This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

Space Propulsion Technology for Small Spacecraft By David K rejci and Paulo Lozano

ABSTRACT | As small satellites become more popular and have proven the high ​Δv ​capability of EP systems for capable, strategies to provide in-space propulsion increase traditional missions with high electrical power available, in importance. Applications range from orbital changes and the recent trend toward miniaturization of spacecraft maintenance, attitude control and desaturation of reaction [11], [12] has resulted in tightened requirements for EP wheels to drag compensation and de-orbit at spacecraft end- systems in terms of mass, volume, and power consump- of-life. Space propulsion can be enabled by chemical or electric tion, while continuing to be competitive in terms of fuel means, each having different performance and scalability efficiency, which is directly related to the system’s capa- properties. The purpose of this review is to describe the working bility to transform electric energy into kinetic energy of principles of space propulsion technologies proposed so far for the exhaust. small spacecraft. Given the size, mass, power, and operational This paper explores the different propulsion prin- constraints of small satellites, not all types of propulsion can be ciples currently developed for small and miniatur- used and very few have seen actual implementation in space. ized spacecraft. While designations for different Emphasis is given in those strategies that have the potential of satellite classes have been somehow ambiguous, a miniaturization to be used in all classes of vehicles, down to the system-mass-based characterization approach will be popular 1-L, 1-kg CubeSats and smaller. used in this work, in which the term “small satellites” KEYWORDS | CubeSat propulsion; micropropulsion; sace will refer to satellites with total masses below 500 kg, propulsion with “nanosatellites” for systems ranging from 1 to 10 kg, “picosatellites” with masses between 0.1 and 1 kg, and “femtosatellites” for spacecraft below 0.1 kg. In this cat- I. INTRODUCTION egory, the popular CubeSat standard [13] will, therefore, Propulsion capability for satellites has been a priority for be characterized as nanosatellite, whereas Chipsats, satellite developers since the early stages of spaceflight Wafersats, and membrane-syle satellites [14], [15] will be to increase spacecraft capabilities [1]–[5], with the first categorized as picosatellite or femtosatellite, depending instances of electric propulsion (EP) occurring in 1964 on their mass and configuration. onboard of the Russian Zond-2 carrying a Pulsed Plasma Different satellite classes result in different power thruster [6] and the American SERT satellites with the budgets available, as in a first approximation, available first operation of an ion engine [7]. Significant other devel- area for solar panels, as well as battery sizes decrease for opment highlights include the development of the Shell decreasing satellite mass. Traditional mass-based power 405 catalyst allowing the utilization of high performance estimation approximations assume specific power per sub- hydrazine-based chemical propulsion systems [8] and the system mass of 25–200 W/kg [16]. Assuming a mass frac- SMART-1 mission showing the capability of an electric Hall tion of 10% for the power subsystem, the available power thruster in an Earth-moon transfer [9], as well as the inter- for small satellites ranges from 1.25–10 kW to 25–200 W, planetary Deep Space-1 mission, which was propelled by an whereas available power on nanosatellites would range NSTAR ion engine [10]. With the ABS-3A and Eutelsat 115 from 25–100 down to 2.5–20 W and power levels for West B, we have seen the first instances of commercial plat- picosatellites and femtosatellites reduced by one and two forms to use EP for orbit raising from a low earth orbit to orders of magnitude, respectively, compared to nanosatel- the final geostationary orbit. However, while these missions lites. However, densification and fast integration of high- efficiency components have shown that these assumptions Manuscript received October 2, 2017; revised November 14, 2017; accepted November 20, 2017. (Corresponding author: David Krejci.) can significantly underestimate the performance of cur- The authors are with the Space Propulsion Laboratory, Department of Aeronautics and rent nanosatellites, especiallcy CubeSats. For example, Astronautics, Massachusetts Institute of Technology, Cambridge, MA 02139 USA (e-mail: [email protected]; [email protected]). current state-of-the-art 3U CubeSats can achieve 50–60 W of total Beginning Of Life (BOL) power when using Digital Object Identifier: 10.1109/JPROC.2017.2778747 deployable solar sails [17]. The available power level can

0018-9219 © 2018 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information. Proceedings of the IEEE 1 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

Krejci and Lozano: Space Propulsion Technology for Small Spacecraft

have significant impact on the propulsive capabilities of a sat- launch vehicle can allow to evenly spread orbits and, ellite platform in the case of EP, both on the choice of thruster therefore, revisit from a shared launch. Depending on principle as well as the resulting propulsive performance. mission architecture and time to deploy the constella- Propulsion is an enabling capability for spacecraft serv- tion, the required Δ​ v​ can range from 1 to hundreds of ing a variety of purposes to enable different applications, meters per second. including the following. • Interplanetary missions: Propulsion may be used for primary orbit transfer and control and adjustment as • Change of orbit altitude, orbit corrections: The pri- well as attitude control, and as means to desaturate mary purpose of propulsive capabilities on spacecraft reaction wheels in the absence of a planetary magnetic is to alter the orbit by changing orbital elements such field. For primary orbit transfer, required Δ​ v​ depends as altitude or inclination. This can include changes in heavily on the initial launch orbit. For autonomous orbital altitude and corrections to maintain an orbit by propulsion to an interplanetary body from a GEO counteracting perturbations. Low earth orbit raising orbit, the necessary Δ​ v ​to reach earth escape velocity maneuvers can range from 50 m/s to 1.5 km/s [18]. is ~1.3 km/s, whereas attitude control is in the order • Life extension by drag compensation: Miniaturized of 1–10 m/s per year. satellites are often deployed in orbit altitudes that guarantee natural decay within a certain timespan Small satellites could be orders of magnitude less costly to comply with orbital debris guidelines, or to lever- than their larger counterparts, with development times in age relatively cheap launch opportunities, such as the one to two year range, which are very short compared resupply missions to the International Space Station. to traditional project cycles, in some instances reaching a During the active phase of the mission, propulsion can decade or more. This represents a dramatic change in the be used to extend the lifetime of the spacecraft, while way satellites can be used, from utility platforms provid- relying on natural orbit decay caused by the rarefied ing traditional services to systems spurring innovation that atmosphere at the end of mission. While Δ​ v​ require- eventually will lead to vigorous scientific and economic ments for drag makeup depend strongly on orbit alti- development. Before that happens, however, the challenge tude and satellite cross section and mass, typical values of high capable and compact propulsion systems needs to be are in the tens to hundreds of meters per second. surmounted. For example, having access to high Δ​ v​ propul- • Deorbiting: Propulsive maneuvers can be used to sion in small satellites would allow them to perform rapid, decrease the obit altitude in low earth orbits, after the frequent, and affordable exploration of a myriad of objects end of a mission to facilitate induced or natural orbit of interest that are relatively close to Earth. This exploration decay. Depending on the initial orbit, typical controlled would increase our understanding of our planetary neigh- reentry ​Δv​ values range from 120 to 150 m/s [18]. borhood, while also could bring new areas of industrial • Formation flight: Much of the attractiveness of min- development. Besides exploration, other areas will benefit iaturized space systems comes from the potential to from the inclusion of propulsion in small satellites, from facilitate the deployment of constellations, for a vari- imaging and communications, to astronomy and fundamen- ety of purposes. Constellations bring benefits, such tal physics. as an increase of revisit times, and the introduction The different nature of these applications results in sig- of redundant or distributed architectures. Spacecraft nificantly differing requirements imposed onto the propul- in accurately controlled formation can also enable sion systems, with a general tradeoff between thrust and advanced instruments, as in the LISA mission for grav- specific impulse as detailed in Section II. itational wave detection [19] and distributed apertures to increase resolution such as the proposed terrestrial II. PROPULSION PRINCIPLES planet finder [20]. In most cases, such applications The ideal change in velocity of a spacecraft ​ ​is described necessitate to maintain the relative spacecraft sepa- Δv by the Tsiolkovsky rocket equation [22] as a function of the ration, requiring the capability for propulsive orbit spacecraft mass before m​​ ​​ and after m​​ ​​ a propulsive maneuver correction, with required Δ​ v​ ranging from 1 m/s to i f m​ ​​ hundreds of meters per second, depending on orbital ​Δv = ​I​ ​ ​g​ ​ ln​ ​ ___i ​ ​ (1) mission parameters. sp 0 (m​ f​​) • Constellation deployment: Ridesharing and cheap where I​​ ​sp​ = ​ue​​ / ​g0​ ​​​ is the specific impulse, which is defined development of larger numbers of miniaturized as the ratio of the propellant exhaust velocity relative to the satellites make constellations of such systems very propulsion system u​ e​​​​ and the standard acceleration due to 2 attractive. Constellation deployment from a shared gravity at Earth’s sea level ​g0​ ​​ ~​ 9.81 m/s​​​​​ ​​. The thrust F​ ​ gen- launch allows to minimize launch cost [21], obviating erated by a propulsion system is given by dedicated launches in different orbits. Even small Δ​ v​ delivered by the satellite after delivery from a shared ​F = ​m ̇ u ​e​ + ​A​e​ ​pe​​ ≈ ​ṁ u e​​​​ (2)

2 Proceedings of the IEEE This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

Krejci and Lozano: Space Propulsion Technology for Small Spacecraft

where m​​ ̇ is the exhaust mass flow,​A e​​​​ is the nozzle exit area, and p​​ e​​​​ is the exhaust flow pressure at the nozzle exit plane, assuming plume expansion to vacuum. The efficiency of a propulsion system can be defined as the fraction of the total source power that is transformed into kinetic power of the exhaust, often called jet power

__​ 1 ​​ ​ 2​ ​ 2ṁue ​η _____ ( 3 ) ≈ P where ​P​ is the total input power, either released from energy stored in the chemical bonds of the propellant, or supplied by an external power source in EP.

