Electric engines, 20-nanometer cones, and why Natalya Bailey would skip the first thousand on Mars

Lex Fridman opens with a story Bailey has told about her childhood: looking up at the stars and dreaming of alien intelligences she might one day talk to. He asks the blunt version. Does she think there is life out there, and is any of it intelligent? She splits the question. Given how long the universe has existed and how many stars have planets, a self-replicating organism seems likely. Intelligent life is trickier. She is hopeful, and a little discouraged that, at least allegedly, nobody has been in touch.

Fridman offers a Carl Sagan-style escape hatch: they may already have tried, on a timescale or through a medium we are too dull to notice. He pushes further. What if the medium is us — the collective intelligence of human beings, the very fact that we can ask the question? Bailey has never thought of it that way. "My train of thought has never gone there, but I like it," she says. Somewhere in the picture, she adds, something would have to travel faster than light, and she is hopeful about that too.

The same cosmic curiosity is why she went into the field: to add to the body of knowledge the species has. Life on Mars or Europa would thrill her, though she is not sure about anything this close to home, and contamination would be hard to untangle. Fridman notes that a positive hit would move a term in the Drake equation and imply basic life on trillions of planets. Bailey is unfazed by a null result in our solar system. Earth is clearly the most habitable world nearby, and life elsewhere might not look like us at all. She still suspects trees may be smarter than humans, on a timescale we cannot appreciate.

How do humans get to Mars? Bailey's answer is unusually firm. SpaceX and Elon Musk will put the first person on the surface, not that long from the conversation, even if his timeline slips. The glib version is a time machine back to the early Cold War, with Mars chosen instead of the Moon. Under the joke is a claim: nothing scientific or technological makes the trip impossible soon. The obstacles are political and financial.

Fridman asks who writes the rules once more than a handful of people live there, and notes the trolly paragraph in SpaceX contracts: Earth governments have no jurisdiction; Martians define their own rules. It sounds to Bailey "very much like the founding fathers for this country." Pure scientific curiosity, she says, produces tiny exploration budgets. Something like disaster or impending doom is what actually funds a permanent presence, and she does not wish for that catastrophe. She also does not see a private-sector business model that can carry a colony.

Then she takes a position she admits is unusual in her industry. Spreading off the planet like bacteria may not be what humans are supposed to be doing. Fragile biology, she suggests, may be an unnecessary obstacle in the way of robotic or silicon-based exploration. Sending people as far as Mars is already hard; farther is worse. If the goal is more knowledge, stop trying to send humans. Send robots, maybe a petri dish of stem cells, maybe not even that. Fridman finds the image sad: a lonely spacecraft still flying after civilization is gone. Bailey says it depends on the goal. Another way to look at it is a time capsule of knowledge and DNA that outlives everyone. "How I sleep at night," she says.

The robot-first argument sits next to a younger ambition: Bailey wanted to be an astronaut. The reason is not especially romantic. She likes working in labs and doing experiments, and she wanted the coolest lab — the International Space Station. She treats it as a research platform in orbit, with power, climate control, and microgravity, which is a dry way of describing the job and also a revealing one.

The station, she says, is due to come out of use around 2025, which she finds sad, though private companies are planning stations of their own. Fridman asks the vacation version: if civilian travel opens up, would she go to Mars? No. Her risk profile has changed. She has three small children. She will not be in the first thousand people to go. Earth, she notes, is still pretty nice.

The basic principle, Bailey says, is conservation of momentum. You throw stuff out the back of the engine, and the spacecraft goes the other way. There are two main families. Chemical propulsion is the loud, fiery one people see on television: a fuel and an oxidizer combust, stored chemical energy heats the resulting gas, and a nozzle directs that gas backward. Solid and liquid fuels both live in this bucket. Electric propulsion uses energy from batteries or solar panels instead, and the stuff leaving the vehicle is charged particles, shoved out by an electrostatic or electromagnetic field. Same momentum exchange. Different costume.

