I've made reference to plans for large-scale habitats before. It's time I write down the big picture as I continue the process of refining the details. The research I've been doing for this project has led to many of my posts here with information about radiation shielding, structural materials, agricultural yields and various life support systems. Perhaps that information will be more useful in context.
There are several large habitats proposed, generally by people who are both smarter and better-educated than I am. Wherever possible I prefer to use solutions proposed or developed by others, but I disagree with some of the fundamental assumptions made for structures like the O'Neill cylinder. That will necessarily result in a different outcome, thanks to several design decisions that go in another direction. #1 on that list: There are no windows. None. Don't even think about it; windows in space are incredibly stupid.
This will be a large post, so I'm continuing after the jump.
The headline results so far are as follows:
Design population: 5,000 people
Maximum population: 5,280 without major changes, up to twice that under emergency conditions
Mass: 142,750 tons shielding, 4,552 tons hull, 2,770 tons air, 350 tons occupants. As-yet unknown masses for furnishings, life support, hydroponics, other systems.
Volume: 2,262,000 m³ (79,882,000 ft³)
Area: 138,000 m² (13.8 hectares / 34 acres) under habitable gravity.
The structure would require the capture and exploitation of 160,000 to 200,000 tons of asteroidal material, or about 67,000 m³ of carbonaceous chondrites. Only about 600 tons (0.3%) needs to be carbon, but nearly 2,000 tons (1%) needs to be nitrogen. A single 50-meter diameter rock should just about do the trick, roughly the size of the Tunguska meteor. An alternative is sixteen 20-meter diameter rocks (Chelyabinsk sized) with the proper composition on average. There are anywhere from hundreds of thousands to tens of millions of near-Earth asteroids in this size range.
The two driving forces in my mind are radiation protection and moderate to full gravity. We know that microgravity is very harmful to humans even with the best medical care available and years of preparation in advance of relatively short (6-12 month) exposures; there is every reason to suspect that microgravity is not survivable in the long term. For this reason gravity or pseudogravity is a fundamental requirement. Radiation exposure is also tremendously harmful; current spacecraft are not survivable over the long term. Shielding that reduces the level of radiation exposure to Earth-normal or lower is also a fundamental requirement. The combination of these two requirements means a series of trades in structural and shielding materials. For free-space habitats I've settled on counter-rotating composite habitat modules on a common axis surrounded by non-spinning shielding made of bulk rock with a metal skin. For habitats in small bodies (Phobos, Deimos, other bodies with a few % Earth gravity at most), the habitat sections would be buried and would rely on bulk material for shielding. Bodies with significant gravity are different enough that they need to be addressed with unique designs.
I'll deal with the free-space version first. Shielding is very expensive in mass terms. There are two primary sources of radiation in free space, the solar wind (particularly solar proton events or SPE) and cosmic rays (GCR).
Cosmic rays are isotropic, so the most efficient shape would be a sphere. The solar wind is highly directional, so the most efficient shape would be a long, thin rod. Spheres are complex and inefficient to turn into living space, so a cylinder is the basic shape of choice. Remember that radiation doesn't turn corners, so we can use unconnected pieces of shielding that allows vessels in and out of the protected area without moving parts.
We can adjust the relative cost of shielding for the Sun vs. shielding for GCR by changing the aspect ratio (length to width ratio) of the cylinder, but the optimum orientation is for one end of the cylinder to point at the Sun and carry somewhat heavier shielding than the rest of the hull. This imposes a stationkeeping burden to keep the end pointed at the sun. For lower dV costs the structure should be vertical, which means the entire shielded hull will rotate into view of the Sun over the course of a year. There does not seem to be a compelling argument either way except for the slim chance that a catastrophic CME or other solar event might be made survivable by taking shelter at the far end of the rod.
