Please login or sign up for a new account.

I forgot my password

Password Reset

“Man Caves: Humanity’s Next Home” by Ken Roy

Image: Shoham Charikar

We’re learning that planets appear to be plentiful, with most stars having a family of such bodies. As we venture into interstellar space it would be nice to think that we would find countless Earthlike planets having good clean breathable air, plenty of liquid water, a little ice, and be free of allergens, toxins, and bacteria that would harm us. A moderate temperature, a reasonable gravity, low background radiation, kindly neighbors, etc., all would be nice to have as well. With such a world, the colony lands and sets up shop, and lives happily ever after.

We shall consider such a world a habitable planet. Stephen Dole of the Rand Corporation estimated that about 1 in 200 stars has such a world. A more recent analysis by William Pollard of the ORAU Institute for Energy Analysis estimated that this number actually ranges somewhere between 1 in 100,000 to 1 in 10,000,000 stars. With Dole’s estimate, our space-faring descendants, on average, would have to search a spherical volume of space having a radius of 25 light-years to find a single habitable planet. With Pollard’s estimate, that radius would range between 200 and 1000 light-years. Habitable planets are probably not as common as we would like and by definition already possess life. This life may not take kindly to alien invaders (us). Indeed, its ecology could lack some essential amino acids that we require and could even be toxic to humans or even Earth life in general. There is much to be learned from alien life and certainly no reason not to study it, but there are real, practical, and even ethical issues involved in trying to colonize such a planet.

Martyn J. Fogg, in the book Islands in the Sky, (edited by Stanley Schmidt and Robert Zubrin) defines two other types of planets that our space-faring descendants would seek out. The first is termed a “Biocompatible Planet.” Fogg defines such a planet as possessing the necessary physical parameters for life to flourish on its surface. It would be initially lifeless but with the introduction of Earth life could host a biosphere of considerable complexity without the need for major planetary engineering. Planetary engineering could involve mega-structures, giant space mirrors, and probably the importation of vast amounts of volatiles such as nitrogen, hydrogen, and water.

The second he terms an easily terraformable planet. Such a planet does require planetary engineering to be rendered biocompatible, and then eventually habitable as Earth life is introduced and a vibrant Earth-based biosphere allowed to develop. Terraforming is a term that applies to both types of planets, one just requires more work than the other.

Habitable planets, biocompatible planets, and easily terraformable planets all must lie in (or at least very close to) a star’s habitable zone. This is a fairly narrow region around a star where the star’s radiation maintains the planet’s temperature above freezing but below the boiling point of liquid water. In our solar system, Venus and Mars are at the edge of our star’s habitable zone and the Earth sits comfortably in the middle. Mercury, the moons of the gas giants, and the dwarf planets need not apply.

According to Dole, stars smaller than 0.72 solar masses won’t have a habitable zone due to the fact that by the time a planet is close enough to the star to be warm enough, they are too close to avoid tidal retardation of the planet’s rotation. Their rotation will slow and eventually the planet will become tidally locked: one face always facing their star. He assumed that a tidally locked planet would be by definition uninhabitable. However, that may not always be the case, as some recent papers suggest that such planets might actually have habitable regions but habitable regions very different from that to which we humans are accustomed.

In any event, sterile planets are probably common, but “traditional” terraforming concepts require that they exist within a star’s habitable zone and that the star possess a light spectrum similar to that of our sun in order for plant life to survive and for the human eye to function properly. But there is one terraforming approach that bypasses these limitations, thus greatly increasing the pool of planets and even star systems available as potential new homes for humanity and any Earth life we choose to bring with us.

The Shell World Approach

The basic idea of a shell world is to build a material shell that totally encloses a planet and its atmosphere. This shell will contain the atmosphere, provide some radiation protection, and serve as a way to regulate heat and light. If we can regulate the light and heat under the shell, then the limitations imposed by the star’s habitable zone vanish. Mercury, Mars, the Earth’s moon, many of the larger moons of our gas giants and even dwarf planets, such as Pluto, become available as potential future homes for humanity. It is reasonable to assume that such planets, moons, and dwarf planets are probably common around other stars, certainly far more common than habitable planets or even easily terraformable planets.

