A key consideration in the interpretation of ground test data of electric propulsion devices purposed for spaceflight is understanding how facility-effects influence thruster operation. This understanding is critical to the prediction of actual thruster performance in space. The necessity of science-based predictions gleaned from ground tests are particularly critical at higher thruster power levels. Operation of engines at higher power levels in vacuum chambers leads to considerable elevation in background pressure, background plasma density, and backsputter rates. This review examines the influence of ground test facility effects on gridded ion thruster operation. Ground test operation is compared with flight data, where available, to obtain a clear picture of operational differences. Mitigation strategies to alleviate facility effects are also commented upon.

Gridded ion thruster technology has been used on missions ranging from commercial communication satellite station-keeping to deep space exploration. This technology has been demonstrated over a broad range of power levels extending from 2.3 kW on the Deep Space 1 mission to the 200-kW testbed engine (Fig. 1) operated at NASA Glenn back in the 1960s.1,2 As plans are formulated for human missions to the moon and Mars, gridded ion thruster technology is an attractive, scalable propulsion option, but high-power operation at 100 kW and above, with associated long operating times, will be required.3 In this section, an overview of gridded ion thruster operation and its subsystems are discussed. This will then be followed by a synopsis of ground testing procedures along with an overview of facility effects considerations for higher power operation and testing. Following the introduction, a detailed review of a range of effects brought on by the interaction of the thruster during operation with the facility and the associated impacts on assessing engine performance. Space flight data are also presented and compared, where possible, with ground test data to provide insight into the magnitude of the facility effects. Potential mitigation and correction strategies are then commented on with identifications of needed areas of research, followed by a summary with concluding remarks. It should be noted that while most of the discussion described herein focuses on direct-current (DC) ion thrusters primarily derived from NASA test and flight data, and to a limited extent Japan Aerospace Exploration Agency (JAXA) microwave engine, the discussion in general, is broad and thus applies to most gridded ion thruster systems.

FIG. 1.

1.5 m mercury ion thruster at NASA VF 6—Engine operated at 200 kW.2 Reproduced with permission from Patterson and Sovey, J. Aerosp. Eng. 26, 300–316 (2013). Copyright 2013 American Society of Civil Engineers.

FIG. 1.

1.5 m mercury ion thruster at NASA VF 6—Engine operated at 200 kW.2 Reproduced with permission from Patterson and Sovey, J. Aerosp. Eng. 26, 300–316 (2013). Copyright 2013 American Society of Civil Engineers.

Close modal

Gridded ion engines are among the most efficient electrostatic engine systems built and flown to date and as such, are an attractive system for high-power electric propulsion. As can be seen in Fig. 2, technology also has the largest performance envelop of any electric propulsion system under investigation with applications extending to nuclear electric propulsion missions. The large performance envelop enables varied throttling for deep space missions. In fact, the first nuclear electric propulsion system ever flown employed a gridded ion thruster on the SNAP 10 A mission.4 

FIG. 2.

Gridded ion performance envelope. Reproduced with permission from M. Patterson, private communication (2022).83 

FIG. 2.

Gridded ion performance envelope. Reproduced with permission from M. Patterson, private communication (2022).83 

Close modal

The ion thruster consists of essentially three subsystems: (1) the discharge chamber, where the beam plasma is formed, (2) ion acceleration grids, where high voltage beam extraction and acceleration takes place, and (3) the neutralizer which injects neutralizing electrons into the extracted ion beam. Figure 3 depicts a two-grid ion engine with its power supply configuration.5 Inside the gridded ion engine's discharge chamber, a plasma is generated. The discharge chamber typically features a magnetic field, either a divergent or consisting of alternating polarity rings forming magnetic cusps that serve to increase ionization efficiency of primary electrons and to reduce the electron loss rate. The function of the discharge chamber is to provide ions that are subsequently extracted to produce the ion beam. The discharge plasma can be produced electrodelessly via radio frequency (RF)6 or microwave excitation—via electron cyclotron resonance7 or via the application of a direct-current (DC) voltage applied between a thermionic emitter, such as a hollow cathode, and the anode, where the anode collects emitted charge. The ions drift toward the exit plane containing closely spaced ion extraction grids. Ions that enter the grid gap are accelerated across the potential difference between closely spaced multi-aperture grids to form the beam. As the beam is exhausted from the thruster, a neutralizer emits electrons that couple to beam, neutralizing it. The potential distribution downstream of the engine is shown in Fig. 4. The effective beam voltage (the ion acceleration voltage) is defined as the voltage between the screen voltage and the beam potential. Thus, a poorly functioning neutralizer results in a reduced beam voltage which leads to reduced thrust. Each of these three subsystems may be affected to varying degrees by the physical presence of the facility (wall effects) and physical processes therein (gas phase).

FIG. 3.

Schematic depiction of gridded ion thruster subsystems.5 Reproduced with permission from J. R. Brophy, “NASA's Deep Space 1 ion engine (plenary),” Rev. Sci. Instrum. 73(2), 1071–1078 (2002). Copyright of AIP Publishing.

FIG. 3.

Schematic depiction of gridded ion thruster subsystems.5 Reproduced with permission from J. R. Brophy, “NASA's Deep Space 1 ion engine (plenary),” Rev. Sci. Instrum. 73(2), 1071–1078 (2002). Copyright of AIP Publishing.

Close modal
FIG. 4.

Potential distribution between grids up to the hypothetical neutralization plane.

FIG. 4.

Potential distribution between grids up to the hypothetical neutralization plane.

Close modal

The neutralizer assures both current and charge neutralization. Current neutralization (current balancing) refers to the emission of equal amounts of electron and ion currents from the thruster such that the thruster does not charge up. Otherwise, thruster charge build-up would lead to beam turn-around—which occurs when the beam partially or fully bends back to the thruster due to potential differences between the thruster and the beam. Charge neutralization refers to equal amounts of electrons and ions in the beam itself such that it remains nearly quasineutral, which facilitates beam transport and prevents beam blow up. In most cases the neutralizer is externally mounted, with the notable exception of variants that emit both alternating negative and positive ions to achieve current neutralization.8 There is typically a coupling plasma present, formed at the neutralizer, that forms a low impedance path or “bridge” to the ion beam.9–13 The externally injected neutralizing electrons enter the beam with greater thermal speeds than that of ion beam ions owing to the mass difference. The actual underlying mechanisms in the beam that give rise to neutralization are complex and remain poorly understood.14 It is hypothesized that the positive potential on axis of the beam traps low energy electrons to enable neutralization. Instabilities are assumed to play some role in the turbulent mixing of electrons and ions to give rise to neutralization.15 Recent studies show evidence that electrostatic solitary waves (ESWs)16,17 excite Trivelpiece–Gould (TG) surface waves, in the beam during the neutralization process.18,19

To qualify a thruster for space, ground testing in a relevant environment is required. The ground test must be carried out in a manner such that both performance and lifetime are accurately assessed before flight. Ground testing is particularly challenging for engine systems proposed to support human missions to Mars. These systems are expected to include arrays of 100 kW-class engines in arrays totaling over a MW in total thruster power.20 This is to be contrasted with the highest power system under development at this time, the 12.5 kW advanced electric propulsion system (Hall thruster) and the 8 kW NEXT engine (gridded ion). Associated with high power operation, is increased propellent throughput, and thus the accompanying need for increased facility pumping speed to maintain appropriate background pressure. Moreover, at higher powers, there is the need to manage the increased backsputter rate from beam-chamber wall interactions—which invariably leads to higher deposition rates on the engine. Past studies recommend that to minimize facility effects such as enhanced background plasma density associated with charge exchange production which scale with the elevated background pressure the engine needs to operate at pressures of less than 1.5 × 10−5 Torr.21 Under such conditions, accelerated wear on grids due to charge exchange ions approaches the charge exchange current levels observed in space. Considering pumping speed limitations of existing facilities, it will be necessary to characterize high power engines at pressures near the upper limit of desired background or perhaps even higher. For example, at high specific impulse operation and power levels up to 2.5 MW, pumping speeds approaching 50 ML/s would be required for such systems, which is currently first order of magnitude higher than the state of the art.22–24 In this respect, to credibly develop flight, high-power gridded ion engines, one must be able to identify and separate facility-related effects present during ground tests from actual thruster related processes which would otherwise skew interpretation of engine performance and lifetime in space. The general design of gridded ion engines has an ion acceleration zone defined between grids takes place over distances that are much smaller ( d grid) than scattering mean free paths of interest ( λ mfp = 1 σ n n) such as that of charge exchange (i.e., λ mfp d grid), resulting in ions being assumed to be collisionless between the grids. Correction for background gas ingestion into the discharge chamber is also straightforward. Nevertheless, there are local and downstream processes such as charge–exchange (CEX) plasma formation that can impact overall operation in facilities and thus can have a major impact on engine performance and lifetime. Additionally, beam contact with chamber walls can augment neutralization processes which would not appear in space as well as give rise to significant backsputtering that can augment apparent engine life by obscuring grid erosion by facility-carbon deposition, giving rise to uncertainty to actual in-space lifetime estimates.

