Interstellar WaferSat Flight: Miniaturization and Direct imaging
Inspired by the Breakthrough Starshot Initiative proposed to send a 1-gram-class spacecraft to Proxima b at 0.2 light speed, I reviewed the feasibility of accelerating object at realistic speeds, the scalable laser power array, and the waferSAT manufacturing techniques that miniaturize the spacecraft design, with special attention to emerging two-dimensional semiconductors that could potentially replace silicon across the board. I also discussed how our data bandwidth allows the spacecraft to send high-resolution pictures back to earth. If the bandwidth is proven to be estimated, we also discussed radio interferometry and transit spectroscopy, two methods already being implemented on earth to study atmosphere and species but would be done in situ with greater accuracy and resolution. Since they would be onboard computing processes, only a very small data bandwidth would be needed.
Technology Roadmap for Interstellar Flight
There is a profound distinction in what we have achieved in accelerating mass by traditional chemical materials, and what we did in facilities like the LHC using electromagnetic materials. We certainly already have the capacity to make individual particles travel at highly relativistic speeds, only bound by the ultimate limit of light speed. On the other hand, chemical processes typically can only supply several eV per bond. To reach a similar relativistic speed, we need a GeV per bond that is not deliverable by chemical materials.
Therefore to be able to propel any form of interstellar flight, we need to redefine our conventional concept of space probes. We need to build a miniature version of a fully functional probe with photon thrusters, cameras, processors, batteries, power management, protective coating, light sails, and transmitters. But each component needs to be rebuilt from the ground up, pushing the weight of the entire probe at the order of grams, or even less.
Even though this may seem extremely fantasy and romanticized, advancements in laser power, semiconductor transistors, CMOS array, and Photovoltaic materials have largely been following variants of ‘’Moores law’’. The co-founder of Intel, Gorden Moore, proposed that the number of transistors in a densely integrated circuit doubles approximately every two years. Therefore we can pack more processing power that scales as the number of transistors into a smaller computing unit while reducing power usage due to lower electrical resistance. We are seeing a similar trend in single laser power doubling in about two years, laser power cost per watt getting cut in half for about two years.
In the following subsections, I will be analyzing technology roadmap for each important component in the system.
Laser Phased Array as Photon Driver
The key to the system lays in the ability to build a photon driver. The photon driver is a laser phased array which eliminates the need to develop one extremely large laser. We replace it with a large number of modest (kilowatt class) laser eliminates the conventional optics. This is because the currently existing kilowatt class Yb fiber-fed lasers and phased-array are already a mature technology
Figure 1 shows that fiber laser cost based on the current Yb fiber amplifier cost per wall chart. Note that the Yb Fiber amp cost is shrinking in half almost every 18 months. So it is highly economical to focus on building a large laser array than building a single laser source.
Speaking in modern computing terms, this is equivalent to modern parallel computing rather than a single supercomputer.
Another obvious advantage is that while a single large laser source is inherently incapable of powering up multiple light sail objects simultaneously. Phased array technology, on the other hand, are designed to multitask. Not to mention much less maintaining costs.
A small array can also be used for orbital debris removal, long-range laser comm., power beaming, ISS defense from space debris as well as stand-on systems for planetary defense so again there is use at practically every level and funding is well amortized over multiple uses.
Wafer Scale Spacecraft
Silicon Wafer
Silicon satellites have been proposed in 1999 to utilize single crystal silicon wafers for electronic substrates, mechanical structure, thermal control and the radiation shield (Janson, S. W, 1999). Janson defined silicon satellites by mass in the range of femtosatellites (1-to-1000 microgram mass) for nanosatellites (1gm to 1kg mass). The key was to use semiconductor batch-fabrication techniques that can produce low-power digital and analog circuits, silicon-based radio-frequency circuits and micro-electromechanical systems (MEMS).
Single-crystal silicon is stronger than aluminum, stainless steel, or titanium yet is less dense. In terms of strength-to-density ratio, silicon is one order-of-magnitude more than that of Titanium and much more than aluminum or stainless steel. But thermally, silicon can be viewed as a brittle metal. Silicon is an excellent heat conductor, which means that we would need intensive thermal coating and isolation, on the other side.