Human physiology limits the speed of spin gravity, thus setting a lower limit on the diameter of the colony; my design uses a radius of 60 meters and rotation speed of 3.89 RPM for 1g Earth-normal gravity at the outer floor. Biology also sets a lower limit on the population of a self-sustaining colony; research differs on the exact value but 5,000 appears to be safe. I specify a set of four habitat modules each 50 meters long, each to house a nominal population of 1250 people. The structure can grow by adding pairs of habitat modules along the common axis, saving the expense of extra endcaps and costing only the shielding mass for the wall of the outer cylinder.
The habitats spin inside a stable shielded hull without contacting it; a physical gap of three meters separates the outer walls of the habitats from the inner walls of the hull so either structure can be maintained while under spin. Internal positioning is maintained with magnetic repulsion as necessary, to minimize the torque applied to the outer hull. No isolation is perfect, so the structure will require thrusters of some kind. Ion would be preferred, using either metals or heavy gases.
Shielding is composed of 2mm of aluminum with a standoff space (Whipple shield for micrometeoroids), then 1cm of nickel-iron (30/70) followed by 104cm of packed regolith (1.5g/cc density, composition similar to lunar soil or stony asteroids). The outer layer of habitat includes 30cm of water as additional shielding, pump-able counterweight, leak indicator and bulk storage. This combines to provide an attenuation of 4.518 (in units of 1/e^x), blocking 98.91% of radiation. The first floor will still experience slightly more radiation than Earth average. For ideal results an attenuation factor of 5.3 is desirable; longer-term projections indicate that the GCR could potentially be as high as 1500mSv (vs. 740mSv recorded so far) which would suggest a factor of 6.0 attenuation to reach Earth-normal levels. In other words, even with this much shielding there may be times where the inhabitants need to evacuate to the inner floors so the material of the outer floors can protect them from radiation spikes.
In order to make the most efficient use of this shielded volume the habitats are built in levels or floors of four meters each. Each floor is structural, capable of functioning as the outer hull during the construction phase, capable of independently supporting its own mass and floor loads, and capable of limiting the spread of damage in case of structural failure or impact. The material is UHMWPE, very long-chain polyethylene (Spectra fiber) with an aluminum film liner. This can be manufactured using plants to process CO2 and water (in turn made from any sources of carbon, hydrogen and oxygen). The habitat's hydroponics system is designed to produce plastic for this purpose in addition to food, though during the construction phase the mix of plants will heavily favor plastics.
The outermost floor is at full Earth gravity, while each floor inward provides progressively less gravity; specific floors can be designed to mimic the gravity of Mars, Venus or the Moon if desired. A microgravity bay in the center of the structure has effectively no gravity; this space can be used as a shirtsleeves environment for building or repairing delicate spacecraft among other things. Total floorspace is 149,540m²; if we define 'habitable' gravity as 0.3g or higher then 138,230m² of that space is habitable. That's eleven floors of useful gravity (including analogs of Earth, Venus and Mars), two more of low gravity (including an analog of the Moon) and one of microgravity.
A nearly identical design would be built into a pit on Phobos or Deimos. Phobos in particular could host a set of habitat modules in an open pit; if the pit is deep enough and placed to face Mars then the disk of Mars will fully block any views to space. No end-cap would be necessary for radiation protection; further, the vastly greater mass of shielding would mean the habitats would see less radiation than on Earth. One application of this might be as a permanent base at the site of a Phobos-Mars transfer tether. The habitats would be built up over time using materials excavated from the pit, with the option of adding more and more hab modules by excavating the pit deeper and deeper. This technique could be applied to main-belt asteroids like Ceres, Vesta, Pallas, etc.; the floors would need to be tilted slightly to accommodate the gravity of the parent body and a covering shield would be necessary but otherwise a very similar prospect.