As an example of the reaches and commingling of facility effects during experimental design and operation of ground test, a brief overview of facility effects on beam neutralization is highlighted here. Being the neutralization process occurs outside the enclosed thruster, the neutralization process is largely affected by facility effects. The sensitivity to the neutralization process to background gas density is not well understood or characterized as one may suspect the electron temperatures to decrease as a function of increased background gas density. It has also been shown that the beam potential profile is affected by the presence of the chamber walls via limitations in beam potential expansion.25 As a result, the potential difference from the beam to the coupling neutralizer-bridge plasma may also alter electron temperatures entering the beam. Premature termination of the beam less than 10 beam diameters downstream is also shown to affect beam expansion26 but has yet to be well characterized experimentally to the knowledge of the authors. Adding to the complexity of facility effects on neutralization is electrical configuration of the thruster relative to the test facility. The thruster system in ground tests is typically isolated from earth-ground using an isolation transformer. The neutralizer cathode serves as the common tie point reference and is also isolated from ground using a Zener diode clamp, allowing the thruster to be floated from earth-ground potential but still maintaining a tethered reference point such that thruster potential does not runaway. However, limited studies have been done to validate this method creates a spacelike test condition, meaning, does the tethered floating condition settle at a spacelike operating condition with reference to the ground potential of the facility such that beam expansion is spacelike? Or should the thruster system be floated at a specified biased during ground tests? Are there additional effects of this engine potential relative to the facility on the neutralization processes and pathways in conjunction with the effects on beam potential expansion? In addition, how do the effects of the engine electrical potential with respect facility ground scale with increased background pressure? The electrical configuration of the thruster adds to the complexity presented by facility effects on neutralization. Understanding and quantifying the processes and mechanisms linked to facility effects on systems, such as on beam neutralization, during ground tests is critical to mimicking a spacelike environment.

Philosophically, ground testing is a vital element of space qualification. The challenge is in understanding the test and how it translates to operation in space. On orbit, gas pressures are less than 10−9 Torr (at 1000 km) and decrease with increasing orbital radius.27 This pressure is some three orders of magnitude lower than typical operating pressures of ion engines on the ground. At elevated background pressure, discharge properties can be altered via ingestion of background gas. Additionally, ion grid lifetime in test facilities is not simply driven by charge exchange erosion derived from neutrals escaping from the engine but rather the balance between increased chamber-derived charge exchange flux associated with elevated background pressure and deposition due to thruster sputtering of facility walls. Ion beam neutralization is also drastically different from the process prevailing in space. In general, charge and current neutralization are dependent on the neutralizer-to-ion beam plasma bridge properties, which are affected by the presence of nearby wall surfaces, background plasma conductivity, and collisions with background gas. The ion engine also consists of high voltage insulators whose tendency to flashover depends on background pressure. It should also be pointed out that the local Earth magnetic field may even influence beam symmetry and the introduction of anisotropy of charge transport in the discharge itself and in the beam, as compared with space, where beyond low Earth orbit (LEO) these are negligible. Test chamber plasma density is much higher than that in space by two–three orders of magnitude. In general, test chambers cannot practically simulate the large length scales over which the beam interacts with the local plasma as occurs in space.28 Indeed, Wang has shown that the potential of the beam and its profile cannot be simulated in finite sized vacuum chambers because the location of the bounding walls impose a limit on plasma expansion of the plume in contrast to space.29 In this respect there are a number of potential areas of concerns regarding interpreting engine operation in Earth-based test facilities. Fortunately, a wealth of knowledge regarding facility effects can be gleaned from in-flight test data. Six flights in particular,—SERT 1, SERT II, Deep Space 1, Hayabusa 2, Dawn, and DART—provide a range of insight into how well operation in ground facilities faithfully replicates actual space operation.

In this paper, we review reported facility related effects on ion engines operation dating back to the 1960s. We also survey experiments and modeling efforts aimed at understanding these effects and pose a range of experiments aimed at clarifying facility-related effects on this technology. We review in-flight engine studies where engine performance in space—the gold standard for reference performance—is compared with its respective ground test operation. The goal of this cursory review that aims to compile credible facility effects associated with gridded ion operation and weigh the relative impact on the interpretation of engine operation in space.

The operating conditions during space flight and ground tests differ drastically. Figure 5 schematically illustrates key thruster plasma-chamber wall interactions that are not present in space. Table I contrasts the differences between space and ground testing environments. In space, background pressure and ambient plasma densities are low, dropping significantly with increasing altitude above LEO, represented by the lighter background purple of Fig. 5. The earth's magnetic field drops nearly an order of magnitude from its value at LEO (roughly half that of the surface magnetic field) at a 7000 km altitude. The neutralizer which establishes the common reference for the engine system is typically tied to spacecraft ground via Zener clamping diodes. In this respect, for the system to faithfully represent space conditions, it must essentially be completely floating. Currents from space plasma can contribute to a limited degree to charge and current balance if the spacecraft chassis surfaces are exposed to the space plasma and are conducting. Yet/However owing to low current densities in LEO (ion current density ∼ μA/cm2)30 and at higher orbital distances, this effect is small. However, off nominal operation of the neutralizer, for example, such as conditions where there is finite conductivity between neutralizer common and the spacecraft chassis can lead to the double probe effects where chassis can be driven negative leading to potentially damaging spacecraft sputtering from back flowing CEX ions. In general, the issues that affect or obfuscate interpretation of ground test data can be categorized into three problem areas: (1) the electrical and magnetic interactions of the thruster plume and neutralizer plasmas with the facility, (2) gas phase collisions with background and thruster derived neutrals, and ionized gas, and (3) physical plasma–material interactions that give rise to processes such as sputtering and subsequent deposition. Often these problem areas are interconnected as well.

FIG. 5.

Comparison between space (a) and ground test (b) engine operation. The darker purple color represents the increased neutral gas and plasma densities. Alternative neutralization pathways through the facility walls are in green. Backsputtered material from the walls are depicted by the blue arrows.

FIG. 5.

Comparison between space (a) and ground test (b) engine operation. The darker purple color represents the increased neutral gas and plasma densities. Alternative neutralization pathways through the facility walls are in green. Backsputtered material from the walls are depicted by the blue arrows.

Close modal
TABLE I.

Comparison between space and ground test engine operation conditions.

Condition (units) Space (LEO) Ground Testing
Background pressure (Torr)  < 10 9  > 10 7 
Background plasma density (1/cc)  < 10 6  < 10 9 
Earth's magnetic field strength (G)  < 0.35  0.5 
Condition (units) Space (LEO) Ground Testing
Background pressure (Torr)  < 10 9  > 10 7 
Background plasma density (1/cc)  < 10 6  < 10 9 
Earth's magnetic field strength (G)  < 0.35  0.5 

Ion beam neutralization processes are difficult to access in a vacuum chamber. The system of circulating currents that can develop during ground testing of a gridded ion engine, as illustrated in Figs. 5 and 6, complicates the interpretation of neutralizer operation. The circulating current systems are dramatically different in the case of ground tests. Perhaps the biggest problem is related to the function of the neutralizer itself. The neutralizer's function is to rid the discharge chamber of excess electrons to prevent engine and spacecraft charging. However principle, it does not matter where these electrons are rejected to. In ground test facilities the neutralizer electrons are collected on grounded structures of the engine housing or on the grounded walls of the facility in addition to the beam, resulting in not all neutralizer electrons traveling directly to the beam as intended. As such, proper experimental design of the electrical configuration of the thruster housing and relative position of the thruster to grounded chamber components must be made to limit these alternate neutralization pathways,31 but carful design will not eliminate these alternate neutralization pathways completely. In principle an ion engine can operate with a poorly neutralized beam in a ground test facility with the beam current neutralized via the grounded target. Ambient charge exchange plasma also supplies electrons along the path of the beam. A key question is, therefore, how does one subtract out these effects to recover or predict how the engine would function in space?

