I haven’t yet read Kim Stanley Robinson’s new novel Aurora (Orbit, 2015), though it’s waiting on my Kindle. And a good thing, too, for this tale of a human expedition to Tau Ceti is turning out to be one of the most controversial books of the summer. The issues it explores are a touchstone for the widening debate about our future among the stars, if indeed there is to be one. Stephen Baxter does such a good job of introducing the issues and the authors of the essay below that I’ll leave that to him, but I do want to note that Baxter’s novel Ultima is just out (Roc, 2015) taking the interstellar tale begun in 2014’s Proxima in expansive new directions.
by Stephen Baxter, James Benford and Joseph Miller
‘Ever since they put us in this can, it’s been a case of get everything right or else everyone is dead . . .’ (Aurora Chapter 2)
This essay is a follow-up to a review of Kim Stanley Robinson’s new novel Aurora by Gregory Benford, which critically examines the case that Robinson makes in the book that ‘no starship voyage will work’ (Chapter 7) – at least if crewed by humans. This is a strong statement, and even if the case is made in fictional form it needs to be backed up by a powerful and consistent argument. Greg criticises Robinson’s book mostly on sociological, political and ethical grounds.
Here, to complement Greg’s analysis, we take a critical look at the science in the book. Is Robinson’s ship a plausible habitat for a centuries-long voyage? Could the propulsion systems function as described? Is the planetary threat encountered by the would-be colonists biologically plausible?
This entry is mainly the initiative of Jim Benford, well known to readers of this blog; Jim is President of Microwave Sciences based in Lafayette, California, and his interests include electromagnetic power beaming for space propulsion. Also contributing has been Joseph Miller, biologist and neuroscientist, previously of the University of Southern California Keck School of Medicine, now at the American University of the Caribbean School of Medicine, with a long-time interest in extraterrestrial life. As for myself, I’m a science fiction writer, part-time contributor to such technical projects as the BIS-initiated Project Icarus, and author of some interstellar fiction myself, such as Ark (2009). And as the full-time writer I’m the one who got the privilege of writing up our conversations. Thanks, guys!
I should start by saying that Stan Robinson has been on my own (very short) list of must-read writers for the last twenty-five years at least, and that Aurora is a key book, as with all Robinson’s work deeply researched and deeply felt. If you haven’t bought the book yet, do so now.
Aurora is a tale of a multigeneration starship mission to Tau Ceti. (Note that Robinson’s starship is unnamed; here I’ve referred to it as ‘the Ship’.) The Ship reaches its target, but when it proves impossible to colonise the worlds there, a remnant of the crew struggles back to Earth.
This review is an analysis of technical and science aspects of this mission, based solely on evidence in the novel’s text. Of course any errors or misreadings of Robinson’s text are our sole responsibility.
We’ll be making comparisons with two classic studies. The BIS’s Project Daedalus (1978) was a study of an uncrewed interstellar probe which used the same fusion-rocket technology as did Robinson’s Ship in its deceleration mode. Daedalus had initial mass 50,000t (tonnes) of fuel (30kt deuterium (D), and 20kt helium-3 (He3)), the dry mass of its two stages amounted to 2700t, the payload was 450t, and the exhaust velocity was about 3.3%c, with cruise velocity 0.12c (c being the velocity of light). The Daedalus propulsion system was used only for acceleration; it couldn’t decelerate, and so was a flyby mission at its target star. In Aurora the Ship uses its fusion rocket only to decelerate.
Meanwhile the ‘Stanford Torus’ space habitat design (Johnson, 1976) was a product of a 1975 workshop involving NASA Ames and Stanford University. The final design was a torus 1790m across with the habitable tube 130m in diameter. Of a total surface area of about 2.3km2, 10,000 people would inhabit a usable surface area of about 0.7 km2. The station, located at L5, would be built of lunar resources. The total mass would be about 10 million t, of which 9.9 million t would be a radiation shield of lunar slag around the habitable ring in a layer 1.7m thick, leaving 0.1 million t as structural mass. The relevance to Aurora is that the Ship looks like two Stanford Toruses attached to a central spine.
Let’s begin by looking at the Ship’s construction and inhabitants.
Most of what we learn about the Ship’s structure is given in Chapter 2. The Ship consists of a central spine 10km long, around which 2 rings of habitable ‘biomes’ spin, torus-like. Each ring consists of 12 cylindrical biomes, each 4km long, 1km diameter. There are also spokes and inner rings. The rings rotate around the spine to give a centrifugal gravity of 0.83g.