Chemical systems are very well understood. In their simplest form they are fireworks, a technology that has been around since around 400 B.C., and even the big engines are quite well mapped. One of the last gaps is the exact products of combustion. Models fall apart in that hot, fast-moving gas, and you have to wander into several scientific fields at once to get closer. What people actually care about, she says, is fuel efficiency. Chemical engines can deliver a lot of instantaneous thrust and very little thrust per unit of fuel. Electric propulsion is the more efficient choice, which is why so much work goes there.

Heating and cooling never leave the problem, because the fire that helps you also melts the rocket. Electric propulsion, by contrast, she finds "refreshingly poorly understood." In college, a propulsion class spent ninety percent of its time on chemical engines and the last ten percent on electric. The professor's line stuck: we only sort of understand how it works, but it works. Even an ion engine, one of the more straightforward electrostatic designs, mixes quantum mechanics, materials science, fluid dynamics, and electrostatics. That combination is what intrigued her.

Bailey sketches the main electric families. Ion engines and Hall thrusters have been around since the 1960s; ion engines are a little simpler because they do not use a magnetic field to make thrust. A third bucket is other plasma machines, including something like a VASIMR engine. A fourth is the class she works on, a colloid thruster, which has a lot in common with an ion engine and enough differences to stand alone.

In an ion engine, a neutral gas such as xenon or argon is injected into an ionization chamber along with a stream of high-energy electrons. Everything moves around at random. The goal is a collision that knocks a secondary electron off a neutral atom and leaves a positive ion — plasma. Some fraction of those ions drift downstream to two grids with holes in them and a voltage between them. The electric field accelerates the lucky ions out the back. About as many ions, by area, hit the walls, lose their charge, and become a main source of inefficiency.

There is a bookkeeping problem. The spacecraft is roughly neutral, and you have just fired a positive beam. Left alone, those ions would eventually come back. An external cathode pumps electrons into the beam and neutralizes it so the exhaust is net-neutral and stays gone. Temperature still matters, especially on the small satellites people now like to launch. The plasma is hot, and those high-energy electrons dump heat into the walls. She thinks the electron energy is on the order of a hundred electron volts, and she would have to convert that to kelvin on paper. Can you stick your hand in it? No.

Mid-century electric-propulsion people already knew there was another source of charged particles besides ionized gas: certain liquids, especially ionic liquids. Ions in a field make force. The catch was scale. You have to manipulate those liquids at nanometers or a few microns, about the diameter of a human hair, and the 1950s had no way to do it. The ideas went into books and went quiet. Silicon MEMS foundries later made the geometry buildable. Bailey's advisor at MIT put the old papers back together and realized the technique could finally be built.

The mechanism is a balance. A strong electric field stresses the liquid; surface tension pulls back. When the two match, the liquid assumes a cone — a Taylor cone. At the tip, the radius of curvature goes toward zero and the field, which scales as one over the radius, would head toward infinity. Instead of a wormhole, a jet of ions issues from the tip. The emission region for a pure-ion cone is something like twenty nanometers. Fridman had pictured something larger. "So this is tiny. Tiny," Bailey says. That is why it could only be done recently.

The engineering trick is not to sculpt a twenty-nanometer cone by hand. They micromachine hundreds or thousands of sharp support tips, with radii of curvature on the order of a micron, and supply liquid to each one from the back of a porous chip. The supports tell you where the beams will form, so the grid above them can have holes in the right places, and they concentrate the field so a much smaller voltage will do. Accion calls the result a thruster chip: one square centimeter, about five hundred emitters today, with a fourfold density increase planned in the coming months. A tank keeps feeding the liquid; power electronics talk to the spacecraft and switch the device. Strictly speaking the work sits closer to ion electrospray than to classic colloid, which meant droplets off a jet. Shrink the cones enough and you get pure ions.