FIG. 6.

Electrical circuit depiction of DC electrical pathways associated with beam neutralization with coupling represented by resistors.

FIG. 6.

Electrical circuit depiction of DC electrical pathways associated with beam neutralization with coupling represented by resistors.

Close modal

In general, a measure of the ease with which neutralizer emitted electrons are transported to the beam is coupling voltage—essentially the potential difference between the neutralizer cathode and the beam. In space, a low coupling voltage corresponds to a well coupled beam. In a vacuum facility, low coupling voltages can arise if neutralizer electrons can flow to nearby grounded surfaces. One issue is that the neutralizer-to-beam coupling voltage can impact the operating mode of the neutralizer. The higher coupling voltage operating mode, which is associated with a condition known as the plume mode, can lead to reduced neutralizer lifetime.32,33 In the plume mode, the neutralizer discharge voltage between the keeper electrode is higher, electrically noisy, and is associated with a luminous plume extending from the cathode, generated largely by energetic electrons ionizing and exciting escaping gas locally. It is therefore desired to have the coupling voltage as low as possible but under conditions where electrons flow exclusively into the ion beam. In this case, the neutralizer tends to operate in the “spot” mode, which is associated with low keeper voltages and low peak-to-peak voltage oscillations. Getting this operating mode correct is a key aspect of engine-neutralizer operation. In general, one increases gas flow to the neutralizer cathode to drive it from plume to spot mode. This flow characterizes the neutralizer's flow margin. A neutralizer operating in spot mode on the ground may give the false impression of greater flow margin than that which would be required for space ultimately leading to a shorter lifetime in space. Ideally, it would be desirable to track where all neutralizer derived electrons go. If this were possible then one could assess true neutralizer flow margin by either eliminating the alternative pathways or computationally predicting neutralizer operation without such pathways.

Perhaps the most obvious difference between operating in space and in a ground test facility is the fate of the thrust producing beam. As discussed previously in Figs. 5 and 6, the beam terminates on a surface or target typically at ground potential. In this regard, the ion beam can be neutralized on grounded surfaces—with ground providing the inflow of neutralizing electrons as well as the emission of secondary electrons and backsputtered material. This assures current neutralization if the target is grounded. The secondary electrons can propagate upstream and contribute to beam charge neutralization. The contribution, however, from secondaries is relatively small since the yield, at least for xenon, for beam energies of state-of-the-art medium power thrusters such as the NEXT ion engine (<2 kV) is less than 0.1, but it increases for high powered, higher beam voltage engines. For example, the beam energy for the HiPEP ion engine designed for a nuclear electric propulsion system was approximately 5500 V which translates to a considerably higher secondary electron emission yield.34 For example, at 5500 V, the yield of argon ions on tungsten is 0.3 and increases linearly with incident ion energy. In general, there is a lack of data on noble gas ion-derived secondary electron emission at relevant energies, which indeed will be a necessary area of research to understand high power engine operation in facilities.35 

In conventional ion engine ground tests, a Zener diode is incorporated to isolate neutralizer common from ground. If the potential difference between neutralizer common and ground exceeds 45 V, the Zener will clamp and current can be passed under these conditions allowing for charge neutralization to occur at the grounded beam dump. The neutralizer coupling voltage for a properly isolated engine is typically around 10 V below ground. For an engine that is not well isolated from ground, current and charge neutralization occur via charge flow to and from ground—the coupling voltage is reduced in this case. Having a well-isolated engine is therefore a key ground test requirement.

The sensitivity of coupling voltage to isolation of engine structures was explored by Patterson.36 In that work, using a switch, he was able to float the plasma screen—the mesh-like outer housing structure which protects engine high potential surfaces from the ambient plasma—of the engine as well as the external enclosure of the neutralizer—both of which otherwise would be grounded. A diagram of electrical conduction paths associated with beam neutralization and coupling in the presence of a chamber for this experiment with external engine surfaces floating or grounded is schematically illustrated in Fig. 7. With the structures grounded, it was found that the neutralizer coupling voltage was insensitive to flow rate changes indicating that the neutralizer was not well coupled to the beam and that other processes were carrying out current and charge neutralization, as shown schematically in the Fig. 7. It was observed that actual beam current neutralization was achieved via the grounded target and partial neutralizer coupling to the beam. In this case, the neutralizer was dumping most of its emitted electrons to nearby ground potential surfaces such as the engine housing and chamber walls. It was energetically easier for the electrons to neutralize the beam through alternative pathways by collecting on nearby grounded surfaces, such as the plasma screen, and be reemitted on the grounded target than to be directly travel to the beam via the plasma bridge. When the exterior neutralizer housing and thruster plasma screen were allowed to float, the coupling voltage was highly sensitive to neutralizer flow rate—which determines the operating mode of the neutralizer (spot or plume) and the plasma conditions—and, thus, impedance of the coupling bridge. The stark difference in coupling voltage sensitivity to floating and grounded screen and neutralizer enclosures conditions is illustrated in Fig. 8. Here, the neutralizer flow rate is given in milliamperes, where 1  sccm is equal to 72 mA for xenon.37 In the floating case, the neutralizer's coupling voltage magnitude decreases with increasing flow as the neutralizer transitions from plume mode to spot mode with increasing flow rate.

FIG. 7.

Apparatus used to study effect of grounded surfaces near neutralizer.36 Adapted with permission from Patterson and Mohajeri, Report No. NASA-TM-105578, 1991. Copyright 1991 Work of the US Gov. Public use permitted.

FIG. 7.

Apparatus used to study effect of grounded surfaces near neutralizer.36 Adapted with permission from Patterson and Mohajeri, Report No. NASA-TM-105578, 1991. Copyright 1991 Work of the US Gov. Public use permitted.

Close modal
FIG. 8.

Difference in neutralizer coupling voltage for floating and grounded structures as related to configures in Fig. 6.36 Patterson and Mohajeri, Report No. NASA-TM-105578, 1991. Copyright 1991 Work of the US Gov. Public use permitted.

FIG. 8.

Difference in neutralizer coupling voltage for floating and grounded structures as related to configures in Fig. 6.36 Patterson and Mohajeri, Report No. NASA-TM-105578, 1991. Copyright 1991 Work of the US Gov. Public use permitted.

Close modal

It is difficult to assess this neutralizer performance in practice if the engine is only partially isolated (nonzero impedance to ground) since part of the neutralizer emission can go to ground and part to the beam, depending on the relative impedance of the plasma bridge to the beam. The indicator of how well the beam is in fact directly neutralized is the plasma potential of the beam itself relative to ground—which should be small in a well neutralized beam. Wall collection area proximity to the neutralizer plasma can therefore have a significant effect on beam plasma potential. Incomplete neutralization (charge) can lead to beam blow up which is an issue for assessing spacecraft plume interactions as well as impacting backsputter rate spatial distribution owing to increased divergence. Thrust can also be impacted since the beam voltage to plasma potential difference is reduced owing to a more positive space potential. Note, as mentioned earlier, and as pointed out by Wang and colleagues, the potential distribution across the beam is markedly different in a chamber of finite size since walls limit the beam expansion process.

Another consideration is the local plasma environment, which varies depending on whether the spacecraft is in low Earth orbit or in interplanetary space. In LEO, ionospheric plasma can also participate in beam neutralization. Figure 9 illustrates these physical processes schematically displaying the local plasma environment and its interaction with the thruster for the SERT II test which featured 2 ion engines T/S-1 and T/S-2. Here the diagram depicts conditions with only one engine operating with beam extraction. The local plasma environment has five contributors: the beam, the charge exchange plasma, the neutralizer plasmas, the space plasma, and the ion engine main discharge plasma associated with one of the engines being operated in discharge-only mode. Discharge-only mode refers to the operation of the thruster without high-voltage beam extraction. In this case, the neutralizer electron emission shortfall could be made up through the collection of ions emitted from the discharge to spacecraft grounded surfaces or the net emission of electrons the discharge. It was shown in the SERT 2 experiment that the presence of spacecraft-produced plasma reduced the coupling voltage of the neutralizer varying from roughly a potential difference of −25 V with one beam on and the second engine's discharge off, to −15 V with one beam on and the other with discharge-only. Additionally, it was shown that the engine could run for nearly an hour without any neutralizers operating at all with a beam current operating at 85 mA, but it did not operate long enough to make spin-rate thrust measurements.38 This was enabled by both CEX ions and returning beam ions to spacecraft ground potential surfaces.39 

FIG. 9.

