The Dipole Drive: A New Concept for Space Propulsion

One reason we look so often at sail technologies in these pages is that they offer us ways of leaving the propellant behind. But even as we enter the early days of solar sail experimentation in space, we look toward ways of improving them by somehow getting around their need for solar photons. Robert Zubrin’s work with Dana Andrews has helped us see how so-called magnetic sails (magsails) could be used to decelerate a craft as it moved into a destination system. Now Zubrin looks at moving beyond both this and solar wind-deflecting electric sails toward an ingenious propellantless solution. Zubrin presented the work at last April’s Breakthrough Discuss meeting, and today he fills us in on its principles and advantages. Read on for a look at a form of enhanced electric sail the author has christened the Dipole Drive.

by Robert Zubrin


The dipole drive is a new propulsion system which uses ambient space plasma as propellant, thereby avoiding the need to carry any of its own. The dipole drive remedies two shortcomings of the classic electric sail in that it can generate thrust within planetary magnetospheres and it can generate thrust in any direction in interplanetary space. In contrast to the single positively charged screen employed by the electric sail, the dipole drive is constructed from two parallel screens, one charged positive, the other negative, creating an electric field between them with no significant field outside. Ambient solar wind protons entering the dipole drive field from the negative screen side are reflected out, with the angle of incidence equaling the angle of reflection, thereby providing lift if the screen is placed at an angle to the plasma wind. If the screen is perpendicular to the solar wind, only drag is generated but the amount is double that of electric sail of the same area. To accelerate within a magnetosphere, the positive screen is positioned forward in the direction of orbital motion. Ions entering are then propelled from the positive to the negative screen and then out beyond, while electrons are reflected. There are thus two exhausts, but because the protons are much more massive than the electrons, the thrust of the ion current is more than 42 times greater than the opposing electron thrust, providing net thrust. To deorbit, the negative screen is positioned forward, turning the screen into an ion reflector. The dipole drive can achieve more than 6 mN/kWe in interplanetary space and better than 20 mN/kWe in Earth, Venus, Mars, or Jupiter orbit. In contrast to the electric sail, the ultimate velocity of the dipole drive is not limited by the speed of the solar wind. It therefore offers potential as a means of achieving ultra-high velocities necessary for interstellar flight.


The performance of rockets as propulsion systems is greatly limited by their need to carry onboard propellant, which adds to the mass which must be propelled exponentially as the extent of propulsive maneuvers is increased. For this reason, engineers have long been interested in propulsion systems that require no propellant.

The best known propellantless system is the solar sail, which derives its thrust by reflecting light emitted by the Sun. Solar sails are limited in their performance however, by their dependence upon sunlight, which decreases in strength with the square of the distance, and the laws of reflection, which dictate that the direction of thrust can only lie within 90 degrees of the vector of sunlight. Moreover, because photons move so swiftly, the amount of thrust that can be derived by reflecting light is at best 0.0067 mN/kW (at 100% reflectance, full normal incidence), which means that very large sails, which necessarily must have significant mass and be difficult to deploy, must be used to generate appreciable thrust. As a result, while solar sails have been studied since the time of Tsiolokovsky [1], we are only now beginning to experiment with them in space.

An alternative to the solar sail is the magnetic sail, or magsail, which was first proposed by Zubrin and Andrews in 1988, and subsequently analyzed extensively by them in a variety of further papers [2,3] in the 1990s. The magnetic sail uses a loop of superconducting wire to generate a magnetosphere to deflect the solar wind. Assuming the development of high temperature superconducting wire with the same current density as existing low temperature superconductors, a magsail should be able to generate significantly higher thrust to weight than is possible with solar sails. However such wire has yet to be developed.

Another propellantless propulsion system of interest is the electric sail [4], which like the magsail operates by deflecting the solar wind, in its case by using an electrostatic charge. As a result, like the magsail, the classic electric sail (electric sail) cannot operate inside of a planetary magnetosphere other than as a drag device, has its thrust decrease with distance from the Sun, and is limited in the potential direction of its thrust. Because of the low momentum density of the solar wind, electric sails must be even bigger than solar sails. However, because only sparsely spaced thin wires are needed to create sail area, higher thrust to mass ratios can be achieved than are possible using solar sails which require solid sheets of aluminized plastic.

Electrodynamic tethers [5] have also been proposed, which use the interaction of a current in a tether with the Earth’s geomagnetic field to produce thrust. In addition to facing a variety of engineering and operational issues, however, such systems can only operate in a planetary magnetic field and can only thrust in a direction normal to the field lines, a consideration which limits their applicability.

Finally, we note recent claims for a system called the EM Drive [6], which according to its proponents can generate about 1 mN/kWe, in any direction, without the use of propellant, an external light source or plasma wind, or magnetic field. Such performance would be of considerable interest. However, as it appears to contradict the laws of physics, there is reason to suspect that the measurements supporting it may be erroneous.

As a result, there clearly remains a need for a new type of propellantless propulsion system, which can operate both inside and outside of a planetary magnetosphere, can thrust in a multitude of directions, and which is not dependent upon sunlight or the solar wind as a momentum source. The dipole drive is such a system.

The Dipole Drive

The principle of operation of the dipole drive while accelerating a spacecraft within a planetary magnetosphere is illustrated in Fig. 1 below.

Fig. 1. The Dipole Drive Accelerating within a Magnetosphere.

In Fig. 1 we see two parallel screens, with the one on the left charged positive and the one on the right charged negative. There is thus an electric field between them, and effectively no field outside of them, as on the outside the field of each screen negates the other. There is also a voltage drop between the two, which for purposes of this example we will take to be 64 volts.

Protons entering the field region from the left are accelerated towards the right and then outward through the right-hand screen, after which they escape the field and experience no further force. Protons entering from the right are reflected towards the right, adding their momentum to that generated by the protons accelerated from left to right. There is thus a net proton current from left to right, and a net proton thrust towards the left.

In the case of electrons, the situation is exactly the opposite, with a net electron current from right to left, and a net electron thrust towards the right. Note that while electrons entering from the right will be greatly accelerated by the field, reflected electrons will only be reflected with their initial velocity. There will also be an electron current through the outside plasma to neutralize the net proton flow to the right.