What is still mysterious? Instabilities — how much liquid, how strong a field, why the emission flickers — have become much clearer in just the past couple of years. Two scales remain poorly connected. At the molecular scale they extract large molecular ions; some of the field energy goes into bonds, the molecule vibrates, and sometimes it comes apart. Zoom out and the beam has its own collective behavior. The gap between those pictures is what would let you predict engine performance. Fridman asks if that is theory or engineering. Both, she says, but Accion is a startup. They build, test, observe, and characterize. They are not diving into the theory right now.

Asked for the most beautiful idea in the space, Bailey keeps running into materials science. For ten years, every problem she wanted to dig into ended there if she went deep enough. Even on her own devices, liquid moves from tank to emitter tip by capillary action, which becomes wetting and surface energy and, further down, quantum mechanics. There is so much at these sizes that is not understood, and she finds that inspiring. More broadly, she remembers learning that the same partial differential equation that describes flow over an airfoil is used to price options — Black-Scholes. "This connectedness I find really magical," she says. Fridman jokes that the simulation is reusing code. Her reaction at the time was to interview for a finance job.

You cannot use an engine like hers to get off the ground. Instantaneous thrust is too small. Given time, that acceleration still adds up to speeds that matter for exploration, even for missions with humans on them. The real limit on electric propulsion, she argues, is the power supply. Solar panels are inefficient and bulky; she does not know when batteries will jump. Nuclear power would let electric systems stay extremely fuel-efficient and still deliver more instantaneous thrust for more interesting missions. The world already launches some nuclear-powered spacecraft, but the reactors and heaters are bulky, suited to large interplanetary work, and wrapped in a rat's nest of politics.

Could orbital-class rockets themselves go electric? Possible, she says, if the power supply is much bigger — meaning nuclear — and if you stage the vehicle the way chemical rockets already do, dropping used engines. People have only been willing to launch nuclear systems in a highly redundant safe mode, switched off until they are far away. Using one during ascent would still bother a lot of people. Everyone who works on nuclear power, she notes, has shown how safe it is; public risk estimates still run the other way. The year before the interview, the Trump administration eased some regulations, and NASA and others were starting to pick the work back up. She thinks it is a good moment to look.

Chemical engines still fly in orbit, but that is also where you might choose electric. Even a satellite whose only goal is to stay in its box needs propulsion. Earth's gravity field changes around the orbit and pulls you out of place; other perturbations do the rest. Most operators want more: a burn to dodge debris, a drop to a lower altitude for a sharper image and then a return, and a responsible end of life. Low Earth orbit usually means burning up in the atmosphere. Geosynchronous satellites sit around forty thousand kilometers up. Pushing them down would take more propellant than the whole mission used, so they go higher, into a graveyard orbit that takes on the order of a million years to decay. "We're also cluttering that higher bit up as well," Bailey says. It is just less urgent than LEO.

A rule of thumb she has heard: some small satellites do about three collision-avoidance maneuvers a year. The planning still lives on the ground; little of it is autonomous. Meanwhile Moore's Law has made spacecraft cheap enough that some teams launch with no ability to maneuver at all — just don't hit me. The commercial version of that logic is to fly twenty small ones instead of one large one, so losing a satellite is affordable. The physics is not. One good-sized satellite, hit, becomes a ballistic event on the order of ten thousand pieces of debris, which then go hit the others.

Before about 1990, she says, the typical satellite was the size of a school bus and cost a couple of billion dollars. Accion's first launches rode on shoebox-sized spacecraft built by high-school students. Big satellites are not going away: broadcasting wants reach, and an optical aperture does not follow Moore's Law. The smallest common form factor is a ten-centimeter cube, small enough to hold a softball; some companies go smaller still for low-bandwidth IoT, such as pinging a ship across an ocean, and those usually just fall out of the sky. Her rule of thumb for fitting an engine: bigger than about twenty kilograms, needed on orbit more than a year, or something someone spent more than about a hundred thousand dollars to build. "Small sat," she notes, is still quite large, and the jargon is a mess.