Beam neutralization processes in space as inferred from SERT II experiments.39 Reproduced with permission from Kerslake and Domitz, Report No. NASA TM-79271, 1979. Copyright 1979 Work of the US Gov. Public use permitted.

FIG. 9.

Beam neutralization processes in space as inferred from SERT II experiments.39 Reproduced with permission from Kerslake and Domitz, Report No. NASA TM-79271, 1979. Copyright 1979 Work of the US Gov. Public use permitted.

Close modal

The neutralization is also affected by stray magnetic field lines which can affect the coupling voltage. Yu-Cai and Wilbur explored the effect of stray magnetic field on neutralizer coupling.40 It was shown that beam plasma potential could be driven to values as high as 40 V, (near Zener clamping voltage) in the presence of 10 G stray field located outside the thruster (for reference, typical magnetic field strength within the discharge volume is less than 60 G).37 As mentioned earlier, neutralizing electrons do not need to couple to the beam directly but could take a lower impedance path via chamber ground potential surfaces. In this regard, ground test structures that augment, weaken, or obstruct the stray magnetic field give rise to uncertainty in the actual coupling voltage that the engine will operate at in space. Additionally, the magnetic field could potentially guide electrons into ground structures or the chamber walls as well favoring these in contrast to deposition into the beam. Such guided flows would not be present in space thus they have the effect of introducing uncertainty into the actual coupling voltage. Additionally, Yu-Cai and Wilbur also found that Earth field strength can impact neutralizer-beam coupling as well, even at large separation distances. For example, they found that the coupling voltage increased by 1V for a neutralizer and beam distance on the order of 1 meter when in the presence of a magnetic field of 0.5G. Remotely located neutralizers on high power systems could potentially be impacted by Earth's field in this case during ground tests and thus, this potential problem should be examined if large neutralizer-beam coupling distances are involved. The importance of this configuration is highlighted by high power thruster variants that feature arrays of engines with a centralized neutralizer or multiple neutralizers.

The SERT II mission explored beam neutralization using a remote neutralizer. Here the neutralizer of one engine is used to neutralize the beam of the second thruster which was located ∼1 m away. In those experiments, it was found that the coupling voltage was actually lower for the case with the distant neutralizer (see Fig. 8). The effect of the residual, thruster-derived magnetic field over the large neutralizer-to-beam separation distance was postulated to be the explanation for low coupling voltages observed on SERT II. It was concluded that the coupling voltage required to neutralize its respective beam was higher because electrons had to crossover field lines to reach the beam whereas the remotely located neutralizer's emitted electrons could travel along field lines into the adjacent beam.41 Stray fields are indeed important to consider when their magnitudes and direction will lead to non-spaceflight like charge particle transport due to reduced impedance pathways.

Finally, the earth's magnetic field can also affect internal discharge uniformity through deflection. It has been shown by Bell and colleagues that stray fields can have significant effect via deflection in mass spectrometer ionizers, impacting overall instrument sensitivity.42 In these instruments, low energy electron beams (<50 eV) were used for ionization. Energies above this are too stiff to be appreciably affected by the Earth's field over the instrument characteristic length scales of interest. In general, as with centrally mounted cathodes inside gridded ion thrusters, a cathode jet is typically apparent. This jet is associated with primary electrons interacting with the background but confined by the modest axial magnetic field. The primary electrons emitted (∼10 eV) would have a Larmor radius of order just over 10 cm, which is of the order or smaller than the dimensions of the discharge chamber (and certainly smaller in larger, high-power engines). In this respect, Earth's field could introduce asymmetry in the discharge profile, especially if the on-axis field is weak, as would be expected far from wall surfaces in large engines, giving rise to nonuniformity. Such nonuniformity can not only perturb the thrust vector which causes the spacecraft to spin (also known as roll torque) but also causes nonuniform erosion of the grids due to charge exchange ions. In this regard, especially for higher power engines with larger characteristic beam diameters, this effect will be more pronounced and thus requires some level of investigation. Figure 10 depicts the operation of a high current cathode operating with a conical discharge chamber without any magnets, only Earth's magnetic field is present. Clearly, a deflection is apparent. Hence, local magnetic field affects must be acknowledged during ground testing and accounted for.

FIG. 10.

In-house observation of curved cathode jet from a high current cathode with conical anode (no magnets present).

FIG. 10.

In-house observation of curved cathode jet from a high current cathode with conical anode (no magnets present).

Close modal

Another difference between in-space thruster operation and ground test facilities is elevated background gas pressure. While space conditions define the ideal pressure to operate an engine, the de facto test pressure standard for non-lifetime impacting operation is in the 2.4–4 × 10−6 Torr range as inferred from NASA's Evolutionary Xenon Thruster (NEXT) and NASA Solar Technology Applications Readiness program thruster (NSTAR) life tests, respectively. It is observed anecdotally that at pressures in 10−5 Torr range can impact engine life not only due to an increased CEX production rate but also due to an increased recycle frequency.21 Standardization is needed regarding acceptable background pressure to provide a more science-based metric for engine operation. Gas phase collisional processes are dependent on mean free paths normalized to the characteristic length-scale of the facility. In this regard, any pressure standard may need to take into consideration facility physical size as well.

In this subsection, we focus on gas phase processes associated with the CEX plasma. The CEX ion production rate is given as
d n CEX d t = n b n n v b σ cex ,
(1)
where n b is the ion density of the beam, n n is the background neutral density, v b is the velocity of the beam ions, and σ cex is the cross section for charge exchange collisions. Furthermore, the background neutral density profile as a function of distance R from a point source r S C behind the thruster exit plane (where background neutral density at the exit plane is n n 0) and θ is the angle from thruster axis to R, where r S C is the radius of grids, has been used in previous simulation models for deep space 1 thruster,43 
n n R , θ = a n n 0 1 1 + r S C R 2 1 2 cos θ .
(2)
Here, a is a correction factor and r S C is the location of the point source one thruster radius upstream of the exit plane of the engine. The ne density at the thruster exit plane as
n n 0 = N ̇ A n c ¯ ,
(3)
where N ̇ is the escaping neutral flux, A n is the ion optics flow-through area, and c ¯ is the mean thermal speed using the thermal wall temperature of neutral (typically approximately 500 K or 0.04 eV).

Even in large facilities, such as NASA Glenn's VF 6 with pumping speed of 400 kl/s on xenon, the background pressure when operating a ∼10 kW ion engine is of order 10−6 Torr—some three orders of magnitude higher than low Earth orbit. The fast ions charge exchange with the background gas give rise to slow ions and fast neutrals. Altogether, these species along with neutralizer plasma, scattered ions, and secondary electrons emitted from the wall, are designated as the background charge exchange plasma. The density of this plasma increases with engine power level. These ions in the vicinity of the engine drive the erosion of the accelerator grid. The combination of Eqs. (1) and (2) can be used to estimate the charge exchange production rate and distribution near the engine.

A key life limiter of ion engines is the charge exchange plasma erosion of the accelerator grid. CEX ion bombardment of the downstream face of the accelerator grid can give rise to the so-called pit and groove erosion as can be seen in Fig. 11.44 Such erosion can lead to the weakening of the webbing which in turn can lead to the cantilevering of portions of the accelerator grid into the screen grid, resulting in a short between the grids. The erosion can lead to rogue holes which can lead to increased backstreaming as well as the accumulation of material which has the potential to bridge the inter-electrode grid gap. While all this is going on, accelerator grid holes are also enlarged by erosion as well, which when sufficiently enlarged, can lead to engine failure via electron backstreaming. It should also be pointed out that charge exchange erosion occurring in facilities with elevated pressure is reasonably well understood and that higher pumping speed and associated reduced background pressure can decrease the accelerator grid current. The primary complicating factor in predicting grid life is the effects of backsputtered material, which can be protective. In this case, one must contend with competition between enhanced erosion due to charge exchange and the sputter yield reducing effect of deposited target material such as carbon on the grids. These facility effects need to be subtracted out to assess true lifetime in space.