A. Chemical Propulsion In chemical propulsion systems, thrust is generated by acceleration of a compressed working fluid by expansion to a low density exhaust stream with increased kinetic energy, typically using a converging–diverging nozzle geometry. Increasing the pressure and temperature of the working medium before expansion increases the resulting kinetic energy of the exhaust, and therefore the achieved specific impulse. Available systems are typically classified according Fig. 1. Some chemical propulsion technologies. (a) Cold Gas. to the principle of energy release in the working medium (b) Warm Gas. (c) Monopropellant. (d) Bipropellant. (e) Solid propellant. (f) Hybrid propellant. before acceleration. 1) Cold and Warm Gas Propulsion: In these basic sys- Using a multiple staged valve architecture, typically with tems, a high pressure working gas is expanded through a an intermittent smaller volume reservoir, between main converging–diverging nozzle to create thrust. This principle valve and thruster selector valve, allows to decrease the risk is limited by the storage pressure achievable in the tank sys- of valve failure leading to failures that can prove critical to tem, typically limited by structural considerations, and often the satellite mission. requires a separate pressure regulator to avoid thrust and Typical specific impulse range from ~10 to ~80 s, and specific impulse decay as the storage pressure decays over thrust levels ​~100​ ​μN​ to ~100 mN. the mission duration as indicated in Fig. 1(a). Typical pro- pellants used are isobutane (C​​4​ ​H10​​​ ​​​), the refrigerants R236fa 2) Monopropellant and Advanced Monopropellant Thruster: and R134a, and sulfur dioxide (SO​​​2​​​) [23]. In warm gas pro- In monopropellant thrusters [Fig. 1(c)], a high energetic pro- pulsion systems, the working fluid is additionally heated pellant is typically decomposed catalytically or thermally before expansion [Fig. 1(b)], increasing the internal energy into a high temperature working gas, before it is expanded of the working medium using an external source, result- through a nozzle to a low temperature and density exhaust ing in an increase in specific impulse of the system when stream with elevated exhaust velocity. This concept requires compared to cold gas systems; see Section II-C1. Hybrid storable, and decomposable propellants, and commonly systems make use of phase changes, or storing a propellant, used fluids include hydrazine (N​2​​​H4​​​ ​) and derivatives, CO​​​2​ or nitrous oxide (N2​​​ ​​​O), in supercritical condition [24]. highly concentrated hydrogen peroxide (H​2​O 2​​​​), and N2​​​ ​O . Although such systems require active components sepa- In these designs, propellant selection is typically governing rating the thruster from the high pressure propellant stor- the design and nature of the decomposition chamber, with age, the general simplicity of the system has allowed a high N​​​2​​​H4​​​ ​​​ requiring preheating of the catalyst bed. Decomposi- degree of miniaturization, especially with the emergence of tion temperatures of monopropellants are, in general, low microelectromechanical system (MEMS)-based valve and enough to allow for radiation cooling without the need for other components. With generally low specific impulse of exotic materials for the chamber designs. Design trade- such systems, the total achievable ​Δv​ is typically limited for offs are usually made between the toxicity and storability miniaturized satellite systems in which space to accommo- of the propellant, which can have a major impact on mis- date large propellant tanks is not available. However, the sion cost. N​2​​​H4​​​ ​​​, which is highly toxic, reactive, and has ability to achieve small impulse bits and a large number of high vapor pressure features provides high performance, impulse cycles make these systems attractive for small and reliability, and long catalyst lifetime compared to available precise positioning and attitude control maneuvers. propellants. However, to comply with reducing mission cost

Proceedings of the IEEE 3 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

Krejci and Lozano: Space Propulsion Technology for Small Spacecraft

typically strived for in small satellite missions, lower toxic- the combustion chamber requires high temperature metals, ity options such as hydrogen-peroxide-based propulsion sys- often refractory metals such as rhodium or platinum alloys, tems have been recently developed [25]–[27]. As the density and can lead to radiative energy transfer from the high- of the stored propellant determines the necessary tank vol- temperature combustion chamber to the satellite structure ume, the density of the propellant in stored condition can which needs to be considered in design. Bipropellant sys- be important in miniaturized designs, leading to increase tems are, therefore, typically used in missions with higher​ volumetric specific impulse of high density propellants such Δv​ requirements such as apogee orbit insertions or maneu- as H​​​2​O 2​​​ ​​​ when compared to liquids with lower density, such vers involving significant orbit changes. as N​​​2​​​H4​​​ ​​​. Another concern specifically for small satellites Typical specific impulses are ~300 s with thrust levels can arise from the fact that high energetic, and potentially usually starting from 10 N. unstable propellant can negatively impact launch opportu- nities for secondary payloads due to safety concerns. B. Solid and Hybrid Propulsion Typical specific impulses range from ~100 to ~230 s and thrust levels range from ~100 mN up to hundreds of By combusting a solid propellant, solid propulsion sys- newtons. tems provide a hot working gas that is then expanded to Originally motivated by the search for less toxic propel- produce thrust [Fig. 1(e)]. Solid propulsion systems can be lant alternatives to hydrazine-based derivatives, monopro- designed without complexity of moving actuators, but gener- pellant systems based on ammonium dinitramide (ADN) ally lack restarting capability and precise controllability, and have been considered as end-of-life deorbiting devices [34]. or hydroxylammonium nitrate (HAN) have been developed Restarting capability and increasing performance are inves- with considerable effort [28]. In such systems, the propel- tigated by introducing hybrid configurations where part of lant is typically thermally and catalytically decomposed, the reactants is fed into the solid combustion chamber using with significantly higher performance than traditional a fluidic system as depicted in Fig. 1(f). While increasing the monopropellants. Such systems are, therefore, similar utility of such propellant systems, this typically comes with to monopropellant systems in terms of necessary fluidic added system complexity and is, therefore, currently consid- components, comprising a single propellant tank and feed ered of limited use for miniaturized space systems. stream, but they provide energy release and, therefore, per- Typical specific impulses for miniaturized solid motors formance similar to traditional bipropellant systems. An range from ~150 to ~280 s, with thrust levels ranging from ADN-based propulsion system has already been success- tens to hundreds of newtons. fully tested in space [29], and the Green propellant infusion mission (GPIM) [30], [31] is currently awaiting launch to increase the maturity of a HAN-based propulsion system by C. Electric Propulsion in-orbit validation. EP systems differ from chemical propulsion devices by Typical specific impulses for such systems are between the means of energy supply for exhaust acceleration. While ~200 and ~250 s, and thrust levels range from hundreds of in chemical propulsion systems the energy is generally millinewtons to tens of newtons. stored within the molecular bonds of the propellants and is 3) Bipropellant Propulsion: In bipropellant systems as released by combustion, decomposition, or expansion, EP shown in Fig. 1(d), combustion of an oxidizer and a fuel systems use an external energy source, supplying electri- are utilized to create a high-temperature, high-pressure cal power that is used to accelerate the exhaust. While any gaseous mixture that can be expanded using a converging– source supplying electrical power such as nuclear reactors diverging nozzle to create a high velocity exhaust stream. or radioisotope thermoelectric generator (RTD) can be Such systems typically show highest performance in terms used, in most cases, especially in small satellites, electrical of specific impulse, but also come with most complex- power is supplied by solar panels. ity due to typically two independent fluidic feed systems 1) Electrothermal Acceleration: Historically, resistojets including two separate tanks and valve sets. Typically were the first instance in which an external energy source used, storable, noncryogenic propellant combinations are was used to electrothermally augment traditional chemical monomethylhydrazine (MMH) or unsymmetrical dimeth- rockets and was originally applied to N2​​​ ​H4​​​ ​​​. In such devices, ylhydrazine (UDMH) with oxidizers such as dinitrogen shown in Fig. 2(a), the propellant gas is typically heated by tetroxide (N​​​2​O 4​​​ ​​​), MON-1 or MON-3, or less toxic combi- an electrically heated surface in various configurations [35] nations such as H2​​​​O 2​​​ ​kerosine, or H​2​​ ​O 2​​​ ​​​methane [32], [33]. to increase the propellant gas beyond the stagnation temper- Bipropellant systems feature the highest performance for ature of the purely chemical propulsion system, and there- chemical systems per stored propellant, but require gener- fore augment the resulting exhaust velocity after expansion. ally complex propellant managements system with multiple Resistojets have been employed over a wide range of power active components. Due to the higher flame temperature levels and propellants, with gas-heater heat transfer usu- compared to monopropellant thrusters, radiative cooling of ally being the premier design challenge. While traditional

4 Proceedings of the IEEE This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

Krejci and Lozano: Space Propulsion Technology for Small Spacecraft

SERT I spacecraft in 1964 [7]. Ionization efficiency in such devices is a function of propellant and electron current as well as residence time of the injected gaseous propellant in the ionization chamber, limiting the miniaturization in terms of volume. Miniaturization of a durable cathode for electron supply presents a challenge with respect to minia- turization, in addition to the ability to manufacture flat grids out of materials that show high resistivity to ion erosion while space charge between the accelerating grids limits the achievable emission current density and therefore thrust. Recent developments in miniaturized ion engines make use of RF ioniziation, obviating the need for an internal elec- tron emitter [36]–[38]. Ion engines have been employed in notable missions such as Deep Space-1 [10], Dawn [39], and GOCE [40], and typically achieve specific impulse between​ Isp​ ​ =​ 2000–3000 s. Hall thrusters present a second form of highly devel- oped electrostatic accelerator systems that was flown Fig. 2. Electric propulsion configurations. (a) Resistojet. (b) Arcjet. among others on the SMART-1 moon transfer mission [9], (c) Pulsed Plasma Thruster. (d) Electrospray/FEEP. (e) Ion thruster. (f) Hall thruster. and has become attractive due to the absence of any accel- eration grids which are typically life limiting components due to erosion and electrical breakdowns. Hall thrusters, resistojet designs have not been found in small satellites so on the other hand, feature an annular ionization and accel- far, warm gas thrusters in which the entire propellant, or eration chamber in which a neutral gas is injected near an a small subset of the propellant is heated in a separate res- upstream anode through a manifold, and is subsequently ervoir before injection into the acceleration chamber, can ionized by electrons that are injected downstream, as be seen as a somehow related thruster type that has been shown in Fig. 2(f). These electrons are attracted to a posi- developed with special focus on small satellite applications. tive upstream anode, while strong magnetic fields perpen- Arcjets [Fig. 2(b)] on the other hand augment the pro- dicular to the gas flow force the electrons on a precessing pellant gas temperature by creating a steady arc discharge, path along the annular chamber, increasing the residence typically through the nozzle throat from a cathode inside the time and, therefore, probability of collision with the combustion chamber that is annually surrounded by propel- injected neutral gas. The ions, due to their higher molecu- lant gas flow. Heating the gas by means of a discharge allows lar mass, feel a proportionally weaker acceleration by the to surpass the maximum working temperature of conven- magnetic force, and are primarily accelerated electrostati- tional heating elements, that typically limit the performance cally by the potential drop between the anode and a cath- of resistojets. While providing higher performance than tra- ode situated at the exit of the otherwise insulating annular ditional monopropellant and resistojet propulsion systems, chamber. While specific impulses achieve by Hall thrusters the added complexity of arcjets necessitating a fluidic and a are generally lower than in ion engines around I​​​sp​ ~ ​ 1500 s, power-intensive electrical subsystem have prevented usage they feature higher thrust densities (up to an order of mag- of such propulsion systems in small satellite applications to nitude higher than in ion engines) and can be adapted over this point. a wide range of power levels. 2) Electrostatic Acceleration: Electrostatic space pro- Colloid-electrospray thrusters and field emission elec- pulsion devices accelerate charged particles, mostly ions, trostatic propulsion (FEEP) [Fig. 2(d)] are propulsion by electrical forces when falling through a potential drop systems with similar acceleration principles, but differ in across two electrodes and are the most evolved EP concepts terms of production of charged particles. In FEEP thrust- originating back to the 1950s. ers, a liquefied metal propellant, such as indium, gallium, Ion engines, also know as Kaufman ion engines or elec- or cesium, is suspended over a sharp emitter structure to trostatic thrusters, shown in Fig. 2(e), accelerate ions that increase the localized electric field at the apex. Balancing are produced in an ionization chamber using a potential the electrostatic pull and the surface tension, the conductive drop between an extractor and an acceleration grid. Ion liquid metal deforms into a Taylor cone [41], which further production can be accomplished by collision with injected increases the local field strength at the apex, where the ioni- electrons and radio-frequency (RF), microwave, or contact zation threshold can be surpassed, leading to the ejection of ionizations. The expelled ion beam is then typically neu- ions. These ions are then accelerated by the same electric tralized by an external electron emitter. The first gridded field used for extraction. In electrospray thrusters, ionic liq- ion thruster has been tested on a suborbital flight on the uids or electrolytes are used as propellants, with the latter

