How we might realistically reach the stars
It is July 20, 2019, the 50th anniversary of the first Apollo moon landing. The teenagers who watched the Apollo landings and dreamed of vacationing on other worlds are retired now, and never had a chance to even orbit Earth. They could not have imagined that in the 50 years hence, the United States has not only made no progress in human exploration of other worlds, but lost the ability for human spaceflight altogether. The Space Shuttle was never a stepping stone to destinations beyond low Earth orbit. The Apollo landings did not lead to human colonies on the Moon or Mars. A manned Mars mission is 20 years away, as it has been for decades, and as it might forever remain.
Will we ever become an interplanetary species — or even an interstellar one? Here, I explore the possibilities. I’ll start with the easily doable ideas for increasing access to space, and get progressively more ambitious until we reach the truly absurd. I’ve restricted this list to only the ideas that do not violate the laws of physics as we currently understand them, and that are possible with current or near-future technology. For this reason, I won’t be talking about warp drives, wormholes, or even space elevators .
The Problem
Why is space travel hard? Why is it that 60 years after Sputnik, our rockets have barely improved while computers and electronics have improved by leaps and bounds? To launch a satellite into space requires a big rocket today, just as it did in 1980, and just as it did in the first years of the Space Age.
The answer, unfortunately, lies in the fundamental laws of physics. These laws determine the maximum efficiency of a chemical rocket. Better technology may make chemical rockets cheaper or slightly more efficient, but the fact that a huge rocket is needed to lift even a small payload into orbit is a fundamental property of our universe that no amount of technology can change.
The most important equation in rocketry is called, appropriately enough, the rocket equation:
This says that the mass of a rocket + payload rises exponentially with v/v_e, the ratio of the desired final velocity to the exhaust velocity. To accelerate a payload to the exhaust velocity, the rocket needs to be 73% fuel. Twice the exhaust velocity requires the rocket to be 88% fuel; three times the exhaust velocity requires it to be 95% fuel. Rockets need 14 km/s to escape Earth’s pull once air resistance is taken into account. With a typical exhaust velocity of 4 km/s, we find that chemical rockets can only deliver exp(-5) = 1% of their mass into interplanetary space. This is the ultimate reason why space travel is difficult and expensive.
Of course, all our problems get exponentially better if we can raise the exhaust velocity by an order of magnitude. To understand why this isn’t possible, we need to understand what sets the exhaust velocity. Chemical rockets work by mixing a fuel and an oxidizer, and burning the mixture. Chemical reactions, including burning, are entirely due to the interactions between electrons. When electrons are transferred from one molecule to another, the final configuration has less energy than the initial configuration, and the extra energy turns into the kinetic energy of the exhaust. The typical energy released by electrons moving around is a few electron-volts. This energy, when given to a typical exhaust molecule weighing several atomic mass units, gives the molecule a speed of a few km/s. This number can be calculated from the kinetic energy equation E = 1/2 mv², or v=sqrt(2E/m) — plug in typical values for E and m and you’ll almost always get a speed of a few km/s. This is the reason that the fastest bullets travel at a few km/s. It is why every fuel and oxidizer combination ever used in a rocket has a exhaust velocity of a few km/s.
Knowing this, there are two possible ways of increasing the exhaust velocity: make electrons lose more energy upon combustion, and decrease the mass of the exhaust particles. The first is difficult because, in very handwavy terms, electrons lose energy when they come closer to atomic nuclei. However, electrons cannot come arbitrary close to atomic nuclei, because quantum mechanics sets a minimum distance (more accurately, a lowest-energy state) that electrons can approach a nucleus. The “quantum” in quantum mechanics refers to the quantization of energy levels, which in turn means there is a lowest possible energy. For hydrogen atoms, this lowest possible energy is -13.6 eV, the Rydberg constant. It is not possible to gain more than 13.6 eV of energy by throwing an electron at a hydrogen nucleus. For most molecules, the maximum obtainable energy is even less than 13.6 eV.
The second approach, decreasing exhaust molecules’ mass, is difficult because we’re already using the lightest possible elements. The highest efficiency engines burn hydrogen, the lightest element, with oxygen, the 8th lightest element. There is no element lighter than oxygen that is suitable as an oxidizer, because none of them react violently with hydrogen. Helium is an inert gas; lithium and beryllium are metals; boron, carbon, and nitrogen don’t react violently with hydrogen. Thus, hydrogen and oxygen is the lightest combination possible. They burn to produce water, with a mass of 10 atomic units — one of the lightest possible combustion products.