FIG. 11.

Comparison between accelerator grid before and after long duration test.44 Reproduced with permission from Soulas, Report No. NASA TM-2001-211275, 2001. Copyright 2001 Work of the US Gov. Public use permitted.

FIG. 11.

Comparison between accelerator grid before and after long duration test.44 Reproduced with permission from Soulas, Report No. NASA TM-2001-211275, 2001. Copyright 2001 Work of the US Gov. Public use permitted.

Close modal

To a limited extent, modeling can predict these deposition and erosion effects, as shown in Fig. 12.45,46 The figure illustrates the predictive ability of the model to track accelerator grid hole wear as a function of time for the NEXT engine long duration test (LDT). Here experimental wear data are obtained optically during the test. The fidelity of the predictive capability of models is heavily dependent on engine and test facility specific parameters such as the neutral distribution near the engine. Even at high utilization, the accelerator grid current does not go to zero as the other physical processes that create background plasma persist (e.g., neutralizer, ion neutral scattering and Coulombic scattering). If these are not accounted for, which are invariably tied to facility conditions, then one can erroneously underpredict erosion rate; even if the engine operates at high utilization.45,47

FIG. 12.

Comparison and agreement between simulation and LDT camera imaging data on accelerator grid erosion.46 Reproduced with permission from Polk et al., J. Electr. Propul. 2(1), 14 (2023).Copyright 2023 Springer Nature.

FIG. 12.

Comparison and agreement between simulation and LDT camera imaging data on accelerator grid erosion.46 Reproduced with permission from Polk et al., J. Electr. Propul. 2(1), 14 (2023).Copyright 2023 Springer Nature.

Close modal

The local plasma and neutral background can impact the engine performance in many ways and, thus, must be accounted for when projecting engine performance in space. Elevated background pressure can also affect neutralizer coupling to the beam particularly in the presence of stray magnetic fields, as was discussed earlier. At elevated background pressure, the coupling of electrons into the beam is modified by collisional transport and thus care must be taken to make sure that fields are quantified along with local pressure so that this enhanced transport can be accounted for example in computational models and thus subtracted out to recover actual space operation. The background CEX plasma can also impact engine performance in other ways as well. The CEX plasma densities in the near field of the engine has been measured as high as 108/cc as observed during the NEXT-3 engine multi-thruster array (>20 kW) test, which is considerably higher than that which is expected in space. For example, ionospheric plasma densities in LEO do not exceed 106/cc and in interplanetary space those densities drop to few particles per cc. Despite the low ambient space plasma density, during engine operation in space it has been found that local densities due to CEX flows from the plume are elevated above background space levels (∼106/cc on Deep Space 1 at ∼1.82 kW) though still lower than what one would expect extrapolating from ground tests. The elevated neutral and CEX plasma densities are due to the discharge chamber and neutralizer efflux.

The backflow of CEX ions rejected from the plume can be guided upstream to spacecraft surfaces and their flow is thus a key spacecraft plume interaction concern. Such ion flows can lead to sputtering or shorting of low potential surfaces and thus can impact the overall functionality of the spacecraft. The requirement for backflow of CEX ions from the plume to the spacecraft surfaces is determined by the ratio of the potential difference between the potential of the plume and spacecraft to the potential difference between the plume and beam edge. The potential difference between the plume and spacecraft must be greater for backflow to occur. Wang and colleagues show that for this backflow condition to prevail,48 the following must be satisfied:
Φ p 0 Φ S C Φ p 0 Φ p 1 1 + L r S C ,
(4)
where subscript 0 refers the potential within the plume, the subscript 1 refers to the plume edge location at the same downstream distance of subscript 0, L refers to the distance from the thruster exit plane that subscripts 0 and 1 are located, subscript S C refers to spacecraft and subscript p refers to the plasma potential. In addition, Φ is the potential at the locations given by the subscripts and r S C refers to the upstream distance behind the thruster exit plane at a distance equal to the radius of the thruster beam. The equation provides a basis for comparing flight and ground test charge exchange flows provided plume potentials are known.

Electrons derived from facility augmented CEX plasma can also be problematic. For example, the diffusion of plasma electrons through the plasma screen are accelerated and terminate on the anode, thus loading the engine. These electrons contribute to heat load and can give rise to uncertainty in beam current since the neutralizer has to pump these additional electrons away. This effect is exacerbated as the CEX density increases and the associated Debye screening is reduced. The tell-tale evidence of electron collection on the external surface of the anode due to electron diffusion through the plasma screen is the presence of burn lines located on anode's plasma screen facing surface at the location of magnetic rings—implying that the external field, which is also cusp like, controls this external flow. These burn lines are routinely observed on the external surface of the anode. Such energetic (beam voltage) electron flux has the potential to burnthrough obstacles in line-of-sight, such as a wiring harness. Figure 13 illustrates burn lines typically observed on internal discharge chamber surfaces associated with center of the magnetic cusp generated by considerably lower energy electrons. Discoloration is associated with electron surface texturing, localized annealing effects, and the cracking of residual oil contaminants. The interaction of beam ions with the background plasma, particularly the electrons, also introduces streaming instabilities that modify neutralization and leads to anomalous divergence.49 Because the background plasma conditions in space and in test facilities are so different, plume characteristics cannot be assumed to be the same.

FIG. 13.

Burn lines associated with electron collection at magnetic cusps.

FIG. 13.

Burn lines associated with electron collection at magnetic cusps.

Close modal

The ingestion of facility background gas associated with operation at elevated pressure increases ion engine ionization efficiency, an effect that would not be apparent in space. The ingestion driven flow rate increases due to background gas uptake is directly proportional to the chamber pressure and so is a first order facility effect correction.50 The additional flow increases the apparent propellant utilization efficiency. Ingestion can be corrected for in a relatively straightforward manner through the calculation of input flow based on grid geometry and background neutral density and temperature. This additional flow would not be present in space and thus it is necessary to lean out the flow to account for the additional flow if one is forced to operate at elevated pressures. This correction is like akin to that used to simulate ion engine operation without beam extraction, as described by Brophy, which in this case accounts for ion recombination on the grid surfaces, which leads to additional effective gas flow into the engine.51 

Ingestion can also lead to changes in production of doubly and triply charged ions. In the presence of increased background pressure, the apparent double to singly charged xenon ion ratio changes—essentially increasing. This increase is largely due to the fact the CEX cross section for singly charged xenon with background neutrals is larger than the corresponding exchange between doubly charged xenon ions and neutrals. This facility effect diminishes the contribution of the singly charged xenon, giving the impression that the engine is producing a higher fraction of doubly charged ions. This doubles to single ratio, if not corrected, can contribute to significant error in modeling internal erosion as well as grid erosion. Typically, the true doubles to singles ratio is determined by measuring the ratio as a function of distance downstream of the engine starting in the near field. The true ratio is found by then extrapolating back to the location of the plane of the optics (z = 0 m). At the plane of the engine's ion optics, it is assumed that the ratio represents that of the plasma in the discharge chamber and it is thus faithful to the true ratio one would expect in space.52 As discussed earlier, the neutral density at the exit plane and downstream varies spatially both radially and axially. This results in the doubles to singles ratio having a spatial dependence as well. More advanced methods can be applied to determine the correct doubles to singles ratio provided one has a realistic model (or measurements) for the true neutral density profile in the plume.52 

Beam ions sputter vacuum chamber surfaces. Elastically scattered beam ions, whose flux on the ion optics increases with background pressure, also participate in sputtering of upstream surfaces as well. The sputtered efflux in turn deposits on engine surfaces, thereby altering physical surface characteristics as well as contributing to shorting of subcomponents including the potential for the formation of a grid-to-grid conducting bridge. Deposition of carbon or exposed accelerator grid surfaces lead to an apparent, reduced wear rate.53,54 If carbon coatings form on the accelerator grid for example, then the wear rate reflects the degree to which the deposits affect the surface sputter yield. This facility effect therefore depends on not only the geometry of the beam target, but also the background pressure which determines the charge exchange rate, the elastic scattering rate and gas phase scattering of the sputtered material derived from chamber walls and the beam target. Additionally, regarding the sputtering of grid and engine surfaces by charge exchange ions or those elastically scattered ions, the actual sputter yield is also sensitive to background pressure. In a ground test facility, background gas can absorb into surfaces. This adsorbed layer has a shielding effect, reducing the sputter yield of that surface. The fraction of surface covered by gases can be calculated. As a rule of thumb, it has been shown that if this coverage ratio is less than 0.1 then one can neglect the impact of this effect. Here, the arrival rate of the neutral gas is one-tenth or less than that of the beam ions,
η n = β n · n ̇ n Y n · n ̇ i 0.1 ,
(5)
where η n is the neutral surface coverage, Y n is the energy dependent sputter yield of the surface, β n is the sticking coefficient of the neutrals, n ̇ n is the arrival rate of the neutrals, and n ̇ i is the arrival rate of beam ions. Monolayer formation time is a function of background pressure and at 10−6 Torr, it is about 1015 atoms per cm2 per second. In this regard, monolayer formation in ground tests at elevated background pressure can also impact erosion rates. It should be noted around 10−6 Torr, the coverage is 0.05.55 High power systems are expected to operate in existing facilities with pressures at least an order of magnitude greater than this, thus highlighting the need for higher pumping speed.

