High specific impulse electric propulsion will allow large (e.g., approximately 0.25 or greater) payload mass fractions (ion rocket payload mass divided by the initial ion rocket mass) to be delivered to deep space at high velocities because of the reduced amount of propellant needed. Once thrusting is complete, an abundance of electrical power also enables high data transmission rates to Earth, enhancing the mission's scientific return. But electric propulsion and deep-space science are viable only if an adequate space-power source exists.  

Radioisotopes have been used successfully for more than 25 years to supply the heat for thermoelectric generators on various deep-space probes. Radioisotope electric propulsion (REP) systems have been proposed as low-thrust ion propulsion units based on radioisotope electric generators and ion thmsters. Payload size will determine the necessary propulsive power, which is selectable in principle by scaling the propulsion unit size or combining a number of smaller modular units in parallel. Payloads of 50 to 100 kg and a total sciencecraft mass, including the REP unit, of 200 to 400 kg are envisioned (Fig. 1).  

 
Figure 1 

The perceived liability of radioisotope electric generators for ion propulsion is their high specific mass (mass per unit power). A higher specific mass results in a lower velocity change since thrust is used primarily to accelerate the massive powerplant (electric generator plus ion engine) and not the payload. The only saving grace for high specific mass powerplants is that for a low rocket in fieldfree space, the thrust time during which a given distance is covered is proportional to the powerplant specific mass to the one-third power. Consequently, to cut the thrust time by half actually requires a factor of eight reduction in specific mass. This relative insensitivity of thrust time, and implicitly flight time (since thrusting will typically take up a large fraction of the flight duration), to specific mass suggests that high specific mass powerplants may be suitable for robotic science missions to the outer Solar System.  

Conventional radioisotope thermoelectric generators (RTGs) have a specific mass of about 200 kg/kW of electric power. Many development efforts have been undertaken with the aim of reducing the specific mass of radioisotope electric systems. Recent performance estimates saggest that specific masses of 50 kg/kW may be achievable for thermophotovoltaic (TPV) and alkali metal thermal-to-electric conversion (AMTEC) generator" (see NN, Apr. 1999, p. 32). The need for low-mass electfic generators using less Power System (ARPS).  

Powerplants constructed from these nearterm radioisotope electric generators and long-life ion thrusters will likely have specific masses in the range of 100 to 200 kg/kW of thrust power if development continues over the next decade. In earlier studies, it was concluded that flight times within the Solar System are indeed insensitive to reductions in the powerplant specific mass, and that a timely scientific program of robotic planetary rendezvous and near-interstellar space missions is enabled by primary electric propulsion once the powerplant specific mass is in the range of 100 to 200 kg/kW. Flight times can be substantially reduced by using hybrid propulsion schemes that combine chemical propulsion, gravity assist, and electrlc propulsion. Hybrid schemes are further explored in this article to illustrate how the performance of REP is enhanced for Pluto rendezvous, heliopause orbiter, and gravitational lens missions.  

REP technology  

Various schemes for radioisotope electric generators have been under development for several years, and some may lend themselves to an evolutionary REP program over the next decade. A reduction in the specific mass of radioisotope electric generators from 200 kg/ kW to 50 kg/kW of electric power now appears likely if evelopment continues.  

Most of these devices are intended to produce total electric powers of tens of watts to a few thousand watts, and their modular designs allow them to be easily scaled up or down.  

Table I summarizes the performance and reliability of some candidate technologies for radioisotope electric generators. The information in the tirst four columns is taken from the references cited below. The last column gives es  
timated specific masses for hypothetical powerplants using a 30-cm derated xenon thruster discussed in this section as an example.  

The first three devices in Table I involve direct conversion of heat to electricity: the conventional radioisotope thermoelectric generator (RTG) using silicon-germanium unicouples, the modular RTG that uses slhcongermanium/gallium-phosphorus multicouples consisting of 20 thermocouples connected in serles, and the radioisotope thermophotovoltaic (RTPV) power system that would use gallium-antimony infrared (IR) photovoltaic cells to directly convert radiant heat from the radioisotope to electricity. With the development of improved IR filters, the estimated efficiency of RTPV units has increased from 13 percent to 23 percent. This, along with better radiator design, has reduced the estimated generator specific mass from 118 to about 60 6 kg/kW of electric power.  

