Colonizing Mars

In remarks at the International Astronautical Congress in Guadalajara, Mexico on September 29, 2016, SpaceX founder and CEO Elon Musk revealed to great fanfare his company’s plans for an Interplanetary Transport System (ITS). According to Musk, the ITS would enable the colonization of Mars by the rapid delivery of a million people in groups of a hundred passengers per flight, as well as large-scale human exploration missions to other bodies, such as Jupiter’s moon Europa.

I was among the thousands of people in the room (and many more watching live online) when Musk gave his remarkable presentation, and was struck by its many good and powerful ideas. However, Musk’s plan assembled some of those good ideas in an extremely suboptimal way, making the proposed system impractical. Still, with some corrections, a system using the core concepts Musk laid out could be made attractive — not just as an imaginative concept for the colonization of Mars, but as a means of meeting the nearer-at-hand challenge of enabling human expeditions to the planet.

In the following critique, I will explain the conceptual flaws of the new SpaceX plan, showing how they can be corrected to benefit, first, the near-term goal of initiating human exploration of the Red Planet, and then, with a cost-effective base-building and settlement program, the more distant goal of future Mars colonization.

As described by Musk, the SpaceX ITS would consist of a very large two-stage fully-reusable launch system, powered by methane/oxygen chemical bipropellant. The suborbital first stage would have four times the takeoff thrust of a Saturn V (the huge rocket that sent the Apollo missions to the Moon). The second stage, which reaches orbit, would have the thrust of a single Saturn V. Together, the two stages could deliver a maximum payload of 550 tons to low Earth orbit (LEO), about four times the capacity of the Saturn V. (Note: All of the “tons” referenced in this article are metric tons.)

At the top of the rocket, the spaceship itself — where some hundred passengers reside — is inseparable from the second stage. (Contrast this with, for example, NASA’s lunar missions, where each part of the system was discarded in turn until just the Command Module carried the Apollo astronauts back to Earth.) Since the second-stage-plus-spaceship will have used its fuel in getting to orbit, it would need to refuel in orbit, filling up with about 1,950 tons of propellant (which means that each launch carrying passengers would require four additional launches to deliver the necessary propellant). Once filled up, the spaceship can head to Mars.

The duration of the journey would of course depend on where Earth and Mars are in their orbits; the shortest one-way trip would be around 80 days, according to Musk’s presentation, and the longest would be around 150 days. (Musk stated that he thinks the architecture could be improved to reduce the trip to 60 or even 30 days.)

After landing on Mars and discharging its passengers, the ship would be refueled with methane/oxygen bipropellant made on the surface of Mars from Martian water and carbon dioxide, and then flown back to Earth orbit.

The SpaceX plan as Musk described it contains nine notable features. If we examine each of these in turn, some of the strengths and weaknesses in the overall system will begin to present themselves.

1. Extremely large size. The proposed SpaceX launch system is four times bigger than a Saturn V rocket. This is a serious problem, because even with the company’s impressively low development costs, SpaceX has no prospect of being able to afford the very large investment — at least $10 billion — required to develop a launch vehicle of this scale.

2. Use of methane/oxygen bipropellant for takeoff from Earth, trans-Mars injection, and direct return to Earth from the Martian surface. These ideas go together, and are very strong. Methane/oxygen is, after hydrogen/oxygen, the highest-performing practical propellant combination, and it is much more compact and storable than hydrogen/oxygen. It is very cheap, and is the easiest propellant to make on Mars. For over a quarter century, I have been a strong advocate of this design approach, making it a central feature of the Mars Direct mission architecture I first laid out in 1990 and described in my book The Case for Mars. However, it should be noted that while the manufacture of methane/oxygen from Martian carbon dioxide and water is certainly feasible, it is not without cost in effort, power, and capital facilities, and so the transportation system should be designed to keep this burden on the Mars base within manageable bounds.

3. The large scale manufacture of methane/oxygen bipropellant on the Martian surface from indigenous materials. Here I offer the same praise and the same note of caution as above. The use of in situ (that is, on-site) Martian resources makes the entire SpaceX plan possible, just as it is a central feature of my Mars Direct plan. But the scale of the entire mission architecture must be balanced with the production capacity that can realistically be established.