This shell is in essence a hollow sphere with a planet or moon in the middle. We know how to calculate the tension in a spherical shell resulting from internal pressure—and with a shell the size of a planet the resulting stress is huge, big enough to rupture even the strongest material that we know of. But there is also a second force in play: gravity. A mass located at the center of a large hollow sphere pulls on the sphere and creates compression stresses in the shell. For a shell around a planet, or large moon, the resulting compression stress due to gravity is huge. Thus, we have two forces acting on the shell, both huge. One is positive and the other is negative. If we’re clever, we can make them cancel out.

Let’s consider Mars. If we constructed a metal shell 10 kilometers above the planet’s surface so as to totally enclose the planet, we would need this metal to be 70 times stronger than steel, assuming no atmosphere, just to withstand the gravitational pull of the planet. But Mars does have a small atmosphere. Suppose we brought it up to Earth-normal pressure (14.7 psi) by adding a lot of oxygen and nitrogen and raised the average atmospheric temperature to 60°F. We also assume that there is no atmosphere above the shell. Now, if we make the shell out of steel that is 2.2 meters thick, then the actual stress within the shell is now 0.0 psi. The tensile forces in the shell induced by atmospheric pressure on the shell counters the compressive forces induced by the planet’s gravity pulling on the shell.

Think about that for a minute. Under the shell Mars now has an Earth-normal atmosphere. The atmosphere is completely contained and will never leak away as long as we can keep the shell airtight. Sure, some leakage through the shell and its airlocks and maybe even outgassing from shell materials is bound to happen but as long as no significant atmosphere is allowed to accumulate on the outside of the shell the forces within the shell will remain balanced. If enough gases do accumulate on the outside of the shell to begin to upset the delicate balance between compression and tension then either more gases will have to be released under the shell or the gases outside the shell will have to be captured and either removed or re-injected under the shell. Another option is to begin to remove some of the mass on the shell, maybe dropping it on the planetary surface. Like a living thing, a shell and its world will need to be constantly managed and defended against asteroids and other space debris. But, as long as the balance can be maintained, the shell itself is under little stress and structurally could survive for millennia.

Once the shell is up and the atmosphere adjusted to Earth-normal composition and pressure you can now comfortably walk the surface of Mars without a spacesuit. Yes, it would be dark, but we will fix that problem later.

The smaller the central world, the thicker and heavier the shell needs to be to allow gravity to induce enough compressive force to counter the tensile force caused by an Earth-normal atmospheric pressure. Solutions for several bodies are shown in Table 1.

Because the shell is under little stress it is possible to build it out of almost anything. The examples shown in Table 1 assume that the shell is entirely steel. However, the actual stress-bearing portion can be relatively small with the rest of the actual shell mass being dead weight, needed to provide the central body with something to tug on and inducing compressive forces within the stress-bearing portion. In the case of Mars, a steel shell one meter thick with about 6.5 meters of regolith (dirt) on top of it could have the same small stress in the thinner shell as a solid steel shell.

Infinite combinations of metal, ice, dirt, and rocks are possible, but the shell must be airtight and the mass must be evenly distributed across the shell. It is possible that some of this non-structural mass can be in the form of vacuum-loving industries, solar collectors, heat exchangers, shuttle landing strips, communication systems, power plants, etc.

Larry Niven’s Ringworld is unstable. A ring around a star, if displaced, will continue to have the nearer edge pulled into the star, leading to a sad, but exciting, end to any ring inhabitants. Niven handled this by having his Ringworld engineers install control rockets. But a spherical shell enclosing a moon or planet is stable as long as atmospheric pressure varies with altitude more than the gravitational field. If a portion of the shell approaches the central world, atmospheric pressure will push it back into place. At the same time, the portion of the sphere moving away from the surface will be pulled back into place by the lower atmospheric pressure that it experiences. Note that placing a solid spherical shell around a star (this is known as a Dyson Sphere) is probably not structurally sound without using materials far stronger than any we know of today.

If we assume Earth-normal pressures and temperature on the central world’s surface, then the central world needs to have a mass of at least two times that of the asteroid Ceres in order to achieve the required pressure gradient to assure shell stability. This is not to say that a shell cannot be constructed around Ceres, but only that such a shell must be stabilized with cables and/or columns or even control rockets.