The last two devices in Table I are thermal engines that use a working fluid to convert radioisotope decay heat to electrical work. The free piston Stirling engine (FPSE) has been proposed as a lightweight dynamic isotope power system (DIPS) for 10 to 1000-watt .space-power applications. The magnet mass m the alternator keeps the specific mass high in these devices. AMTEC generators electrochemically convert the isothermal expansion of sodium or potassium vapor to electrical work via a charge exchange of the liquid metal in a beta-alumina solid electrolyte and electron recombination at a porous metal electrode. Recent design improvements have led to slight reductions in the estimated specific mass of hypothetical AMTEC generators.Because of their low specific mass, both the RTPV and AMTEC generators are potential candidates for the ARPS on the proposed Pluto-Kuiper Express and the Europa Orbiter missions. If successfully developed, either of these generators could become the prototypical generator for future radioisotope electric propulsion systems.  

A reliable ion thruster makes up the other essential part of a propulsive unit. A variety of thrusters has been researched for different applications over the years. Significant progress has been made in reducing the mass and extending the lifetime of low-power, inert-gas (xenon, krypton, argon) ion thrusters in recent years. Between June 1996 and September 1997, a NASA 30-cm-diameter, 2-kW xenon ion engine underwent an 8000-hour endurance test at the Jet Propulsion Laboratory in Pasadena, Calif. This ion engine design was then used on the Deep Space 1 mission launched October 24, 1998. The electric power source was based on solar cells for the Deep Space 1 demonstration mission, but it is a natural evolution to consider radioisotope electric generators for electric propulsion missions far from the Sun.  

The specific mass of a complete propulsive powerplant has contributions from both the electric generator and ion thruster. For example, the mass of a NASA developmental 30-cm thruster, excluding the power processing unit (PPU), is estimated at 7 kg.  
  

(The PPU converts the voltage and current produced by a space electric generator into the adjustable and regulated voltages and currents needed for the ion engine.) An added gimbal assembly for maintaining proper thrust vectoring has a mass of 34 percent of the thruster mass (2.4 kg, in this case). The specific mass of present PPUs is 8 kg/kW of input power, although a reduction by a factor of rive to 10 may occur in the next decade.  
   
  

Electric Generator	kg/Kw	Lifetime (years)	Limiting Component	Estimated powerplant specific mass
Radioisotope thermoelectric generator (RTG)	197	> 20 demonstrated	Dopant Migration	224
Modular RTG (MOD-RTG)	127	> 8 estimated   1.7 demonstrated	Dopant Migration	154
Radioisotope thermophotovoltaic (RTPV)	60	> 10 estimated	Neutron Damage to GaSb cells	87
Free piston Stirling engine dynamic isotope power system (FPSE-DIPS)	118	> 10 estimated	Two moving parts	145
Alkali metal thermal to electric converters (AMTEC)	60	> 10 estimated	Porous electrode grain growth	87

Table I. Comparison of radioisotope electric generators and estimated powerplant specific masses using a 0.S-kW, 30-cm xenon thruster (Source: Fermilab)  

The specific mass of the thruster unit will depend on the input power, the total number of thrusters (including spares), the thruster mass, and the gimbal mass.  

In the last column of Table I are listed the estimated specific masses of hypothetical powerplants constructed from a NASA 30cm thruster with 0.5 kW of input power supplied by the different radioisotope electric generators. These are probably conservative estimates since PPU and ion thruster specific masses will probably decrease with development. A 30-cm ion thruster operating at 0.5 kW input power has a projected lifetime of about nine years. Assuming a total thruster efficiency of 75 percent, all the technologies beyond the standard RTG can give effective powerplant specific masses in the 100-200kg/kW range, suitable for propelling small sciencecraft.  
  

Pluto rendezvous missions  

Planetary rendezvous missions are very demanding due to the need to carry extra propellant for deceleration at the target planet.  

Aerobraking is useful for planets with atmospheres, but this option is not available for the Pluto-Charon system. Because of the high transit velocity, electric propulsion is perhaps the only viable near-term solution to achieve fast Pluto rendezvous. Electric propulsion also enables an extended science mission with orbital transfers between Pluto and Charon (its moon) for closer investigations.  