Proceedings of the IEEE 5 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

Krejci and Lozano: Space Propulsion Technology for Small Spacecraft

tending to produce significant ratios of slow moving drop- material, and have been specifically developed for low lets, whereas ionic liquid electrosprays are able to operate power nanosatellites [45], [46]. in pure ionic mode [42]. While electric conductivities, and therefore ultimately emission current per emission site, of D. Propellantless Propulsion such propellants are generally lower when compared to liq- uid metals ion sources, they do not require energy to liquefy In addition to traditional propulsion system that oper- the propellant by heating and, therefore, have the potential ate by accelerating neutral or charged particles to produce for higher system efficiency, assuming similar beam proper- thrust, a variety of propulsion concepts utilizing different ties. In addition, no ionization is required due to the nature physical mechanisms have been proposed and tested. of the ionic liquid propellant since molecular ions are read- 1) Solar Sail Propulsion: It uses low density solar radia- ily extracted from the propellant bulk [43]. In addition, tion pressure to generate thrust, by typically employing ionic liquid electrospray thrusters are capable of producing lightweight, expandable structures to obtain very high charged particles of both polarities, obviating the need for surface-to-mass ratio, contracting the low forces due to low an external neutralization device as required in other EP radiation pressure. This concept has been notably tested by systems [44]. Due to the absence (in case of electrosprays) the Venus bound IKAROS spacecraft [47] and the CubeSat or locally confined (FEEP) ionization process, both FEEP LightSail-1 [48]. In addition to radiative pressure propul- and electrospray systems feature uniquely small ionization/ sion, these expandable structures have been used as end-of- acceleration chambers and lend themselves favorably life deorbit devices to lower a low earth orbiting spacecraft toward miniaturization. by significantly increasing the atmospheric drag, as proven 3) Electromagentic Acceleration: While in electrostatic in the NanoSail-D mission [49]. propulsion systems, the force that is acting on the accelerated Electrodynamic tethers are long, conductive wires or charged particle is reciprocally felt by the accelerating elec- structures deployed in orbit in the presence of a planetary 2 trode, the definition of electrostatic pressure T​ / A = 1 / 2 ​ϵ0​ ​ ​E​​ ​​ magnetic field and ionosphere, producing a force by the with ​T / A​ the thrust density, ϵ​​ 0​ ​ the vacuum permittivity, and ​ interaction of the external magnetic field and the charges E​ the accelerating electrical field strength, imposes a limit to moving along the tether which, for propulsion purposes, is the achievable thrust per area in real devices where ​E ​ cannot held at a different potential with respect to the spacecraft typically be increased indefinitely. In electromagnetic devices using a power supply, thus adding kinetic energy to the orbit. however, the force transfer from the accelerated particle beam To remove kinetic energy, for example, to lower the orbit to the structure occurs via magnetic fields in such devices, altitude, the ends of the tether can be electrically shortened. allowing for higher thrust densities. Motion of the spacecraft with respect to the magnetic field Magnetic plasma dynamic (MPD) thrusters are high will induce a voltage difference, drawing a current across power propulsion systems with self-induced magnetic the tether, which interacts with the external magnetic field, fields and operate at power levels incompatible with small effectively reducing the orbital velocity. An extreme form satellite technology, but would be interesting for larger of high specific impulse propulsion is photonic propulsion, spacecraft due to their ability to reach multi newton which utilizes the photonic pressure to produce thrust. thrust levels. While thrust levels are generally too low for power levels Pulsed plasma thrusters (PPTs) and vacuum arc thrust- achievable in small satellites with respect to traditional solar ers (VATs), on the other hand, are a type of propulsion system bound missions, this concept has attracted interest system that is operated in an unsteady regime, in which a either by using external high power laser beams to propel solid propellant, typically Teflon, is ablated by an induced nanosatellites [50], or in the context of small propulsion discharge, and acceleration of ablated material occurs by maneuvers in very long mission durations, such as attitude the Lorentz force. As the magnetic fields in the accelera- corrections in interplanetary femtosatellites [50]. tion electrodes interact with the magnetic field induced in Further concepts of propellantless propulsion include the perpendicular discharge, the ablated material is accel- magnetic sail propulsion, in which a loop structure erated perpendicular to the discharge surface as indicated induces a magnetic field that interacts with solar wind, in Fig. 2(c). The pulsed operation, in which a capacitor is leading to an exchange of momentum to the magnetic slowly charged and then rapidly discharged, suits itself well sail structure [51], and electric sail propulsion, in which a to small power budgets, and operation frequency can easily large conductive mesh is kept at a positive potential rela- be adjusted to available power. With adaptability to available tive to a solar wind plasma, blocking positive particles, power budget, generally small impulse bits and mechanical therefore leading to a theoretical exchange of momentum simplicity due to the solid propellant trade favorable for toward the spacecraft [52]. Due to the small order of mag- PPTs, they generally feature very low efficiency <10% [6]. nitude of such interactions which necessitates large-scale VATs are similar to PPTs in terms of mechanical design, structures, such concepts have not yet evolved beyond but initiate a lower power discharge that ablates anode conceptual studies.

6 Proceedings of the IEEE This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

Krejci and Lozano: Space Propulsion Technology for Small Spacecraft

III. PERFORMANCE METRICS OF terms of thrust and specific impulse. For space qualified SMALL SATELLITE PROPULSION systems, the satellite name is included in brackets. While TECHNOLOGY this plot is intended to provide a general classification of the A plethora of propulsion systems suited for small satel- different propulsion solutions, care should be taken with the lites and beyond has been developed and researched in the individual performance metrics of the propulsion systems, recent years, with multiple different propulsion principles as these rely on data provided by suppliers, and are deter- matured through inspace testing or ready for commerciali- mined with a varying degree of accuracy and detail. In this zation, from both chemical propulsion systems and EP sys- plot, the data shown for each propulsion system correspond tems. While chemical propulsion systems used as primary to the nominal operational point provided by the suppliers. propulsion means on small satellites showed some overlap It should, however, be noted that especially EP systems are with existing technologies used for station keeping on larger capable of throttling over significant ranges, as discussed space platforms, miniaturization of propulsion devices later. Propellantless systems are not included in this plot to smaller platforms such as nanosatellites and beyond as a specific impulse becomes a meaningless metrics, and required significant effort to achieve miniaturization with- bipropellant engines are omitted due to their limited util- out significant decrease in performance, and led to a variety ity for miniaturized satellites. A notable exception is the of new developments altogether. 30 HYDROS propulsion system [66], which is based on the Fig. 3 shows a classification of the different propulsion combustion of H​​ ​2​ and ​​O​2​​​ provided by electrolysis of stored systems for small satellites discussed in Section II, supple- water, and does not require two separate propellant streams mented with data from commercially existing systems in as do traditional bipropellant systems. While there is significant overlap for electrical propul- sion systems based on different acceleration principles, sev- eral trends can be clearly identified. • Cold and warm gas propulsion systems occupy the lower specific impulse range, but are available over a large range of thrust. As expected, cold gas propul- sion systems are found on the lower bound in terms of specific impulse, while heated systems are capa- ble of increasing the specific impulse to ​~ <​ 100 s. The ability to generate large numbers of small impulse bits makes these systems attractive for small orbit and attitude correction applications. • Chemical propulsion systems based on decomposition of an energetic compound, such as HAN-, ADN-, or hydrazine-based systems are found on the high thrust region of the plot, with specific impulse significantly higher than cold and warm gas systems, but lower than the bulk of EP systems. • EP systems feature increased specific impulse, and Fig. 3. Commercial micropropulsion systems: Thrust versus specific range from thrust levels comparable to cold gas sys- impulse at nominal operation point. 1ÐPPT, Austrian Institute of tems to micronewtons levels. Inside this group, Technology [53]; 2ÐPPTCUP, Mars Space [54]; 3ÐBmP-220 Busek the highest performing systems in terms of specific (Falcon-Sat 3); 4Ðμ​ ​CAT, George Washington University (BricSat-P) impulse are FEEP thrusters, which generally show a [55]; 5ÐVAT, University of Illinois; 6ÐRFT, Phase Four [56]; 7ÐBHT- 200, Busek; 8ÐBET-1mN, Busek; 9ÐBET-100 Busek; 10ÐiEPS, wide range of throttleability. A variety of commer- Massachusetts Institute of Technology (Aerocube-8) [57]; 11ÐMiXi, cially available ion engines are available, with thrust Jet Propulsion Laboratory; 12ÐRIT-μ​ X​, Airbus; 13ÐBIT-1, Busek; levels comparable or higher than electrospray thrust- 14ÐBIT-3, Busek; 15ÐIFM Nano, Enpulsion/Fotec [58]; 16ÐNANOPS, ers, with lower specific impulse than FEEP thrusters. University of Toronto (CanX-2) [59]; 17ÐMEPSI, The Aerospace • Pulsed plasma thrusters are found on the lower thrust Corporation (MEPSI-3); 18ÐCold gas micropropulsion system, Micro and lower specific impulse region of the studied EP Space (POPSAT-HIP1) [60]; 19Ð​T​​ 3​ μ​PS, TNO (Delfi-n3Xt) [61], [62]; 20ÐCPOD MiPS, VACCO; 21ÐPUC, VACCO; 22ÐBevo-2 cold gas, systems. University of Texas [63]; 23ÐCubeProp, NanoSpace (STU-2B) • One miniature Hall thruster was considered, to be on [64]; 24ÐMPS-120, Aerojet Rocketdyne; 25ÐMPS-130, Aerojet a mature level compared to the other systems inves- Rocketdyne; 26ÐBGT-5X, Busek; 27ÐBGT-1X, Busek; 28ÐADN Micro tigated. Hall thrusters are capable of providing the Propulsion System, VACCO/ECAPS; 29ÐLMP-103S, ECPAS (PRISMA) [29], [65]; 30ÐHYDROS, Tethers Unlimited [66]; 31ÐCHIPS, CU highest thrust of all EP systems considered, but fea- Aerospace; 32ÐResistojet, Busek. Unless otherwise noted, data are ture lower specific impulse compared to the available taken from vendor data sheets or [23]. ion engines.