Thruster components erode due to normal processes such as accelerator grid ion impingement and ion-driven cathode erosion. In the case of the NSTAR thruster, the keeper was constructed of tantalum, whereas the keeper is made of graphite in the NEXT engine, which is more resistant to sputter erosion. Obviously, the engine service life assessment of an ion engine developed for an actual flight requires the separation of facility-related deposition effects and their implications from normal engine wear of engine subcomponents. Such an assessment is required to validate the engine for a particular mission and thus it is at the heart of the facility effects problem—how do we assess life for high power systems in that are being tested in environments far different from space? Deposition processes derived from direct engine erosion, the redeposition of grid material, and the deposition of backsputtered material due to beam impact on chamber surfaces are shown schematically by Groh and colleagues, as shown in Fig. 14.56 Such processes can lead to flake formation inside the engine and on the surface of the grids.

FIG. 14.

Test facility derived deposition and normal engine wear/deposition processes. Deposition processes derived external to discharge chamber shown on left with discharge chamber deposition processes derived from engine wear is shown on the right.56 Reproduced with permission from de Groh et al., Report No. NASA/TM-2005-213195, 2005. Copyright 2005 Work of the US Gov. Public use permitted.

FIG. 14.

Test facility derived deposition and normal engine wear/deposition processes. Deposition processes derived external to discharge chamber shown on left with discharge chamber deposition processes derived from engine wear is shown on the right.56 Reproduced with permission from de Groh et al., Report No. NASA/TM-2005-213195, 2005. Copyright 2005 Work of the US Gov. Public use permitted.

Close modal

A great deal of insight can be gained from the examination of the NSTAR and the NEXT extended wear tests of 30 and 51 kHrs, respectively, regarding the effect of backsputtered material deposition. In both cases, deposition on the discharge chamber wall and free flakes inside the discharge chamber were observed. Figure 15 illustrates deposition on the NEXT discharge chamber wall. The coating consisted of grid material due to ion bombardment, which occurs in space as well. These deposited nodules, as can be seen in Fig. 15, also contained carbon.57 The composition of flakes observed, as determined from scanning electron microscope (SEM) analysis, revealed contributions primarily from the chamber carbon panels and molybdenum derived from accelerator grid erosion. It was found that some flakes were in fact layered suggesting spalled carbon is sputter coated with molybdenum in the discharge chamber. In this case, the flakes were possibly derived external to the discharge chamber and electrostatically attracted into the engine. The very nature of flake formation is therefore convoluted in facility effects. Indeed, admixtures of carbon and molybdenum have different physical properties than adherent molybdenum flakes. Thus, spalling leading to flake formation of such mixtures is peculiar to facility testing. Such processes would not happen in space but nonetheless they confound interpretation in ground test facilities. In the NEXT extended-life test (ELT), deposition reduced impedance between anode and screen grid. The reduction in impedance was attributed to the deposition of molybdenum and carbon, which gave rise to the formation of a bridge between the screen connector and the anode insulating gromet. It is not clear that such an effect would occur in space due to normal grid wear derived deposition alone—such a question may be addressable via computational modeling.

FIG. 15.

Carbon deposition on discharge chamber just upstream of ion optics at high resolution.57 A point to carbon nodules. Adapted from with permission Shastry and Soulas, Report No. NASA TM-20210009633/AIAA 2016-4630, 2016. Copyright 2016 Work of the US Gov. Public use permitted.

FIG. 15.

Carbon deposition on discharge chamber just upstream of ion optics at high resolution.57 A point to carbon nodules. Adapted from with permission Shastry and Soulas, Report No. NASA TM-20210009633/AIAA 2016-4630, 2016. Copyright 2016 Work of the US Gov. Public use permitted.

Close modal

The presence of deposition on the grids is particularly problematic in that it (1) can lead to grid-to-grid shorting and (2) depending on composition can affect overall grid lifetime. In both NSTAR and NEXT ELTs, the screen grid and accelerator grid mass loss were masked by deposition from backsputtered carbon from the facility. Mass from accelerator grid erosion also deposits on the screen grid. In both engine tests, the accelerator grid downstream surface displayed significant deposition derived from the vacuum facility. Figure 16 depicts heavy carbon flake buildup on the outer perimeter of the accelerator grid.56 Again, it is postulated that the source of some internal discharge flakes derived from the external coatings on the accelerator grid. Such flakes can also deposit into screen apertures and can lead to ion beamlet deflection which in turn may result in rogue hole formation on the accelerator grid.

FIG. 16.

The 30 000-h Extended-Life Test of the Deep Space 1 Flight Spare Ion (NSTAR) Thruster outer perimeter accelerator grid downstream face heavily coated with backsputtered material.56 Reproduced with permission from de Groh et al., Report No. NASA TM-2005-213195, 2005. Copyright 2005 Work of the US Gov. Public use permitted.

FIG. 16.

The 30 000-h Extended-Life Test of the Deep Space 1 Flight Spare Ion (NSTAR) Thruster outer perimeter accelerator grid downstream face heavily coated with backsputtered material.56 Reproduced with permission from de Groh et al., Report No. NASA TM-2005-213195, 2005. Copyright 2005 Work of the US Gov. Public use permitted.

Close modal

In space, accelerator grid erosion is primarily due to charge exchange ion bombardment or direct impingement if the beam is poorly focused. In this regard, one would, therefore, expect net erosion to take place on this surface over time monotonically and reflect the spatial distribution of the beam. However, during ground testing, deposition of backsputtered material from vacuum chamber walls is also present. It was found that during the NEXT ELT, a transition in the spatial distribution of charge exchange flux took place at the downstream surface of the accelerator grid after approximately 36.5 kH.58 Beyond this time point, at inner radii where presumably ion current density is highest, net deposition was observed. Even the pits and grooves in the central region experienced net deposition. This puzzling transition can be observed in Fig. 17 where an early time operating point is compared with a post transition point. Here the clean central region of the grid indicates net erosion at early times in the figure on the left while at later times, clearly only an annular region at larger radii on the downstream surface experiences net erosion. This puzzling observation suggests that the external charge exchange flux spatial distribution may have changed. At the time of this publication, the underlying mechanism that gave rise to this transition is not known or fully understood. Additionally, posttest analysis of the accelerator grid erosion was 75% of the grid thickness instead of the expected 35% predicted by the model. This difference is attributed to net carbon deposition. For comparison purpose at full power operation the facility carbon backsputter rate was 3  μ m / kH and the accelerator grid current density was approximately 50  μ A / c m 2.59 

FIG. 17.

NEXT ELT accelerator grid surface deposition transition from net erosion near center to net deposition on-axis.58 Lighter regions indicate net erosion. Reproduced with permission Soulas and Shastry, Report No. NASA TM 20210009717, 2021. Copyright 2021 Work of the US Gov. Public use permitted.

FIG. 17.

NEXT ELT accelerator grid surface deposition transition from net erosion near center to net deposition on-axis.58 Lighter regions indicate net erosion. Reproduced with permission Soulas and Shastry, Report No. NASA TM 20210009717, 2021. Copyright 2021 Work of the US Gov. Public use permitted.