Because space plasmas are electrically neutral, the number density of both electrons and ions (which for the moment we will consider to be protons, but may which – advantageously – be heavier species, as we shall discuss later) will be the same, so the proton and electron electrical currents will be equal, as will the power associated with each of them. However because the mass of a proton is about 1842 times as great as the mass of an electron, the thrust of the proton current will be about 43 times greater than the opposing electron current thrust (because the momentum of particles of equal energy will scale as the square root of their mass, sqrt(1842)=43) and the system will generate a net thrust. The acceleration of the electrons is a form of drag, which is provided for by loss of spacecraft kinetic energy. It therefore could, in principle be used to generate electric power, partially compensating for the power consumed to accelerate the protons. In the following examples, however, we will assume that there is no provision for doing this, i.e. that the efficiency of any such energy recovery is zero.

To see what the performance of a dipole drive might be, let us work an example, assuming a 500 W power source to drive the system. The electron current negates about 2% of the thrust (1/43rd) produced by the proton current. The maximum possible jet power is thus about 490 Wj. Assuming additional inefficiencies, we will round this down to 400 Wj, for a total system electrical to jet power efficiency of 0.8.

A Coulomb of protons has a mass of 0.011 milligrams. If the jet power is 400 W, and the potential difference is 64 V, so the proton current will be 6.25 A, and have a mass flow of 0.0652 mg/s.

The relationship of jet power (P) to mass flow (m) and exhaust velocity (c) is given by:

P = mc2/2                                                                         (1)

Taking P = 400 W and m = 0.0652 mg/s, we find that c= 110,780 m/s. Since thrust (T) is given by T=mc, we find:

T = mc = 7.2 mN                                                            (2)

This is a rather striking result. It will be recalled that the electrical power driving this system is 500 W. So what we are seeing here is thrust to power ratio of 14.4 mN/kWe, more than ten times better than that claimed for the EM Drive, but done entirely within the known laws of physics!

If it is desired to deorbit (decelerate) a spacecraft, the direction of the screens would be reversed, with the negative screen leading in the direction of orbital motion. In this case, the screens would become a proton reflector. An electric sail could also be used as a drag device to serve the same purpose. However, because the dipole drive doesn’t merely create drag against passing protons, but reflects them, it would create twice the drag of an electric sail of the same area. If the dipole drive is positioned obliquely to the wind angle, it can reflect protons, with the angle of incidence equaling the angle of reflection. For example, if it is tilted 45 degrees to the wind, a force will be generated perpendicular to the wind, that is “lift” will be created. Such maneuvers could also be done with the dipole drive in acceleration mode, deflecting protons to combine lift with thrust. Using this capability, a dipole drive propelled spacecraft in orbit around a planet could execute inclination changes.

To summarize, in contrast to the electric sail which can only create drag against the wind to lower its orbit, the dipole drive can thrust in any direction, raising or lowering its orbit or changing its orbital inclination. In addition, when used as a drag device, the dipole drive can create twice the drag per unit area as the electric sail.

The Dipole Drive in Planetary Orbit

Let us therefore analyze the system further. The dipole drive exerts no field outside of its screens, so the only plasma it collects is the result of its own motion through the surrounding medium. So how big does its screen need to be?

We consider first the case of the above described dipole drive system operating in LEO at an altitude of 400 km, being used to thrust in the direction of orbital motion. It is moving forward at an orbital velocity of 7760 m/s. The average density of ions at this altitude is about 1,000,000 per cc. Assuming (conservatively) that all the ions are protons, the required ion mass flow of 0.0652 mg/s would be swept up by a screen with a radius of 127 m.

It may be noted however, that at 400 km altitude there are also O+ ions, each with a mass 16 times that of a proton, with a numerical density of about 100,000/cc. These therefore more than double the ion mass density provided by the protons alone. If these are taken into account, the required scoop radius would drop to about 80 m.

Another way to reduce the scoop size would be by going to higher voltage, so that more power can be delivered to a smaller number of ions. If, for example, we quadrupled the voltage to 256 volts, the exhaust velocity would double, to 222 km/s, allowing us to cut the mass flow by a factor of four, and the scoop radius by a factor of two, to just 40 m. The thrust, however, would be cut in half, giving us 3.6 mN/kWe.

As we go up in altitude, the plasma density decreases, as does the orbital velocity, requiring us to go to larger scoops. Examples of 500 W dipole drive systems operating at a variety of altitudes are provided in Table 1. In Table 1, Vo and C are orbital velocity and exhaust velocity, in km/s.

Table 1. Dipole Drive Systems Operating in Earth Orbit (Power=500 W)

It can be seen that the dipole drive is a very attractive system for maneuvering around from LEO to MEO orbits, as the high ion density makes the required scoop size quite modest. It should be emphasized that the above numbers are for a 500 W system. If a 5 W dipole drive thruster were employed by a microsatellite, the required scoop areas would be reduced by a factor of 100, and the radius by a factor of 10.

It may be noted that Mars, Venus and Jupiter all have ion densities in low orbit comparable to those above. For example, Mars has 500,000/cc at 300 km, Venus has 300,000/cc at 150 km, and Jupiter has 100,000/cc at 200 km, making the dipole drive attractive for use around such planets as well. Many of the moons of the outer planets also have ionospheres, and the dipole drive should work very well in such environments.

As one ascends to higher orbits, the density of ions decreases dramatically, while the orbital speed decreases as well. For example, in GEO, the ion density is only about 20/cc, while the orbital velocity is 3 km/s. These two factors combine to make much larger scoops necessary. So, for example, in GEO, a 500 W dipole drive operating at 1024 volts would need a scoop 3.6 km in radius.

Because the effectiveness of the dipole drive decreases at higher altitudes while operating within the magnetosphere, the best way for a dipole drive propelled spacecraft to escape the Earth is not to continually thrust, as this would cause it to spiral out to trans GEO regions where it would become ineffective. Rather, what should be done is to only employ it on thrust arcs of perhaps 30 degrees around its perigee, delivering a series of perigee kicks that would raise its apogee on the other side of its orbit higher and higher until it escaped the magnetosphere and became able to access the solar wind.