Proceedings of the IEEE 7 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

Krejci and Lozano: Space Propulsion Technology for Small Spacecraft

In order to compare propulsion systems, it is impor- storage density as a fundamental parameter is highlighted tant to consider, besides propellant mass, the physical for EP systems as well. parameters of the propellant during storage, most notably The first important conclusion evident from the data the propellant density. It is, therefore, useful to introduce shown Fig. 4 is the importance of considering the thermody- the concept of volumetric specific impulse, defined as namic state in which the propellant is stored under storage vol 3 I​​ ​ sp​ ​ = ​ρprop​ ​ I sp​ ​​​ given in [kg s/m​​​ ​​ ​], with ρ​​ prop​ ​ the propellant conditions, evident from the significantly lower volumetric density at storage conditions, allowing to draw conclusions specific impulse for the only system considered that features of the specific impulse achievable per required storage vol- propellant stored in gaseous state, compared to the systems ume. Fig. 4 gives a comparison of the volumetric specific that guarantee liquid propellant storage conditions. Except impulses for commercial chemical propulsion systems as a for the gaseous argon-based system, most cold gas systems function of nominal thrust. If no information on storage feature very similar propellant densities, therefore being pressure was available, storage under liquid conditions was more similar than when compared on specific impulse level, assumed, unless information indicating otherwise. For the except for the systems using S​ ​F​6​​​, which has densities com- TNO ​T3μ​PS it is noted that due to the gas generation prin- parable to HAN- and ADN-based chemical propellants. The ciple from a solid material, the volumetric specific impulse only higher performing chemical systems with lower den- has only limited significance, and therefore a generic sity compared to other systems are the N2​​​​H4​​​ ​-based MPS- propellant density was calculated based on gas generator 120 and the water-based HYDROS thruster. volume and propellant mass stored, conservatively attribut- The discussion so far has neglected to mention the power ing 12.5% of the volume to the propellant (assuming the consumption of EP systems, which, with increased minia- same density for propellant and gas generator). It should turization of the spacecraft itself, can act as a main factor be noted that volumetric specific impulse can be an impor- limiting usability. Fig. 5 plots the nominal thrust as a func- tant metric for EP systems as well, however in most cases, tion of the system input power for a variety of EP systems. the auxiliary system mass such as the mass of the power For pulsed thrusters, the nominal impulse bit is plotted. The processing unit can significantly alter the total system systems operational at lowest power levels are pulsed plasma mass, and therefore inhibits a systems comparison solely on thrusters and electrospray thruster systems, with the latter volumetric specific impulse. As the PPU masses vary con- achieving higher thrust for the same power consumption siderably for the EP systems considered, a comparison of compared to pulsed plasma thrusters, which corresponds EP systems based on volumetric specific impulse is not con- well to the generally lower total efficiencies found for pulsed sidered here. Nevertheless, the importance of propellant plasma thrusters. It should be noted that due to the adjusta- bility of the duty cycle, pulsed systems such as PPTs or VATs

Fig. 4. Commercial chemical micropropulsion systems: Thrust versus volumetric specific impulse at nominal operation point. Fig. 5. Commercial EP systems: Nominal thrust as a function of 16ÐNANOPS, University of Toronto (CanX-2) [59]; 17ÐMEPSI, The system power, with most systems showing significant throttling Aerospace Corporation (MEPSI-3); 18ÐCold gas micropropulsion capability not indicated. 1ÐPPT, Austrian Institute of Technology [53]; system, Micro Space (POPSAT-HIP1) [60]; 19Ð​T​​ 3​ μ​PS, TNO (Delfi- 2ÐPPTCUP, Mars Space [54]; 3ÐBmP-220 Busek (Falcon-Sat 3); n3Xt) [61], [62]; 20ÐCPOD MiPS, VACCO; 21ÐPUC, VACCO; 22ÐBevo-2 4Ðμ​ ​CAT, George Washington University (BricSat-P) [55]; 6ÐRFT, cold gas, University of Texas [63]; 23ÐCubeProp, NanoSpace (STU- Phase Four [56]; 7ÐBHT-200, Busek; 8ÐBET-1mN, Busek; 9ÐBET-100 2B), [64]; 24ÐMPS-120, Aerojet Rocketdyne; 25ÐMPS-130, Aerojet Busek; 10ÐiEPS, Massachusetts Institute of Technology (Aerocube-8) Rocketdyne; 26ÐBGT-5X, Busek; 27ÐBGT-1X, Busek; 28ÐADN Micro [57]; 11ÐMiXi, Jet Propulsion Laboratory; 12ÐRIT-​μX​, Airbus; 13ÐBIT-1, Propulsion System, VACCO/ECAPS; 29ÐLMP-103S, ECPAS (PRISMA) Busek; 14ÐBIT-3, Busek; 15ÐIFM Nano, Enpulsion/Fotec [58]. Unless [29], [65]; 30ÐHYDROS, Tethers Unlimited [66]. otherwise noted, data are taken from vendor data sheets.

8 Proceedings of the IEEE This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

Krejci and Lozano: Space Propulsion Technology for Small Spacecraft

can be operated at even smaller average power consump- tions, albeit at the cost of further decreasing the achieved impulse bit. Electrospray thrusters generally exhibit throt- tleability without significant impact on performance, and it is noted that they can be operated well below the indicated power and thrust levels. FEEPs using liquid metals show excellent throttleability in terms of thrust, but typically can- not be steadily operated below a certain power threshold required for heating the propellant. Ion engines, on the other hand, occupy the higher power spectrum, showing compara- ble or higher thrust levels than FEEPs or colloid thrusters, but cannot typically be operated below 10 W. Miniature Hall thrusters are found at the high power spectrum of this com- parison, with the potential for increased thrust levels. Fig. 7. Throttling range for sample EP thrusters showing The investigated EP systems are available over a wide dependency of specific impulse and thrust on input power. range of power levels, and it is, therefore, useful to look 3ÐBmP-220 Busek (Falcon-Sat 3); 4Ðμ​ ​CAT, George Washington at the relationship of specific impulse and the thrust-to- University (BricSat-P) [55]; 10ÐiEPS, Massachusetts Institute of Technology (Aerocube-8) [57]; 14ÐBIT-3 full system including power ratio. As the relationship between system power, neutralizer, Busek [67]; 15ÐIFM Nano full system including thrust, and specific impulse is typically not trivial for neutralizer, Enpulsion/Fotec [58], [68]. the individual systems, a comparison of the stated per- formance at the nominal operational point is plotted in as the Phase Four RFT thruster, can achieve similar thrust- Fig. 6. In this comparison of specific thrust per power to to-power ratios, they feature significantly reduced specific specific impulse, it is again evident that ion engines and impulse compared to other EP systems. FEEPs, providing highest specific impulse, trade in the Figs. 3, 5, and 6 compare the specific impulse of differ- middle in terms of thrust per input power, whereas elec- ent propulsion systems at the nominal operational point. trospray, colloid, and Hall thrusters feature the highest However, it should be noted that especially EP systems are thrust-to-power ratio. often capable to operate over significant throttling ranges. Comparison of Figs. 5 and 6 also shows that while thrust- Fig. 7 shows the throttling capability of selected EP systems, ers with predominantly electrothermal acceleration, such plotting the thrust and specific impulse as a function of input system power. The systems compared show the MIT iEPS sys- tem, which features constant specific impulse over the thrust and power range tested [57], the University of Washington VAT thruster that increases delivered impulse bit by increas- ing the pulsing frequency, the Busek BeP-220 Pulsed plasma thruster with similar operational properties compared to the VAT, the Busek BIT-3 ion thruster with linear specific impulse and thrust increase with increasing power [37], [67], and the Enpulsion IFM Nano thruster with variable thrust/ power tradeoff for increasing specific impulse [58].

IV. FLIGHT EXPERIENCE WITH SMALL SATELLITE PROPULSION SYSTEMS A variety of propulsion systems have been flown on small satellites and smaller platforms, and many nanosatellite missions are currently in preparation that will be using some type of propulsion. In this section, some notable flight Fig. 6. Commercial EP systems: Thrust to power versus specific tests of propulsion systems will be discussed. impulse. 1ÐPPT, Austrian Institute of Technology [53]; 2ÐPPTCUP, Mars Space [54]; 3ÐBmP-220 Busek (Falcon-Sat 3); 4Ðμ​ ​CAT, George Washington University (BricSat-P) [55]; 6ÐRFT, Phase Four [56]; A. Small Satellite Case: ADN- and HAN-Based 7ÐBHT-200, Busek; 8ÐBET-1mN, Busek; 9ÐBET-100 Busek; 10ÐiEPS, Advanced Monopropellants Massachusetts Institute of Technology (Aerocube-8) [57]; 11ÐMiXi, While commercial propulsion systems for small satel- Jet Propulsion Laboratory; 12ÐRIT-μ​ X​, Airbus; 13ÐBIT-1, Busek; 14ÐBIT-3, Busek; 15ÐIFM Nano, Enpulsion/Fotec [58]. Unless lites, both chemical and electrical, are readily available otherwise noted, data are taken from vendor data sheets. and have been used in space, two missions are highlighted

Proceedings of the IEEE 9 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

Krejci and Lozano: Space Propulsion Technology for Small Spacecraft

in this section: a reduced toxicity, ADN-based propul- microsatellite. A version of the BHT-200 using iodine sion system was tested onboard of the PRIMSA satellite, as propellant is planned to be used in the iodine satellite featuring two ECAPS LMP-103S thrusters, providing (iSAT) mission, a 12U CubeSat intended as a rapid orbital 1 N at 252-s specific impulse, providing a Δ​ v​ capacity of demonstration of the iodine Hall thruster technology ~60 m/s to the satellite. The mission was able to prove [74], [75]. The mission is intended to demonstrate small successful operation of this thruster system in space, and spacecraft maneuverability and mitigate concerns regard- the thruster’s capability to be used in autonomous forma- ing iodine deposition on spacecraft structures. Featuring tion flying, and allowed propulsion system performance a 200-W thruster, this mission will demonstrate relatively measurements [29]. high power propulsion in a CubeSat envelope. This program The Green Propellant Infusion Mission (GPIM) includes testing of a higher power BHT-600 iodine thruster, [30], [31] is noted for testing another reduced toxicity, and in space validation of the iodine Hall thruster technol- advanced monopropellant propulsion system on a small ogy is seen as a precursor for nanosatellite and small satellite satellite platform with <180 kg total mass shown in Fig. 8. missions with high Δ​ v​ capability, enabling future Near Earth The mission, which is scheduled for launch in spring 2018, Asteroid Orbiter and Lunar Orbiter missions [76]. will carry four Aerojet GR1 1-N thrusters propelled by AF-M315E, a HAN-based propellant, maturing this type of high-performance propulsion system, which is expected to C. Nanosatellite Case 1: Cold Gas and Warm provide a specific impulse of 235 s. The total impulse of the Gas Thrusters propulsion system is stated as 23 kNs, and while this mis- A variety of cold gas systems which are commercially sion is intended as a technology demonstration only, it will available have been flown on small satellites, and recently prove the ability to achieve considerable ​Δv​ using a system on nanosatellites, including the 3U “PRopulsion Operation that can be implemented in smaller spacecraft as primary Proof SATellite -High Performance 1” (POPSAT-HIP1) mis- propulsion as well. sion, demonstrating attitude control (4°) using microfabri- Various entities are currently maturing propulsion sys- acted nozzles with a specific impulse estimated from orbital tems based on AF-M315E [30], [70], [71] or similar ADN- data, of 32 s [60]. based propellants [29] with applications ranging from small Delfi-n3Xt, a 3U CubeSat by Delft University of satellites to CubeSats, including the planned utilization of Technology, verified a cold gas propulsion system, in which a VACCO MiPS system featuring four 100-mN ADN-based the propellant is stored in a solid phase, and operated in a blow thrusters on the Lunar Flashlight CubeSat mission [72]. down mode with continuously decreasing thrust level [62]. CubeProp, another cold gas system, developed by B. Small Satellite Case EP: Hall Thruster NanoSpace, was tested onboard the STU-2B CubeSat (launched in 2015), raising the orbital altitude by ~600 m, The Busek BHT-200 hall thruster, operated with Xenon and performing attitude control tasks [64]. propellant at 100–300 W was flown on TacSat-2, a ~380-kg While still awaiting launch, the JPL Marco mission is a satellite, becoming the first U.S. built Hall thruster to be notable example of a CubeSat mission using cold gas propul- operated in space. TacSat-2 was launched in 2006, fol- lowed by successful thruster operation [73]. The same sion for trajectory correction maneuvers on its flight path thruster is scheduled to be flown on the 180-kg FalconSat-6 to a Mars flyby. The 6U CubeSats will use four R-236FA propelled thrusters for minor trajectory correction, and four canted thrusters for attitude control maneuvers, all fed by a single propellant tank [77].

