore and Gilchrist, PhD students Louis Musinski (EECS) and Thomas Liu (Aerospace) teamed up to construct a model with particles in the millimeter range. They put a layer of silicone oil between two electrically conductive metal plates a centimeter apart. Aluminum rods about one millimeter long took the place of nanoparticles. As researchers applied voltage across the plates, the rods became negatively charged, stood on end, then shot towards the positively charged plate. There they acquired a positive charge, and the negative plate immediately pulled them back. The rods continued to jump back and forth between the plates but they wouldn’t break out of the oil and enter the thruster channels – the oil’s surface tension held them in. At these very small scales, the surface tension of fluids is relatively strong – insects can walk on water, for example. Through experimentation, Musinski and Liu found that it took a charge of an appropriate intensity to pull the rods through the surface and into the thruster channels. A greater or lesser charge didn’t work.

The concept was sound. The next question was: Would the model work at micro and nano levels?

The answer turned out to be “no.” Decreasing the size by a factor of 1,000 turned the model into a jumble of problems. Transporting the nanoparticles with a fluid wasn’t efficient. Delivering particles to each one of the millions of channels was a particularly knotty problem. Preventing the particles from clumping and blocking channels took some very clever nanotech. Finding the best material for the nanoparticles required not only experimentation but engineering intuition.

Of those and other problems, clumping might have been the biggest challenge. Gallimore said that at a very small scale, atomic integration becomes an issue. “This is an incredibly strong cohesive force that made the particles stick together so firmly that it seemed impossible to pull them apart.”

The word “impossible” is Red Bull for engineers. “That kept a lot of us up in the wee hours of the morning,” Gallimore said. “First of all, the size of the particles makes them hard to work with – they’re about 50 nanometers in diameter. Each nanometer is only three to five atoms wide. Working with particles of that size is what Nik Kotov and Mike Solomon do.”

Professors Kotov (ChE, BME, MSE), Solomon (ChE, MacroE) and the rest of the team came up with a particle that’s a semiconductor in an envelope of proteins. “They tailor-made a cadmium-telluride particle with cytochrome C protein on top of it,” Gallimore said. “It’s an amazing particle. They found that a cadmium-telluride semiconductor can hold a large charge. And the protein outer shell acts like a soft buffer that separates the particles, making it easier to peel them off each other and from other surfaces – and that prevents clumping.”

Saying these particles are easy to separate is one thing, doing it is another thing altogether. Liu found that they could use piezo-electrics to create agitation to separate particles like a sifter removes clumps from flour. At that point, the nanoFET needs to transport the nanoparticles to the channels in the chip. “We started by moving them in a fluid environment, as we did with the aluminum particles in the model,” Gallimore said. “That didn’t work well. We eventually discovered that the particles could move freely by themselves; they didn’t need a transport medium. The solution was using back pressure – think of it as a plunger that pushes the particles through the piezo-electric agitator and then between two plates, or electrodes, as we did in the model. The chip will actually be a stack of layers that we charge to create a powerful electric field which, in turn, charges the particles. Once they have a charge – positive or negative – we adjust the polarity of the field, and they shoot into the micro channels that we’ve bored in the chip. The channels accelerate them until they come out of the chip at high speed.”

More specifically, the plates create an electric potential of about 10,000 volts and accelerate the particles from a stand-still to thousands of meters per second over the course of a micron. But instead of applying all of the voltage in one shot, the nanoFET breaks it up. “The device uses a series of stacked, micron-thick gates,” Gilchrist said. “The gates alternate between conductive and insulating layers to create electric fields. These small but powerful electric fields – each gate provides about 1,000 volts – charge and accelerate a reservoir of nanoparticles, shooting them out into space at about 22,000 miles per hour (10 kilometers a second) to create thrust.”

“Because we want to shoot out trillions of these particles at a time – that’s what it would take to propel a small spacecraft – we need millions of these channels in one chip,” Gallimore said. “That problem fell primarily to Professor Joanna Millunchick (MSE). She was able to bore these holes using nanofabrication techniques that exist today.”

NanoFETs will someday provide the thrust for nanosatellites such as Michigan Engineering’s RAX CubeSat, a cubic nanosat, 10 centimeters per side. The CubeSats could fly aboard a mothership – a large spacecraft that’s difficult and expensive to maneuver and probably in the multi-billion dollar range, unlike CubeSats that are small, agile and inexpensive. Instead of maneuvering the mothership to observe and explore, mission operators would deploy a small fleet of CubeSats. A proposed design puts four of the postage-stamp-sized nanoFETs on top of the CubeSat, and one nanoFET on each side, creating a main thruster to propel the craft, and smaller thrusters to orient it.

“To date, we haven’t made a complete nanoFET,” Gallimore said. “We’re finding that the biggest challenge is the integration of its various elements. The plates must be able to apply the proper charge. The piezo-electrics have to vibrate at the proper frequency. The gates have to charge fast and repeatedly. And it all has to be on one chip, which means the structure has to withstand up to 40,000 volts, the vibration of the agitators, the back pressure… it has to be tremendously strong. We’ve made the particles, the piezo-electrics and the back-pressure system that pushes the particles into the chip; that is, we’ve made major strides, but we still have some very big challenges ahead of us.”

As work continues, the nanoFET team has turned to Professor Khalil Najafi (BSE EE ’80, MSE ’81, PhD ’86), chair, Electrical and Computer Engineering; Schlumberger Professor of Engineering; and Arthur F. Thurnau Professor – his experience with microelectromechanical systems is opening doors that might otherwise have gone unnoticed.

Researchers other than space scientists are also waiting for a complete nanoFET. As the project evolved, these researchers identified a multiplicity of areas in which nanoFET technology might be useful. David Liaw, a graduate student research assistant in the Department of Electrical Engineering and Computer Science, said that nanoFET technology could be used for “precise printing of nano-dimensional conducting lines or the attachment of nanoparticles on a substrate for advanced electronics.”

Liu, now a research fellow, added that researchers might someday use nanoFET technology to create an instrument that injects nanoparticles through cellular walls to deliver chemotherapeutic drugs exactly where they need to go. “That would eliminate side-effects like nausea and anemia. You can see how beneficial nanoFET technology could be.”

“There’s no end to nanoFET’s potential,” Gallimore said. “It’s a device that’ll make interplanetary research more efficient – and there’s no telling what secrets worlds beyond ours might reveal. We’re getting an awful lot from something so little.”


Bill Clayton
Magazine Editor

Michigan Engineering

The outside of the Ford Robotics building

U-Michigan, Ford open world-class robotics complex

The facility will accelerate the future of advanced and more equitable robotics and mobility | Medium Read