Gridded Ion Drives
This is the best-known of the electric propulsion thrusters of any type, and is often shortened to “ion drive.” Here, the thruster has four main parts: the propellant supply, an ionization chamber, an array of conductive grids, and a neutralizing beam emitter. The propellant can be anything that is able to be easily ionized, with cesium and mercury being the first options, these have largely been replaced by xenon and argon, though.
The type of ionization chamber varies widely, and is the main difference in the different types of ion drive. Particle beams, radio frequency or microwave excitation, in addition to magnetic field agitation, are all methods used in different gridded ion drives over the years and across the different manufacturers. The first designs used gaseous agitation to cause electrons to be stripped, but many higher-powered systems use particle (mostly electron) beams, radio frequency or microwave agitation, or cyclotron resonance to strip the electrons off the atoms. The efficiency of the ionization chamber, and its capacity, define how much propellant mass flow is possible, which is one of the main limiting factors for the overall thrust possible for the thruster.
After being ionized, the gas and plasma are then separated, using a negatively charged grid to extract the positively charged ions, leaving the neutral gas in the ionization chamber to be ionized. In most modern designs, this is also the beginning of the acceleration process. Often, two or three grids are used, and the term “ion optics” is often used instead of “grids.” This is because these structures not only extract and change the acceleration of the plasma, but they also shape the beam of the plasma as well. The amount of charge, and the geometry of these grids, defines the exhaust velocity of the ions; and the desired specific impulse produced by the thruster is largely determined by the charge applied to these screens. Many US designs use a more highly charged inner screen to ensure better separation of the ions, and a charge potential difference between this grid and the second accelerates the ions. Because of this, the first grid is often called the extractor, and the second is called the accelerator grid. The charge potential possible on each grid is another major limitator of the possible power level – and therefore the maximum exhaust velocity – of these thrusters.
These screens also are one of the main limitators for the thruster’s lifetime, since the ions will impact the grid to a certain degree as they’re flowing past (although the difference in charge potential on the plasma in the ionization chamber between the apertures and the structure of the grid tends to minimize this). With many of the early gridded ion thrusters that used highly reactive materials, chemical interactions in the grids could change the conductivity of these surfaces, cause more rapid erosion, and produce other problems; the transition to noble gas propellants has made this less of an issue. Finally, the geometry of the grids have a huge impact on the direction and velocity of the ions themselves, so there’s a wide variety of options available through the manipulation of this portion of the thruster as well.
At the end of the drive cycle, after the ions are leaving the thruster, a spray of electrons is added to the propellant stream, to prevent the spacecraft from becoming negatively charged over time, and thereby attracting some of the propellant back toward the spacecraft due to the same electrostatic effect that was used to accelerate them in the first place. Problems with incomplete ion stream neutralization were common in early electrostatic thrusters; and with the cesium and mercury propellants used in these thrusters, chemical contamination of the spacecraft became an issue for some missions. Incomplete neutralization is something that is still a concern for some thruster designs, although experiments in the 1970s showed that a spacecraft can ground itself without the ion stream if the differential charge becomes too great. In three grid systems (or four, more on that concept later), the final grid takes the place of this electron beam, and ensures better neutralization of the plasma beam, as well as greater possible exhaust velocity.
Gridded ion thrusters offer very attractive specific impulse, in the range of 1500-4000 seconds, with exhaust velocities up to about 100 km/s for typical designs. The other side of the coin is their low thrust, generally from 20-100 microNewtons (lower than average even for electric propulsion, although their specific impulse is higher than average), which is a mission planning constraint, but isn’t a major show-stopper for many applications. An advanced concept, from the Australian National University and European Space Agency, the Dual Stage 4 Grid (DS4G) thruster, achieved far higher exhaust velocities by using a staged gridded ion thruster, up to 210 km/s.
Past and Current Gridded Ion Thrusters
These drive systems have been used on a number of different missions over the years, starting with the SERT missions mentioned in the history of electric propulsion section, and continuing through on an experimental basis until the Deep Space 1 technology demonstration mission – the first spacecraft to use ion propulsion as its main form of propulsion. That same thruster, the NSTAR, is still in use today on the Dawn mission, studying the minor planet Ceres. Hughes Aircraft developed a number of thrusters for station-keeping for their geosynchronous satellite bus (the XIPS thruster).
JAXA used this type of drive system for their Hayabusa mission to the asteroid belt, but this thruster used microwaves to ionize the propellant. This thruster operated successfully throughout the mission’s life, and propelled the first spacecraft to return a sample from an asteroid back to Earth.
ESA has used different variations of this thruster on multiple different satellites as well, all of which have been radio frequency ionization types. The ArianeSpace RIT-10 has been used on multiple missions, and the Qinetiq T5 thruster was used successfully on the GOCE mission mapping the Earth’s magnetic field.