D. Nanosatellte Case 2: Vacuum Arc Thrusters and Pulsed Plasma Thrusters A type of VAT called microcathode arc thruster (​μ​CAT) of the University of Washington [78], [79] has flown on the BRICSat-P mission, a 1.5U CubeSat launched in spring 2015, featuring four μ​ ​CAT heads. Fig. 9 shows an image of the satellite. After satellite deployment, the thrusters were successfully used in the detumbling of the satellite, but a premature loss of communication to the satellite prevented further investigation of the propulsion system. The VAT Fig. 8. GPIM small satellite featuring 1-N advanced monopropellant propulsion system is planned to be further tested on an thrusters using AF-M315E; from [69]. upcoming GWSat mission in 2018.

10 Proceedings of the IEEE This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

Krejci and Lozano: Space Propulsion Technology for Small Spacecraft

tens and hundreds of emitter arrays, achieving unique con- trollability and redundancy.

V. FLIGHT-READY HIGH- PERFORMANCE NANOSATELLITE PROPULSION SYSTEMS A. Ion Engines Various entities have developed miniature RF ion engines targeting 6-12U CubeSats, based on their power level and volume required to house thruster, neutralizer cathode, as well as propellant tank and feed system [36], [37]. Busek is currently maturing the BIT-3 propelled by iodine, which allows for an unpressurized tank system and, Fig. 9. The 1.5U BRICSat-P featuring four VAT heads developed by the University of Washington, indicated by white arrows; from [78]. therefore, reduces the structural requirements of such a sys- tem, providing 1.4 mN at 60-W beam power, which reduces to ~0.8 mN at 60-W PPU input power when including the PPTs developed by Busek were flown on the 50-kg neutralizer, and 1600-s specific impulse [67]. Fig. 11 shows Falconsat-3, featuring three MPACS units [80]. Each the thruster firing with iodine propellant together with a thruster had a discharge energy of ~2 J at an average specific neutralizer cathode. Various missions are currently in the impulse of ~820 s. While this system was tested on a small planning stage that intend to use the BIT-3, including the satellite, the size and suitability to small power budgets by 6U Lunar Icecube [81] and the 6U LunaH-Map CubeSat reducing the discharge energy and, therefore, the impulse mission [82]. bit per discharge, or the duty cycle of the system, make PPTs relatively compatible to nanosatellites. B. FEEP A CubeSat-compatible liquid indium FEEP propulsion E. Nanosatellite Case 3: Electrospray Thrusters module featuring 28 emission sites has been developed MEMS-based electrospray thrusters developed at the by FOTEC/Enpulsion [58], and is scheduled for flight in Space Propulsion Laboratory, Massachusetts Institute January 2018. Fig. 12 shows the flight propulsion system of Technology (MIT SPL) have been flown on several during acceptance testing. A prototype of this propul- Aerocube-8 satellites, which are 1.5U CubeSats launched in sion system is currently undergoing a long duration firing 2015 and 2016. Each of the satellites carried a PPU capa- tests, firing in a laboratory vacuum chamber continuously ble of firing eight individual thrusters, each consisting of for over 2.5 years. The technology has been derived from a micromachined emitter array featuring 480 emitter tips single emitter ion sources with considerable flight heritage, in a 1-cm​​​​​ 2​​ area. Fig. 10 shows engineering units of a fully including NASA’s MMS and ESA’s Rosetta mission [84]. The integrated electrospray propulsion system. In this system, compact size and moderate power requirements together the propellant is fed passively by capillary forces from a zero with very high specific impulse and the ability to throttle pressure propellant reservoir. The highly miniaturized elec- over a large range of thrust and specific impulse make this ​ ​ small satellite mis- trospray emitter array allows for scaled up systems featuring system additionally attractive for high Δv sions, with multiple FEEP units being operated in parallel to increase thrust levels.

Fig. 10. Engineering units of the MIT electrospray propulsion system featuring eight individually controllable micromachined emitter arrays. Fig. 11. Busek BIT-3 iodine thruster; from [83].

Proceedings of the IEEE 11 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

Krejci and Lozano: Space Propulsion Technology for Small Spacecraft

losses due to convective heat transfer scale according to​ 1.6 D​ t​ ​​, whereas the thrust, which is linear to the mass flow, 2 scales according to T​ ∝ ​D​ t​ ​​. This indicates the nonlinear increase of thermal losses, and therefore reduced efficiency, with increased miniaturization of propulsion systems that convert thermal energy into kinetic energy of the exhaust by expansion. In addition, this relationship indicates increased heat load on the throat region of such systems with increased miniaturization, which can become cru- cial for higher performing thrusters: As active cooling, as

Fig. 12. Fotec/Enpulsion IFM nanothruster during acceptance accomplished in larger chemical propulsion devices, is not testing. expedient for miniature systems due to a number of reasons, including size constraints, complexity, and cost, (5) can dic- VI. SCALING LAWS FOR SMALL tate the choice of an inefficient operational regime to lower SATELLITE PROPULSION TECHNOLOGY the total heat load by decreasing the hot gas temperature ​​T​g​, for example, by operating in a nonstoichiometric combus- A. Gas Dynamic Acceleration tion regime in bipropellant engines. The miniaturization of Direct scaling of existing propulsion systems to smaller reaction chambers in propulsion systems utilizing decom- envelopes in terms of volume and power is generally limited position or combustion of compounds, such as monopropel- by a nonlinear increase in the inefficiency beyond a certain lant and bipropellant thrusters, is additionally limited by point, as inefficiencies do not scale linearly with devices the residency time ​τ​ that the reactants take to complete the [85]. Perhaps the biggest challenge in maintaining high effi- intended reaction, which can be formulated as ciency while miniaturizing gas dynamic systems is the scal- L LρA ​τ = __ ​ = ____ (6) ing of the Reynolds number, according to [86] v m​ ̇ ​

1.2…1.5 where v​​ is the mean propellant velocity, L​​ and A​ ​ are the ​Re ∝ ​p0​ ​ ​Dt​​/ T​ ​ 0​ ​ (4) chamber length and cross section, respectively, ρ​ ​ is the mix- ture density, and ​​m ​​ is the mass flow. This relationship estab- where D​​ t​​​​ is the throat diameter and subscript 0 refers to ̇ stagnation conditions. In general, for a nozzle to scale with- lishes a minimum chamber volume necessary for a given out viscous losses, the Reynolds number would need to propellant mass flow m​ ̇,​​ and thus thrust, which is required remain constant [86]. For this to be possible, the stagnation in order to avoid inefficiencies due to incomplete reaction. pressure of a system would need to increase proportionally While this provides a minimum volume to avoid losses from to the decrease in thrust, which is hardly ever possible in incomplete reaction, the increased thermal losses described in (5) will become increasingly decisive for miniature devices. actual devices for engineering reasons. Experimental data This trend is amplified by the typical temperature depend- for Reynolds numbers of micromachined nozzles with throat ency of chemical reactions, which amplifies the negative diameters in the order of hundreds of micrometers, ranging impact of thermal losses by decreasing the rate of reaction. from ​​10​​ 2​ to ​​10​​ 4​​, have been reported, with the extreme case of Re<1000 for nozzles ​~20 ​​μm ​, where a significant portion of the flow is found in the subsonic boundary region, lead- B. Electric Acceleration ing to significant viscous losses and, therefore, significant 1) Gas Phase Ionization: With the exception of electro- decrease in nozzle efficiency [86]. spray and FEEP thrusters, most EP systems rely on ionization Another example of a physical loss mechanism not of a neutral gas, by either electron collision, RF ionization, decreasing linearly with decreased thruster size are ther- or contact ionization. In these gas phase ionization systems, mal losses in chemical propulsion systems. Considering namely ion engines and Hall thrusters, the propellant utili- exemplarily the widely accepted correlation from Bartz for zation efficiency is therefore primarily governed by the abil- the convective heat transfer coefficient at the throat of a ity to ionize the propellant within the acceleration chamber. nozzle [87] In electron collision ionization, the probability for a collision 1.6 0.8 to take place is dependent on the mean-free path λ​ ​ ____D​ ​ t​ ​ ___P​ c​​ q ​ ̇ = ​h​g(​ T​ g​​ − ​T​w)​ ​, hg​​ ∝ 1.8 ​ ​​ ​ ​ ​ ​ ) (5 D​ ​​ ​(c *) 1 ​λ = ____ ​​ ( 7 ) n​ e​Q​ where ​​P​c​​​ is the chamber pressure, ​c *​ is the characteristic velocity, and D​ ​is the diameter at which the equation is where Q​ ​is the collision cross section and ​n​e​ is the number evaluated. This relation assumes constant Reynolds num- density of electrons in the plasma. As the mean-free path ber scaling, which from the discussion above, is evidently relates to the probability of a neutral particle to experi- an optimistic assumption. From (5) it is evident that the ence a collision and ionization, miniaturization of the

12 Proceedings of the IEEE This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

Krejci and Lozano: Space Propulsion Technology for Small Spacecraft

ionization chamber needs to scale with the mean-free path where ϵ​​ 0​ ​​​ is the vacuum permittivity, E​​ is the local field to retain thruster efficiency. This typically means operating strength, ​γ​ is the liquid surface tension, and ​r *​ is the radius in a higher density regime, which can have a detrimental of curvature of the emitting region that is formed by the impact by causing higher losses, such as higher ion fluxes to mass flow balance of depleted and replenishing ion flow. The the chamber walls. In Hall thrusters, permanent magnetic beam composition of FEEP is found to be largely dominated fields are used to force the externally injected electrons on by singly charged ions, in the presence of slower microdrop- a precessing orbit along the cylindrical discharge chamber lets, reducing the average beam velocity and, therefore, spe- to increase residence time and therefore probability of colli- cific impulse [90]. While similar in terms of the engineering sion with the injected neutral propellant. The radius of pre- implementation, electrospray emitters operate by extracting cession r​​ e​​ is given by charged particles from an ionic liquid propellant, capable of m​ ​​ ​v​​ operating in the pure ionic mode, in which only ions and ​​ ​​ ____ e e (8) re = eB ions solvated to the ​n​th degree are emitted [42]. In current versions of thrusters with larger total emission currents where m​​ e​​, v​​ ​e​, and e​ ​ are the electron mass, velocity, and achieved by multiplexing the number of emission sites, charge, respectively, and B​ ​ is the magnetic field strength. operation is typically in an ion–dropled mixed regime [57], As the chamber dimensions need to be matched to the elec- in which droplets with lower charge-to-mass ratio are accel- tron precessing radius, increased miniaturization in such erated to lower exhaust speed, lowering the specific impulse devices without efficiency deficiency can only be achieved but increasing the thrust produced. by increasing the magnetic field strength B​ ​, which is either Charged particle extraction processes require local field a material parameter of the permanent magnets used, or strengths in the order of 10​​​​​ 9​​ V/m for ion emission. In the determines the scaling of power for electromagnets, which case of an ionic liquid emission site, (9) allows to estimate increases with increasing miniaturization, reducing the the size of the region of ion extraction of r​​​​ *​ ~​ 15 nm (for an overall efficiency. ionic liquid with 1 Si/m conductivity, γ​ ~ ​ 0.05 N/m). While Note that, in addition, miniaturization of ionization the characteristic dimension of the Taylor cone itself could chambers, especially in light of increasing the plasma den- be orders of magnitudes larger than the region of actual sity, comes with increasing particle flux toward the chamber charge emission described by ​​r​​ *​, it should be noted that, walls, leading to increased erosion rates, which has a nega- theoretically, thrust densities in the order of MN/m​​​ ​​2​ would tive effect on lifetime. be possible, for experimentally found emission currents of 2) Electromagnetic Pulsed Thrusters: In PPTs and VATs hundreds of nanoamperes per emission site, which trans- ionization is accomplished in a small envelope along a usu- lates to ~0.05–0.1 nN. However, due to the high local field ally solid propellant surface, whereas the energy transfer strength required to surpass the threshold for ion evapora- into the plasma discharge, and therefore the final kinetic tion, and the need to amplify the applied electric potential energy of the exhaust, is proportional to the change in the using field enhancing structures to support the Taylor cone, circuit impedance and, therefore, proportional to the area current technology does not allow for such dense packing of the discharge circuit is encompassing during acceleration. emission sites in engineering implementations. The current Ziemer and Choueiri [88] showed a quasi linear decrease MIT electrospray thrusters feature 480 emitter structures in thruster efficiency with decreasing pulse energy, showing per square centimeter, while test emitters with up to four the increasing significance of plasma resistance for decreas- times the emitter density have been successfully fabricated ing pulse energy. to date. 3) Liquid Phase and Field Emission Ionization: Ion evapo- ration from a liquid phase or field emission ionization are C. Systems Consideration processes occurring when a conductive liquid is electri- Increasing miniaturization of the thruster heads with- cally stressed above a threshold where particle extraction out compromise in efficiency is considered advantageous is achieved. While the field strength required for such pro- regardless of acceleration principle, as it not only decreases cesses is generally high, the electrostatic stressing forces the the structural mass and volume of the system, but may also conductive liquid into a sharp structure, such as a Taylor allow redundant systems. However, any miniaturization cone (FEEP), [41] or similar electrified meniscii [89], which only achieves its full potential if its merit is not outweighed amplifies the local field strength at the tip of the structure, by the mass and volume required by auxiliary propulsion allowing for a region of high enough field strength for ioni- components in the propulsion system. While advances have zation or ion evaporation. While dependent on the geom- been made in the miniaturization of auxiliary components etry and liquid properties, the force balance at the apex can such as the Busek Microvalve used for the colloid thruster be expressed by system on LISA Pathfinder [91], the propellant stored __1 2 __γ itself constitutes a natural barrier toward miniaturization. ​​ ​ ​ ϵ0​ ​ ​E​​ ​ ~2 ​ ​​ (9) 2 r​​​ *​ Any meaningful miniaturization of a thruster technology