Will there be millions of satellites? Tens of thousands, she can see. Imaging should plateau: one satellite or constellation can sell the same picture to an unlimited number of customers. Communications does not work that way. Broadband and satellite phones are limited by transponders, so more users means more spacecraft, and data consumption is not slowing down. Fridman, having watched a flat-Earth documentary, asks for the most convincing evidence that the planet is round. Bailey picks photographs taken from the Moon — the lunar lander, the surface, Earth in the distance. First you think how tiny and insignificant we are. Then you notice that people actually landed on another body. Fridman calls it a perspective shifter, the pale blue dot. She likes that juxtaposition of smallness and a cool thing we did.

Fridman is a sucker for humans on Mars and teases her first-thousand refusal. Bailey's threshold is closer to two million — roughly the population of the United States when the Declaration of Independence was signed. She wants to arrive when they are writing the documents. It is not the health risk or the engineering of the flight that worries her. Early colonization looks like a prison: a tight enclosed space, a stressful selection problem, and drama, because there is always drama. She wants some rules already in place. She also, she admits, wants Netflix.

Fridman turns that into a rarer opportunity: the first people get to define a society from scratch. The financial system could be cryptocurrency; the politics could try a more direct democracy instead of a representative one. Bailey still wants to participate as a citizen, not to be the person cutting the trail. The first thousand is out. The first million is still too early. Two million is about right.

The real game-changer, Bailey says, would be propulsion that does not throw mass overboard. Every couple of years a startup or a paper claims a device that does it. Conservation of momentum is still there: you carry the propellant, or you mine it from an asteroid. She mentions the Casimir effect — two plates close enough that vacuum fluctuations might produce a net force — and admits she may be mixing several propellantless ideas. It could be real eventually. She does not know how to do it.

The latest episode was the EmDrive, debunked around March of that year. The claim was that microwaves in a resonant cavity produced thrust. People jumped from a loose concept to hardware, measured force on a thrust balance, and other labs built copies and saw similar numbers. Nobody could explain it. Supporters said the microwaves were unchanged but the speed of light inside the cavity changed, so momentum still closed. Finally someone, she thinks at NASA, built the device, measured the same thrust, unhooked it, flipped it backward, and turned it on again. The thrust came out in the same direction. It was an interaction with the test stand or the chamber. It would be wonderful if propellantless thrust worked. She does not know how.

A candidate signal from the direction of Proxima Centauri, 4.2 light-years away, is the cue for speed. What propulsion gets a probe there in a lifetime? She points to Breakthrough Starshot: gram-scale chip probes, accelerated by a laser while they are still near Earth. She loves the project because it is a stack of independent unsolved problems — the laser array, and then enough power at the destination to shout a signal home. No part of it is possible right now, and that is part of the appeal. A crewed ship near light speed leaves the realm she can describe with credibility. Special relativity makes rest mass infinitely heavy as you approach *c*. In her childlike heart, she still thinks we are missing something, whether dark matter or other dimensions. Quantum tunneling is her everyday reminder: an electron is here with this energy, then there with that energy, and we use the effect while barely understanding it.

Building Accion, Bailey says, was relatively straightforward in the early going — she refuses the word easy in rocket science — because an ion engine is a sexy vision. That vision has to stay exciting, because when everything goes to crap it is what gets people out of bed. Early employees matter as much: people with their own networks, so the company is more than the sum of the first team. Her thread is a genuine network, including unsolicited introductions and "weird, uncomfortable conversations" at conferences that skip small talk. She is introverted enough to want to poke her eyes out rather than work a room, so she huddles in a corner with one person and talks about aliens. Those relationships, she says, are the ones you can later call on.