4. All flight systems are completely reusable. This is an important goal for minimizing costs, and SpaceX is already making substantial advances toward it by demonstrating the return and reuse of the first stage of its Falcon 9 launch vehicle. However, for a mission component to be considered “reusable” it doesn’t necessarily need to be returned to Earth and launched again. In general, it can make more sense to find other ways to reuse components off Earth that are already in orbit or beyond. This idea is reflected in some parts of the new SpaceX plan — such as refilling the second stage in low Earth orbit — but, as we shall see, it is ignored elsewhere, at considerable cost to program effectiveness. Furthermore the rate at which systems can be reused must also be considered.

5. Refilling methane/oxygen propellant in the booster second stage in Earth orbit. Here Musk and his colleagues face a technical challenge, since transferring cryogenic fluids in zero gravity has never been done. The problem is that in zero gravity two-phase mixtures float around with gas and liquid mixed and scattered among each other, making it difficult to operate pumps, while the ultra-cold nature of cryogenic fluids precludes the use of flexible bladders to effect the fluid transfer. However, I believe this is a solvable problem — and one well worth solving, both for the benefits it offers this mission architecture and for different designs we may see in the future.

6. Use of the second stage to fly all the way to the Martian surface and back. This is a very bad idea. For one thing, it entails sending a 7-million-pound-force thrust engine, which would weigh about 60 tons, and its large and massive accompanying tankage all the way from low Earth orbit to the surface of Mars, and then sending them back, at great cost to mission payload and at great burden to Mars base-propellant production facilities. Furthermore, it means that this very large and expensive piece of capital equipment can be used only once every four years (since the feasible windows for trips to and from Mars occur about every two years).

7. The sending of a large habitat on a roundtrip from Earth to Mars and back. This, too, is a very bad idea, because the habitat will get to be used only one way, once every four years. If we are building a Mars base or colonizing Mars, any large habitat sent to the planet’s surface should stay there so the colonists can use it for living quarters. Going to great expense to send a habitat to Mars only to return it to Earth empty makes no sense. Mars needs houses.

8. Quick trips to Mars. If we accept the optimistic estimates that Musk offered during his presentation, the SpaceX system would be capable of 115-day (average) one-way trips from Earth to Mars, a somewhat faster journey than other proposed mission architectures. But the speedier trips impose a great cost on payload capability. And they raise the price tag, thereby undermining the architecture’s professed purpose — colonizing Mars — since the primary requirement for colonization is to reduce cost sufficiently to make emigration affordable. Let’s do some back-of-the-envelope calculations. Following the example of colonial America, let’s pick as the affordability criterion the property liquidation of a middle-class household, or seven years’ pay for a working man (say about $300,000 in today’s equivalent terms), a criterion with which Musk roughly concurs. Most middle-class householders would prefer to get to Mars in six months at the cost equivalent to one house instead of getting to Mars in four months at a cost equivalent to three houses. For immigrants, who will spend the rest of their lives on Mars, or even explorers who would spend 2.5 years on a round trip, the advantage of reaching Mars one-way in four months instead of six months is negligible — and if shaving off two months would require a reduction in payload, meaning fewer provisions could be brought along, then the faster trip would be downright undesirable. Furthermore, the six-month transit is actually safer, because it is also the trajectory that loops back to Earth exactly two years after departure, so the Earth will be there to meet it. And trajectories involving faster flights to Mars will necessarily loop further out into space if the landing on Mars is aborted, and thus take longer than two years to get back to Earth’s orbit, making the free-return backup abort trajectory impossible. The claim that the SpaceX plan would be capable of 60-day (let alone 30-day) one-way transits to Mars is not credible.