NASA certainly hasn’t given up on further developing this technology. The NEXT thruster is three times as powerful in terms of thrust compared to the NSTAR thruster, although it operates on similar principles. The testing regime for this thruster has been completed, demonstrating 4150 s of isp and 236 mN of thrust over a testing life of over 48,000 hours, and it is currently awaiting a mission for it to fly on. This has also been a testbed for using new designs and materials on many of the drive system components, including a new hollow cathode made out of LaB6 (a lanthanum-boron alloy), and several new screen materials.
HiPEP: NASA’s Nuclear Ion Propulsion System
Another NASA project in gridded ion propulsion, although one that has since been canceled, is far more germane to the specific use of nuclear electric propulsion: the High Power Electric Propulsion drive (HiPEP) for the Jupiter Icy Moons Observer mission. JIMO was a NEP propelled mission to Jupiter which was canceled in 2005, meant to study Europa, Ganymede, and Callisto (this mission will get an in-depth look later in this blog series on NEP). HiPEP used two types of ionization chamber: Electron Cyclotron Resonance ionization, which combines leveraging the small number of free electrons present in any gas by moving them in a circle with the magnetic containment of the ionization chamber with microwaves that are tuned to be in resonance with these moving electrons to more efficiently ionize the xenon gas; and direct current ionization using a hollow cathode to strip off electrons, which has additional problems with cathode failure and so is the less preferred option. Cathode failure of this sort is another major failure point for ion drives, so being able to eliminate this is a significant advantage, but the microwave system ends up consuming more power, so in less-energy-intensive applications it’s often not used.
One very unusual thing about this system is its’ shape: rather than the typical circular discharge chamber and grids, this system uses a rectangular configuration. The designers note that not only does this make the system more compact to stack multiple units together (reducing the structural, propellant, and electrical feed system mass requirements for the full system), it also means that the current density across the grids can be lower for the same electrostatic potential, reducing current erosion in the grids. This means that the grid can support a 100 kg/kW throughput margin for both of the isp configurations that were studied (6000 and 8000 s isp). The longest distance between two supported sections of grid can be reduced as well, preventing issues like thermal deformation, launch vibration damage, and electrostatic attraction between the grids and either the propellant or the back of the ionization chamber itself. The fact that it makes the system more scalable from a structural engineering standpoint is one final benefit to this system.
As the power of the thruster increases, so do the beam neutralization requirements. In this case, up to 9 Amperes of continuous throughput are required, which is very high compared to most systems. This means that the neutralizing beam has to be both powerful and reliable. While the HiPEP team discuss using a common neutralization system for tightly packed thrusters, the baseline design is a fairly typical hollow cathode, similar to what was used on the NSTAR thruster, but with a rectangular cross section rather than a circular one to accommodate the different thruster geometry. Other concepts, like using microwave beam neutralization, were also discussed; however, due to the success and long life of this type of system on NSTAR, the designers felt that this would be the most reliable way to deal with the high throughput requirements that this system requires.
HiPEP consistently met its program guidelines, for both engine thrust efficiency and erosion studies. Testing was conducted at both 2.45 and 5.85 GHz for the microwave ionization system, and was successfully concluded. The 2.45 GHz test, with 16 kW of power, achieved a specific impulse of 4500-5500 seconds, allowing for the higher-powered MW emitter to be used. The 5.85 GHz ionization chamber was tested at multiple current loads, from 9.7 to 39.3 kW, and achieved a maximum specific impulse of 9620 s, and showed a clear increase in thrust of up to close to 800 mN during this test.
Sadly, with the cancellation of JIMO (a program we will continue to come back to frequently as we continue looking at NEP), the need for a high powered gridded ion thruster (and the means of powering it) went away. Much like the fate of NERVA, and almost every nuclear spacecraft ever designed, the canceling of the mission it was meant to be used on spelled the death knell of the drive system. However, HiPEP remains on the books as an attractive, powerful gridded ion drive, for when an NEP spacecraft becomes a reality.
DS4G: Fusion Research-Inspired, High-isp Drives to Travel to the Edge of the Solar System
The Dual Stage 4 Grid (DS4G) ion drive is perhaps the most efficient electric drive system ever proposed, offering specific impulse well over 10,000 seconds. While there are some drive systems that offer higher isp, they’re either rare concepts (like the fission fragment rocket, a concept that we’ll cover in a future post), or have difficulties in the development process (such as Orion derivatives, which run afoul of nuclear weapons test bans and treaty limitations concerning the use of nuclear explosives in space).