Thus, we run into a conundrum. The fundamental properties of matter limit the exhaust velocity to a few km/s. The rocket equation says that we can’t propel anything to more than a few times the exhaust velocity without expending hundreds of times the payload’s mass in fuel. The laws of gravity say that to escape Earth’s gravity, we need a speed of 11 km/s; add in air resistance and that number is closer to 14 km/s. From this list of orbital rocket engines, we can read off the exhaust velocity — just multiply the specific impulse in seconds by 9.8 m/s² to get the exhaust velocity in m/s. The most efficient engine ever made burned hydrogen and oxygen and had an exhaust velocity of 4.7 km/s. The engine was first flown in 1962, showing just how little progress has been made in the past 50 years.
Cutting costs
The simplest and most obvious way to cut the cost of space travel is to accept that rockets will always be enormously inefficient, but make them cheaper. The traditional ways of cutting costs are always available. Governments can encourage competition among private space launch companies. The companies themselves can adopt automation, new technologies, innovative industrial methods, and vertical integration. These traditional cost-cutting approaches are the main reason that SpaceX can offer low launch costs.
The other way of cutting costs — also adopted by SpaceX — is reusable rockets. Instead of throwing away the rocket, the different stages of the rocket can be brought back to Earth and reused. SpaceX is only reusing the first stage of the rocket, and there are no serious proposals by any company to reuse anything more than the first stage. The fundamental reason behind this is that in order for a rocket to reach Earth orbit, it must accelerate to at least 7.9 km/s. When the first stage of the Falcon 9 detaches, it is still inside the atmosphere and moving at less than 2 km/s. When the second stage engine cuts off, the rocket is at orbital velocity. Since kinetic energy goes up as the square of the speed, the first stage has only 5% the kinetic energy per unit mass as the second stage. This means that while the first stage can return to Earth without significant heat shielding, the second stage must dissipate 20 times the amount of energy — an impossible task without a heavy heat shield.
While a reusable rocket can in theory cut costs, even that is not guaranteed. After the first stage lands, it needs to be inspected, refurbished, and refueled before it can be used for the next launch. Rockets suffer high temperatures and pressures in the combustion chamber, residue from incomplete combustion, and vibrations during ascent. These can damage components and limit the lifespan of the stage. The Space Shuttle, for example, was reusable — but the fastest ever refurbishment time was 54 days, and each mission cost 400–1500 million dollars due to the complexity of refurbishment. Although it is possible that a reusable first stage would save SpaceX money in the long run, this remains to be demonstrated.
Skyhook
In order for cheap access to space to become a reality, we need to move beyond chemical rockets. When discussing non-rocket space launch, the skyhook is often unfairly neglected in favor of the more famous space elevator. This is unfortunate, because no known material is strong enough to build a space elevator with — not even carbon nanotubes. The skyhook, on the other hand, can be built with currently available materials.
The skyhook consists of a massive space station in low Earth orbit — say at 600 km altitude — with a long (say 500 km) rope that can almost reach the ground. The space station + rope combination is rotating fast enough that the end of the rope is stationary relative to the ground at the moment of its closest approach to the ground. That way, a specially designed spaceplane can grab onto the rope at the moment it swoops down. The massive space station then drags the relatively light plane up into orbit. All the plane needs to do is hang on for dear life as it’s accelerated by the force exerted through the rope. After the rope accelerates the payload to orbital speed, the spaceplane lets go, and fires its own engines to make final orbital corrections. In this process, the space station loses altitude, because it sacrificed some of its own momentum to accelerate the payload. The space station then has to recover this altitude, which it can do without propellant by running an electric current through an electrodynamic tether.
In 2000 and 2001, Boeing Phantom Works studied the rotating skyhook in detail and concluded that the space station would need to be 1000–2000 times more massive than the payload. Otherwise, the payload would drag the space station into a low enough orbit that it re-enters the atmosphere. To avoid such a massive space station, Boeing Phantom Works considered a skyhook with half the rotation rate, so that the tip velocity is 4.1 km/s relative to ground rather than 0 km/s. A hypersonic spaceplane would then need to move at Mach 10 to sync up with and catch the skyhook. A Mach 10 spaceplane does not exist, but if DARPA’s XS-1 spaceplane works, it will exist by the end of this year.
Skyhooks require a massive upfront investment in the form of the space station, tether, and spaceplane. However, they are possible with current materials. A skyhook would cut the delta-V required by 10 km/s, the amount required to get into orbit with a rocket. Assuming typical chemical propellants, this corresponds to a 10–20 fold reduction in fuel mass, making interplanetary travel much cheaper.
Nuclear thermal rocket
Streamlining business operations and reusing first stages can only get us so far. To truly revolutionize rocketry, we have no choice but to increase the exhaust velocity.