9. The use of supersonic retropropulsion to achieve landing on Mars. This is a breakthrough concept for landing large payloads, one that SpaceX has demonstrated successfully in landing the first stages of its Falcon 9 on Earth. Its feasibility for Mars has thus been demonstrated in principle. It should be noted, however, that SpaceX is now proposing to scale up the landing propulsion system by about a factor of 50 — and employing such a landing techniques adds to the propulsive requirement of the mission, making the (unnecessary) goal of quick trips even harder to achieve.

Taking the above points into consideration, some corrections for the flaws in the current ITS plan immediately suggest themselves:

Changing the plan in the ways described above would greatly improve the performance of the ITS. This is because the ITS in its original form is not designed to achieve the mission of inexpensively sending colonists and payloads to Mars. Rather, it is designed to achieve the science-fiction vision of the giant interplanetary spaceship. This is a fundamental mistake, although the temptation is understandable. (A similar visionary impulse influenced the design of NASA’s space shuttle, with significant disadvantage to its performance as an Earth-to-orbit payload delivery system.) The central requirement of human Mars missions is not to create or operate giant spaceships. Rather, it is to send payloads from Earth to Mars capable of supporting groups of people, and then to send back such payloads as are necessary.

To put it another way: The visionary goal might be to create spaceships, but the rational goal is to send payloads.

To get a sense of some of the benefits that would come from making the changes I outlined above, let’s make some estimates. In the table below, I compare six versions of the ITS plan, half based on the visionary form that Elon Musk sketched out (called the “Original” or “O” design in the table) and half incorporating the alterations I have suggested (the “Revised” or “R” designs).

Our starting assumptions: The ship begins the mission in a circular low Earth orbit with an altitude of 350 kilometers and an associated orbital velocity of 7.7 kilometers per second (km/s). Escape velocity for such a ship would be 10.9 km/s, so applying a velocity change (DV) of 3 km/s would still keep it in a highly elliptical orbit bound to the Earth. Adding another 1.2 km/s would give its payload a perigee velocity of 12.1 km/s, sufficient to send it on a six-month trajectory to Mars, with a two-year free-return option to Earth. (In calculating trip times to Mars, we assume average mission opportunities. In practice some would reach Mars sooner, some later, depending on the launch year, but all would maintain the two-year free return.) We assume a further 1.3 km/s to be required for midcourse corrections and landing using supersonic retropropulsion. For direct return to Earth from the Martian surface, we assume a total velocity change of 6.6 km/s to be required. In all cases, an exhaust velocity of 3.74 km/s (that is, a specific impulse of 382 s) for the methane/oxygen propulsion, and a mass of 2 tons of habitat mass per passenger are assumed. A maximum booster second-stage tank capacity of 1,950 tons is assumed, in accordance with the design data in Musk’s presentation.

Concept A is the original ITS concept as presented by Musk, with a 130-day transit from Earth to Mars. The plan is technically feasible, but it has the downsides discussed above, including the glaring problem marked in red: no payload is delivered along with the people, leaving the colonists at Mars with no supplies or equipment or housing.

Concept B gives Musk’s original plan only a slight twist: the trip to Mars is longer — by fifty days — which means a lower DV is required for the journey, which in turn means (as marked in blue) that 210 tons of cargo can be delivered along with the colonists, for 2.1 tons of payload per colonist.

Concept C incorporates another of my suggested improvements from above, leaving the second stage of the launch vehicle near Earth. In such an arrangement, the second stage needs to do only 3 km/s DV, with the remaining 2.5 km/s DV needed to reach Mars done by the (now separate) spaceship’s own much smaller propulsion system. Concept C then leaves the 200-ton habitat behind on Mars, along with a further 190 tons of cargo, for a total of 4.1 tons per colonist, double that of Concept B.

Concept C has another even greater advantage over Concepts A and B: it requires only 465 tons of propellant to go back from Mars to Earth, less than a third of that needed by Concepts A or B. Furthermore, because of its rapid reuse of the launch vehicle’s second stage, the in-space propulsion system required to support a rate of five missions per opportunity in Concept C is also less than a third of that in Concepts A or B. If we combine these advantages, we see as a bottom line (as marked in green) that during each launch window, Concept C would allow for the delivery of about six times the payload to Mars as Concept B per each unit of transport system mass or per each unit of propellant produced on Mars.