The 24 biomes contain samples of ecospheres from 12 climatic zones: Old World versions in one ring, New World in the other. Each biome has a ‘roof’ with a sunline, which models the required sunlight and seasonality, and a ‘floor’ on the side away from the spine. The liveable area in each cylinder is given as about 4 km2, which is about a third of the cylinder’s inner surface area: 96km2 total. In each biome there are stores under the ‘floor’, including fuel; we’re told this is used as a radiation shield during the cruise.
The total habitable space is allocated as 70% agricultural; 5% urban / residential; 13% water; 13% protected wilderness. The wilderness areas are meant to be complete ecologies.
The crew numbers given appear contradictory; in some places Robinson states there are about 2100 total, but elsewhere is given a number of 300 people per biome which would total 7200. The crew numbers do vary through the centuries-long mission, with births and deaths.
How reasonable are these numbers, given the mission’s objectives? Could the Ship support that many people? Are they enough to found a human population at the target? And is there room for true wilderness?
We don’t yet know how to maintain closed ecologies for long periods. The Ship’s biomes would suffer from small-closed-loop-ecology buffering problems, as Robinson illustrates very well in the text; we see the crew having to micro-manage the biospheres, and dealing with such problems as the depletion of key trace elements through unexpected chemical reactions. In some ways this may prove to be an even more daunting obstacle to interstellar exploration than propulsion systems.
If there are 300 people per biome, and given a total of 96km2 habitable area, that’s a population density of 75 /km2. Compare this with Earth’s global average of 13 /km2 ; crowded southern England is 667 / km2. In terms of the ability of the agricultural space (70% of total) to support the crew, that seems reasonable to us.
But if only 5% of the space is used for residential purposes, the effective living density is high, at 1500 per km2 – comparable to densely populated urban areas such as Hong Kong. Such densities would seem problematic on a long-duration mission, though of course the crew do have access to the other 95% of the habitable areas; people hike the wildernesses.
This group is of course meant to be sufficient to found a new human breeding population on a virgin world. What is the minimal population size to maintain the species without an evolutionary bottleneck? Something like 1000 is a good guess. Robinson’s original population was at least twice that. If that population size was maintained, genetic diversity would plausibly be sufficient.
We’re told (Chapter 2) that each biome has about 4km2 of living space and that 13% of that space is given over to ‘wilderness’, that is 0.52 km2 per biome. The ecologies can include apex predators. In a biome called Labrador, for instance, ‘In the flanking hills sometimes a wolf pack was glimpsed, or bears’ (chapter 2).
This idea is explored in more depth in Robinson’s 2312, in which mobile habitats called ‘terraria’, hollowed-out asteroids, are used as reserves for species threatened on a post-climate-change Earth. But even these terraria are not very large in terms of the space needed by wildlife in nature. A wolf pack, consisting of about 10 animals, may have a territory of 35 km2 (Jędrzejewski et al, 2007). A 2312 terrarium with an inner surface area of about 160 km2 would have room for only about 4 packs, or about 40 individual animals, a small population in terms of genetic diversity.
It seems clear that the much smaller biomes of the Ship, though large in engineering terms, would be far too small to be able to host meaningful numbers of many animal species in anything resembling a natural population distribution. A wilderness needs a lot of room.
We are given a mass breakdown for the Ship as a whole. We’re told that during the Ship’s cruise phase, when it is fully laden with fuel, the total mass is 76% fuel, 10% each biome ring, and 4% the spine.
We aren’t told the Ship’s total mass, however, and to study the propulsion system’s performance we’ll need at least a guesstimate. This is derived by a comparison with the Stanford Torus design.
Each torus-like biome ring consists of 12 pods of length 4km, diameter 1km. So the surface area of 1 pod is 14.1 km2, including end caps. And the surface area of one biome ring is 170 km2 (which is much larger than the Stanford Torus).