One idea for doing so is the nuclear thermal rocket. This idea is exceptionally old and simple. A nuclear reactor vaporizes liquid hydrogen, the gaseous hydrogen leaves the nozzle at high speed — and that’s it.. There is no fuel, no oxidizer, no combustion. This idea is so old and simple that the US started development in 1955 under the name NERVA, with the first engine being completed several years later. Although no NTR ever flew, the engines developed by Project Rover are by far the most efficient engines ever developed with a thrust to weight ratio greater than 1. Hydrogen-powered NTRs achieve an exhaust velocity of 8.5–10 km/s, twice as fast as the best chemical engines.
Plug in this number to the rocket equation and it would naively appear that a NTR can boost 35% of its mass into low Earth orbit — far better than chemical rockets, which can only achieve 13%. If we want to go beyond Earth orbit — say to Mars, which takes an additional delta-V of 4.7 km/s — the difference becomes even starker: 5% for a chemical rocket, 21% for a nuclear rocket.
So if NTRs are so good, why isn’t everyone using them for orbital launches? The catch is that nuclear engines are heavy and don’t produce much thrust. Pewee, the smallest engine developed by Project Rover, weighed 3.2 tons and could only lift 11 tons off the ground. While it might seem from the previous paragraph that you can launch a 1-ton communication satellite into orbit with a 3 ton rocket, 3 tons isn’t even enough to contain the engine. Just to get the engine into orbit without any payload, we need 9 tons of rocket — and that’s assuming the fuel tanks, payload, electronics, and other structural elements weigh nothing. Add them in and we exceed the 11 tons that the engine can lift. Using two engines doesn’t help. In that case, we would have 6.4 tons of engine that can lift 22 tons off the ground, but lifting just the engine into orbit takes 18 tons of rocket — again just 4 tons shy of the total lifting capacity of the engine.
To get around this problem, we have two choices: greatly increase the thrust-to-weight ratio of the engine, or use the engine only for the upper stage. The former is theoretically possible, even with 60 year old technology, although whether it is possible in practice remains to be demonstrated. Dumbo, a competitor to NERVA that was not selected, could manage a ratio of 70 with an engine mass of 5 tons and an exhaust velocity of 8 km/s. To illustrate how well Dumbo can do, let’s suppose we have a rocket with two engines that weighs 500 tons when fueled (the same as a Falcon 9) and that the dry mass is 33 tons (10 tons more than the Falcon 9). A 500 ton NTR can bring 150 tons into orbit. If 33 tons are the fuel tanks, engines, and other structure, we’re still left with 117 tons of payload that can be sent into low Earth orbit. This is significantly better than the 13 ton payload capacity of Falcon 9. More importantly, we achieved this in a single stage to orbit (SSTO) design. Even if returning the entire rocket to Earth for reuse is too difficult due to re-entry heating, we could return the nuclear reactor and core — probably the most expensive part of the rocket — and reuse those portions only.
The second option, namely using a NTR for the upper stage, is fully achievable without any further technological development. In April of this year, Congress gave NASA $125 million to develop nuclear thermal propulsion with the eventual goal of shortening journey times to the Moon and Mars. A NTR would shorten the trip to Mars by 20% — not a huge decrease in itself, but it would greatly expand the launch window for journeys to Mars and back, giving more mission flexibility and abort options in case of emergency.
Whenever anything nuclear is discussed, visions of mushroom clouds and cancer inevitably dance in everyone’s heads. Usually, the danger is grossly exaggerated. Although radiation really is dangerous, nuclear reactions are a million times more energetic than chemical reactions, meaning that the amount of nuclear material needed is a million times less than the amount of kerosone, UDMH, or other chemical propellants. The proper question is not whether nuclear material is dangerous — for the sake of argument, let’s grant that it is — but whether it is more than a million times more dangerous than chemical reagents. For NTRs, the answer is almost always“no”.
NTRs are significantly safer than chemical rockets. Chemical rockets contain both a fuel and an oxidizer in close proximity, and become enormous bombs when the two find each other in unexpected ways. While NTRs do have hydrogen, they have no oxidizer for the hydrogen to mix with, significantly slowing down the pace of any reaction. For the hydrogen to burn, it would first need to escape the rocket, vaporize, and mix with the oxygen in the air. This is a significantly slower and less violent process than the explosion of a chemical rocket.