However, Concepts A, B, and C all embrace an optimistic aspect of Musk’s proposal: the estimate of propulsion systems with dry-mass fractions of 0.08. The “dry-mass fraction” is the mass of a rocket or stage “wet” (that is, filled with fuel) divided by its mass “dry” (that is, empty). A dry-mass fraction of 0.08 means that the mass of the empty rocket would be 8 percent the mass of the filled rocket. For the remaining concepts, we will assume a more conservative dry-mass fraction of 0.12.

So Concept D repeats Musk’s plan (the slower version described in Concept B), but assumes a higher dry-mass fraction. And Concept E repeats my revised version (the slower and staged Concept C), but assumes a higher dry-mass fraction. Using these more conservative assumptions, the Revised version performs an order of magnitude better than the Original in all the relevant figures of merit. The advantages of employing the Revised design with the six-month trip to Mars are thus decisive.

But the relevant issue is not how these ideas might be implemented in a future Mars colonization program, but how we might put them to use in the sort of nearer-term Mars exploration and base-building program to be conducted by our own generation. Such a possibility is illustrated in Concept F. Like Concept E, Concept F adopts the revision suggestions I described above, and assumes the more conservative dry-mass fraction. However, in Concept F, the design is scaled down by an order of magnitude, so that instead of requiring a launch vehicle that can put about 500 tons into low Earth orbit, a launch vehicle able to put 50 tons into low Earth orbit will suffice. This is a critical distinction because, in contrast to 500-ton-to-orbit launchers — which at this point are the stuff of science fiction — at least three different launchers with capabilities of 50-tons-to-orbit or more may soon be available, including SpaceX’s own Falcon Heavy (54 tons to orbit, scheduled for first flight in 2017), as well as NASA’s Space Launch System (75 tons to orbit, first flight in 2018), and the Blue Origin New Glenn (about 65 tons to orbit, first flight by 2020). The improvements and revisions I’ve described make it possible to accomplish a Mars exploration mission using a 50-ton-to-orbit launch vehicle. Indeed, the mission presented in Concept F is comparable in crew size and capability to the Mars Direct or Mars Semi-Direct mission plans that I’ve described elsewhere, but with the advantage of using a 50-ton-to-orbit launcher instead of the 120-ton-to-orbit launcher employed by those concepts. This is a very exciting prospect.

Consider what this revised version of the ITS plan would look like in practice, if it were used not for settling Mars but for the nearer-at-hand task of exploring Mars. If a SpaceX Falcon Heavy launch vehicle were used to send payloads directly from Earth, it could land only about 12 tons on Mars. (This is roughly what SpaceX is planning on doing in an unmanned “Red Dragon” mission “as soon as 2018.”) While it is possible to design a minimal manned Mars expedition around such a limited payload capability, such mission plans are suboptimal. But if instead, following the ITS concept, the upper stage of the Falcon Heavy booster were refueled in low Earth orbit, it could be used to land as much as 40 tons on Mars, which would suffice for an excellent human exploration mission. Thus, if booster second stages can be refilled in orbit, the size of the launch vehicle required for a small Mars exploration mission could be reduced by about a factor of three.

In all of the ITS variants discussed here, the entire flight hardware set would be fully reusable, enabling low-cost support of a permanent and growing Mars base. However, complete reusability is not a requirement for the initial exploration missions to Mars; it could be phased in as technological abilities improved. Furthermore, while the Falcon Heavy as currently designed uses kerosene/oxygen propulsion in all stages, not methane/oxygen, in the revised ITS plan laid out above only the propulsion system in the trans-Mars ship needs to be methane/oxygen, while both stages of the booster can use any sort of propellant. This makes the problem of refilling the second stage on orbit much simpler, because kerosene is not cryogenic, and thus can be transferred in zero gravity using flexible bladders, while liquid oxygen is paramagnetic, and so can be settled on the pump’s side of the tank using magnets.