The Ship’s biomes seem to lack a Stanford-like cloak of radiation-shielding material. Robinson says that ‘fuel, water and other supplies’ are stored under the biome floors to provide shielding; the ceilings are shielded by the presence of the spine. Elsewhere Robinson says that during the voyage, the fuel is ‘deployed as cladding around the toruses and the spine’ (Chapter 2)
Assume then that if a Ship biome ring has the same structural properties as the Stanford torus, and if most of its mass is in the hull, then a guesstimate for a single ring mass (without the fuel cladding) can be obtained by multiplying Stanford’s 0.1m tons structure mass (without shielding) by a factor to allow for the Ship ring’s larger surface area. The result is (0.1 * 170 / 2.3 =) 7.4 million tons per biome ring. We know this is 10% of the Ship’s total mass, which therefore breaks down as
76% fuel = 56.2 million tons
20% biome rings = 14.8 million tons
4% spine = 3 million tons
Total = 74 million tons.
These numbers shouldn’t be taken seriously, of course, except as an order of magnitude guide. Maybe they seem large – but remember that Daedalus needed 50,000t of fuel to send a 450t payload on a flyby mission to the stars, a payload comparable to the completed mass of the ISS. By comparison the Ship will be hauling two habitat rings each fifteen kilometres across. This is not a modest design.
Notice that if the Ship’s propulsion follows the Daedalus ratio, the fuel would consist of 60% D = 33.7m tons, 40% He3 = 22.5m tons.
And notice that since this fuel is used for deceleration only, the acceleration systems need to push all this mass up to ten per cent of lightspeed. These numbers do illustrate the monstrous challenges of interstellar travel, with a need to send very large masses to very large velocities, and decelerate them again.
On that note, let’s consider the propulsion systems.
The Ship is a generation starship. Launched in 2545, it travels 11.8ly (light years) to Tau Ceti at cruise 0.1c (chapter 2). According to the text the journey consists of a number of phases.
- The Ship is accelerated to the cruise speed of 0.1c by means of electromagnetic ‘scissors’ slingshot at Titan, imposing a brief’ acceleration of about 10g, and then a laser impulse for 60 years.
- The Ship decelerates at the Tau Ceti system using its on-board fusion propulsion system. The technology, like that used by Daedalus, is known as ‘inertial confinement fusion’ (ICF), in which pellets of fuel are compressed, perhaps with laser or electron beams, until they undergo fusion; the high-speed products provide a rocket exhaust. For twenty years the Ship is decelerated by the detonation of fusion pellets at a rate of two per second. The fusion fuel is a mix of D and He3, as was the case for Daedalus (Chapter 1).
- We’re told that the total journey time is about 170 years (Chapter 3), consistent with the profile given.
- Colonisation in the Tau Ceti system is attempted and fails (this will be considered below).
- A section of the crew chooses to return to the Solar System. The ICF system is refuelled at Tau Ceti, and used to accelerate the Ship to 0.1c (Chapter 5).
- As the Ship’s systems break down, the surviving crew completes the final leg of the journey in cryosleep.
- The Ship has no onboard way to decelerate at the Solar System (Chapter 6). The ICF fuel was exhausted by the acceleration from Tau Ceti, save for a trickle to be used during Oberth Manoeuvres (see below). The laser system reduces the Ship’s speed, but not to rest: from 10%c to 3%c. We’re told that the Ship then sheds the rest of this velocity mostly with 28 Oberth Manoeuvres, using the gravity wells of the sun, Jupiter, and other bodies. This process takes 12 years before crew shuttles are finally returned to Earth.
We can consider these phases in turn.
Acceleration from Solar System
In considering the acceleration system, it should be borne in mind that what we need to do is to give a very large, fuel-laden Ship sufficient kinetic energy for it to cruise at 0.1c. And because of inevitable inefficiencies, the energy input to any acceleration system will have to be that much greater.
In fact the launch out of the Solar System is a combination of two methods, vaguely described, neither of which is remotely efficient. There’s a ‘magnetic scissor’ that accelerates the ship over 200 million miles: ‘…two strong magnetic fields held the ship between them, and when the fields were brought together, the ship was briefly projected at an accelerative force equivalent to 10 g’s’.
(Of course such acceleration would stress the crew, even though in tests humans have survived such accelerations for very short periods – indeed the book claims five crew died. And such acceleration could stress lateral structures, such as the spars to the biome rings. Perhaps the stack is launched with its major masses in line with the thrust, and reassembled later.)