Close modal
In general, it is difficult to predict the actual erosion pattern of the grids in the presence of backsputtered material because the backsputtered material coverage, backsputtered material flux, and CEX ion backflow to the grid profiles are usually not all known. Attempts have been made to account for the coverage. Following the NSTAR thruster analysis, the time-rate coverage can be expressed as
N 0 d θ C d t = γ C j a e Y C M o θ c ,
(6)
where N 0 is the surface density of substrate adsorption sites, θ c is the total fractional surface coverage of substrate adsorption sites by adsorbate, γ C is the carbon back-sputter flux, Y C M 0 is the sputter yield for single carbon adsorbate adlayer on a molybdenum substrate, j a is the average accelerator current density in the pit and groove erosion pattern, and e is the electron charge. A weakness of the model is it assumes only a single carbon layer, but multiple carbon layers may be present. A more rigorous approach was developed by Soulas to give a more realistic picture of surface conditions on the grids in the presence of sputtering CEX ions.59 This model requires extensive experimental inputs, such as spatial variation of backsputtered flux, accelerator grid current, and ion impact energy. Measured spatial backsputtered carbon rates have been made by Crofton et al.60 

Pollard and colleagues measured the spatial distribution of the NEXT engine's doubles to singles ratio which has an annular shape just downstream of the grid exit, as shown in Fig. 18.61 The charge exchange cross section for Xe2+ + Xe → Xe + Xe2+ at these tested beam energies is comparable to that of Xe+ charge exchange62 and thus may play a role in the formation of this annular pattern. At 1100 V the Xe+ cross section with Xe is approximately 45  Å 2 whereas Xe2+ is of similar magnitude of around 20  Å 2 . The asymmetric reaction Xe2+ + Xe → Xe+ + Xe+ is also possible.62 Interestingly, in the very near field, the local plasma potential radial profile also reflects this profile as observed by Arthur and Williams,63 suggesting the need to understand how local CEX plasma spatial variations particularly in the near field is influenced by the facility. Additionally, is it possible that the target underwent some form of change that altered its relative sputter rate? Or did the neutral gas distribution near the engine somehow change? The mechanisms underlying these questions remain not well understood.62 

FIG. 18.

Spatial profile of doubles to singles ratio for NEXT ion engine taken —82 cm downstream as a function of engine throttle level.61 Reproduced with permission from Pollard et al., AIAA Paper No. AIAA 2010-6779, 2010. Copyright 2010 Work of the US Gov. Public use permitted.

FIG. 18.

Spatial profile of doubles to singles ratio for NEXT ion engine taken —82 cm downstream as a function of engine throttle level.61 Reproduced with permission from Pollard et al., AIAA Paper No. AIAA 2010-6779, 2010. Copyright 2010 Work of the US Gov. Public use permitted.

Close modal

Near field assessment (within a few beam diameter) of neutral density field and the plasma conditions is critical for assessing the grid erosion related facility effects. The doubles-to-singles ratio is highly dependent on this spatial profile. Additionally, assessing the spatial distribution of backsputtered carbon flux in the near field is also critical to accessing true lifetime. These represent critically needed diagnostics to assess grid failure (direct short or electron backstreaming), as it is the primary failure mode of an ion engine. At high power, these effects are all exacerbated and thus are a high priority.

Another impact of backsputtering is the deposition of conductive films of carbon on insulators. This was observed during the 30 000 h. Deep Space 1 life test.33 It was found that neutralizer keeper to ground and neutralizer common to keeper impedance degraded over time. In fact, the keeper to ground potential dropped from 10 GΩ to only 40 kΩ. While these levels were still sufficiently high to mitigate any issues with performance, the drop in impendence would interfere with the clamping function of the Zener diode that isolates the neutralizer from spacecraft ground as well as obscure total current collected by the keeper. This purely facility effect also requires that attention is paid to adequate shadow shielding to prevent ambiguous neutralization, which in turn, affects coupling voltage and ultimately the inferred neutralizer performance and lifetime.

A lesser-known gas phase surface process that is associated with facility effects is related to the actual composition of gas in the chamber at base pressure. Base pressure represents for all practical purpose the pressure associated with the residual air. The interaction of the thruster plasma with this background gas generates nitrogen ions that can find their way to negatively biased surfaces such as the accelerator grid or to cathode potential surfaces inside the discharge chamber. The impact of the ionized nitrogen on surfaces chemically gives rise to the formation of metal nitrides. Such nitrides have a lower sputter yield and so the formation of these monolayers can lead to error in the interpretation of actual subcomponent lifetime. The magnitude of this effect is significant and should be studied computationally. Wilbur developed a chemisorption model for atomic nitrogen and nitrogen ion deposition on cathode potential surfaces in the discharge chamber.64 One goal of the model was to determine the effect of chemisorbed nitrogen on sputter erosion processes in the discharge chamber. The effect of nitrogen on the erosion of negatively biased thruster components was also confirmed by Rawlin and Mantenieks.65 As shown in Fig. 19, nitrogen was added to the background to augment background pressure during thruster operation. A clear reduction in the wear rate was observed. However, the actual dependence on base pressure has yet to be determined. It may require revisiting as thruster power is increased giving rise to higher pumping load and associated higher operating background pressure. An example of this that may need to be revisited is if the relative concentration of nitrogen increases under these conditions due to desorption from cryo-surfaces because of elevated heat load, then these chemical effects may be a factor. Additionally, if the test is long duration, then even at low levels, the nitriding effect may manifest over time. In either case, background nitrogen should always be monitored for long duration tests to account for this possibility.

FIG. 19.

Molybdenum grid erosion rate variation with background pressure.67 Reproduced with permission from V. Rawlin and M. Mantenieks, AIAA Paper No. AIAA 78–665, 1978. Copyright 1978 Work of the US Gov. Public use permitted.

FIG. 19.

Molybdenum grid erosion rate variation with background pressure.67 Reproduced with permission from V. Rawlin and M. Mantenieks, AIAA Paper No. AIAA 78–665, 1978. Copyright 1978 Work of the US Gov. Public use permitted.

Close modal

Several confounding physical processes germane to engine operation in ground test facilities are possible. By subtracting out these effects, it may be possible to predict engine operation in actual flight. These require careful measurement of neutral and ion species and well as the potential distribution in the plume and in near vicinity of the engine. These experiments provide critical data for predictive model validation and performance prediction. Comparison between a flight spare engine tested on the ground with the actual flight engine operated in space allows one to put context to observations made on the ground—providing some guidance as to what the actual key considerations in are making the ground test more space-like. Flight data, though limited quantity, can give insight into processes obscured by test facility effects. We next review key flight data derived from the SERT, NSTAR, μ 10, and NEXT engines.

Understanding facility effects has been an important consideration in the research and development of ion engine technology since the inception of the technology. The Solar Electric Rocket Test I and II (SERT I-1964 and SERT II 1970) missions were aimed at verifying beam neutralization/thrust production and long life on-orbit operation, respectively. At the time, it wasn't clear from ground tests that neutralization approaches would even work in space. The flight tests provided some of the earliest data sets comparing on-orbit and ground test performance. The Deep Space 1 mission was a proof-of-concept demonstration of solar electric propulsion technology for deep space science missions and was equipped with a suite of diagnostics that was used to characterize the particle and field environment around the ion engine. The Dawn and Hayabusa missions also provided engine flight data that can be compared with ground test performance. The findings of these missions are reviewed in this section with differences between ground testing and space-flight data highlighted.

The SERT I mission was the first demonstration of ion beam neutralization in space. The mission involved a 31-min engine test in space as the spacecraft proceeded along a ballistic trajectory for a total of 47 min at altitudes above 400 km.66 A caesium contact engine and an electron bombardment mercury ion thruster were tested. The flight provided insight into the interpretation of how facility effects impact actual engine operation. These included understanding the role secondary electron emission produced by the beam at the target, the role of dilute plasma produced on the ground during engine operation and its role in neutralization, and the role of coupling effects between beam and chamber walls. Thrust production was used as the indication of beam neutralization, which to first order, was in good agreement with ground tests. Engineers observed a 44% higher beam current on-orbit than during ground tests—suggesting degraded beam neutralization in ground tests. On the other hand, they found beam spreading in facility was like that on-orbit meaning that the divergence observed was mainly a function of the ion optics. Another difference between in space operation and ground testing was electron flow to discharge chamber surfaces. They also found that electron flow through thruster screens gave rise to higher indicated beam current for ground tests suggesting effect of background plasma requires robust screening. The comparative ground tests took place in a 5 m diameter, 20 m long chamber with an operating test pressure of 5 × 10−6 Torr.67 In terms of chamber dimensions and background pressure, the ground testing parameters of SERT I meet ground testing requirements but the elevated beam currents indicate that even at acceptable ground conditions, facility effects can alter thruster performance from that of its in space performance.