Proceedings of the IEEE 13 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

Krejci and Lozano: Space Propulsion Technology for Small Spacecraft

therefore needs to consider the entire propulsion mass, D. Pushing the Envelope highlighting the importance of maintaining, or increasing, 1) Chipsats and Membrane Spacecraft: Pushing the the propulsion system efficiency with advancing minia- boundaries of miniaturization of spacecraft, Kicksat [14] turization. This becomes most important for technologies was an early instance of radical miniaturization of semiau- used in space missions with high ​Δv​ requirements, as in tonomous systems, with a 3U CubeSat designed to deploy such cases, the stored propellant mass easily outweighs 128 subsatellites called Sprites, intended to perform basic the mass of the thruster and auxiliary components and it functionality including communication with the dispensing may be more productive to increase the specific impulse of satellite. While limited in capability, and ultimately not suc- the thruster, than striving to decrease the structural mass cessful due to a malfunction in the dispensing mechanism, of the thruster and components without considering the this and followup missions in which individual chipsats impact on the overall propulsion system. Based on the dis- were launched as independent piggyback payloads to other cussion in Sections III–VI, two promising approaches are satellites, highlighted the theoretical capability of extremely highlighted. miniaturized spacecraft, especially in large swarm constel- 1) Technologies that miniaturize without impact on lation concepts. A related concept, called Wafersat, was efficiency: Liquid phase ionization technologies such as the undertaken by the MIT Lincoln Laboratory and the MIT ionic liquid electrospray emitters do not suffer from an Space System Laboratory, in which a satellite platform inherent decrease of performance with advancing miniatur- with full capabilities similar to a larger sized CubeSat will ization of engineering devices beyond submillimeter scale. be reduced to a mass-producable silicon wafer form fac- Such systems scale, therefore, linearly in terms of thrust, tor. While such missions pose challenges to a variety of and are, in theory, not subject to an adverse impact on the subsystems, including thermal and communications, spe- specific impulse. Such systems thus allow miniaturization of cific challenges are imposed on future propulsion systems the thrusters, without increasing the amount of stored pro- that are deemed an enabling technology for such missions, pellant to accomplish a given mission requirement, there- both for orbit and attitude controls. Using the baseline of fore leading to an overall decrease in the mass and volume Wafersat, requirements for propulsion systems regarding of the propulsion module. mass, volume, power, and total impulse can be derived for a propulsion system usable for minor orbit correction and 2) Technologies with novel aspects that have positive attitude control. Based on a standard 8-in wafer form factor, systems impact: Novel technologies can outweigh inefficien- the total impulse required for a year of drag compensation cies that originate from miniaturization of a thruster by ranges from ~22.5 to ~3.35 kNs in a 400-km altitude orbit, having a net-positive impact when considering the over- depending on the angle of attack. For an angle of attack of all propulsion system. The HYDROS system developed by zero, which is with minimal drag cross section, a state-of- Tethers Unlimited [66] is a bipropellant system which uses the-art electrospray emitter with specific impulse of 1500 s electrolysis of stored water to produce the reactants. Such a could perform a year of drag compensation in such an orbit system could provide specific impulse similar to a bipropel- using ​<​ 0.2 g of propellant. Another example are ultrathin lant engine in the future, but may come with significantly spaceraft such as the membrane-based Brane Craft, stud- lower structural mass for the propellant storage system com- ied by the Aerospace Corporation [15]. In this concept, pared to a traditional bipropellant system. Even for a modest the spacecraft itself consists of a flexible, thin membrane decrease in efficiency of the thruster, caused by the small with large surface-to-volume ratio, imposing strict volume size of the bipropellant combustion chamber and consider- requirements on the propulsion system that is required ing the added electric power required, the overall system to fit within the μ​ 30- ​m envelope. Janson [15] concluded can favorably trade off in terms of the overall system mass that electrospray emission from sharp emitter structures and volume required. in combination with a separate extractor and acceleration An independent systems aspect that can become grid could be accomplished within the scale of 10 μ​ ​m, fed significant with increased miniaturization is potential passively with propellant that could be distributed around electromagnetic interference (EMI), especially in the case of the emitters in a gap between the spacecraft membrane lay- miniaturized EP systems. Such interference can become an ers, leading to significant propellant mass per emitter, with issue for thrusters which require components with strong, 40 h of accumulated firing time identified as a baseline. In unsteady magnetic fields, such as RF thrusters or thrust- addition, Janson’s analysis showed that acceleration of up to ers with unsteady operation principle, such as PPTs, which 0.1 m/s​​​​​ 2​​ could be achievable, which is orders of magnitude require a high current discharge. While shielding measures higher than typical EP systems, partly enabled by the favora- can, to some extent, decrease such unwanted interactions, ble large surface area, enabling large solar panel area and, miniaturization necessarily requires closer spacing and therefore, a high power per spacecraft mass configuration. smaller margins on components, and care needs to be taken As outlined in Section VI, the only currently available to avoid undesired interaction of high voltage components system that theoretically allows controlled propellant accel- with sensitive subsystems such as onboard computers. eration at sufficient efficiency within submillimeter scale

14 Proceedings of the IEEE This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

Krejci and Lozano: Space Propulsion Technology for Small Spacecraft

the purpose of increasing the spacecraft atmospheric drag area for deorbiting purposes [95]. The 14-m​​​​​ 2​​ solar sail was deployed in June 2010 onboard the 300-kg IKAROS satel- lite and was able to accumulate a change in orbital velocity of approximately 100 m/s until December 2010 [95], [96]. Despite a variety of technical issues on Lightsail-1, it was able to successfully deploy the 32-m​​​ ​​2​​ solar sail, (shown in Fig. 13), acting as a precursor for the followup Lightsail-2 mission [48].

Fig. 13. Deployed sail of the Lightsail 1 CubeSat; from [48]. VII. CONCLUSION The success of small satellites has led to increased mission and is a liquid phase ionization system, and both the MIT scientific capabilities of miniaturized spacecraft, spurred by Wafersat as well as the Brane Craft studies use electrospray the decreased mission cost and faster development schedules. emitters as their current baseline technology [15]. As mission complexities evolve, the need for high capability propulsion, previously mostly reserved for traditional large 2) Infinite Specific Impulse Propulsion: Infinite specific satellite platforms with high power budgets, has driven the impulse propulsion, that is propulsion without the need for development of a plethora of propulsion solutions for small propellant stored onboard the spacecraft, is attractive for long duration, very high Δ​ v​ missions, potentially enabling inter- satellites, CubeSats, and beyond. This review discusses the planetary, asteroid encounter (including the planned NASA different propulsion principles applicable to small satellites, NEA Scout CubeSat mission [92]) and deep space explora- and presents a classification of available propulsion solutions, tion missions, such as the scientifically rewarding targets in including a variety of different chemical and EP systems of the region of the outer planets and Oort cloud, positioning a varying complexity and performance. A classification of these spacecraft in the gravitational focus point of the sun to form propulsion systems is given, based on the tradeoffs in perfor- a future powerful telescope [93] or even interstellar flight mance parameters, including thrust, specific impulse, and attempts as proposed by the project Starshot [94], which input power. A review of selected space qualified propulsion attempts to propel a miniaturized spacecraft to Proxima systems is presented, highlighting key technologies for spe- centauri using a man made high power light source. Besides cific satellite classes. A discussion of the predominant scal- mission studies of such long-term goals, technology dem- ing laws for miniaturization of different propulsion systems onstration missions of solar radiation pressure-based sail is given, identifying the limiting factors in miniaturization on technology have been launched so far, including Jaxa’s thruster, and system level is given. Based on this, the propul- IKAROS sail [95] and the LightSail-1 CubeSat mission by sion related requirements and state-of-the-art technologies the Planetary Society [48]. In addition, similar technology for the extreme cases of chip and membrane satellites, and has been tested on NASA’s Nanosail-D2 CubeSat, but for infinite specific impulse missions are discussed.