To maximize the payload delivered for a mission along a low-thrust trajectory, both the rocket configuration (relative masses of powerplant and propellant) and the thrust program (thrust magnitude and direction versus time) are typically optimized. For small sciencecraft, ion thrusters with constant propellant velocity and thrust magnitude are adopted here becanse of their simplicity and long life.  

Previous work on planetary rendezvous missions to the outer Solar Systemm established a relatively simple rocker configuration and constant-thrust program that gives approximately the same flight rime for a given payload as programs quoted in the lirerature but uses a smaller powerplant. This may be advantageous for small sciencecraft where the cost of the radioisotope electric generator may be a large fraction of the total mission cost.  

The thrust program is illustrated in Fig. 2 and has a constant thrust directed at a constant angle relative to the velocity vector during each powered phase. A coast period, during which there is no thrust, is introduced to minimize the flight rime for a given payload fraction.  

Table II dëscribes flight rime, total thrust rime, and optimal coast period for Pluto rendezvous starting from Earth orbit. To simplify the calculations, Pluto's orbit was taken as being in the same orbital plane as the Earth with an eccentricity of 0.25. Launch from Earth occurs in 2010 for all cases.  

Powerplant specific masses of 100 and 200 kg/kW are used, characteristic of the range for near-term REP systems. A fixed payload mass fraction of 0.25 for the ion rocket is assumed since flight rimes quickly increase for larger values and do not decrease much for smaller values. Gravitational effects of the PlutoCharon system are not included in the trajectory analysis, so the propellant mass of the ion rocket refers only to that needed on the lowthrust trajectory to achieve rendezvous. Any mission-specific propellant needed for orbital entry and maneuvers is assumed to be included in the payload mass allocation.  
 

 
Figure 2: Constant thrust program for planetary rendezvous missions in the ecliptic between Earth (1 AU - Astronomical Unit) and an orbital point at distance R from the Sun. During the accelaration phase tauA, the thrust vector T is parallel to the velocity vector v.There is no thrust during the coast period tauC.During the orbitmatch phase tauM, the thrust vector and velocity vector are separated by a aconstant angle Omega (Source:FermiLab)

Escape from a low Earth orbit (LEO) of 320 km is achieved by chemical propulsion for small probes, and ion thrusting commences immediately after Earth escape. The minimal Earth escape velocity (from a 320-km orbit) is about 10.9 km/s and results in no hyperbolae excess velocity relative to Earth (the object's remaining velocity after it has escaped from the gravitational sphere of influence).  

Providing a hyperbolic excess velocity with the chemical rocket at Earth escape reduces the amount of time needed by the ion rocket to spiral out of the inner Solar System.  

Pluto rendezvous can be achieved by small REP sciencecraft in only 10 to 20 years when excess velocities of several km/s are supplied at Earth escape. As the hyperbolic excess velocity is increased, there is proportionally less improvement in flight rime with lower specific mass. Thus, there is no reason to delay a Pluto science prograin in the hope of getting very low specific mass powerplants. Table II also gives the mass ratio in LEO of the chemical rocket system (oxygen-hydrogen, dry mass of 15 percent) and the REP sciencecraft for the different hyperbolic excess velocity values applied at Earth escape. Even for a 400-kg sciencecraft (100-kg payload, 100-kg REP unit, and 200 kg of propellant) given a 10 km/s excess velocity, the LEO mass is only about 5000 kg.  

Heliopause orbiter missions  
The Voyagerprobes (NN, Apr. 1999, p. 29) launched in the 1970s to explore the outer planets are leaving the Solar System at about 3 astronomical units (AUs) per year. (An AU is the distance from Earth to Sun, and 1 AU/ yr = 4.74 km/s.) They may reach the heliopause (the boundary between the Sun's magnetosphere--the solar wind--and interstellar gas, ~100 AU) between 2005 and 2010.  

Missions have been proposed to send dedicated probes out of the Solar System on fast escape trajectories to explore the heliopause and near-interstellar space.  

