A Sketch of a Self-Replicating von Neumann Probe "Thistledown" Introduction Interstellar colonization and exploration is a hard problem due to the immense distances involved; sending a crewed spaceship across a distance of lightyears would entail solving several large engineering problems: accelerating a spaceship to relativistic speeds using known sources of energy appears very hard to implement, which in turn implies a trip duration of at least decades. This in turn suggests the need for reliable suspended animation or the design of a long-term viable habitat, which would likely increase the necessary mass of the expedition significantly and create problems with keeping many interconnected life-support systems running optimally for decades. Sending unmanned probes appears to be more feasible, especially since they can be built smaller than crewed expeditions and hence require less energy to accelerate. Sending individual probes to nearby stars appears to be feasible using modest extrapolations of current technology [Daedalus, starwisp etc], and could give valuable information about the systems. Unfortunately most designs proposed so far do not provide for long-range exploration or colonization, just quick flybys. This paper describes one possible architecture for a general purpose exploration and colonization probe, employing self- replicating machine technology (von Neumann machines) to both establish a beach-head in the remote system and to send copies of itself to unexplored systems. The use of von Neumann probes have previously been suggested by others [Refs, refs in Tipler]. The Basic Lifecycle of a Probe The lifecycle consists of nine different phases: 1. Launch. Using a laser-driven solar sail the probe is launched towards the destination system. Over a period of several years it is accelerated to a high albeit sub-relativistic speed. 2. Coasting phase. The probe passively travels the distance between the systems. 3. Braking phase. [Magsail?] The laser is switched on. The solar sail divides into a small central disc and a large outer annulus. Reflected light from the annulus is used to brake the central disc as it approaches the destination system. 4. Navigation phase. The probe navigates the destination system using its solar sail. Using its sensors surveys the system, especially looking for carbonaceous chondrites. 5. Seeding. Once the probe has found a suitable asteroid, it approaches and plants one javellin-like seed into its surface. This seeding process may be repeated with other asteroids in order to reduce risks. 6. Sprouting. Self-replicating nano- or micro-machines are released from the javellin-seed. They begin to colonize the asteroid surface, replicating and building a photovoltaic covering which also acts as a protective shield against ultraviolet light. 7. Maturation. Once the probe "biomass" is large enough growth is changed from replication to building an antenna system. 8. Flowering. The antenna is directed towards the departure system, and a signal is sent back. The probe goes into a hibernating state where it waits for a response. 9. Reproducing. Eventually a responce arrives from the departure system, containing new instructions about what to build. This could involve building new probes and launching laser systems, or the creation of a habitat suited for colonization. A radio beacon is activated proclaiming the system as inhabited (to avoid having other systems send redundant probes). Note that in this design the probe does not contain all the information necessary for full replication (the solar sail and laser arrangement) until specifically given it in the reproduction phase. This makes it impossible for the probe (or any simple mutation of the basic template) to replicate out of control; instead the probe- colonized systems form a communication network where each new generation has to be deliberately launched. Of course, if an independently spreading probe is desired, the design could be extended by adding the necessary information and behavior programming at a slight increase in complexity. The time per generation of this probe design appears to be on the order of decades; slow by current standards but still spreading exponentially. It should be noted that for the cost of a single initial probe and its launch indefinite growth becomes possible. Technological Assumptions These steps involve reasonable extrapolations from current technology or theoretical applied science. The most "risky" assumptions are the feasibility of nanotechnological (or microtechnological) replicators, sufficiently complex robotic behavior and interstellar laser-driven solar sails. Nanotechnological replicators appear to be physically possible [Nanosystems] and planty of research is directed towards this area [REF]. That robotic behavior of sufficient complexity (navigating a solar system, finding a suitable asteroid, nanoconstruction) is possible appears reasonable, at least in the case of the first two problems (the complexity of the third problem is currently unknown). Laser-driven solar sails have been analysed in [REF FORWARD?]