REFERENCES Space 1,” in Proc. 35th AIAA/ASME/SAE/ASEE Available: http://www.sciencedirect.com/ Joint Propulsion Conf. Exhibit, Los Angeles, science/article/pii/S0094576511003742 [1] M. Martínez-Sánchez and J. E. Pollard, CA, USA, 1999. “Spacecraft electric propulsion—An [13] J. Puig-Suari, C. Turner, and W. Ahlgren, overview,” J. Propulsion Power, vol. 14, no. 5, [8] H. Greer, “Vacuum startup of reactors for “Development of the standard CubeSat pp. 688–699, 1998. catalytic decomposition of hydrazine,” deployer and a CubeSat class PicoSatellite,” J. Spacecraft, vol. 7, no. 5, pp. 522–528, 1970. in Proc. IEEE Aerosp. Conf., Big Sky, MT, [2] V. Kim, “Electric propulsion activity in Russia,” USA, Mar. 2001, pp. 1/347–1/353. in Proc. Int. Electr. Propulsion Conf., 2001. [9] G. Racca, “Smart-1 mission description and development status,” Planetary Space Sci., [14] Z. Manchester, M. Peck, and A. Filo, “KickSat: [3] G. P. Sutton, “History of liquid propellant vol. 50, no. 14, pp. 1323–1337, 2002. A crowd-funded mission to demonstrate the rocket engines in the United States,” [Online]. Available: http://www. world’s smallest spacecraft,” in Proc. Small J. Propulsion Power, vol. 19, no. 6, sciencedirect.com/science/article/pii/ Satellite Conf., Logan, UT, USA, 2013. pp. 978–1007, 2003. S003206330200123X [15] S. W. Janson, “Brane craft: Phase I results,” [4] E. Y. Choueiri, “A critical history of electric [10] J. R. Brophy, “NASA’s Deep Space 1 ion in Proc. Small Satellite Conf., Logan, UT, propulsion: The first 50 years (1906–1956),” engine (plenary),” Rev. Sci. Instrum., vol. 73, USA, 2017. J. Propulsion Power, vol. 20, no. 2, no. 2, pp. 1071–1078, 2002. [Online]. [16] J. K. McDermott, “Power,” in Space Mission pp. 193–203, 2004. Available: http://dx.doi. Analysis and Design, W. J. Larson and [5] R. L. Sackheim, “Overview of United States org/10.1063/1.1432470 J. R. Wertz, Eds. Norwell, MA, USA: Kluwer, rocket propulsion technology and associated [11] A. M. Baker, A. da Silva Curiel, J. Schaffner, 1999, ch. 11.4, pp. 407–427. space transportation systems,” J. Propulsion and M. Sweeting, “‘You can get there from [17] D. Selva and D. Krejci, “Capabilities of Power, vol. 22, no. 6, pp. 1310–1332, 2006. here’: Advanced low cost propulsion concepts CubeSats for earth science,” J. Small [6] R. L. Burton and P. J. Turchi, “Pulsed plasma for small satellites beyond Leo,” Acta Satellites, to be published. thruster,” J. Propulsion Power, vol. 14, no. 5, Astronautica., vol. 57, no. 2, pp. 288–301, 2005. [18] R. L. Sackheim and S. Zafran, “Space pp. 716–735, 1998. [12] D. Selva and D. Krejci, “A survey and propulsion systems,” in Space Mission [7] J. S. Sovey, V. K. Rawlin, and M. J. Patterson, assessment of the capabilities of CubeSats for Analysis and Design, W. J. Larson and “A synopsis of ion propulsion development earth observation,” Acta Astronautica., J. R. Wertz, Eds. Norwell, MA, USA: Kluwer, projects in the United States: SERT I to Deep vol. 74, pp. 50–68, Jun. 2012. [Online]. 1999, pp. 685–718.

Proceedings of the IEEE 15 This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

Krejci and Lozano: Space Propulsion Technology for Small Spacecraft

[19] K. Danzmann, “LISA—An ESA cornerstone [34] H. Burkhardt, “Evaluation of propulsion [50] B. Fernando, “Planning our first interstellar mission for the detection and observation of systems for satellite end-of-life de-orbiting,” journey,” Astron. Geophys., vol. 58, no. 1, gravitational waves,” Adv. Space Res., vol. 32, in Proc. 38th AIAA/ASME/SAE/ASEE Joint pp. 1.28–1.30, 2017. no. 7, pp. 1233–1242, 2003. [Online]. Propulsion Conf. Exhibit, Indianapolis, IN, [51] R. M. Zubrin and D. G. Andrews, “Magnetic Available: http://www.sciencedirect.com/ USA, 2002. sails and interplanetary travel,” J. Spacecraft science/article/pii/S0273117703903231 [35] R. G. Jahn, Physics of Electric Propulsion. Rockets, vol. 28, no. 2, pp. 197–203, 1991. [20] P. R. Lawson, “Terrestrial planet finder New York, NY, USA: McGraw-Hill, 1968. [52] P. Janhunen, “Electric sail for spacecraft interferometer: 2007–2008 progress and [36] H. Leiter, “Development of a miniaturized propulsion,” J. Propulsion Power, vol. 20, plans,” Proc. SPIE, vol. 7013, p. 70132N, RF ion engine system for commercial and no. 4, pp. 763–764, 2004. Jul. 2008. scientific applications,” in Proc. 47th AIAA/ [53] D. Krejci, B. Seifert, and C. Scharlemann, [21] T. Werkerle, J. B. P. Filho, L. E. V. L. da ASME/SAE/ASEE Joint Propulsion Conf. “Endurance testing of a pulsed plasma Costa, and L. G. Trabasso, “Status and trends Exhibit, San Diego, CA, USA, 2011. thruster for nanosatellites,” Acta Astronautica., of smallsats and their launch vehicles—An [37] M. Tsay, J. Frongillo, and K. Hohman, vol. 91, pp. 187–193, Nov. 2013. [Online]. up-to-date review,” J. Aerosp. Technol. “Iodine-fueled mini RF ion thruster for Available: http://www.sciencedirect.com/ Manage., vol. 9, no. 3, pp. 269–286, 2017. CubeSat applications,” in Proc. 34th Int. science/article/pii/S0094576513002026 [22] G. P. Sutton and O. Biblarz, Rocket Electr. Propulsion Conf., Hyogo-Kobe, [54] S. Ciaralli, M. Coletti, F. Guarducci, and Propulsion Elements, Hoboken, NJ, USA: Japan, 2015. S. Gabriel, “Pptcup lifetime test results,” in Wiley, 2010, 8th ed. [38] R. E. Wirz, “Miniature ion thrusters: Proc. 33rd Int. Electr. Propulsion Conf., [23] K. Lemmer, “Propulsion for CubeSats,” Acta Miniature ion thrusters: A review of modern Washington, DC, USA, 2013. Astronautica., vol. 134, pp. 231–243, May technologies and mission capabilities,” in [55] M. Keidar, “Micro-cathode arc thruster for 2017. [Online]. Available: http://www. Proc. 34th Int. Electr. Propulsion Conf., Hyogo- small satellite propulsion,” in Proc. Aerosp. sciencedirect.com/science/article/pii/ Kobe, Japan, 2015. Conf., Mar. 2016. S0094576516308840 [39] J. R. Brophy, “The ion propulsion system for [56] M. U. Siddiqui, C. Cretel, J. Synowiec, [24] D. R. Lev, J. Herscovitz, D. Kariv, and dawn,” in Proc. 39th AIAA/ASME/SAE/ASEE A. G. Hsu, J. A. Young, and R. Spektor, “First I. Mizrachi, “Heated gas propulsion system Joint Propulsion Conf. Exhibit, Huntsville, AL, performance measurements of the phase conceptual design for the Samson nano- USA, 2003. four RF thruster,” in Proc. 35th Int. Electr. satellite (propulsion),” J. Small Satellites, Propulison Conf., Atlanta, GA, USA, 2017. vol. 6, no. 1, pp. 551–564, 2017. [40] E. Canuto and L. Massotti, “All-propulsion design of the drag-free and attitude control [57] D. Krejci, F. Mier-Hicks, R. Thomas, T. Haag, [25] J. Lee, S. An, and S. Kwon, “Development of of the European satellite GOCE,” Acta and P. Lozano, “Emission characteristics of a liquid propellant rocket utilizing hydrogen Astronautica., vol. 64, no. 2, pp. 325–344, passively fed electrospray microthrusters peroxide as a monopropellant,” in Proc. 44th 2009. [Online]. Available: http://www. with propellant reservoirs,” J. Spacecraft AIAA/ASME/SAE/ASEE Joint Propulsion Conf. sciencedirect.com/science/article/pii/ Rockets, vol. 54, no. 2, pp. 447–458, 2017. Exhibit, Hartford, CT, USA, 2008. S0094576508002683 [58] D. Jelem, B. Seifert, R. Sypniewski, [26] D. Krejci, A. Woschnak, C. Scharlemann, [41] G. Taylor, “Disintegration of water drops in N. Buldrini, and A. Reissner, “Performance and K. Ponweiser, “Structural impact of an electric field,” Proc. R. Soc. Lond. A, mapping and qualification of the IFM Nano honeycomb catalysts on hydrogen peroxide Math. Phys. Sci., vol. 280, no. 1382, thruster FM for in orbit demonstration,” in decomposition for micro propulsion,” Chem. pp. 383–397, 1964. Proc. 53rd AIAA/SAE/ASEE Joint Propulsion Eng. Res. Design, vol. 90, no. 12, Conf., Atlanta, GA, USA, 2017. pp. 2302–2315, 2012. [Online]. Available: [42] P. Lozano and M. Martínez-Sánchez, “Ionic http://www.sciencedirect.com/science/ liquid ion sources: Characterization of [59] S. Mauthe, F. Pranajaya, and R. E. Zee, “The article/pii/S0263876212002158 externally wetted emitters,” J. Colloid design and test of a compact propulsion Interface Sci., vol. 282, no. 2, pp. 415–421, system for Canx nanosatellite formation [27] R.-J. Koopmans, J. S. Shrimpton, 2005. [Online]. Available: http://www. flying,” in Proc. Small Satellite Conf., Logan, G. T. Roberts, and A. J. Musker, sciencedirect.com/science/article/pii/ UT, USA, 2005. “Dependence of pellet shape and size on S0021979704008422 [60] G. Manzoni and Y. L. Brama, “CubeSat pressure drop in H2O2,” J. Propulsion Power, vol. 30, no. 3, pp. 775–789, 2014. [43] P. C. Lozano, “Studies on the ion-droplet micropropulsion characterization in low mixed regime in colloid thrusters,” Ph.D. earth orbit,” in Proc. Small Satellite Conf., [28] B. A. Palaszewski, M. L. Meyer, L. Johnson, dissertation, Dept. Aeronautics Astron., Logan, UT, USA, 2015. D. M. Goebel, H. White, and D. J. Coote, Massachusetts Inst. Technol., Cambridge, [61] C. Müller, “Implementation of the T3PS in In-Space Chemical Propulsion System MA, USA, 2003. Roadmap. Cham, Switzerland: Springer- the Delfi-n3Xt satellite,” in Proc. Small Verlag, 2017, pp. 655–671. [Online]. [44] F. Mier-Hicks and P. Lozano, “Spacecraft- Satellite Missions Earth Observat., Berlin, Available: https://doi.org/10.1007/978-3-319- charging characteristics induced by the Germany, 2010, pp. 411–424. 27748-6_26 operation of electrospray thrusters,” [62] J. Guo, J. Bouwmeester, and E. Gill, “In-orbit J. Propulsion Power, vol. 33, no. 2, [29] K. Anflo and R. Möllerberg, “Flight results of Delfi-n3Xt: Lessons learned and pp. 456–467, 2017. demonstration of new thruster and green move forward,” Acta Astronautica., vol. 121, propellant technology on the PRISMA [45] M. Keidar, “Magnetically enhanced vacuum pp. 39–50, May 2016. [Online]. Available: satellite,” Acta Astronautica., vol. 65, arc thruster,” Plasma Sour. Sci. Technol., http://www.sciencedirect.com/science/ nos. 9–10, pp. 1238–1249, 2009. [Online]. vol. 14, no. 4, pp. 661–669, 2005. article/pii/S0094576515004373 Available: http://www.sciencedirect.com/ [46] M. Keidar, “Electric propulsion for small [63] S. Arestie, E. G. Lightsey, and B. Hudson, science/article/pii/S0094576509001970 satellites,” Plasma Phys. Controlled Fusion, “Development of a modular, cold gas [30] R. A. Spores, R. Masse, S. Kimbrel, and vol. 57, no. 1, p. 014005, 2015. propulsion system for small satellite applications,” J. Small Satellites, vol. 1, no. 2, C. McLean, “GPIM AF-M315E propulsion [47] Y. Tsuda, “Flight status of IKAROS deep pp. 63–74, 2012. system,” in Proc. 49th AIAA/ASME/SAE/ASEE space solar sail demonstrator,” Acta Joint Propulsion Conf. Exhibit, San Jose, CA, Astronautica., vol. 69, nos. 9–10, [64] S. Wu, W. Chen, and C. Chao, “The STU-2 USA, 2013. pp. 833–840, 2011. [Online]. Available: CubeSat mission and in-orbit test results,” [31] R. Masse, M. Allen, R. Spores, and http://www.sciencedirect.com/science/ in Proc. Small Satellite Conf., Logan, UT, E. A. Driscoll, “AF-M315E propulsion system article/pii/S0094576511001822 USA, 2016. advances and improvements,” in Proc. 52nd [48] R. W. Ridenoure, “Testing the lightsail [65] B. Crowe and K. Anflo, “In-space AIAA/SAE/ASEE Joint Propulsion Conf., Salt program: Advancing solar sailing technology demonstration of high performance green Lake, UT, USA, 2016. using a CubeSat platform,” J. Small Satellites, propulsion (HPGP) and its impact on small [32] C. Scharlemann, “GRASP—Analysis of green vol. 5, no. 2, pp. 531–550, 2016. satellites,” in Proc. Small Satellite Conf., 2011. propellant candidates,” in Proc. 62nd Int. [49] L. Johnson, M. Whorton, A. Heaton, [66] K. James, T. Moser, A. Conley, J. Slostad, and Astron. Congr., Cape Town, South Africa, 2011. R. Pinson, G. Laue, and C. Adams, R. Hoyt, “Performance characterization of [33] D. Krejci, A. Woschnak, C. Scharlemann, “NanoSail-D: A solar sail demonstration the hydros water electrolysis thruster,” in and K. Ponweiser, “Performance assessment mission,” Acta Astronautica., vol. 68, Proc. Small Satellite Conf., Logan, UT, USA, of 1 n bipropellant thruster using green nos. 5–6, pp. 571–575, 2011. [Online]. 2015. propellants H2O2/kerosene,” J. Propulsion Available: http://www.sciencedirect.com/ [67] M. Tsay, J. Frongillo, J. Model, J. Zwahlen, Power, vol. 29, no. 1, pp. 258–289, 2013. science/article/pii/S0094576510000597 C. Barcroft, and C. Feng, “Neutralization