Becanse of the scientific interest in the temporal evolution of plasmas and particle transport in the region of the heliopause, as well as  

NASA's Deep Space I demonstration mission (shown here in an artist's conception), launched October 24, 1998, used a 30-cm-diameter, 2-kW xenon ion engine powered by solar cells, tt is a natural evolution, however, to consider radioisotope electric generators for electric propulsion missions far from the Sun. The spacecraft flew within an estimated 26 km (I 6 miles) of asteroid 9969 Braille on July 28, but a tracking problem affected the taking of black-and-white photos during the fiyby. On July 30, the spacecraft's ion engine started a burn that would continue thrusting almost continuously for about three months. This phase is in preparation for flybys of two comets during a possible extended mission after the September 18, 1999 conclusion of Deep Space I's primary mission.  

Because of their Iow specific mass, both the radioisotope thermophotovoltaic (RTPV) and alkali metal therrnal-to-electric conversion (AMTEC) generators are potential candidates for the ARP5 on the proposed Pluto-Kuiper Express mission, depicted here in an artist's conception (shown is Pluto and its moon Charon). If successfully cleveloped, either of these generators could becorne the prototypical generator for future radioisotope electric propulsion systems. (Source: NASAEJPL/Caltech)  

in extending stellar distance measurements using the parallax technique with a deep-space telescope over longer baselines, this article considers the use of REP for sending a sciencecraft to a 100-AU orbit for an extended mission.  

REP enables economical deceleration of a rapidly moving robotic craft far from the Sun.  

For high specific mass REP powerplants, however, using low-thrust propulsion alone for both the solar escape and deceleration phases can resuit in long mission times due to the long distances. For example, the rendezvous flight time from Earth to I00 AU is 50 years for a specific mass of 200 kg/kW and 36 years for 100 kg/kW, assuming minimal Earth escape velocity and a payload mass fraction of 0.25 for the ion rocket.

These times seem excessively long for such a mission. To reduce the fiight time to the 100 AU distance, this article uses a hybrid propulsion scheme in which chemical propulsion and gravity assist are applied in the inner Solar System, and electric propulsion is used in the outer Solar System.  

The Earth-Jupiter gravity assist (EJGA) maneuver9 is adopted here for the 100 AU orbiter missions. Application of low-thrust propulsion occurs only after the Jupiter encounter, once the chemical rocket has been jettisoned. Transfer to Jupiter is achieved by placing the craft on a so-called  resonance orbit (period 2 years, perihelion [orbital point nearest to the sun] of 1 AU) and performing a small velocity change maneuver at the aphelion (orbital point farthest from the sun) of  

2.175 AU. This enables an Earth gravity assist (EGA--which involves pickup of additional velocity via Earth's gravitational field during a flyby past Earth) 1.82 years after launch, thereby saving considerable chemical propellant compared to a conventional Hohmann transfera to Jupiter. The EGA is a well-established technique that was used for the Galileo mission to Jupiter and the Cassini mission to Saturn.  

The EGA transfer also makes it economical to take along extra chemical propellant to Jupiter for a powered gravity assist maneuver there. Performing a velocity change maneuver deep in Jupiter's gravitational potential can result in a large increase in hyperbolic excess velocity. For example, an unpowered JGA (Jupiter gravity assist) results in a 2.7AU/yr excess velocity, while adding a 3-krn/s periapsis (the orbital point closest to the planet) burn ives a remarkable 6 AU/yr excess velocity. Because a larger periapsis burn at Jupiter requires more storable propellant to be boosted out of Earth orbit, velocity changes at Jupiter greater than 3 km/s resuit in very high mass ratios in LEO (the ratio of the LEO chemical booster rocket mass---including the ion rocker and its payload--and the ion rocket mass) and excessive LEO masses even for small sciencecraft (see Table III).  

Table III describes results on flight time, total thrust time, and optimal coast period for 100 AU rendezvous missions assuming REP powerplants with a specific mass in the 100to 200-kg/kW range. The payload mass fraction ofthe REP sciencecraft is 0.25. The cases for no EJGA and EJGA are presented. Even though the Earth-Jupiter transfer takes 3.54 years, the benefit of the EJGA in reducing flight time is substantial.  