The Dipole Drive in Interplanetary Space

The dipole drive can also operate in interplanetary space. Compared to planetary orbit, the ion densities are lower, but this is partially compensated for by much higher spacecraft velocities relative to the plasma wind. As a result, the required scoop sizes are increased compared to planetary orbital applications, but not by as much as considerations of ion density alone might imply.

Let us consider the case of a dipole drive traveling in heliocentric space at 1 AU, positioned at an angle of 45 degrees to the wind, with its negative screen on the sunward side. It would thus reflect solar wind protons 90 degrees, thereby accelerating itself forward in the direction of orbital motion. A diagram showing the dipole drive operating as a sail in interplanetary space is shown in Fig. 2.

Fig. 2 The Dipole Drive Operating as a Sail in Interplanetary Space.

The solar wind has a velocity of 500 km/s, so to insure reflection, we employ a voltage of 2028 volts, sufficient to reverse the motion of a proton moving as fast as 630 km/s. With a density of 6 million protons per cubic meter, the wind has a dynamic pressure of 1.25 nN/m2. As the sail is positioned 45 degrees obliquely to the wind, its effective area will be reduced by a factor of 0.707, with the thrust reduced to 0.9 nN/m2. In this case, virtually all of the protons hitting the sail will be coming from the sunward side, and since they are reflected without adding any kinetic energy, no power is required to drive them. However, we still have an electron current coming from the sunward side being accelerated outward. This requires power. With 500 W, total radial thrust would be 1.27 mN, with 1.27 mN also delivered in the direction of orbital motion, for a L/D ratio of 1. The total effective screen area would therefore need to be 1,414,000 m2, with an actual area of 2,000,000 m2, requiring a radius of 798 m. Total thrust to power would be 3.6 mN/kWe.

If instead we had not concerned ourselves with obtaining complete deflection of each particle, we could have used a lower voltage. This would increase the thrust per unit power, but increase the required sail area for a given amount of thrust. So, for example, if we chose 512 volts, we would have a total thrust of 3.6 mN, for a thrust/power ratio of 7.2mN/kWe, but need a sail radius of 1127 m.

It may be noted that all of these results are for a 500 W dipole drive. A microsatellite might employ a 5 W dipole drive, in which case the required scoop radii would drop by a factor of 10.

The thrust and diameter of a 1 kWe dipole drive system operating as a solar wind sail in interplanetary space at 1 AU is shown in fig. 3.

Fig. 3. Thrust and Diameter of a 1 kWe dipole drive system operating as a solar wind sail in interplanetary space.

Use of the Dipole Drive for Interstellar Flight

In contrast to the electric sail, the dipole drive can be used to accelerate a spacecraft at velocities greater than that of the solar wind. For example, consider a spacecraft moving away from the Sun at a velocity of 1000 km/s. The solar wind is following it at a velocity of 500 km/s, so relative to the spacecraft there is a wind moving inward towards the sun at a velocity of 500 km/s. In this case, to accelerate the spacecraft would direct its positive screen away from the sun. This would cause it to accelerate protons sunward, while reflecting electrons outward, for a net outward thrust. At 500 km/s the protons are approaching the spacecraft with a kinetic energy equal to 1300 volts. It can be shown that employing a screen voltage difference that is about triple the kinetic voltage produces an optimal design for an accelerating system, while one using a voltage difference equal to the kinetic voltage is optimal for deceleration. This is illustrated in figs 4 and 5 which respectively show the kinetic voltage as a function of velocity, and the relative power/ thrust and area/thrust ratios of the spacecraft as a function of the dimensionless parameter Z, where Z=(engine voltage)/(kinetic voltage.)

Fig 4. Kinetic Voltage as a function of spacecraft velocity.

Fig 5. Relative Power/Thrust and Area/Thrust as a function of Z=(engine voltage)/(kinetic voltage.) There is a step factor of 2 increase in thrust during deceleration when Z reaches 1, because protons are reflected. For acceleration, Power/Thrust ~ 1 + sqrt(1+Z), while Area/Thrust ~ 1/(-1 + sqrt(1+Z)).

If we add 3900 volts to the incoming protons, quadrupling their energy, we will double their velocity relative to the spacecraft, thereby providing an effective exhaust velocity of 500 km/s. The solar wind has a density of 6 million protons/m3 at 1 AU, with ambient density decreasing to 1 million/m3 in interstellar space. If we take the former value, we get a thrust of (1.67e-27 kg/proton)(500,000m/s)2(6,000,000/m3) = 2.5 nN/m2. If we take the latter value, it would be 0.42 nN/m2. The proton current at the smaller value would be 80 nA/m2, which at 3900 volts works out to 0.312 mW/m2. The thrust to power ratio would therefore be 1.35 mN/kW. (This ratio would also hold true at the 1 AU value, but the magnitudes of both the thrust and power per unit area would be six times greater.)

If a dipole drive powered spacecraft were receding 500 km/s directly away from the Sun, it would see no relative wind and thus produce no thrust. However, like a modern sailboat that can sail faster crosswind than downwind, because it can generate lift, the dipole drive can get to speeds above 500 km/s by sailing across the wind. As the spacecraft’s crosswind speed increases, it becomes advisable to turn the sail to ever greater angles to the solar wind and increasingly normal to the crosswind. As this occurs, the L/D resulting from solar wind reflection increases while the total solar wind thrust decreases. At the same time, however, thrust resulting from the acceleration through the screens of crosswind protons increases, maintaining total thrust constant at ever higher L/D (relative to the solar wind) levels. Once the crosswind velocity exceeds the solar wind velocity the solar wind becomes increasingly irrelevant and the dipole drive becomes a pure acceleration system, driving the incoming crosswind plasma behind it to produce thrust,

As the speed of the spacecraft increases relative to the wind, it is necessary to increase the voltage in order maintain thrust/power ratio efficiency. For example, let’s say we want to achieve 3000 km/s, or 0.01c. Then the kinetic energy equivalent voltage of the approaching protons would be 47 kV. So, to double this velocity we need to quadruple the total voltage, or add a sail voltage drop of 141 kV. The proton current would have a value of 480 nA/m2, with a power of 68 mW/m2. The thrust would be 15.1 nN/m2, for a thrust to power ratio of 0.22 mN/kW.