Throughout my design process I have sought to use Earth-normal design standards. This phase of space exploitation is far beyond the early adventurers, meant to be constructed using by-then-proven technologies and inhabited by average people (doctors, mechanics, teachers). Average people like having creature comforts; it's good for our physical and mental health. For example, I assume that each resident requires about 100 m² of space: roughly 32 m² of personal space, 36 m² of public space and 32 m² of work space. This is about 344 ft², or the size of a smallish studio apartment. A family of four would enjoy 128 m² or about 1,378 ft², modest by American standards but ranging from generous to lavish in many other places. Smaller is certainly possible, but that's a size that most people would accept. My research suggests that floor space for growing food, clothing and furnishings will be roughly 20 m² per person; this space is categorized as work space. That leaves another 12 m² per person or 15,000 m² per habitat (160,000 ft²) or roughly one large-ish office building at 32mx32m and 15 floors. The public 36 m² accounts for hallways, engineering spaces (including life support), schools, public eating and/or meeting space and parks. Some of that could be considered 'work' space; I didn't try to stick to some particular metric, but rather used that as a basic assumption. If it turns out that the mix is more like 20 m² public and 48 m² work it makes no difference.
The internal structure of the habitat starts with the structural hull layer, 7.6mm of PE fiber with an aluminum foil liner to make it gas-tight. This layer is strong enough to handle two atmospheres of pressure; the entire four-meter floor could be flooded with water and still fall within structural limits. This is primarily formed of fibers wound around the cylinder, with secondary layers wound at 45° intervals to provide strength on the axis. These secondary layers flow into the endcap walls or sidewalls and carry the stress of the endwalls through the hull.
Next is a protective layer. The outer floor uses plastic liners filled with water and plastic-cushioned isogrid aluminum ribs which function like joists. Inner floors use plastic-cushioned planks or deck plates of aluminum or bamboo, potentially over their own isogrid layer for heavy-use areas. This layer protects the hull against abrasion and distributes loads to prevent punctures.
On top of that floor fits standard furnishings. Walls are thin metal or even Shoji-style paper with foam cores for noise attenuation. This is less about minimizing weight and more about reducing the amount of material used. Each floor has its own unique radius of curvature, so furniture is designed to bear weight along the flat axis and accommodate multiple radii along the curved axis. Still, the structure is designed to handle a floor load of 100kN/m² (nearly 15 psi / 2,160 psf), which is strong enough to drive heavy machinery over without damage. Most objects are made out of 'foamed' aluminum or titanium, or bamboo; the choice of material depends on what resources are most plentiful at the time. Bamboo sinks carbon and nitrogen while foamed metals sink metals; either can be reprocessed back into base material if necessary.
Furnishings can be painted, anodized or covered in fabrics as appropriate. Bathroom and kitchen facilities will be similar to Earth counterparts, with added efficiency features like auto-off taps. Plumbing would generally be plastic and wiring would generally be aluminum with plastic insulation. Ceramic materials will serve the same purpose as on Earth, so a table setting is likely to have very familiar dishes. Silverware will likely be titanium since that is less complex to produce than stainless steel. Environmental systems will consist of air supply and return, with user-settable temperature and humidity controls. Each compartment will include emergency oxygen supplies and patch-sealing kits, but this is largely a formality.
The layout of the habitat depends on a number of factors. I've examined dedicated residential and occupational floors but in order to balance the load on environmental systems it may be better to make each floor carry a mix of spaces. There should be a lower limit for gravity in living spaces, and those living in low g should work in higher (and vice versa) if commuting is required. This supports an argument for above-normal gravity in parts of the habitat, but doing this for the entire outer floor would take up a lot of living space at increased gravity. At any rate, some mix of residential and work space will exist, with hallways to connect them and public spaces such as parks to break things up.
Access between floors would be by elevator at the ends of the cylinder. There should be no penetration of each floor's structural hull; all connections pass through the sidewall at either end and then in or out as needed. The stresses involved are significantly less. In any case, the longest commute to work would be only a few hundred meters at worst; an easy walk.