For the EJGA trajectories, it is interesting to note how little reduction in flight time is obtained by reducing the powerplant specific mass by a factor oftwo or adding a Jupiter periapsis burn with a velocity change of 3 km/s alone. Only the combination makes a significant reduction in flight time. For the purpose of carrying out a timely science prograin, it probably makes little sense to wait for powerplants with specific masses below 100 kg/ kW to be developed since the savings in mission rime will likely be lost in the development time. With EJGA, near-term REP sciencecraft can reach heliopause orbits within 20 to 30 years.  

Gravitational lens missions  
Sending sciencecraft farther out of the Solar System and into near-interstellar space with short mission times becomes increasingly difficult because of the relative insensitivity of flight time to the powerplant specific mass. In Ref. 9, the use of REP for sending robotic probes on fast solar escape trajectories out to several hundred AU was studied. In particular, flight times for getting to the tirst gravitational lens focusb of the Sun (550 AU) were presented for REP augmented by powered Jupiter gravity assist. A range of powerplant a specific masses from 20 to 200 kg/kW was used in the previous study (Ref. 9) to determine the dependence of flight time on this parameter. For specific masses of 100 to 200 kg/ kW, it was found that large periapsis burns at JGA were required to reduce che time to fly out to 550 AU. The payload mass fraction of the REP probe was taken as 0.25.  

To investigate the tradeoff in payload mass fraction versus flight time, Fig. 3 describes the results for the flight time to 550 AU for payload fractions of 0.1, 0.25, and 0.4. It shows che flight time as a function of powerplant specific mass for these payload mass fractions. A 3-km/s periapsis burn is used during the JGA, and the flight cimes include the 3.54-year transit cime between Earth and Jupiter. Note that payload mass fractions above 0.25 are probably too costly in flight time unless che specific mass is below 70 kg/ kW. Reducing the payload mass fraction from 0.25 to 0.1 reduces the flight time by only about 12 percent. For a payload mass fraction of 0.25, Che flight time ranges from 45 to 52 years for specific mass of 100 to 200 kg/kW, and falls below 30 years only for specific masses of less than 20 kg/kW.  

For the results in Fig. 3, the thrust duration of the ion rocker (commencing after che Jupiter encounter) was fixed at 10 years, assuming that this was the reliability limit of the propulsion unit. Typically for low-thrust escape from the Solar System, there is an optimal thrust duration to minimize the flight Cime out to a particular distance. This optimum arises because of the competing effects of initial acceleration and final velocity. For shorter thrust Cimes, the ion rocket's design acceleration would be higher (reducing the initial escape period), while for longer thrust times, the final velocity would be higher (reducing the final coast period). For a payload mass fraction of 0.25, the optimal thrust time was found to vary from 15 years to 26 years in the specific mass range of 20 to 200 kg/kW for flights to 550 AU. But Che actual flight Cime was in fact quite insensitive to this choice of optimum, being reduced by only 3 to 8 percent relative to that for a 10-year thrust period in this specific mass range.  

Because of their low specific mass, both the radioisotope thermophotovoltaic (RTPV) and alkali metal thermal-to-electric conversion (AMTEC) generators are potential candidates for the ARPS on the proposed Pluto-Kuiper Express mission, depicted here in an artist's conception (hown is Pluto and Charo). If successfully developed, either of these generators could become the prototypical generator for future radioisotope electric propulsion systems (soucrce NASA Caltech)

Nasa's Deep Space I demonstration mission (shown here in an artist's conception),launched October 34, 1998, used a 30 cm diameter, 2 kw xenon ion engine powered by solar cells. It is a natural evolution, however, to consider radioisotope electric generators for eletric propulsion missions far from the Sun. The spacecraft flew within an estimated 26 km (16 miles) of asteroid 9969 Braille on July 28, but a tracking problem affected yhe taking of black and white photos during the flyby. on July 30, the spaceccraft's ion engine started a burn that could continue thrusting almost continuously for about three months. This phase is in preparation for flybys of two comets during a possible extended mission after Sept. 18, 1999 conclusion of Deep Space I's primary mission.