Shell World Design Considerations

The mass and size of the planet, or moon, will determine the gravity and surface area of the new world. The asteroid Ceres with a mass of approximately 0.00016 that of Earth and having a radius 0.07 that of Earth represents, perhaps, the smallest world that could be “terraformed” with a shell. The shell would be very thick and it would have to be stabilized with large columns and/or heavy cables. The builders would have to provide 363 metric tonnes of mass for every square meter of shell area. This doesn’t have to be all steel. In this case it could be 2 meters of steel, 354 meters of ice (Ceres has a lot of water ice) and 20 meters of regolith. This is a lot of mass but also a lot of radiation shielding. A Ceres colony with such a shell could survive a very intense gamma ray burst, one intense enough to destroy most life on Earth, the type of which may account for several mass extinctions in Earth’s history.

Because Ceres has a rocky core covered by an icy mantle many kilometers thick, the atmospheric temperature under the shell would probably be below freezing. Another possibility is to melt the ice (yes, we’re talking about a lot of energy) and create a water world with the ocean hundreds of kilometers deep. Any cities would be floating structures but with an Earth-normal atmosphere, and human powered flight would be possible.

The height of the shell above the surface of the world is a design choice. If the shell is intended to rotate relative to the planet’s surface, then it needs to be high enough that any mountain would be unlikely to rip a gash in the shell. If the shell is stationary with respect to the planetary surface, then mountains could actually project through the shell. This presents some engineering challenges as the shell/mountain interface would have to be effectively sealed. For example, Olympus Mons on Mars is a shield volcano with a height of 22 km above the surface. If the shell is only 10 km above the surface the large volcano would extend through the shell and would make a good space port, as it would have no mass loading limitations.

But beyond that, the limit is only how much atmosphere you have available to import. It should be noted that atmospheres with pressures other than 14.7 psi (Earth-normal) and compositions other than 21 percent oxygen, 78 percent nitrogen, and 1 percent argon are certainly possible, but all atmospheres discussed here are limited to Earth-normal compositions. Planetary engineering encompasses the importation of atmospheric gases to make up the atmosphere. With a shell world, the amount of gas needing to be imported can be significantly reduced. Using Mars as an example, oxygen could probably be manufactured using local sources, but nitrogen would probably have to be imported from the outer planets, or maybe Venus. A 10-km high shell requires only 36 percent of atmospheric gases needed for an uncontained atmosphere.

Oceans are nice. Once you have an Earth-normal atmosphere and temperature, the addition of water will result in lakes and maybe oceans fairly quickly, even on a low gravity world. Rains will wash salts from the soil into the new oceans, making the soil more fertile. Oceans can be useful for regulating heat and are vital to numerous ecological cycles. And if we’re thinking in terms of providing a new home to all of Earth life, then oceans become necessary. The size and extent of the ocean(s) is a design choice. To make the required amount of water, hydrogen may have to be imported, assuming that oxygen can be produced locally. And make no mistake, we’re talking about lots of hydrogen.

Climate can be a design choice. The entire world could be temperate, or it can be configured to have frozen poles and a tropical equator with resulting weather patterns. It could be a hot desert world, or a cold frozen world, or as we have discussed already, an ocean world.

There are no reasons structures can’t be hung from the underside of the shell, provided that the deadweight above the shell at that point is reduced to compensate. The entire underside of the shell might be utilized as living space, effectively doubling the living area of our Shell World. Heavy and dirty industry could be located on the outside of the shell.

Heat and Lighting

Because of the shell, everything beneath it would be in total darkness without some provision for artificial lighting. Structural penetrations in the shell need to be kept to a minimum and an alien star might not provide the correct spectrum of light required for Earth life. But artificial lighting offers many choices and could effectively duplicate our sun’s light spectrum along with the Earth’s 24-hour, day-night cycle regardless of the rotation rate of the planet/moon. Even a tidally locked planet/moon could have a 24-hour day-night cycle under the shell.

Humans need light to see, but plants need light to live. Earth plants use less than 250 W/m2 out of the 1400 W/m2 available from the Sun, and that mainly in the blue-violet and orange-red ends of the spectrum. This much is required over agricultural and forest areas. How much more than this would be provided for the humans? The choice of colors and intensities is almost infinite. Ultraviolet light could be provided over beaches to allow tanning and sunburns. Infrared could be provided to control temperatures within the shell. Some areas could be left in eternal night and others would be forever bright. The energy bill would be rather large and of course all of that energy would need to be radiated away as waste heat after use.