16 Proceedings of the IEEE This article has been accepted for inclusion in a future issue of this journal. Content is final as presented, with the exception of pagination.

Krejci and Lozano: Space Propulsion Technology for Small Spacecraft

demo and thrust stand measurement for [78] J. Kolbeck, G. Teel, J. Lukas, and M. Keidar, transfer coefficients,” J. Jet Propulsion, BIT-3 RF ion thruster,” in Proc. 53rd AIAA/ “A vacuum arc thruster with ablatable vol. 27, no. 1, pp. 49–53, 1957. SAE/ASEE Joint Propulsion Conf., Atlanta, GA, anode,” in Proc. 52nd AIAA/SAE/ASEE [88] J. K. Ziener and E. Y. Choueiri, “Scaling laws USA, 2017. Joint Propulsion Conf., Salt Lake, UT, for electromagnetic pulsed plasma [68] A. Reissner, B. Seifert, N. Buldrini, D. Jelem, USA, 2016. thrusters,” Plasma Sour. Sci. Technol., vol. 10, D. Krejci, J.Gonzales del Amo, and L. Massotti, [79] J. Lukas, G. Teel, J. Kolbeck, and M. Keidar, no. 3, pp. 395–405, 2001. “Development, simulation and testing of the “High thrust-to-power ratio micro-cathode [89] C. Coffman, M. Martínez-Sánchez, mn-feep thruster technology,” Acta arc thruster,” AIP Adv., vol. 6, no. 2, F. J. Higuera, and P. C. Lozano, “Structure of Astronautica., to be published. p. 025311, 2016. the menisci of leaky dielectric liquids during [69] C. H. McLean, “Green propellant infusion [80] V. Hruby, “Review of electric propulsion electrically-assisted evaporation of ions,” mission program development and activities in the US industry,” in Proc. 39th Appl. Phys. Lett., vol. 109, no. 23, p. 231602, technology maturation,” in Proc. 50th AIAA/ AIAA/ASME/SAE/ASEE Joint Propulsion Conf. 2016. [Online]. Available: http://aip. ASME/SAE/ASEE Joint Propulsion Conf., Exhibit, Huntsville, AL, USA, 2003. scitation.org/doi/abs/10.1063/1.4971778 Cleveland, OH, USA, 2014. [81] D. C. Folta, N. Bosanac, A. Cox, and [90] M. Tajmar, A. Genovese, and W. Steiger, [70] R. Masse, R. A. Spores, and C. McLean, “The K. C. Howell, “The lunar IceCube mission “Indium field emission electric propulsion advancing state of AF-M315E technology,” in design: Construction of feasible transfer microthruster experimental Proc. 50th AIAA/ASME/SAE/ASEE Joint Propulsion trajectories with a constrained departure,” in characterization,” J. Propulsion Power, Conf. Exhibit, Cleveland, OH, USA, 2014. Proc. 26th AAS/AIAA Spaceflight Mech. vol. 20, no. 2, pp. 211–218, 2004. Meeting, Napa, CA, USA, 2016. [71] M. Tsay, D. Lafko, J. Zwahlen, and W. Costa, [91] J. Ziemer, “Colloid micro-Newton thruster “Development status and 1U CubeSat [82] M. Tsay, J. Frongillo, J. Model, J. Zwahlen, development for the ST7-DRS and LISA application of Busek’s 0.5N green and L. Paritsky, “Maturation of iodine fueled missions,” in Proc. 41st AIAA/ASME/SAE/ monopropellant thruster,” in Proc. Small BIT-3 RF ion thruster and RF neutralizer,” in ASEE Joint Propulsion Conf. Exhibit, Tucson, Satellite Conf., Logan, UT, USA, 2013. Proc. 52nd AIAA/SAE/ASEE Joint Propulsion AZ, USA, 2005. Conf., Salt Lake, UT, USA, 2016. [7 2] Lunar Flashlight Propulsion System, VACCO, [92] L. McNutt, L. Johnson, D. Clardy, J. Castillo- El Monte, CA, USA, 2017. [83] M. Tsay, J. Frongillo, J. Model, J. Zwahlen, Rogez, A. Frick, and L. Jones, “Near-earth and L. Paritsky, “Flight development of asteroid scout,” in Proc. AIAA Space Conf., [73] C. J. Finley and N. Peck, “Techsat-2: A story iodine BIT-3 RF ion propulsion system for San Diego, CA, USA, 2014. of survival,” in Proc. Small Satellite Conf., SLS EM-1 CubeSats,” in Proc. Small Satellite Logan, UT, USA, 2007. [93] G. A. Landis, “Mission to the gravitational Conf., Logan, UT, USA, 2016. focus of the sun: A critical analysis,” in Proc. [74] K. A. Polzin, “Propulsion system development [84] A. Reissner, “Lifetime testing of the AIAA SciTech Forum, Grapevine, TX, USA, for the iodine satellite (ISAT) demonstration mN-FEEP thruster,” in Proc. 52nd AIAA/SAE/ 2017. mission,” in Proc. 34th Int. Electr. Propuls. ASEE Joint Propulsion Conf., Salt Lake, UT, [94] Z. Merali, “Shooting for a star,” Science, Conf., Hyogo-Kobe, Japan, 2015. USA, 2016. vol. 352, no. 6289, pp. 1040–1041, 2016. [75] T. D. Smith, “Overview of NASA iodine Hall [85] P. Lozano, MEMS for Automotive and [95] L. Johnson, D. Alhorn, M. Boudreaux, thruster propulsion system development,” in Aerospace Applications. Urbana, IL, USA: J. Casas, D. Stetson, and R. Young, Proc. Space Propuls. Conf., Rome, Italy, 2016. Woodhead, 2013. “Solar and drag sail propulsion: From [76] J. Dankanich, J. J. Szabo, B. Pote, [86] A. D. Kedsdever, “System considerations and theory to mission implementation,” in S. R. Oleson, and H. Kamhawi, “Mission and design options for microspacecraft Proc. Space Propulsion Conf., Cologne, system advantages of iodine Hall thrusters,” propulsion systems,” in Microporpulsion for Germany, 2014. in Proc. 50th AIAA/ASME/SAE/ASEE Joint Small Spacecraft, vol. 187, M. M. Micci and [96] C. Garner, B. Diedrich, and M. Leipold, Propulsion Conf., Cleveland, OH, USA, 2014. A.≈D. Kedsdever, Eds. Washington, DC, “A summary of solar sail technology [77] A. Klesh and J. Krajewski, “MarCO: USA: AIAA, 2000, ch. 4. developments and proposed demonstration CubeSats to Mars in 2016,” in Proc. Small [87] D. Bartz, “A simple equation for rapid missions,” in Proc. 35th AIAA Joint Propulsion Satellite Conf., Logan, UT, USA, 2015. estimation of rocked nozzle convective heat Conf., Los Angeles, CA, USA, 1999.

ABOUT THE AUTHORS David Krejci received the M.S. degree in applied Paulo Lozano received the M.S. degree in physics physics and the Ph.D. degree in mechanical engi- from Cinvestav, Mexico and the M.S. degree in aer- neering from Vienna University of Technology, onautics and astronautics and the Ph.D. degree in Vienna, Austria, in 2008 and 2012, respectively, space propulsion from the Massachusetts Insti- and the M.A. degree in political science from Uni- tute of Technology (MIT), Cambridge, MA, USA. versity of Vienna, Vienna, Austria, in 2012. He is the M. Alem�n-Velasco Professor of He is a Research Scientist in the Department Aeronautics and Astronautics and the Director of of Aeronautics and Astronautics, Massachu- the Space Propulsion Laboratory at the Massa- setts Institute of Technology (MIT), Cambridge, chusetts Institute of Technology (MIT), Cambridge, MA, USA, working on the design and characterization of micromachined MA, USA. He has published over 100 conference and journal publications ionic liquid electrospray thrusters, including multiple flight experiments. and has delivered over 40 invited talk presentations on his work. He teaches Before joining MIT's Space Propulsion Laboratory in 2014, he was working subjects in space and rocket propulsion, fluid mechanics, and plasma phys- on various types of electric and chemical propulsion systems, including ics. His main interests are in plasma physics, space propulsion, ion beam pulsed plasma thrusters for CubeSats, and chemical green bi-propellant physics, small satellites and nanotechnology. thrusters for small satellites at the Austrian Institute of Technology (later: Prof. Lozano is an Associate Fellow of the American Institute of Fotec GmbH). Aeronautics and Astronautics and a member of the American Physical Society. He has served on the Asteroid Mitigation, NASA Technology Road- maps and Scientific CubeSats panels of the National Research Council.

Proceedings of the IEEE 17 本文献由“学霸图书馆-文献云下载”收集自网络，仅供学习交流使用。

学霸图书馆（www.xuebalib.com）是一个“整合众多图书馆数据库资源，

提供一站式文献检索和下载服务”的24 小时在线不限IP 图书馆。 图书馆致力于便利、促进学习与科研，提供最强文献下载服务。

图书馆导航：

图书馆首页 文献云下载 图书馆入口 外文数据库大全 疑难文献辅助工具