It may be observed that since the necessary voltage increases as the square of the velocity, with power increasing with voltage but thrust increasing with velocity, the thrust to power ratio of the dipole drive decreases linearly with velocity. This puts limitations on the ultimate velocity achievable. For example, the most optimistic projections for advanced large space nuclear power systems project a mass to power ratio of 1 kg/kW. If we accept this number, then, neglecting the mass of any payload or the dipole drive system itself, then the system described in the previous paragraph performing with a thrust to power ratio of 0.22mN/kilowatt at 3000 km/s would have an acceleration of 0.00022m/s2, or 7 km/s per year. The average acceleration getting up to 3000 km/s would be twice this, so the spacecraft would take 214 years to reach this speed. During this time it would travel 1.07 light years. To reach 6000 km/s (0.02 c) starting from negligible velocity would require 857 years, during which time the spacecraft would travel 8.57 light years. The performance of such a system is shown in Table 2. Note 63,000 AU = 1 light year. The performance shown assumes an advanced 1 kg/kWe power supply. If a more near-term power system with a higher mass/power is assumed, the time to reach any given distance increases as the square root of the mass/power ratio. So for example, if we assume a conservative near-term space nuclear power reactor with a mass/power ratio of 25 kg/kW, the time required to reach any given distance would increase by a factor of 5.

Table 2. Advanced Dipole Drive Performance for Ultra High-Speed Missions (1 kg/kW power)

It can be seen that advanced dipole drive spacecraft could be quite promising as a method of propulsion for missions to near interstellar space, for example voyages to the Sun’s gravitational focus at 550 AU. Unless much lighter power systems can be devised than currently anticipated however, they would still require centuries to reach the nearest stars. Power beaming may provide an answer. However such technologies are outside the scope of this paper.

If a spacecraft has been accelerated to interstellar class velocities, whether by means of the dipole drive or any alternative technology, the dipole drive provides a means of deceleration without power (it could actually generate power) by creating drag against the relative plasma wind. This feat can also be done by a magnetic sail or an electric sail. However because it can also create lift as well as drag, the dipole drive offers much greater maneuverability during deceleration as well as a means to freely maneuver within the destination solar system after arrival.

Dipole Drive Design Issues

Let us consider the case of a 2 kg microsatellite operating in LEO, with 5 W of available power to drive a dipole drive. (Note, a typical CubeSat has a mass of 1.3 kg. At 20 kg/kWe, a 5 W solar array should have a mass of about 0.1 kg.) If we operate it with a voltage of 16 Volts, it will produce 28.8 mN/kWe, or 0.144 mN thrust over all. It would have an acceleration of 0.000072 m/s2. This would allow it to generate a ?V of 2288 m/s in a year, sufficient to provide extensive station keeping propulsion, substantially change its inclination, or to raise it from a 400 km altitude orbit to a 700 km orbit in 1.6 months. To generate this much thrust at 400 km would require a scoop with a radius of 16 m, while doing so at 700 km would require a scoop with a radius of 58 m. Let us assume that the scoop is made of aluminum wire mesh, using wires 0.1 mm in diameter separated by distances of 2 m. Each square meter of mesh would thus have about 1 m length of wire. This needs to be doubled as there are two meshes, one positive and one negative. Therefore, a scoop with a radius of 16 m would have a mass of 32 grams. If the propulsion system were used simply for station keeping, inclination change, or deorbit functions at the 400 km altitude, that’s all that would be needed. To operate at 700 km, a 116 gram scoop would be required. From these examples we can see that the use of the dipole drive to provide propulsion for microsatellites in LEO could potentially be quite attractive, as the modest scoop sizes required do not pose major deployment challenges.

Now let us consider a 100 kg interplanetary spacecraft in interplanetary space, operating with 500 W at a voltage of 2028 volts. From the discussion above it can be seen that this would generate about 2.54 mN of thrust in the direction of orbital motion. The scoop would need to have a radius of about 800 m. In interplanetary space, the Debye shielding length is ~60 m, and so a screen with a 20 m mesh would suffice. Such a screen would have a mass of about 8.5 kg, which would be well within the spacecraft mass budget. The 2.54 mN thrust would accelerate the spacecraft at 0.000025 m/s2. It could thus impart a V to the spacecraft of about 804 m/s per year. Higher accelerations could be provided by increasing the spacecraft power to mass ratio.

The deployment of large scoops composed of two parallel, oppositely charged meshes poses operational and design issues. Prominent among these is the fact that the two opposite charged screens will attract each other. However the total force involved is not that large. For example, let us consider a configuration consisting to two sails of 500 m radius separated by 500 m with a 2 kV potential difference. Then the electric field between them will be 4 volts/m. The area of each screen will be 785,400 m2. From basic electrostatics we have EA = Q/?, so Q, the charge of each screen will be given by Q=(4)(785,400)(8.85 e-12) = 0.000028 coulombs. The electrostatic force on each sail is given by F=QE, so the total electrostatic force of each sail will be 0.1 mN. This is about a tenth the thrust force exerted by the screens themselves. Nevertheless, as small as they are, both of these forces will need to be negated. This can be done either with structural supports or by rotating the spacecraft and using artificial gravity to hold the sails out perpendicular to the axis of rotation. An alternative is to use the self-repulsion of the charge of each sail to help hold it out flat. In such a configuration two sails held separate from each other by a boom attached to their centers could be expected to curve towards each other at their edges until the stiffening self-repulsive force on each sail from its own charge balanced the bending forces exerted by the spacecraft’s acceleration, the push of the wind, and the attractive force of the opposite sail.