 Fig. 3. Flight time to reach 550 AU as a function ofthe powerplant specific mass for different payload rnass ratios of the ion rocker (ratio of the ion rocket payload mass tothe initial ion rocket rnass). The change in velocity at Jupiter periapsis is 3 km/s. The Iow-tbrust period is I0 years, commencing after the Jupiter encounter. The flight time includes the 3.54 years needed for the Earth-to-Jupiter transit.  (Source: Fermilab)  

placed at a 100-AU parking orbit in 20 to 30 years when REP is combined with JGA. Finally, REP and powered JGA enable a sciencecraft to fly out tothe tirst gravitational lens focus of the Sun (550 AU) in 40 to 50 years. In ail cases, flight times are relatively insensitive to powerplant specific mass.  

For the purpose of performing a timely scientific program, the sure of hardware development time and mission cime is an important consideration. It may be îaster to simply use high specific mass REP systems rather than delay space science programs while waiting for low specific mass propulsion systems to be developed.  

References  

1. G. L. Bennett, R. J. Hemler and A. Schock, Space Nuclear Power: An Overview, J. Propulsion and Power, Vol. 12, No. 5, pp. 901-910, 1996.  

2. R. F. Hartman, J. R. Peterson and W. Barnet, Modular RTG Status, Ninth Symposium on Space Nuclear Power and Propulsion, A.I.P. Conf. Proc.  No. 246, pp. 177-181, 1992.  

3. M. D. Morgan, W. E. Home and P. R. Brothers, Radioisotope Thermophotovoltaic Power System Utilizing the GaSb IR Photovoltaic Cell, Tenth Symposium on Space Nuclear Power and Propulsion, A.I.P. Conf. Proc. No. 271, pp. 313 318, 1993.  

4. D. J. Bents, J. G. Schreiber, C. A. Withrow, B. I.  McKissock, and P. C. Schmitz, Design of Small Stirling Dynamic Isotope Power System for Robotic Space Missions, Tenth Symposium on Space Nuclear Power and Propulsion, A.I.P. Conf. Proc. No.  271,pp. 213 221, 1993.  

5. R. K. Sievers, T. K. Hunt, J. F. Ivanenok, J. E.  Pantolin, and D. A. Butkiewicz, Modular Radioisotope AMTEC Power System, Tenth Symposium on Space Nuclear Power and Propulsion, A.I.P. Confi Proc. No. 271, pp. 319-324, 1993.    

6. A. Schock, Comparison of Radioisotope Space Power Systems Using Alternative Conversion Options, AIAA Paper No. 96-0120, 34th Aerospace Sciences Meeting, Reno, Nev., 1996.  

7. A. Schock, Design and Performance of Radioisotope Space Power Systems Based on OSC Multitube AMTEC Converter Designs, in Proc. of Che 32nd Intersociety Energy Conversion Engineering Confercnce, Paper No. 97530, July 27-August 1, 1997.  

8. R. J. Noble, Radioisotope Electric Propulsion of Small Payloads for Regular Access to Deep Space, AIAA Paper No. 93-1897, 29th Joint Propulsion Conf., Monterey, Calif., 1993.  

9. R. J. Noble, Radioisotope Electric Propulsion for Robotic Science Missions to Near-lnterstellar Space, J. British Interplanetary Society, Vol. 49, pp.  322-328, 1996.  

10. R. J. Noble, Radioisotope Electric Propulsion for Small Robotic Space Probes, J. British Interplanetary Society, Vol. 49, pp. 455-468, 1996.  

11. J. S. Sovey, J. A. Hamley, M. J. Panerson, V.  K. Rawlin, and T. R. Sarver-Verhey, lon Thruster Development at NASA Lewis Research Center, Tenth Symposium on Space Nuclear Power and Propulsion, A.I.P. Conf. Proc. No. 271, pp.  1309-1316, 1993.  

12. G. Saccoccia, European Electric Propulsion Activities in the Era of Applications, J. British Interplanetary Soc., Vol. 48, pp. 487-500, 1995.  

13. P. M. Latham et al., The UK-25 lon Thruster, J.  British Interplanetary Soc., Vol. 48, pp. 501-506, 1995.  

14. See the NASA Jet Propulsion Laboratory Wcb page for information on the Deep Space I mission at the URL address <http://www.jpl.nasa.gov>.  

15. V. R. Eshleman, "Gravitational Lens of che Sun: Its Potential for Observations and Communications over Interstellar Distances," Science, Vol. 205, pp.  1133-1135 (14 September 1979).