If a NTR doesn’t explode, its exhaust is pure hydrogen, with no radioactive isotopes. Even if a NTR does explode, its radiation poses negligible safety risks. Uranium, by itself, is only barely radioactive — it decays with a half life of 4.5 billion years for U-238, and 700 million years for U-235. Uranium decay emits alpha particles, which can be blocked by a piece of paper, a few inches of air, or the layer of dead skin cells on your body. You can safely hold uranium in your hands, with toxicity being more of a concern than radioactivity. It is only when the reactor has been operating for a while that short-halftime radioisotopes are produced which emit significantly more radiation. If a NTR is used as an upper stage in an interplanetary mission, it will never be ignited in Earth’s atmosphere, making widespread dispersal of radioactive material impossible.
Fission fragment rocket
Alert readers might have noticed that even though nuclear energies are a million times larger than chemical energies, a nuclear thermal rocket isn’t *that* much better than a chemical rocket. Its exhaust velocity is higher by a factor of two, not by a factor of a million. The reason for this is we can only heat the hydrogen to 3500 K before the engine starts to melt. After all, the engine is held together by chemical bonds. We know that chemical propellants reach temperatures of thousands of degrees when burned, and it’s no coincidence that materials start to melt at the same temperature.
There are many proposals to get around this problem to harness the true power of the atomic nucleus. For example, there are open cycle gas core designs that use magnetic confinement to keep the fissioning uranium away from the engine walls. These have theoretical exhaust velocities of 20–30 km/s, but none of them have left the drawing board.
Enter the fission fragment rocket. Conceptually, it’s dead simple. A magnetic field is used to suspend particles of fissile material, preventing them from touching and melting the engine walls. The fission products, which emerge at a few percent the speed of light, are channeled backwards. This gives the engine an effective exhaust velocity of up to 10,000 km/s, making it possible to reach relativistic speeds.
Unfortunately, the fission fragment rocket has not progressed beyond the design stage, and whether the practical challenges can be solved is unknown. For example, is it possible to confine the fission products sufficiently to prevent melting of the engine walls? What’s the maximum thrust that can be sustained? Only further work can answer these questions.
Project Orion
Project Onion is simultaneously the craziest and most realistic plan for interstellar travel. Originating in the 1960s, it is the only plan that is both possible with current technology and has undergone some level of field testing. Speeds of up to 10% the speed of light are possible, making it possible to visit Alpha Centauri in 50 years. It is, nevertheless, mind-boggling in its audacity.
The idea is to build a gigantic spaceship filled with nuclear bombs. Bombs are ejected one at a time and detonated, causing the ejecta to hit a giant pusher plate. The pusher plate is attached to the spaceship via a dampening mechanism to convert the sudden impulse from the blast to a more gradual pushing force. This design avoids the inherent limitation of the nuclear thermal rocket — the fact that the propellant can only be heated so much before it melts the engine — by having the nuclear reactions entirely outside the spacecraft. Having the reactions occur outside also avoids the inherent limitation of the fission fragment rocket , namely that nuclear reactions can only happen so fast before they melt the reactor, greatly limiting the thrust.
So how big does the Orion spaceship need to be? One scenario proposed by Freeman Dyson, a member of the original Orion team, called for a 400,000 ton spaceship with 300,000 one-megaton hydrogen bombs. Three bombs would be detonated per second, accelerating the ship to 3% the speed of light over 10 days.
Unfortunately, Partial Test Ban Treaty of 1963 killed the project and ended all hopes that any version of Project Orion — even a far more realistic interplanetary version — would reach the launch pad. Before this happened, the Orion team reported no show-stoppers that would prevent this plan from being realized, and had even field-tested a pusher plate propelled by conventional explosives.
The Future
Humanity has come a long way. In less than a century, we progressed from inventing the steam engine, to building the first airplane, to putting a man on the moon. Unfortunately, progress in human spaceflight has since stalled, with little visible progress in the past 50 years.
Progress has stalled due to fundamental physics, not technological limitations. A molecule’s chemical bonds only hold so much energy; that energy can only accelerate the molecule to a few km/s. The rocket equation makes it impossible to propel a payload to more than several times the exhaust velocity without carrying thousands of times more fuel than payload. No technological breakthrough will ever get around this fundamental physical limitation.
In order to make interstellar travel or fast interplanetary travel a reality, we need to go nuclear. Nuclear energy densities are millions of times higher than chemical energy densities; hence, nuclear exhaust velocities are thousands of times higher than the maximum possible with chemical propellants. The nuclear thermal rocket is a baby step in the right direction, but it is limited by the melting point of the reactor walls. Melting, being a chemical process, limits the effective exhaust velocity to around 9 km/s. In the future, we will need to build a fission fragment rocket, Project Orion, or another type of nuclear rocket to have any hope of reaching the nearest stars within a century.