In Jim Benford’s grad school days, he ran some actual experiments on this effect, using a single turn coil. The energy in the capacitor bank driving it was about 1 kJ and the subject of the acceleration was a screwdriver sitting on a piece of wood in the coil centre. The coil current pulsed to peak in 2 µs. The screwdriver was accelerated across the room to a target at about 10 meters per second. The kinetic energy of the screwdriver was about 5 J and therefore the efficiency of transfer was less than 1%. It seems unsafe to assume an efficiency much better than this.
For the Ship, there then follows a laser driven acceleration. While lasers can certainly accelerate light craft, as has been shown experimentally, they can’t accelerate the enormously massive vehicle that the novel describes. The power required to accelerate by reflection of the laser photons can be calculated from the Ship mass (74 million tons), final velocity and acceleration time (to 0.1c in 60 years, so 0.17% g). The amount of power is about 100,000 TW, a truly astronomical scale. (Earth’s present electrical power output is 18 TW.) The efficiency of power beaming is low because only momentum is transferred from the photons to the ship. Efficiency is the time-averaged ratio of velocity to the speed of light. Therefore the efficiency of this process is about 5%.
The Ship and its mission would have to be a project of a very wealthy and very powerful interplanetary civilisation. It seems unlikely that they would resort to such a hopelessly inefficient system, if it could be made to work at all.
Deceleration at Tau Ceti
The Ship uses its onboard fusion rocket to decelerate.
We’re told the ICF deceleration phase takes 20 years at 0.005g, starting from 10%c cruise speed, with a Ship with an initial fuel load of 76% total mass. These numbers enable us immediately to calculate one critical number, the exhaust velocity of the fusion rocket. A ship with 76% fuel mass has a mass ratio (wet mass / dry mass) of (100/24=) 4.17. The rocket equation tells us that given that mass ratio and a total velocity change of 0.1c, the exhaust velocity must be 7%c. This is twice that of Daedalus, but perhaps not impossible for an advanced ICF system.
Our mass guesstimate above allows us to assess the performance of the rocket. Consuming 56.2mt of fuel in 20 years gives a mass usage rate of 94 kg/sec (cf Daedalus first stage 0.8 kg/sec). (Notice that the two fusion ‘pellets’ consumed per second are pretty massive beasts; in the Daedalus design pellets a few millimetres across were delivered at a rate of hundreds per second. This detail may be implausible. Indeed 49kg may be larger than fission critical mass!)
You can find the rocket’s thrust by multiplying mass usage by exhaust velocity, to get about 2000 MN (megaNewtons). This is much larger than the Daedalus first stage’s 8 MN. And the rocket power is 20,000 TW (the Daedalus first stage delivered 30 TW). Note that this power number is comparable to the launch figures.
Again, these numbers can be taken only as a guide. But you can see that the power generated needs to be maybe three orders of magnitude better than Daedalus, and exceeds our modern global usage by four orders of magnitude.
Meanwhile this system would consume a heck of a lot of fusion fuel. Where would you acquire that fuel, and where would you store it?
The storage is the easy part, relatively. Daedalus’s 50 kt of fuel was stored in six spherical cryogenic tanks with total volume 76,000 m3. At similar densities to store the Ship’s fuel load would require 860 million m3. That sounds a lot, but the volume of a biome ring is about 38 billion m3, so the fuel volume is only 2% of this, making it plausible that it could be stored, as Robinson says, in cladding tanks on the biome rings and spine, without requiring large separate structures. The Ship is big but hollow. It’s not immediately clear however how effective a layer of fuel would be as a cosmic radiation shield.
And note that the need for cryogenic store over centuries before use would be a challenge – as would the need to store any short-half-life propulsion components such as tritium, which has a half-life of 12.3 years, and would decay away long before the 170-year mission was over.
Getting hold of the fusion fuel, meanwhile, is the tricky part. It’s hard to overstate the scarcity of He3 in the Solar System, and presumably at Tau Ceti. Even Daedalus’s 20,000t would deplete the entire inventory of the isotope on Earth (37,000t), and the Ship’s 22.5mt would dwarf the Moon’s store (1 million t); only the gas giants could reasonably meet this demand (the Daedalus estimate was that the Jovian atmosphere contains about 1016 t). The Daedalus design posited acquisition from Jupiter, but estimated that to acquire Daedalus’s fuel load in 20 years would require that the Jovian atmosphere be processed at a rate of 28 tonnes per second. So again the challenge for the Ship’s engineers will be three orders of magnitude more difficult.