Each floor balances its center of gravity by pumping water between storage bags at various points. The elevator system does the same, measuring the mass of occupants for a trip and moving mass on the other side to counterbalance the effect. This allows the structure to respond to shifting mass distribution and maintain balance without requiring significant amounts of power.
The ends of each cylinder hold other utility connections: air, water, power. Airlock connections to neighboring modules are here. The end habitats have larger airlocks to allow for large equipment to be moved in or out. Each floor handles as much of its own life support and environmental load as possible, but power needs to come in from outside and heat needs to go back out. Power can be transferred through a rotation surface (like a DC brush motor), but it is more efficient to provide power connections at either end of the cylinder and run electricity through the axis.
There is no 'central shaft'. Loads are carried through a ring connection between each habitat. This interface is where motors can spin up both members of a habitat pair without needing to use reaction mass. A short nonrotating access tunnel between modules can also accept concentrated sunlight from outside the shield and forward it via waveguide to growing areas. This section carries the load of any position adjustments made using magnetic pads connected to the shield, as well as sensors necessary to determine position and rotation of each module and the shield itself.
Hydroponics make the most of the available space. Aquaculture tanks will typically occupy the bottom half-meter to meter of the space, with stacked layers of high-intensity hydroponics above. Methods will be specific to each species; leafy greens like lettuce may be grown in float rafts directly in the fish water or in NFT channels while most vegetables will use flood and drain. Grains will generally use sub-irrigation (capillary transport). Racks will be built from extruded aluminum cross-sections that assemble without fasteners. Trays will be either aluminum or plastic sheets, all built to a standard size. Some areas will use motorized tray systems where each tray is passed automatically through a series of racks ending in a harvesting machine. Lighting will be from guided sunlight where feasible and from LED lighting otherwise. LED light sources will be manufactured in the colony's semiconductor facility.
Edible produce will be routed to food distribution. The specifics depend on culture: do the inhabitants cook their own meals, eat at a cafeteria or some combination of the two? There are some economic questions involved as well, but let's deal with that later. Any harvest waste or surplus that cannot be stored will be converted into feed; part of that stream will be concentrated into high-protein insect meal then mixed back into the rest as appropriate. For production of plastic, sugars and carbohydrates will be fermented into alcohol as a feedstock. The waste from that process is also suitable animal feed. Fibrous production (cotton, bamboo, flax) will be processed as necessary and sent on for spinning and weaving.
The hydroponics sections will receive waste air from the environmental system. This is high-humidity high-CO2 air that helps maintain an adequate CO2 level for growth (around 1000 ppm). Since plant growth in this system will fix much more carbon than is available from life support, additional carbon must be introduced. All inedible biological waste (including sewage) will be passed through a supercritical water oxidation reactor, which will produce CO2, water and ash (mostly mineral salts). This captures virtually all carbon in the system as CO2, except that which is fixed into structural plastics and similar uses. As a further purification step, the output of these systems will be fed to Spirulina in order to capture any dissolved minerals. The harvested algae can be used as a dietary supplement for people and animals or can be autolyzed into liquid nutrient solution.
Life support / environmental systems start at the user-facing side with air exchangers. Their purpose is to pass return air through a zeolite bed to extract nearly all CO2, then provide proper humidity and temperature for the supply air. A carbon bed removes odors and VOCs; this is formed from charred plant material and is recycled by burning in the catalytic reactor. Atmospheric bulk is made up with compressed nitrogen tanks as necessary. The hydroponics sections will include oxygen concentrators to provide oxygen-rich air back to residential areas.
Power to perform this task comes from external systems, and heat extracted plus heat generated by extraction must be rejected to outside systems. Phase-change refrigerants are a convenient way to manage heat transport, so it is expected that either ammonia or light hydrocarbons could be used as working fluids.