One way to avoid such issues would be to design the system as a literal dipole, with a rod holding a positive charge at its end to the front of the spacecraft, and a rod holding the negative charge pointing to the rear of the spacecraft. Seen from a distance, such a configuration is electrically neutral and would exert negligible field. However, in the zone between the charges, there is a strong field from one pole to the other. Particles entering this field along the rod center lines would experience the full voltage drop. Particles entering the field at some distance from this central axis would experience a lower voltage drop. The overall functional voltage of such a system, from the point of view of power consumption and exhaust velocity, would be an average over many particles entering the dipole field at all distances from its axis. This is obviously a more complex configuration to analyze than that of the two parallel screens discussed so far, but it may be much simpler to implement in practice on an actual spacecraft.

A critical issue is the material to be used to create the dipole drive. In his original paper on the classic electric sail [4], Pekka Janhunen suggested using copper wires with diameters between 2.5 and 10 microns. This is not an optimal choice, as copper has a much lower strength to mass ratio than aluminum, and such thin strands would be quite delicate. For this reason, in the above examples we specified aluminum wire with 100-micron diameters. A potentially much better option, however, might be to use aluminized Spectra, as spectra has about 10 times the yield strength of aluminum, and roughly 1/3 the density (Aluminum 40,000 psi, 2700 kg/m3, compared to Spectra 400,000 psi, 970 kg/m3.). Spectra strands with 100-micron diameters and a coating of 1 micron of aluminum could thus be a far superior material for dipole drive system, and classic electric sails as well. An issue however is Spectra’s low melting point of 147 C. Kevlar, however, with a yield strength of 200,000 psi, a density of 1230 kg/m3, and a melting point of 500 C could provide a good compromise. Still another promising option might be aluminized strands made of high strength carbon fiber, such as the T1000G (924,000 psi, 1800 kg/m3) produced by Toray Carbon Fibers America.

Some options for dipole drive spacecraft configurations are show in in Fig. 6. As can be seen, small dipole drive systems can be used for spacecraft control, for example as an empennage. Such small dipole drive units could also be used for attitude control on non-dipole drive spacecraft, such as solar sails.

Fig. 6. Options for dipole drive spacecraft configuration. Small dipole drive systems can be used for attitude control.

As with the electric sail, the dipole drive must deal with the issue of sail charge neutralization caused by the attraction of ambient electrons to the sail’s positive screen. In reference 4, P. Janhunen showed that the total such current that an electric sail would need to dispose of would be modest, entailing small power requirements if ejected from the spacecraft by a high voltage electron gun. In the case of the dipole drive, the current would be still smaller because the spacecraft has no net charge. In addition electrons acquired by the positive screen could be disposed of by using the power source to transport them to the negative screen. Alternatively, if an electron gun were used, its required voltage would be less than that needed by an electric sail because external to the screens, the dipole drive’s field is much weaker and falls off much more quickly. For these reasons, the issue of sail charge neutralization on the dipole drive should be quite manageable.

Because the dipole drive does not interact with plasma outside of the zone between its screens, the issue of Debye shielding of its screen system to outside charges is not a concern. Debye shielding of its individual wires within screens can be dealt with by means of adequately tight wire spacing. As shown by Janhunen [4], such spacing may be quite liberal (~60 m in near Earth interplanetary space), enabling sails with very low mass to area ratios. [7]


The dipole drive is a promising new technological concept that offers unique advantages for space propulsion. Requiring no propellant, it can be used to thrust in any direction, and both accelerate and decelerate spacecraft operating within planetary magnetospheres, in interplanetary space, and interstellar space. Unlike magnetic sails and electric sails, it can generate both lift and drag, and its maximum velocity is not limited by the speed of the solar wind. Near-term dipole drives could be used to provide a reliable, low cost, low mass technology to enable propellantless movement of spacecraft from one orbit to another, to provide station keeping propulsion, or to deorbit satellites, as required. Then dipole drive could also be used as a method of capturing interplanetary spacecraft into orbit around destination planets, or of lowering the orbits of spacecraft captured into initial elliptical orbits using high thrust propulsion. The latter application is particularly interesting, because it could enable a small lightweight lunar ascent vehicle to carry astronauts home from the Moon by launching directly from the lunar surface to trans-Earth injection and then subsequently lower itself to LEO to rendezvous with a space station or reentry capsule spacecraft without further use of propellant. Such an approach could potentially reduce the mass of a manned lunar mission to within the launch capacity of a single Falcon Heavy. Because it needs no propellant, the dipole drive offers the unique advantage of being able to provide its propulsion service to any spacecraft indefinitely. While the dipole drive is most attractive in orbital space whether ambient plasma is thickest, it can be used in interplanetary space and even enable interstellar missions as well, becoming more attractive for such applications as ancillary technologies, such as power generation evolve.

There are many technical issues that need to be resolved before practical dipole drive spacecraft can become a reality. However both the theory of dipole drive operation and it potential benefits are clear. Work should therefore begin to advance it to flight status. The stars are worth the effort.


1. Jerome Wright (1992), Space Sailing, Gordon and Breach Science Publishers

2. D. G. Andrews and R. Zubrin, “Magnetic Sails and Interstellar Travel”, IAF-88-553, 1988

3. R. Zubrin and D.G Andrews, “Magnetic Sails and Interplanetary Travel,” AIAA-89-2441, AIAA/ASME Joint Propulsion Conference, Monterey, CA July 1989. Published in Journal of Spacecraft and Rockets, April 1991.

4. Pekka Janhunen, “Electric Sail for Spacecraft Propulsion,” J. Propulsion, Vol. 20, No. 4: Technical Notes, pp763-764. 2004.

5. Cosmo, M.L., and Lorenzini, E.C., Tethers in Space Handbook, NASA Marshall Space Flight Center, 1997

6. D. Hambling, “The Impossible EM Drive is Heading to Space,” Popular Mechanics, September 2, 2016.

7. “Debye Length,” Plasma, accessed Feb 18, 2018.