And regarding the return journey, although the Ship is stripped down, a fuel load of similar order of magnitude must be acquired from the Tau Ceti system, and without the assistance of a Solar-System-wide infrastructure. Of this huge project, Robinson says only that ‘volatiles came from the gas giants’ (Chapter 4).
Deceleration at Solar System
At the end of the novel, the Ship returns to Earth, decelerating mostly using what is called the ‘Oberth Manoeuvre’, invented by Hermann Oberth in 1928. This is a two-burn orbital manoeuvre that would, on the first burn, drop an orbiting spacecraft down into a central body’s gravity well, followed by a second burn deep in the well, to accelerate the spacecraft to escape the gravity well. A ship can gain energy by firing its engines to accelerate at the periapsis of its elliptical path.
Robinson wants to use this to decelerate from 3% of light speed down to Earth orbital velocity. 3% of lightspeed is 9,000 km/s. For reference, Earth’s orbital velocity is 30 km/s. Several deceleration mechanisms are referred to in the book. An unpowered gravity assist, passing by the sun and reversing direction, can steal energy from the sun’s rotational motion around the centre of the galaxy. That’s worth about 440 km/s. Other unpowered gravity assists can be used once the ship is in a closed orbit in the sun’s gravitational well. Flybys for aerobraking in the atmospheres of the gas giants are referred to as well. Altogether, these can get you <100 km/s.
But the key problem with using the Oberth Manoeuvre for deceleration of this returning starship is that this craft is on an unbound orbit. That means that, on entering the Solar System its trajectory can be bent by the sun’s gravity, but will then exit the System because it has not lost enough velocity to be bound to the Solar System. To be bound would require velocity decreased down to perhaps 100 km/sec, which is 1% of the incoming velocity. Therefore 99% of the deceleration has to take place in the first pass. And you can’t get that much from an Oberth Manoeuvre.
As the Ship’s systems collapse, the returning crew gets from Earth plans to build a cryonic cold sleep method, which allows the viewpoint characters to survive until they reach the Earth.
This technology logically undermines most of the problems the early parts of the novel confront, and therefore undermines most of Robinson’s point about the difficulty of interstellar travel: If only the colonists had waited a few centuries for cryo technology, it would all have been so much easier! But this contradicts Robinson’s thesis.
Having arrived at Tau Ceti, the colonists’ target planet, called Aurora, is judged lifeless but habitable from a remote sensing of an oxygen atmosphere – presumed created by non-biological process billions of years ago – but in the event the environment proves lethal for humans because of the presence of a deadly ‘prion’.
In a sense this is the point of the novel, that even if we reach the stars we will find only dead or hostile worlds: ‘I mean, they [alien worlds] are all going to be dead or alive, right? If they’ve got water and orbit in the habitable zone, they’ll be alive. Alive and poisonous . . . What’s funny is anyone thinking it [interstellar colonisation] would work in the first place’ (chapter 3). And as Greg noted in his essay this reflects recent misgivings expressed by Paul Davies and others about the habitability by Earth life of exoplanets.
Is this reasonable? And is Robinson correct that this could be the solution to Fermi’s famous paradox?
Robinson seems to be saying ‘alive’ worlds will be toxic to all possible biological explorers (there is a little wiggle room here since non-biological automated probes might still survive such worlds). This is a bold statement, but plausible since we lack any relevant data. However Robinson also says ‘dead’ worlds, essentially rocky Earth-size planets in the Goldilocks zone, could be terraformed but that project would take thousands of years. But why should that matter in a galaxy that is billions of years old? There should be plenty of time to terraform such planets, either by biological explorers or perhaps some type of self-replicating von Neumann probes or seed ships. There appears to be no solution in Aurora to Fermi’s question.
Oxygen and Biosignatures
(See Sinclair et al (2012) for a relevant reference.)
It seems implausible that oxygen in Aurora’s atmosphere might not be a biosignature: that is, that it could credibly be created by non-biological processes. Without some continual input into the atmosphere, you would expect any oxygen to rust out, as on Mars. Robinson says the oxygen on Aurora is due to the ultraviolet breakdown of water. We haven’t run the numbers, but that would be a hell of a lot of UV (which itself could make the planet uninhabitable). That might actually work better as a mechanism for oxygen production on Mars, at least long ago when Mars had liquid water. Indeed, UV is how Mars lost its water and atmosphere, and the same would happen on a dead Earthlike world. So Aurora can’t have oxygen; it gets blown off after the hydrogen from water.