However, it is very important to point out that almost the entire mass and volume of spacecraft electronic systems are determined by packaging such as enclosures, connectors, circuit boards, and chip carriers. Therefore, we use multi-chip modules to place multiple electrically connected die into a single carrier to reduce wasted space and increase active device area density. Wafer scale integration can take that approach even further by integrating all electronics on a single silicon wafer.
But we do recognize that we have a long way to go for this SpaceChip project. The satellite-on-a-chip idea was first proposed in 1994 by A. Joshi in an interview, and the idea has not yet been realized experimentally. In parallel to the SpaceChip feasibility study, a rapid PCB prototype of a fermotosatellite was also introduced in the hope of guiding the SpaceChip architecture.
Two-dimensional materials as a silicon replacement
Since the mass density of silicon is largely constant, the idea of SpaceChip or making all necessary electronics on a small silicon wafer is, essentially, reducing the volume of silicon and packaging weight. On a wafer, the weight of enclosures, connectors and circuit boards are not relevant because they are simply not present on a wafer.
Due to the quantum tunneling effect, when the distance from the source to the drain of a silicon-based transistor is down to about 5nm, the voltage would leak from one side to another preventing normal functionality. To put this fundamental physics limit in perspective with the current status of silicon fabrication techniques, 10nm-based computing units have been massively produced. Samsung flagship mobile phones of 2017 are using 10-nm CPUs. 7nm- based architecture is being produced for testing and optimization purposes at Intel and Samsung. We are, in fact, very close to the 5nm limit.
Scientists and engineers are highly motivated to find alternatives to silicon. The two- dimensional semiconductor is a type of natural semiconductor with thicknesses on the atomic scale. In 2004, Geim Novselov reported new semiconducting material graphene, which is a single layer of carbon in the geometry of honeycomb lattice (Novoselov, Kostya S., et al 2004). Graphene is known for very high electron mobility so the electrons move extremely fast from one side to the other side of graphene as currents. Since electrons also carry phonons, graphene is extremely good thermal conductivity. But it does not have a bandage as the difference between valance band and a conduction band. They cannot serve just like silicon directly.
Transition metal dichalcogenides, or more commonly known as TMDs, are another class of emerging two-dimensional materials that are similar to graphemes but have bandgaps depending on the number of layers. Single-layer TMDs have direct bandgaps where minimum points of valance band and maximum points of conduction band are sitting at same momentum values. Such direct bandgaps are exactly what we need for realizing transistors from two-dimensional materials.
TMDs work very well with graphene when stacked together as heterostructures, finding next-generation electronics and optoelectronics devices. TFT( Thin-film Transistor) for two-dimensional materials is similar to how FET (Field Effect Transistor) is the central device for virtually all users of semiconductor technology). High-performance two-dimensional TFT based on MoS2 as a type of TMD at room temperature has been achieved, featuring the characteristic high on/off current rain and current saturation expected from high equality TMDs. Combined with saturation velocity and mechanical strength, TMDs are very attractive for low- power TFT which is exactly what we need on board a 1-gram-level spacecraft.
TMDs also are gaining attention as electrode materials for energy storage such as supercapacitors and Li-ion batteries. The battery design is one of the most challenging aspects of any interstellar mission. The Breakthrough Starshot project considers radioisotope materials such as Plutonium-239 or Americium-241 for storage. When the device interacts with the interstellar medium, the battery would be charged by the heat supply of the frontal surface of the nano-craft. Due to their atomically layered structure, high surface area, and excellent electrochemical properties, TMD layered structures provide more sites for irons in energy storage while maintaining structural stability during charge and discharge cycles. In fact, the graphene has a surface area density of 2630m²/g which is the highest among carbon materials.
Light Sail
A traditional solar cell is a photovoltaic device with a bandage that covers most of the sunlight spectrum. However, since we are using a laser power source, we need a sail that is optimized for a single frequency light source. Also, since the plan was to focus a large laser array on a very small array, the flux can be as high as 100MW/m² which is about 10⁵ Sun intensity. So the approach to design such a high-efficiency light single frequency sail should be completely different. Also, high reflectivity is highly desired for the purpose of photon thruster.