‘Oumuamua: New Data Point to a Comet

New evidence for the nature of interstellar object ‘Oumuamua is in, making it far more likely that the unusual interloper is a comet rather than an asteroid. The data come from an array of instrumentation — the Hubble Space Telescope, the Canada-France-Hawaii Telescope, ESO’s Very Large Telescope and the Gemini South Telescope — and show that `Oumuamua is slowing down slightly less than expected. We are talking about a tiny force, about 1/1000 as strong as the pull of the Sun’s gravity, according to this overview of new work in Nature.

The science paper on this work, which also appears in Nature, looks at a variety of possible explanations for the velocity change. The one the authors think most likely is that `Oumuamua (pronounced “oh-MOO-ah-MOO-ah”), now moving at some 114,000 kilometers per hour, has vented material during its pass through our system, behaving the way many comets do. Marco Micheli (ESA), lead author of the paper, puts it this way: “We can see in the data that its boost is getting smaller the farther away it travels from the Sun, which is typical for comets.”

Co-author Karen Meech (University of Hawaii) concurs. It was Meech who led the initial discovery team characterisation in 2017. Although the scientists found no visual evidence for outgassing, it remained true that the composition of its surface resembled a cometary nucleus. Meech adds: “We think that ‘Oumuamua may vent unusually large, coarse dust grains.” If passage through interstellar space had eroded smaller dust grains on the surface of the object, a cloud of larger particles would not have been bright enough for Hubble to detect.

Image: This artist’s impression shows the first interstellar object discovered in the Solar System, ?Oumuamua. Observations made with the NASA/ESA Hubble Space Telescope, CFHT, and others, show that the object is moving faster than predicted while leaving the Solar System. Researchers assume that venting material from its surface due to solar heating is responsible for this behavior. This outgassing can be seen in this artist’s impression as a subtle cloud being ejected from the side of the object facing the Sun. Because outgassing is a behavior typical for comets, the team thinks that ?Oumuamua’s previous classification as an interstellar asteroid should be changed to a comet. Credit: ESA/Hubble, NASA, ESO, M. Kornmesser.

For my part, I like the description offered by Michele Bannister (Queen’s University Belfast) in the article cited above. Bannister compares the object to a “‘baked Alaska’ dessert, with a frozen heart and warm exterior.” An object like this, approaching the Sun, would begin outgassing, though the rate here is tiny. Let me quote from the Nature article:

The outgassing rate is small compared to what typical comets experience, says Jessica Agarwal, an astronomer at the Max Planck Institute for Solar System Research in Göttingen, Germany. ‘Oumuamua also emits relatively little debris, perhaps because its dust particles are too large and heavy for the weak outgassing to carry aloft. That could explain why ‘Oumuamua never developed a visually stunning, comet-like tail.

Comets can be affected by non-gravitational accelerations, however, as ‘Oumuamua now apparently shows. From the paper:

After ruling out solar-radiation pressure, drag- and friction-like forces, interaction with solar wind for a highly magnetized object, and geometric effects originating from ‘Oumuamua potentially being composed of several spatially separated bodies or having a pronounced offset between its photocentre and centre of mass, we find comet-like outgassing to be a physically viable explanation, provided that ‘Oumuamua has thermal properties similar to comets.

The paper also considers solar radiation pressure, the Yarkovsky effect (in which thermal variations on the surface of a rotating object like an asteroid can lead to asymmetric forces), and the possibility of a collision with another object, none of which fit the bill. The unlikely idea that `Oumuamua is an alien spacecraft is rejected because the object is tumbling on all three axes. Our short-term interstellar guest, in any case, has been nudged a bit faster than expected.

Assuming we’re dealing with a comet, the outgassing may make our attempts to trace its home star that much more difficult. The new observations were carried out to help make that call, but team member Olivier Hainaut (European Southern Observatory, Germany) now wonders whether we will ever know its true home. As to its apparently cometary nature, he adds:

“It was extremely surprising that ?Oumuamua first appeared as an asteroid, given that we expect interstellar comets should be far more abundant, so we have at least solved that particular puzzle. It is still a tiny and weird object that is not behaving like a typical comet, but our results certainly lean towards it being a comet and not an asteroid after all.”

The paper is “Non-gravitational acceleration in the trajectory of 1I/2017 U1 (`Oumuamua)”, Nature 27 June 2018 (abstract).


Hayabusa 2 Arrives at Ryugu

The asteroid game is heating up. The Japanese probe Hayabusa 2 has arrived at asteroid 162173 Ryugu, the plan being to reach the surface with landers later this year and bring back samples in 2020. We also have ORISIS-REx, launched in 2014, on course to 101955 Bennu in December, with a sample return planned for 2023. Assuming both missions are successful, scientists will have the opportunity to compare the composition of the two. Both are C-type (carbonaceous) asteroids, darker than previously explored asteroid Itokawa. The current Hayabusa is similar to the probe that first returned an Itokawa sample to Earth in 2010.

JAXA confirmed the arrival of Hayabusa 2 at 9:35 (Japan time) on the 27th: “The National Research and Development Corporation Japan Aerospace Exploration Agency (JAXA) announces that we have confirmed the arrival at asteroid Ryugu (Ryugu) of the asteroid explorer ‘Hayabusa 2′”, adding that the distance between the spacecraft and the asteroid is about 20 kilometers.

From a JAXA statement:

The confirmation of the Hayabusa2 rendezvous made at 9:35 a.m. (Japan Standard Time, JST) is based on the following data analyses:

* The thruster operation of Hayabusa2 occurred nominally
* The distance between Hayabusa2 and Ryugu is approximately 20 kilometers
* Hayabusa2 is able to maintain a constant distance to asteroid Ryugu
* The status of Hayabusa2 is normal

Itokawa is an S-type asteroid, so this will be our first chance to sample a C-type, the closest previous encounter with the latter being the NEAR Shoemaker flyby of 253 Mathilde in 1997, which could not close to less than 1000 kilometers. Hayabusa 2 has already released an image of Ryugu from about 40 kilometers out. I liked the way project manager Yuichi Tsuda described the boulder-filled object:

“The shape of Ryugu is now revealed. From a distance, Ryugu initially appeared round, then gradually turned into a square before becoming a beautiful shape similar to fluorite [known as the ‘firefly stone’ in Japanese]. Now, craters are visible, rocks are visible and the geographical features are seen to vary from place to place. This form of Ryugu is scientifically surprising and also poses a few engineering challenges.”