Robinson also cites a failure to detect CH4 and H2S, possible markers of life, in Aurora’s air as ruling out a biological origin for the oxygen. However the interpretation of the presence of methane (CH4) in the Martian atmosphere has been a bone of contention for well over 15 years. Is it a biomarker or an index of geological activity? And as far as hydrogen sulphide goes, it sure as hell is not a biomarker on Io!
The most significant biological problem in Robinson’s scenario is the organism that was so toxic to humans on Aurora. This is said to be ‘something like a prion’, and is apparently an isolated organism: as far as the explorers could tell there simply was no wider biosphere on Aurora.
For a biologist, that sounds really weird. This is a satellite a couple of billion years older than Earth and the only evolved organism is a prion? In addition we are not sure what ‘something like’ really means, but if it was indeed like a prion one must ask: where on Aurora are the proteins capable of being misfolded by a prion action? That’s what prions do; they cannot exist in isolation. And then why was it that human proteins, from a different biosphere altogether, were such a good match to the prion’s mechanisms?
Of course you can say it was ‘something like’ a prion but not really a prion. But then, what makes it ‘like’ a prion if not protein-folding?
It would take a lot more detail to make this strange single-organism biosphere a plausible ecosystem. Maybe if Robinson ever revisits Aurora and the stayers we could find out! Joe Miller thinks that an Andromeda Strain-like organism, inimical to Earth biology, is no more or less likely than ET organisms which simply find Earth biology indigestible. We don’t know, but the possibility that ET biology would be simply oblivious to Earth biology is a plausible situation, though not treated very much in SF because it is not very dramatic!
Robinson’s Aurora is a finely crafted tale of human drama and interstellar exploration. Its polemic purpose appears to be to demonstrate, in Robinson’s words, that ‘no [human-crewed] starship voyage will work’. There is much of the science and technology we haven’t explored in this brief note; there’s probably a master’s thesis here – indeed I’ve recommended the book to Project Icarus as a study project.
However, to summarise our conclusions:
- The human crew transported to Aurora may plausibly be large enough to support a new breeding population. And the Ship’s dimensions seem adequate to support the crew through their centuries-long mission.
- The challenge of maintaining small closed biospheres is depicted credibly, but the ‘wilderness’ areas of the biome arks are too small for their purpose.
- Of the elements of the propulsion system, the electromagnetic / laser Solar System acceleration system needs to be so powerful it stretches credibility, while the Oberth Manoeuvre return-deceleration system as depicted is impossible. The ICF fusion rocket system appears generally credible, but would require the acquisition of heroic amounts of helium-3 fuel, a challenge especially at Tau Ceti.
- Regarding Aurora itself, the notions of a non-biogenic oxygen atmosphere, and of a single-organism biosphere, and that an extraterrestrial organism as described might necessarily be inimical to humans, all lack credibility.
In summary, while Aurora is an intriguing combination of literary, political, scientific and technical notions, and while it reflects many current speculations about the difficulty of interstellar travel, in many instances it lacks the supporting credible scientific and technical detail required to make its polemic case that human interstellar travel is impossible. The journey is not plausible, and nor is the destination.
What Aurora illustrates very well, however, at least at an impressionistic level, is the tremendous difficulty of mounting such a voyage. Interstellar travel is a challenge for future generations, which will bring both triumph and tragedy.
Kim Stanley Robinson, Aurora, Orbit, 2015.
Kim Stanley Robinson, 2312, Orbit, 2012.
Bond et al, Project Daedalus Final Report, British Interplanetary Society, 1978.
Johnson, Richard D. and Holbrow, Charles, (editors), ‘Space Settlements: A Design Study’, NASA SP-413, 1977.
Jędrzejewski W, Schmidt K, Theuerkauf J, Jędrzejewska B, Kowalczyk R. 2007. ‘Territory size of wolves Canis lupus: linking local (Białowieża Primeval Forest, Poland) and Holarctic-scale patterns’. Ecography 30: 66–76.
Sinclair, S., Schulze-Makuch, D., Radley, C., Papazian, A., Miller, J., Marzocca, P. Lee, J., Gaviraghi, G., How to Develop the Solar System And Beyond: A Roadmap to Interstellar Space, Kindle Books, 2012.