So how does this design work? Traditional ion drives use either two grids (like the HiPEP drive) combining the extraction and acceleration stages in these grids and then using a hollow cathode or electron emitter to neutralize the beam, or use three grids, where the third grid is used in place of the hollow cathode. In either case, these are very closely spaced grids, which has its’ advantages, but also a couple of disadvantages: the extraction system and acceleration system being combined forces a compromise between efficiency of extraction and capability of acceleration, and the close spacing limits the acceleration possible of the propellants. The DS4G, as the name implies, does things slightly differently: there are two pairs of grids, each pair is close to its’ partner, but further apart from the other pair, allowing for a greater acceleration chamber length, and therefore higher exhaust velocity, and the distance between the extraction grid and the final acceleration grid allows for each to be better optimized for their individual purposes. As an added benefit, the plasma beam of the propellant is better collimated than that of a traditional ion drive, which means that the drive is able to be more efficient with the mass of the propellant, increasing the specific impulse even further.
This design didn’t come out of nowhere, though. In fact, most tokamak-type fusion reactors use a device very similar to an ion drive to accelerate beams of hydrogen to high velocities, but in order to get through the intense magnetic fields surrounding the reactor the atoms can’t be ionized. This means that a very effective neutralizer needs to be stuck on the back of what’s effectively an ion drive… and these designs all use four screens, rather than three. Dr. David Fearn knew of these devices, and decided to try and adapt it to space propulsion, with the help of ESA, leading to a 2005 test-bed prototype in collaboration with the Australian National University. An RF ionization system was designed for the plasma production unit, and a 35 kV electrical system was designed for the thruster prototype’s ion optics. This was not optimized for in-space use; rather, it was used as a low cost test-bed for optics geometry testing and general troubleshooting of the concept. Another benefit to this design is a higher-than-usual thrust density of 0.86 mN/cm^2, which was seen in the second phase of testing.
Two rounds of highly successful testing were done at ESA’s CORONA test chamber in 2005 and 2006, the results of which can be seen in the tables above. The first test series used a single aperture design, which while highly inefficient was good enough to demonstrate the concept; this was later upgraded to a 37 aperture design. The final test results in 2006 showed impressive specific impulse (14000-14500 s), thrust (2.7 mN), electrical, mass, and total efficiency (0.66, 0.96, and 0.63, respectively). The team is confident that total efficiencies of about 70% are possible with this design, once optimization is complete.
There remain significant engineering challenges, but nothing that’s incredibly different from any other high powered ion drive. Indeed, many of the complications concerning ion optics, and electrostatic field impingement in the plasma chamber, are largely eliminated by the 4-grid design. Unfortunately, there are no missions that currently have funding that require this type of thruster, so it remains on the books as “viable, but in need of some final development for application” when there’s a high-powered mission to the outer solar system.
Cesium Contact Thrusters: Liquid Metal Fueled Gridded Ion Drives
As we saw in our history of electric propulsion blog post, many of the first gridded ion engines were fueled with cesium (Cs). These systems worked well, and the advantages of having an easily storable, easily ionized, non-volatile propellant (in vapor terms, at least) were significant. However, cesium is also a reactive metal, and is toxic to boot, so by the end of the 1970s development on this type of thruster was stopped. As an additional problem, due to the inefficient and incomplete beam neutralization with the cathodes available at the time, contamination of the spacecraft by the Cs ions (as well as loss of thrust) were a significant challenge for the thrusters of the time.
Perhaps the most useful part of this type of thruster to consider is the propellant feed system, since it can be applied to many different low-melting-point metals. The propellant itself was stored as a liquid in a porous metal sponge made out of nickel, which was attached to two tungsten resistive heaters. By adjusting the size of the pores of the sponge (called Feltmetal in the documentation), the flow rate of the Cs is easily, reliably, and simply controlled. Wicks of graded-pore metal sponges were used to draw the Cs to a vaporizer, made of porous tungsten and heated with two resistive heaters. This then fed to the contact ionizer, and once ionized the propellant was accelerated using two screens.
As we’ll see in the propellant section, after looking at Hall Effect thruster, Cs (as well as other metals, such as barium) could have a role to play in the future of electric propulsion, and looking at the solutions of the past can help develop ideas in the future.
MIT Open Courseware Astronautics Course Notes, Lecture 10-11: Kaufmann Ion Drives https://ocw.mit.edu/courses/aeronautics-and-astronautics/16-522-space-propulsion-spring-2015/lecture-notes/MIT16_522S15_Lecture10-11.pdf
NSTAR Technology Validation, Brophy et al 2000 https://trs.jpl.nasa.gov/handle/2014/13884
The High Power Electric Propulsion (HiPEP) Ion Thruster, Foster et al 2004 https://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/20040139476.pdf
Dual stage 4 Grid thruster page, ASU: https://physics.anu.edu.au/cpf/sp3/ds4g/
Dual Stage Gridded Ion Thruster ESA page: http://www.esa.int/gsp/ACT/projects/ds4g_overview.html
The NASA Evolutionary Xenon Thruster: The Next Step for U.S. Deep Space Propulsion, Schmidt et al NASA GRC 2008 https://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/20080047732.pdf
Deflectable Beam Linear Strip Cesium Contact Ion Thruster System, Dulgeroff et al 1971 https://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/19710027900.pdf