Image: Closing on destination. This is asteroid 162173 Ryugu from a distance of roughly 40 kilometres. The image was taken by the spacecraft’s ONC-T (Optical Navigation Camera – Telescopic) on June 24, 2018 at around 00:01 JST. Credit : JAXA, University of Tokyo, Kochi University, Rikkyo University, Nagoya University, Chiba Institute of Technology, Meiji University, Aizu University, AIST.

I like the idea of a ‘firefly stone,’ and also note that the name Ryugu means ‘dragon’s palace’ in Japanese, giving the imagination plenty to play with. The science payoff from both Ryugu and Bennu could be significant. C-type asteroids are intriguing because they are assumed to be the source of carbonaceous chondrite meteorites and are thought to contain organic material as well as water. With a diameter of about 900 meters and a low-reflectance surface, Ryugu may not win any beauty contests, but it is conceivable that asteroids like this could be sources of water or oxygen that future space missions will tap.

Moreover, asteroids like these are expected to give us insights into objects in the early Solar System, and the samples gained by Hayabusa 2 and its landers should be able to tell us whether the asteroid’s dark surface is as rich in carbon as we assume. Of particular interest is the question of water. The isotopic and chemical analysis of Ryugu should help us gain insights into the formation of Earth’s oceans through incoming asteroids or comets. Landing site selection will be critical as the mission teams plans a 30 cm per second touchdown in October.

Tsuda goes on to note some of the challenges the early imaging is beginning to reveal. While the rotation axis of the asteroid is perpendicular to its orbit, there is a peak near the equator and various small craters, factors that will have to be considered as landing sites are assessed. The object’s unusual shape also introduces some issues:

Globally, the asteroid also has a shape like fluorite (or maybe an abacus bead?). This means we expect the direction of the gravitational force on the wide areas of the asteroid surface to not point directly down. We therefore need a detailed investigation of these properties to formulate our future operation plans.

Image: It looks like a big space diamond — but with craters. It’s 162173 Ryugu (Dragon’s Castle), and Japan’s robotic Hayabusa 2 mission is now arriving at this near-Earth asteroid. Ambitious Hayabusa 2 is carrying an armada of separable probes, including two impactors, four small close-proximity hoverers, three small surface hoppers, and the Mobile Asteroid Surface Scout (MASCOT) which will land, study, and move around on Ryugu’s surface. Most of these are equipped with cameras. Moreover, Hayabusa 2 itself is scheduled to collect surface samples and return these samples to Earth for a detailed analysis near the end of 2020. Pictured, a series of approach images shows features suggestive of large boulders and craters. Credit & Copyright: ISAS, JAXA, Hayabusa 2 Team.

Hayabusa 2’s distance from Earth is now 1.9 AU, with a round-trip light time of 1895 seconds.


Exoplanet Hunt: Speeding Up the Data Pipeline

The K2 mission’s C16 and C17 observing campaigns — each containing observations of one patch of the sky for an 80-day period — have proven fruitful for astronomers at MIT. The institution’s Ian Crossfield, working with graduate student Liang Yu, has brought new software tools developed at MIT to work, producing results just weeks after the mission’s raw data for these observing runs were made available. Now we have nearly 80 new exoplanet candidates from C16, but we also have a method of fast analysis that should benefit future missions.

Let’s pause on method. The idea here is to speed up the analysis of light curves, the graphs showing the intensity of light from a star. Between them, C16 and C17 tracked about 50,000 stars, the analysis of whose light would normally take at least several months and perhaps as long as a year. Speed is of the essence because a faster planet identification process makes it possible for quick ground-based radial velocity follow-ups that might otherwise prove challenging.

The reason: K2’s position in a trailing orbit as Earth moves around the Sun means that, depending on its orientation, some stars are not observable by scientists on the ground until the Earth returns to the same patch of sky. So-called ‘rear-facing’ campaigns, as this MIT news release points out, have little need for fast data analysis because the follow-up cannot happen for almost a year. But C16 and C17, in ‘forward facing mode,’ examined stars that remained within Earth’s field of view for several more months. The confirmation process could begin.

Image: The Kepler Space Telescope, now operating in its K2 extended mission, diagrammed here in ‘forward-facing mode.’ “Forward-facing” implies looking towards Earth in the spacecraft’s orbit, in the direction of the spacecraft’s velocity vector. Credit: This is a slide from a talk by Geert Barentsen and Tom Barclay for the 20th Microlensing Workshop in Paris in 2016.

The researchers’ goal, then, was to work through the light curves, which had been released on February 28, to let their algorithms narrow the field. In the case of C16, which is discussed in a paper on this work in The Astronomical Journal, that means paring 20,647 targets down to 1,097 stars of particular interest. The paper presents a catalog of interesting C16 targets identified through photometry, candidates for rapid identification and follow-up. 30 high-quality planet candidates emerged from the process, along with 48 lower-quality candidates, 164 eclipsing binaries, and 231 other periodically-variable astrophysical sources.

One find stands out as particularly interesting, a planet around the star HD 73344 that orbits its primary every 15 days. HD 73344 is a high proper-motion F6 star, its planet (if confirmed) orbiting the brightest planet host K2 has yet discovered. Crossfield and Yu’s data point to a planet of 2.5 Earth radii and 10 Earth masses circling its star every 15 days — Crossfield refers to it as “a smaller, hotter version of Uranus or Neptune.” At 114 light years from Earth, HD 73344 b could be a prime candidate for atmospheric composition studies.

The paper argues that fast planet searches should be of value for TESS (Transiting Exoplanet Survey Satellite), which will study nearby stars in 30-day periods, ultimately covering 85% of the sky. The data trove from this mission should be enormous, and as it is received, there will be a period of months before the stars examined during that month set for the year. Says Crossfield:

“If we get candidates out quickly to the community, everyone can start immediately observing systems discovered by TESS, and doing a lot of great planetary science. So this [analysis] was really a dress rehearsal for TESS.”