Ultimately all heat produced inside the colony must be rejected to space. This will require a substantial radiator structure that must be sun-shielded. On its way off-station this can be used as process heat for tasks like melting ice or pre-heating ores. The radiators must be highly modular and should limit the amount of coolant in each loop; damage is inevitable so repairs should be straightforward and resources lost to a puncture should be feasible to replace.
Power for all of this activity comes from concentrating solar PV panels. Reflectors will be built of aluminum, either as sheet or as thin film over plastic. PV cells will be produced in the colony's semiconductor facility and will be actively cooled using the same heat rejection system as environmental (including any process heat applications).
All of this structure requires a significant industrial base. Materials must be extracted from rock and converted to useful forms, a couple hundred thousand tons of it. The primary tool for this is large-area solar reflectors, so the equipment necessary to build these will be some of the first on the scene. A pre-screen to extract useful metal nodules, volatiles and ice will be done first. The remaining material will be heated and separated into component elements; this process can be stopped while specific oxides are still intact if desired. That yields a stream of nickel-iron, one of volatiles, one of light metals and one of heavier oxides. Trace elements, rare earths and other useful things will be accumulated in the process. Nickel-iron will be separated vial the Mond process, leaving pure iron, pure nickel and assorted iron-phase materials like platinum and rare earths. Rare earth elements and silicon will be processed via zone refining into semiconductor-grade bars and passed to the semiconductor facility to be turned into PV cells, LEDs and microprocessors. Volatiles will be further processed into atmosphere components and water; any hydrocarbons will be fed through the catalytic reactor for reclamation. The leftover slag will be used as shielding material.
The initial construction of the colony will begin with this kind of industrial equipment, where the first shielding sections will be assembled and the first structural plastics will be prepared. If construction is manned then the initial habitat and greenhouse will start producing plastic as quickly as possible; the greenhouse section would be much larger than necessary for just the construction crew. For an automated 'seed' system, fully chemical means could be used to develop the necessary structural fibers and build the first few floors; with suitable gravity and environmental systems a crew could live and work at the station during the middle construction phase and handle the more complex process of outfitting the internal volume properly as each floor is built. Each new floor would provide higher gravity; the facility could even be spun faster during construction as long as the crew can tolerate it.
Once a colony is established, it can continue to grow by capturing and processing asteroids into useful material. A wide variety of structures and devices can be built in addition to base materials like water and oxygen. Extra space in the colony can be devoted to food production. Such a facility could serve as a port offering food, medical care, repairs and restocking to spacecraft. Any parts or materials the colony cannot build or acquire for themselves could be bought or traded for valuable metals (platinum, uranium, thorium), bulk PV cells, water, etc.
The people of the colony will live lives much like we do on Earth, but with a strong reliance on advanced technology for survival. By the time a colony like this is possible, people will be able to live in the worst conditions on Earth: glaciers, deserts and to some extent underwater. It is difficult to predict what might change between now and then; perhaps we will see significant advances in robotic automation to the point that the facility's cleaning and basic maintenance are fully automated. Since I can't guess what the future will look like I've assumed a mostly Earth-normal distribution of occupations, with an emphasis on medical and scientific positions.
A basic economic system might value a good or service at the number of hours required to create it; time itself then becomes the basis of currency. In one sense, a colony is only possible if the time required to sustain a person is less than or equal to the time it takes them to do their daily work. Using production time as the basis for exchange would provide a starting point; several economic and governmental approaches could be taken from there to make any necessary adjustments and to ensure that enough of the necessary work gets done so the colony survives.
Construction of such a habitat is likely to be driven for a purpose, whether that be corporate profit, national prestige or religious ideology. It won't be until a significant number of these are built and populations routinely travel among them that people will begin to consider life in space to be normal. Still, there is enough material in the asteroid belt to build colonies for perhaps a few tens of billions of people. Once we include outer bodies and Oort cloud objects that number goes up by a factor of 100 or more. Add in material from various moons and accessible planets and our race's voracious expansion can continue for the next few centuries unabated.