The industry trend to watch is cost, and you have to get ahead of it further than you think. Since the start of the U.S. space industry, people paid twenty to twenty-five million adjusted dollars for an ion engine; customers now want to pay something like ten thousand. NASA does not make those engines with a batch, mass-producible process. They have one person who has been building them for twenty years, like bespoke jewelry. Accion's unconventional move was silicon MEMS manufacturing. Fridman worries that cost pressure, as in the car industry, can squeeze out the margin for real technical risk; Bailey agrees the same pattern shows up in space. She does not treat SpaceX as a competitor. She hopes it is a customer before long. Her MIT advisor, she adds, used to wear a thruster chip as a lapel pin.

2020 was grim on the ground and, for her field, oddly bright: SpaceX and NASA flew the first commercial crew missions, twice, successfully. Accion put the launch on an outdoor screen. Commentators kept saying "Bob and Doug," which felt casual for astronauts, closer to Chris Hadfield playing guitar than to a single human representing the species. She thinks the shift is for the better. Crewed trips to the station have become common enough that people barely notice the splashdowns, which is itself magical. The private sector, she says, stepped up where the U.S. government pulled back — at about the right time for commercial work, and too soon for exploration and science. Other countries are spending more. She wants more public money on the science, even if NASA is not the fastest vehicle for it. "We're just kind of at the very starting point," she says.

Advice for people who want to start something: make sure the vision will get you out of bed and let you rally others, because opportunism does not survive the crappy stretches. Stay close to first principles and fundamentals as long as you can — a deep grasp of the problem, the customer, and how to move a thing forward beats every undergraduate entrepreneurship class, hardware or not. Accion's guiding star has not changed; inside that star they flip-flop constantly, including whether to ship a smaller first product or swing at the park. Culture is the piece she sees treated as an afterthought. Founders rarely carve out time for it, then try to repair a bro culture later. You cannot wait five or ten years and slap an HR person on the mess.

She had assumed you could announce values as something distinct from who the founders are. That is not how it works. "We just kind of like ooze out our behaviors and then the company grows around that." You have to do the introspection, or you get a bad culture. Fridman talks about losing his temper at work in a Steve Jobs register and having to decide, deliberately, whether that signal is worth the cost. The first project Bailey took to a leadership coach was how not to snap at people who were being idiots. Step one was getting good at apologizing, because the behavior would take longer to change. The coach's drill was almost silly: every time you walk through a doorway, think calm, take a breath, then respond. Habits moved a little. She is better at it now.

Three books. F. Scott Fitzgerald's *Tender Is the Night* marked the start of her adult reading; going back, she does not think it is that good, but the flowery prose made the classics feel less intimidating once she noticed they were mostly love stories. David Deutsch's *The Beginning of Infinity* put physics rigor around memes, DNA as the ultimate meme, infinity, and objective beauty, and it also, she says, had something to do with marrying her husband. Harry Potter is the third, without apology: the first three books came out when she was ten, she grew up on Harry's clock, and she has read all seven nine or ten times. Anything that keeps you reading counts, including a trashy detective novel in a dry spell. The official story of Accion is accelerate plus ion. The real one is Accio, the summoning charm, plus an *n*. Harry uses it to call his broom from the dorm while fighting a dragon.

Fiction, for her husband, became a way to practice empathy; for Fridman it is travel without a plane. Bailey told her three-year-old they read in order to see places in the mind, and the child preferred *Frozen* with popcorn. Movies decide the pictures. A book mixes with your own hopes and fears, which is why adaptations sometimes hurt. On the meaning of life — a question you do not end up in space propulsion without asking — she thinks survival and reproduction are no longer the whole story. The pursuit and generation of knowledge is. If an asteroid takes us out, the endeavor was still meaningful if the next civilization does not have to start over. She loved early Google's mission to make the world's information searchable, and she thinks everyone should donate to Wikipedia. A lonely USB drive of the encyclopedia would mean the civilization was well spent. The other half, cooperation and love, she has not fully connected yet. Day to day it boils down to two things: "love and knowledge." Fridman closes with Carl Sagan: all civilizations become either spacefaring or extinct.