And this is from the paper:

The release of planet catalogs has occurred only irregularly during the K2 mission, but this paradigm will change once TESS operations begin in earnest. Data from TESS will be released and processed on a 27-day rhythm for most of the two-year mission duration. With the shorter observing windows, ephemeris decay is also a much larger problem for TESS and therefore the importance of securing planet candidates in the same season is even higher. If interesting objects could be rapidly gleaned from TESS data and circulated to the community, follow-up observations and analyses could begin a full season earlier and so the full impact of that mission could more quickly be achieved.

As to TESS itself, it’s hard to believe we’re already 68 days past the launch, a period in which the spacecraft has been undergoing commissioning operations before beginning the collection of scientific data.

The paper is Yu et al., “Planetary Candidates from K2 Campaign 16,” The Astronomical Journal Vol. 156, No. 1 (21 June 2018). Abstract / preprint.


The Importance of an Eclipsing Charon

The quality of the image below isn’t very high, but consider what we’re looking at. This is the ‘night side’ of Pluto’s moon Charon as viewed against a star field by the New Horizons spacecraft. We’re looking at reflected light from Pluto –’Plutoshine’ — as the sole illumination of most of the surface. Who would have thought, in the 88 years since Clyde Tombaugh’s discovery of Pluto, that we would see a Plutonian moon’s dark side by Pluto’s light?

I wonder if there would have been a mission to Pluto at all if it hadn’t been for James Christy. Working with astronomer Robert Harrington at the U.S. Naval Observatory Flagstaff Station (Arizona), Christy was situated just miles away from Lowell Observatory, where Pluto was discovered, when he noticed a strange elongation in images of the world. That was forty years ago, on June 22, 1978, during an effort to tighten up estimates of Pluto’s orbit around the Sun.

I suppose an astronomical analogue to this odd ‘blob’ on Christy’s plates would be the elongation Galileo puzzled over when he looked at Saturn with his earliest optics. In both cases, it would take a bit of imagination and better viewing to make out the real cause, which in Christy’s case was the moon that would be called Charon. It was Christy who proposed the name, and it stuck because the ferryman who carried souls across the river Acheron fit right in with the Plutonian milieu. Acheron was one of five rivers in mythical Pluto’s underworld.

Image: Jim Christy points to the photographic plate on which he discovered Pluto’s largest moon, Charon, in 1978. Credit: U.S. Naval Observatory.

A careful scientist, Christy compared other images to note that the ‘bump’ on Pluto seemed to move from one side to another. Bear in mind that it had been almost a half century since the discovery of Pluto, and in that time no moon had been found. Finding still more images showing an ‘elongated’ Pluto, Christy and Harrington worked out orbital parameters of a moon that could explain what they were seeing, and subjected them to yet more data via a confirmation by the Naval Observatory’s 61-inch telescope. The discovery was announced on July 7, 1978.

And the personal note: We’re told that Christy first chose the name because its first four letters matched the name of his wife, Charlene, known as ‘Char’ to her friends..”A lot of husbands promise their wives the moon,” Charlene Christy would later say, “but Jim actually delivered.”

Image: Forty years after his important discovery, Jim Christy holds two of the telescope images he used to spot Pluto’s large moon Charon in June 1978. A close-up photo of Charon, taken by the New Horizons spacecraft during its July 2015 flyby, is displayed on his computer screen. Credit: NASA/Johns Hopkins University Applied Physics Laboratory/Southwest Research Institute/Art Howard/GHSPi.

What a fine tribute that the New Horizons science operations center at JHU/APL, dedicated in 2002, is named after Christy. For the discovery of Charon marked a huge step in getting New Horizons bumped along the difficult terrain of mission approval and launch. Charon is aligned to make eclipses of Pluto possible (for a period of several years) just twice every 248 year Plutonian orbit of the Sun. The eclipses would start in 1985, allowing good science to flow.

Surface brightness could be measured, compositions inferred, possible atmospheres. As David Grinspoon and Alan Stern note in Chasing New Horizons:

In the pantheon of strange and lucky coincidences, the fact that Charon was discovered just before it was about to swing into place for this mutual event season — after all, another such set of eclipses wouldn’t occur for more than a century — is right up there with the grand-tour alignment of the planets appearing just when humans were ready to take advantage of it by mastering spaceflight.

Suddenly a six-year observing window was opening, one that would put Pluto at the top of every major scientific conference, and given the trials of getting New Horizons flown, it’s more than likely that without this string of events, the mission would never have been built. As it is, those of us who were startled to read about Christy’s discovery years ago now have the opportunity to look at images like the one below, showing what we saw then and what we see now.

Image: What a difference 40 years makes. An enhanced color image of Charon from data gathered by the New Horizons spacecraft in 2015 shows a range of diverse surface features, significantly transforming our view of a moon discovered in 1978 as a “bump” on Pluto (inset) in a set of grainy telescope images. Credit: U.S. Naval Observatory; NASA/Johns Hopkins University Applied Physics Laboratory/Southwest Research Institute.

The Charon image that follows shows the fractures and canyons of Charon’s Pluto-facing hemisphere, a belt that stretches fully 1600 kilometers across the face of the moon and most likely onto the far side. Stark evidence of geological upheaval. the terrain is four times as long as the Grand Canyon and in places twice as deep, as this JHU/APL news release points out. New Horizons deputy project manager Cathy Olkin (SwRI) sees this tectonic belt as evidence that Charon once had a subsurface ocean whose eventual freezing cracked the surface.

Image: This great canyon system stretches more than 1,600 kilometers across the entire face of Charon and likely around onto Charon’s far side. Four times as long as the Grand Canyon, and twice as deep in places, these faults and canyons indicate a titanic geological upheaval in Charon’s past. Credit: JHU/APL.

So this post is all about who we were 40 years ago and who we are now, what we knew back then and how dogged persistence and extraordinary coincidence could propel us to today’s stunning views of Charon. “When you go from this little blur in which you don’t actually see anything, to the enormous detail New Horizons sent back,” says Christy, “it’s incredible.”