Traditionally, people use multilayer dielectric on the metabolized glass to achieve reflections of 99.999%. This is almost an industry standard. We tune the reflectivity curve to have a maximum exactly at the laser wavelength. Under the extreme flux of the interstellar probes, which use small reflectors, pure dielectric reflection coating on the ultra-thin glass is proposed. The metabolized substructure is not reflective enough so the thermal management becomes a
critical problem. Then if we remove the metals and use a fully dielectric reflector instead, we would avoid the heat buildup.
On the other hand, some researchers are using metamaterial monolayer with near-unity reflectivity and negligible absorptivity (Slovick et al. 2013). Unlike Bragg reflectors and photonic band gap materials, which require multiple layers for high reflection, this metamaterial is both highly reflective and subwavelength in thickness. They developed a 0.45-um- thick, silicon-based meta material monolayer with normal-incidence reflectivity over 99.999% (even though they did not mention the exact weight).
Direct Imaging
In the simulation software Universal Sandbox, we present three potentially interesting exoplanet systems to be investigated by our space probes. Proxima Bis by far our closest neighbor, and it is the most logical choice of our first target system. We should, at the same time, leave the opportunity open for newly observed systems such as Ross 128 which receives much less UV and x-ray radiation from its host star.
If we can make sure our probes travel within roughly 1AU distance and be able to rotate onboard cameras to point at the planet, capture an image that includes basic compositions or even surface features, and be able to compress the image to a data flow that can be sent back to earth by onboard laser array, that would be equivalent to have a 300km telescope on Earth, according to Breakthrough Starshot (which is obviously economically unfeasible).
If we send thousands of crafts at once from earth towards the system, there is a good chance that multiple crafts would be within 10–100 Au distance from the target planet, then we are enabled to do a lot more. Having multiple probes in the system that can individually sense radiation will confirm the location of the star and the planet. Using computer vision, the crafts would take a certain number of pictures and brightness analytics would help craft locate exactly where the planet is in real time through cross verification.
Computer vision algorithm based interferometry
Inspired by the efforts of the Event Horizon Telescope, where Computer vision algorithm is used to fill in the gaps of black hole telescope data across the globe, and approximate a 10,000-kilometer-wide antenna, we know that we do not have to rely on a single picture being captured and sent by a single craft. The Event Horizon Telescope is essentially a radio interferometry experiment. A radio signal reaches two telescopes located far apart on earth at slightly different times. Visual information can be extracted from the difference. But due to the atmosphere where the index of reflection is not only different from the vacuum but has a gradient vertically and generally not uniformly distributed, the two signals would both be delayed for sometimes that cannot be estimated well and throwing off the calculation. But if multiple measurements are taken, they can each form pair with another that can cancel
out the atmospheric noise. Also, given how sparse the camera’s field of view might be, we would need such an algorithm to assemble an image by blurs together bright points near each other, or simply restore continuity to the image.
In the next figure where flight parameters including data rates for a 1 gram wafer spacecraft and 1m sail. Craft achieves 0.2 light speed in 10 mins and takes 20 years to the target system. (Lubin, 2016). Communications rate are assumed to be a class 4 drive array with a 1-watt short burst. At 4 light years, which is 3.78e16m away, the data rate received at the Earth is about 0.65 k bits per second. At this rate, merely one thousand seconds are needed to send a reasonably sized picture.
Speaking about planets and their habitability themselves, the question of whether such a planet can sustain an atmosphere and liquid water on its surface is a matter of intense debate (Anglada-Escude, et al. 2016). The most common arguments against habitability are tidal locking, strong stellar magnetic fields, strong flares, and high ultraviolet and X-ray fluxes. But none of this evidence has been proven to be definitive.
Global atmospheric circulation can still allow a tidally locked planet to sustain the atmosphere. Even though Proxima b suffers from X-radiation fluxes that are about 400 times what reaches on the Earth, that that would not be a deal-breaker. Direct imaging would, and answer most of these questions in the most definitive manner.
Direct Imaging based transit spectroscopy
The first studies of exoplanetary atmospheres were made with the transit spectroscopy method conducted on planets that transit their host star (D. Charbonneau et al. 2002). But since only planets that are close to their host star would leave stronger transit signals, planets under transit spectroscopy studies tend to fall in the category of planets that suffer from too much heat or radiation from their host stars. In addition, this method can sample a rather limited number of light wavelengths due to low spectral resolution. If we could target the exoplanet directly with relatively high- resolution spectrographs, this would avoid confusion of the exoplanet with the host star when we compare observations made during planetary transit with observations made when the planet is out of transit.