Using such a system, a manned expedition of Mars could be carried out any number of ways. For example, it could be done in a manner similar to the Mars Direct mission plan, with the first trans-Mars payload delivering an unfueled Earth Return Vehicle with an onboard propellant factory to make methane/oxygen propellant on Mars, and the second delivering a habitat module with a crew of astronauts aboard who land near the ERV, using their hab as their house on Mars. After 1.5 years of exploration they would return in the ERV, leaving their hab behind on Mars to add incrementally to the facilities of a growing Mars base as the missions proceed.

Or a different plan, closer in spirit to the SpaceX ITS, could be adopted, in which a single payload combining the hab and the ERV is sent, with the hab above and the ERV below. The ERV would use a limited amount of methane/oxygen propellant to perform supersonic retropropulsion of the combined payload upon Mars entry, bringing the assembly to subsonic speeds. Once this is done, the hab would pop a parachute, or possibly a parasail, to lift it off the ERV and then land nearby using a very small terminal landing propulsion system. The first such mission could send such an assembly out with no crew, allowing the ERV to be fueled in advance of the first piloted launch, which would then arrive two years later provided with a redundant hab and plentiful extra supplies. Once the base is well-established, the hab and ERV modules could be landed together, with the hab subsequently lifted off the ERV by a crane.

The number of such potential variations is endless. Another: In initial missions, the Falcon Heavy second stage could perform the full burn, allowing it to coast out to Mars in company with the piloted spacecraft, which could then use it as a counterweight on the opposite end of a tether to provide the crew with artificial gravity on their way to Mars (just as in the standard Mars Direct plan). This would entail expending the second stage, but it could be worth it for the first missions to have their crews in top physical strength, as they will reach a Mars with minimal support facilities. In later missions, the Falcon Heavy second stage could be left behind just short of Earth escape for ready reuse (as in the revised ITS plan I described above), and the crew be allowed to fly to Mars in zero gravity, since they would by that point have plenty of ample base facilities to provide local support for recovery from zero-gravity weakening once they reach the Red Planet.

Toward the end of his presentation, Musk briefly suggested that one way to fund the development of the ITS might be to use it as a system for rapid, long-distance, point-to-point travel on Earth. This is actually a very exciting possibility, although I would add the qualifier that such a system would not be the ITS as described, but a scaled-down related system, one adapted to the terrestrial travel application.

The point is worthy of emphasis. For three thousand years or more, people have derived income from the sea, for example by fishing — but far more by using the sea as a favorable comparatively low-drag medium for transport. Similarly, while there is money to be made by human activities in space, there is potentially much more to be made by human travel across space, taking advantage of the drag-free quality of space for rapid travel. It has long been known that a rocketplane taking off with a high suborbital velocity could travel halfway around the Earth (that is, reaching anywhere else on the planet) in less than an hour. The potential market for such a capability is enormous. Yet it has remained untouched. Why?

The reason is simply this: Up till now, such vehicles have been impractical. For a rocketplane to travel halfway around the world would require a DV of about 7 km/s (6 km/s in physical velocity, and 1 km/s in liftoff gravity and drag losses). Assuming methane/oxygen propellant with an exhaust velocity of 3.4 km/s (it would be lower for a rocketplane than for a space vehicle, because exhaust velocity is reduced by surrounding air), such a vehicle, if designed as a single stage, would need to have a mass ratio of about 8, which means that only 12 percent of its takeoff mass could be solid material, accounting for all structures, while the rest would be propellant. On the other hand, if the rocketplane were boosted toward space by a reusable first stage that accomplished the first 3 km/s of the required DV, the flight vehicle would only need a mass ratio of about 3, allowing 34 percent of it to be structure. This reduction of the propellant-to-structure ratio from 7:1 down to 2:1 is the difference between a feasible system and an infeasible one.

In short, what Musk has done by making reusable first stages a reality is to make rocketplanes possible. But there is no need to wait for 500-ton-to-orbit transports. In fact, his Falcon 9 reusable first stage, which is already in operation, could enable globe-spanning rocketplanes with capacities comparable to the DC-3, while the planned Falcon Heavy (or New Glenn) launch vehicles could make possible rocketplanes with the capacity of a Boeing 737.