. As Landis [Landis, Small laser propelled probe] points out the propulsion system is achievable entirely within currently known laws of physics, but it requires engineering advances in the field of large laser systems, the ability to fabricate large gossamer structures with high position control and the ability to fabricate very thin reflective films. The assumption of nanotechnology makes the last problem comparatively simple, and will likely make it possible to build the needed Fresnel lens using methods similar to the methods suggested in [chapter Flowering] using asteroidal material. Probe Design The basic probe design is a solar sail with a core package and sensors distributed along the rim. The core package contains instrumentation, processing and a number of javellins to seed the asteroids. Each javellin contains, beside the necessary armor and launch equipment, solar collectors, an initial population of replicators and extra instructions which are activated once when needed. solar sail The solar sail is intended to both be launched using a laser system and to navigate in the destination system using solar light pressure. The solar sail structure is determined by how short wavelength laser can be used for launch, the maximum operating temperature of the sail and the availability of the necessary elements in other solar systems. Landis [ref] considers both reflective sails and dielectric sails, and diamond or silicon carbide dielectric sails appears to have many desirable properties. However, their absorbtion and emissivity may be less than optimal and hence Landis concentrates mainly on reflective sails. However, with nanotechnology it appears very likely that the material properties can be fine-tuned on the molecular scale to become optimal, which would suggest a diamondoid sail. I will assume that these properties at least can be improved to correspond to aluminum, which is the otherwise most likely sail material. It should be noted that the initial sail could use the best possible sail materials (like beryllium) since it is built in the solar system and by being extra fast its cost can be amortized earlier. aluminum The launch system 500 nm ca 40% efficiency To provide the energy needs of the core package an annulus of photovoltaics is placed near the center. Actuators are assumed to be distributed across the sail, for example piezoelectric strands that allow control over the sail shape- sensors core package The core package will contain the javellins, a launching system, processors and some energy storage. javellins The objective of the javellins is to penetrate the upper regolith layer of the selected asteroids, deploy solar collectors and let loose the replicator systems. Their armor can likely be made out of diamondoid (it only has to survive impact; after that it is even desirable if the armor is shattered). If we assume a density 75% of diamond (half of the probe is diamond armor, the remaining packing material, processors, energy stores, replicators and photovoltaics), a cone shape, a maximum radius of one centimeter and a length of one meter the mass becomes 0.28 kilograms for each javellin. To ensure that the javellins stop at a desired depth in the asteroid they can be provided with extendable barbs which extend during impact at a certain depth; this also provides a suitable point of egress for the replicators. The photovoltaics telescope out of the rear end, extending fan-like to provide energy for the first generations of replicators. Since their cross-section is cylindric they will produce constant energy regardless of the angle of the sun as long as they are in light. By using a telescope-like structure similar to that suggested by Bishop [accelerator] for interstellar accelerators their length could be made many times the probe length, covering a much larger area. Assuming a length of 4 meters and a radius of 0.1 millimeter the total area becomes 25.13 m^2 (thinner photovoltaics have less surface area, but more can be packed into the javellin). The resulting hemisphere will produce on the order of 10 kW as long as it is in sunlight (assuming a solar constant similar to Earth orbit). A = length * 2Pi * 10^-4/radius kolla buckling memory launchers processor ID code navigation program weight size use present elements, contain different recipes. Launch This, and the following two phases are essentially equivalent to the proposal of Robert Forward [REF]. Coasting The main problem during this phase is damage from interstellar dust and radiation. Braking Navigation It appears virtually certain that there are asteroids around most stars [27 28 29, planet formation], most likely similar in composition to the asteroids in the solar system. The probe will navigate through the new system and attempt to identify one or more suitable seeding places, ideally carbonaceous asteroids of the right size. Seeding NASA probe. CRAm Sprouting nanotech ecology. Shield/photovolt replic, burrowers, gatherers. Auxons theory replicating machines The regolith environment of a carbonaceous asteroid is rich in carbon, oxygen, magnesium, silicon, sulphur and iron, with smaller amounts of sodium, aluminum, calcium and chromium. These are the basic building materials for the growing probe. It appears likely that carbon in a diamondoid form is the best material for the assemblers themselves, it is versatile, forms strong bonds and is relatively common. The model used here is somewhat reminiscent of an ant colony, with central assemblers supported by a population of specialized nanodevices such as energy-gatherers and material transporters. It is closely related to the auxon system [REF], which is intended to build solar energy collectors in desert areas. In the initial phase of growth both systems are doing exactly the same, replicating and extending solar panels, although in different environments and on different scales. Two problems are movement and energy distribution. The devices have to move through the regolith, and the energy from the surface has to be distributed into the regolith. Create solvent? Store energy in chemical form (solvent molecule?) Sulphur. Koldisulfid? vilka temperaturer i regolit? scavengers kemiska signaler assembler överbefolkning timer Maturation slime mold system Flowering problem with rotation: several antennas Reproducing cannot become obsolete Summary The probe described here is just one possible solution to the problem of self-replication in space and its use for exploration/colonization.