But due to very small angular separations of exoplanets and the extreme difference in planet and star brightness, this what we would call ‘direct imaging’ technique for exoplanet atmosphere studies has been quite complex. However, if any crafts can be within the proximities of the planets, the angular separations would be extremely distinct. We could see the noticeable brightness difference between the planet and the host stars. As a result, if we are proven to be overconfident in the communication data
rates, we can do in-situ studies of transit spectroscopy using direct imaging techniques. Note that this would require that we carry additional adaptive optics systems on board and sufficient processing power to separate the light from a planet and its host star at near-IR wavelengths.
Inspired by this approach, we could also carry single frequency analytics onboard. This would allow species-specific analytics such as iron and potassium which have their characteristic peak frequency. If our pixel density is high enough, we could even do spatial- or temporal-dependent studies of the distribution of such species on the target planet. Again, we would highly desire the presence of multiple cameras from different locations at the same time in the target system, such that we would be able to confirm and cross-verify our conclusions.
References
Anglada-Escudé, Guillem, et al. “A terrestrial planet candidate in a temperate orbit around Proxima Centauri.” Nature536.7617 (2016): 437–440.
Benedict, G. Fritz, et al. “Interferometric Astrometry of Proxima Centauri and Barnard’s Star Using HUBBLE SPACE TELESCOPE Fine Guidance Sensor 3: Detection Limits for Substellar Companions.” The Astronomical Journal 118.2 (1999): 1086.
Bhimanapati, Ganesh R., et al. “Recent advances in two-dimensional materials beyond graphene.” (2015).
Charbonneau, David, et al. “Detection of an extrasolar planet atmosphere.” The Astrophysical Journal 568.1 (2002): 377.
Demory, Brice-Olivier, et al. “Hubble Space Telescope search for the transit of the Earth-mass exoplanet α Centauri B b.” Monthly Notices of the Royal Astronomical Society 450.2 (2015): 2043–2051.
Doeleman, Sheperd, et al. “Imaging an event horizon: submm-VLBI of a super massive black hole.” arXiv preprint arXiv:0906.3899 (2009).
Lubin, Philip. “A roadmap to interstellar flight.” arXiv preprint arXiv:1604.01356 (2016).
Lu, Ru-Sen, et al. “Imaging the supermassive black hole shadow and jet base of M87 with the event horizon telescope.” The Astrophysical Journal 788.2 (2014): 120.
Moore, Gordon E. “Cramming more components onto integrated circuits.” Proceedings of the IEEE 86.1 (1998): 82–85.
Novoselov, K. S., Geim, A. K., Morozov, S. V., Jiang, D., Zhang, Y., Dubonos, S. V., … & Firsov, A. A. (2004). Electric field effect in atomically thin carbon films. science, 306(5696), 666–669.
Janson, Siegfried. “Mass-producible silicon spacecraft for 21st century missions.” Space Technology Conference and Exposition. 1999.
Kasper, Markus, et al. “EPICS: direct imaging of exoplanets with the E-ELT.” Ground-based and Airborne Instrumentation for Astronomy III. Vol. 7735. International Society for Optics and Photonics, 2010.
Shanmugam, Mariyappan, et al. “Two-dimensional layered semiconductor/graphene heterostructures for solar photovoltaic applications.” Nanoscale 6.21 (2014): 12682–12689.
Slovick, Brian, et al. “Perfect dielectric-metamaterial reflector.” Physical Review B 88.16 (2013): 165116.
Wang, Qing Hua, et al. “Electronics and optoelectronics of two-dimensional transition metal dichalcogenides.” Nature nanotechnology 7.11 (2012): 699–712.
Acknowledgments
This is individual work with no collaborators. I would like to give special thanks to Prof. Courtney Dressing for teaching this graduate Astronomy course during Fall 2017 at University of California, Berkeley, for her advice on adding scientific parts in addition to technical review of this paper. This paper will not be completed without her advice. I also thank the Astronomy Department and its staff for supporting the operation of the course.