There are two types of heating problems. If the world is too close to a star, such as Mercury, the problem becomes one of minimizing heat input into the world and providing means for adequate dumping of waste heat. If the world is far out in the Oort Cloud the problem is one of retaining heat. These are engineering problems. The solutions will no doubt require large quantities of energy and power, but a civilization that can build planetary shells probably has the ability to solve those problems.

Radiation Protection

Radiation exposure is an issue that any proposed space settlement must address. On Earth, the average human absorbs 30 millirem (mrem) annually from space radiation. (A rem is a unit of damage done by radiation to the human body.) By comparison, the total average annual dose for each human from all sources is about 360 mrem. Most of this is from terrestrial sources such as radon. Most charged particles from the Sun and some cosmic radiation are deflected by Earth’s magnetic field. Some do get through and hit Earth but most of those are attenuated by our atmosphere. Thus, the low 30 mrem dose from space radiation for Earth dwellers. Note that someone standing on the lunar surface has no magnetic field or thick atmosphere for protection, and with only a few centimeters of aluminum as protection would receive a dose of radiation far higher than their twin on Earth.

One advantage of a shell world is that the shell itself can serve as a radiation shield. Then there is the atmosphere below the shell to attenuate any radiation that penetrates the shell. Small celestial bodies generally have a slight to nonexistent magnetic field, making them prone to high levels of space-based radiation at the surface. But if one were to be transformed into a shell world, the shell could be used to reduce the radiation levels to some degree—perhaps below Earth-normal levels, and this on a planet with no magnetic field to protect it.

Construction of a Shell World: The Earth’s Moon

Construction of a shell world requires energy generation and material fabrication and transport on a vast scale. It could be done only by an advanced civilization capable of producing and using large quantities of energy and of safely moving large quantities of material around a solar system.

Let’s consider Earth’s moon. We could equally well choose Mars or Pluto, but let’s assume we’ve decided on Earth’s moon and somehow obtained the “rights” to it. The designers decide on a shell only five kilometers above the surface. They want an ocean that covers a quarter of the surface and has an average depth of 100 meters. Future locations for cities and forests are laid out. The initial sculpting is done with kinetic energy: rocks from space. The ocean basins are carved out, crater rims erased, and mountains and hills created. Final sculpting is done by automated dozers.

Now we have a lot of ice delivered and stored in the future ocean’s basins. This ice can come from the moons of the gas giants such as Jupiter, or perhaps from the Oort Cloud where countless icy comets have waited billions of years. We also need oxygen. This can probably be obtained by processing lunar materials and be stored in large depots. There is limited nitrogen on the Moon so it will probably have to be imported from Jupiter or Titan or Venus and also stored in large depots. Argon is entirely optional; plant and animals don’t seem to need it, but we can add a little to make our breathing air truly Earthlike.

Fabrication plants now begin producing a carbon-based fabric, perhaps using nanotubes or graphene sheets. We know how to make Kevlar® and, in the future, should be able to make large sheets of graphene. Both are incredibly strong in tension. This fabric must be strong, airtight, durable, and corrosion resistant. We cover the entire surface of the Moon with this fabric and seal up any seams. It is 25cm thick and designed to allow a stretch of one part in a thousand. On top of this we drop steel plates, each 25cm thick. These plates are connected in such a way as to allow the same stretch as the fabric. They are not intended to be structural but they will protect the fabric below and distribute any surface loads over a large area. On top of this we now dump regolith to reach a total loading of 52 metric tonnes per square meter. Assuming that regolith has a density of dry dirt, this requires about 36 meters of rock and dust. Industrial facilities intended for vacuum operation are sometimes substituted for regolith.

Now we slowly release the oxygen, nitrogen and argon under the shell. The shell rises off the surface of the Moon. The pressure under the shell stays at 14.7 psi. As we release more gas the shell rises. Eventually the shell is 5 kilometers above the lunar surface and the fabric structure goes into a slight tension. Now additional gases will result in increased pressure.

Next we start heating the oceans, converting them into liquid. This initiates a hydraulic cycle and rains begin to fall on the land, washing salts into the new oceans. Lighting and communications systems are installed on the underside of the shell, all run by power plants located on the outside of the shell.