Such flight systems could change the world.

In his talk introducing the ITS, Musk suggested that a Mars colonization program using thousands of such systems could be used to rapidly transport a million people from Earth to Mars. This would be done to provide a large enough population to allow the colony to be fully self-sufficient. In subsequent interviews, he also said that none of these colonists would include children, since having kids around would be a burden upon the colony.

My own ideas on how the colonization of Mars could be achieved are different. Rather than a massive convoy effort to populate the planet, I see the growth of a Mars colony as an evolutionary development, beginning with exploration missions, followed by a base-building phase. As the series of missions proceeds, additional elements of the flight-hardware set would become reusable, causing transport costs to drop. Furthermore, as the base grows, its capability to produce more and more necessary items, including water, food, ceramics, glasses, plastics, fabrics, metals, wires, tools, domes, and structures, would expand — progressively reducing the amount of materials that needs to be transported across space to support each settler. This will provide the material basis for an expanding Martian population, which will grow exponentially as families are formed and children are born.

That said, Mars is unlikely to become autarchic for a very long time, and even if it could, it would not be advantageous for it to do so. Just as nations on Earth need to trade with each other to prosper, so the planetary civilizations of the future will also need to engage in trade. In short, regardless of how self-reliant they may become, the Martians will always need, and certainly always want, cash. Where will they get it?

A variety of ideas have been advanced for potential cash exports from Mars. For example, Mars might serve as a source of food and other useful goods for asteroid-mining outposts which themselves export precious metals to Earth. Or, since the water on Mars has six times the deuterium concentration as Earth’s, that potentially very valuable fusion-power fuel could be exported to the home planet once fusion power becomes a reality. Or maybe precious metals will be found on Mars, which, with a fully reusable interplanetary transportation system, it might be profitable to mine and export to Earth.

While such possibilities exist, in my view the most likely export that Mars will be able to send to Earth will be patents. The Mars colonists will be a group of technologically adept people in a frontier environment where they will be free to innovate — indeed, forced to innovate — to meet their needs, making the Mars colony a pressure cooker for invention. For example, the Martians will need to grow all their food in greenhouses, strongly accentuating the need to maximize the output of every square meter of crop-growing area. They thus will have a powerful incentive to engage in genetic engineering to produce ultra-productive crops, and will have little patience for those who would restrict such inventive activity with fear-mongering or red tape.

Similarly, there will be nothing in shorter supply in a Mars colony than human labor time, and so just as the labor shortage in nineteenth-century America led Yankee ingenuity to a series of labor-saving inventions, the labor shortage on Mars will serve as an imperative driving Martian ingenuity in such areas as robotics and artificial intelligence. Such inventions, created to meet the needs of the Martians, will prove invaluable on Earth, and the relevant patents, licensed on Earth, could produce an unending stream of income for the Red Planet. Indeed, if the settlement of Mars is to be contemplated as a private venture, the creation of such an inventor’s colony — a Martian Menlo Park — could conceivably provide the basis for a fundable business plan.

To those who ask what are the natural resources on Mars that might make it attractive for settlement, I answer that there are none, but that is because there is no such thing as a “natural resource” anywhere. There are only natural raw materials. Land on Earth was not a resource until human beings invented agriculture, and the extent and value of that resource has been multiplied many times as agricultural technology has advanced. Oil was not a resource until we invented oil drilling and refining, and technologies that could use the product. Uranium and thorium were not resources until we invented nuclear fission. Deuterium is not a resource yet, but will become an enormous one once we develop fusion power, an invention which future Martians, having limited alternatives, may well be the ones to bring about. Mars has no resources today, but will have unlimited resources once there are people there to create them.

Martian civilization will become rich because its people will be smart. It will benefit the Earth not only as a fountain of invention, but as an example of what human beings can do when they rise above their animal instincts and invoke their creative powers. It will show to all that infinite possibilities exist — not to be taken from others, but to be made.

No one will be able to look upon it without feeling prouder to be human.