Now comes the hard part. We introduce life from Earth to its new home and carefully nurture what once was a sterile world into a vibrant living ecology. It will have been a long and expensive undertaking, but at the end we will have a new home for humanity and all the other forms of life that shared Earth with us—except perhaps for certain pathogens and parasites. We’ve just increased the available habitable land area available to humans by about 20 percent. If we consider the underside of the shell as habitable area as well, then we’ve doubled this to 40 percent. True, we can’t walk on the underside of the shell but we can build structures that attach to the underside or even hang from it. Think of pictures from the Graf Zeppelin; and we have the entire underside of the shell to work with, not just a few gas bags. Lighting, communications, sensors, and other utilities would all have to be integrated into this suspended structure but it could become the preferred location for humans with the lunar surface left to grow wild with Earth life. And in the light gravity and Earth-normal atmosphere human powered flight is possible. And the view . . .

No doubt, there will be mistakes made and lessons learned from both the shell construction and the following terraforming work that will be available to future terraforming efforts.


Shell worlds could be constructed at any star that has something orbiting it with a mass greater than two Ceres. The type of star doesn’t matter. The location of the star’s habitable zone doesn’t matter. The radiation environment does matter but can be dealt with. In high radiation fields we want smaller bodies. This results in thicker and more massive shells. Such shell worlds could last for many thousands of years, and with proper maintenance and gradual improvements could last far longer, perhaps for geologic time scales.

The most common type of star in our galaxy is the small, unassuming red dwarf. These are M-class stars with a mass less than 0.45 that of our sun. They are generally not considered good targets for SETI. This may be a mistake. Such stars probably have many bodies suitable for being made into shell worlds. These stars have lifespans many times longer than our sun.

Alien civilizations having the technology to construct shell worlds may consider such stars for their long-term residence or even their retirement homes. There are probably even more brown dwarfs (or failed stars) in our galaxy than the numerous red dwarfs, and many probably have something akin to planets that could be transformed into shell worlds.

The shell world approach offers us the possibility of converting virtually any solar system with some orbital debris into a habitable star system, not just for humans but for as much Earth life as we care to import. When we finally head to the stars, we won’t be limited to finding habitable worlds or terraforming planets just in the habitable zone of a star, we’ll be able to transform lifeless planets well outside the habitable zone into Earthlike worlds. Surely the elder races can’t fault us for that.


Corliss, W.R., Space Radiation, U.S. Energy Research and Development Administration, Washington, 1968, pp.47-51.

Dole, S., Habitable Planets for Man, Blaisdell Publishing, New York, 1962, pp.6-22.

Fogg, M.J., Terraforming: Engineering Planetary Environments, SAE International, Warrendale, Pennsylvania, 1995

Fogg, M.J., “A Planet Dweller’s Dreams” in Islands in the Sky, S. Schmidt and R. Zubrin (eds), Wiley Popular Science: John Wiley & Sons, Inc, New York, 1996

Goody, R. M. and Walker, James C.G., Atmospheres, Prentice-Hall, Englewood, 1972, pp.126-141.

Mader, S., Inquiry into Life, 4th edition, Wm. C. Brown Publishers, Dubuque, 1985, pp.120-121.

National Council on Radiation Protection and Measurements, Exposure of the Population in the United States and Canada from Natural Background Radiation, NCRP Report No. 94, 1987, pp.16-21.

Oberg, J. O., New Earths, Stackpole Books, Harrisburg, 1981.

Roy, K.I., Robert G. Kennedy, and David E. Fields, “Shell Worlds: An Approach to Terraforming Moons, Small Planets and Plutoids,” Journal of the British Interplanetary Society, 62, pp.32-38, 2009.

Roy, K.I., Robert G. Kennedy, and David E. Fields, “The Question of Shell Stability,” Journal of the British Interplanetary Society, 67, pp.364-368, 2014.

Zubrin, R., The Case for Mars, The Plan to Settle the Red Planet and Whey We Must, The Free Press, New York, 1996

Copyright © 2019 Ken Roy

Ken Roy is an engineer who lives and works amid the relics of the Manhattan Project in Oak Ridge. He has published technology speculation pieces in such venues as the Journal for the British Planetary Society, and the United States Naval Institute Proceedings. His current interests include terraforming and geoengineering.