SERT II, shown via artist conception in Fig. 20, was a testbed for assessing ion rocket performance in space.68 The engine test ran from 1970 to 1981. The spacecraft was placed in a 1000 km polar orbit. Two 15-cm diameter mercury ion engines (1 kW each) with a plasma potential emissive probe onboard for plume sweep were tested. One engine was a spare. In this flight test, the neutralizer potential was varied relative to spacecraft ground to demonstrate control over spacecraft potential relative to space. Figure 21 shows that as neutralizer bias voltage was varied, the spacecraft potential in space was approximately 10 V more negative than was observed on the ground.69 Here a negative neutralizer bias with respect to the spacecraft potential is important as it aides in driving electrons into the beam. Therefore, if the neutralizer is biased more negatively with respect to spacecraft potential, the beam should become more neutralized and the potential of spacecraft should approach that of the local space potential. This appears to occur in space but during ground testing the spacecraft potential is driven more positive indicating that some fraction of the electrons from the neutralizer are being collected on grounded surfaces instead of the beam, requiring the neutralizer to emit more electrons than it would in space (as neutralizer electrons make their way to nearby ground potential surfaces, they complete the circuit and reduce the coupling voltage required by using the grounded surfaces as an alternative lower impedance path to the beam).70 Note the dependence on orbital location, which suggests a local space plasma condition effect.

FIG. 20.

Artist conception of SERT II spacecraft.68 Reproduced with permission from W. Kerslake and L. Ignaczak, Report No. NASA TM 105636, 1993. Copyright 1993 Work of the US Gov. Public use permitted.

FIG. 20.

Artist conception of SERT II spacecraft.68 Reproduced with permission from W. Kerslake and L. Ignaczak, Report No. NASA TM 105636, 1993. Copyright 1993 Work of the US Gov. Public use permitted.

Close modal
FIG. 21.

Variation in spacecraft potential with neutralizer bias voltage. Nominal operation at neutralizer bias of 0 V.69 Reproduced with permission from S. G. Jones, J. V. Staskus, and D. C. Byers, Report No. NASA-TM-X-52856, 1970. Copyright 1970 Work of the US Gov. Public use permitted.

FIG. 21.

Variation in spacecraft potential with neutralizer bias voltage. Nominal operation at neutralizer bias of 0 V.69 Reproduced with permission from S. G. Jones, J. V. Staskus, and D. C. Byers, Report No. NASA-TM-X-52856, 1970. Copyright 1970 Work of the US Gov. Public use permitted.

Close modal

Accelerator grid current observed on the ground and in space were similar in magnitude, suggesting a similar charge exchange plasma environment. The test also found that on orbit, under certain conditions, there was a shortfall in neutralizer emission current; that is, it was less than the beam current. Experiments on-orbit observed an shortfall that was made up by potentially one of three sources: (1) collection of charge exchange ions along with background plasma ions falling back to grounded spacecraft surfaces, (2) plasma from the spare engine operating in discharge-only mode (no high voltage applied), or (3) ions from the neutralizer plasma collecting on grounded surfaces. The on-orbit data suggests that background plasma conditions modify emission requirements of the neutralizer—thus highlighting the importance of accounting for facility-derived CEX plasma in neutralizer performance characterization.13 

Additionally, the SERT II experiment demonstrated that the engines generate thrust in discharge mode-only operation and that it was possible to run the engines in this plasma beam mode without a neutralizer operating suggesting return currents to the spacecraft were balancing the charge flow. It was also demonstrated that with two neutralizers operating, one could vary the emission contribution from one neutralizer by varying its potential relative to spacecraft ground such that the total emission current was constant and commensurate with the beam current. It was also observed that at high negative bias, the neutralizer current was higher than the beam current, suggesting spacecraft collection of space plasma current or neutralizer current directly contacting spacecraft surfaces. Furthermore it was also observed that it was possible to run the engine without any neutralizer at all. In this case, the beam attained very high potential–an indication of a poorly neutralized beam. Neutralization was facilitated by the returning of the beam and CEX ions to spacecraft surfaces, and the spacecraft charged to very large negative voltages.

It should be pointed out that SERT II neutralizers were directly biased relative to spacecraft ground and not allowed to float as is the case in current NASA thrusters. It was also found that the current emitted by the neutralizer at a fixed bias voltage was larger in ground tests than in space, indicating the facility was playing some role. This indicates that higher bias voltages (more negative) are necessary for space coupling as opposed to that needed on ground. This would be the consequence of plasma wall interactions or the elevated background pressure local to the engine in ground tests. In summary, while the SERT tests demonstrated that for the most part, the engine's performance on the ground closely represented space, it did highlight the need to correctly assess neutralizer coupling and the charge exchange background plasma in ground tests to truly account for and understand deviations from space behavior. Any credible qualification test must assess the underlying drivers that give rise to the measured neutralizer coupling voltage so that the in-space value can be predicted.

The Deep Space 1 mission featured onboard diagnostics that allowed for the measurement of the in-flight plasma environment associated with engine operation. In this regard, the CEX environment experienced by an engine in deep space could be compared directly to ground tests of the same engine. The DS1 spacecraft was a technology demonstration mission under the NASA Solar Technology Applications Readiness program (NSTAR) where several mission critical technologies—the ion engine among them—was demonstrated for space science missions. The single 30-cm gridded ion engine onboard was used as the primary space propulsion system for a space science fly by mission that ultimately included both an asteroid and a comet.

The objective of the DS1 thruster flight test was to verify that its total impulse capability and overall lifetime was compatible with deep space missions of interest. A secondary goal was to demonstrate control technology for the solar electric propulsion system. At full power, 2.3 kW the 30-cm engine could develop 92 mN of thrust with a 3100 s specific impulse at a total efficiency of 62% with an initial design goal of 87 kg of throughput.5 Thrust developed by the DS1 ion engine was measured using Doppler shift method which was accurate to less than 0.5 mN. Measured thrust was compared to the thrust calculated using beam voltage and current. The difference between experimentally measured in-flight thrust and calculated thrust was compared with the ground test data (measured ground test data derived from thrust stand) over the engine's throttle range. In all cases, it was found that measured thrust was less than calculated thrust for both DS1 engine and ground test engines. There is good fidelity between ground testing and flight measurements in this instance. However, the calculated thrust deviates from the measured thrust at high throttle levels, shown in Fig. 22. Some insight into the deviation can be explored by revisiting the thrust equation, given as
T = α β I b 2 m i e ( V s V g ) ,
(7)
where α is the correction factor associated with the fraction of beam ions that are doubly charged, β is the correction factor for beam divergence, I b is the beam current, V s is the voltage of beam power supply, and V g is the coupling voltage, the potential between neutralizer common and ground (facility ground during ground test and spacecraft ground in space flight). Plume divergence is sensitive to background pressure and thus tends to increase with increasing background pressure owing to elastic scattering and charge exchange production. In general, a plume current density probe typically cannot distinguish between beam ions and CEX ions, thus at elevated pressures, the plume profile is artificially broadened. The doubles to singly charged ion ratio is also dependent on background pressure and in general, must be corrected for owing to the large difference in the cross section for charge exchange for each species with background neutrals. In general, a higher doubles fraction tends to reduce the overall thrust since the contribution of a double is less than that of two individual ions accelerated across the grid gap. The correction factor for double ions is given below using J + is the current density of singly charged ions and J + + is the current density of doubly charge ions, both should be evaluated at the thruster exit plane.
α = 1 + 2 2 J + + J + 1 + J + + J + .
(8)
FIG. 22.

Ground and flight thrust comparison with calculated thrust.71 Reproduced with permission from Polk et al., “In-flight performance of the NSTAR ion propulsion system on the Deep Space One mission,” in IEEE Aerospace Conference Proceedings (Cat. No. 00TH8484), Big Sky, MT (IEEE, 2000), Vol. 4, pp. 123–148. Copyright 2000 IEEE.

FIG. 22.

Ground and flight thrust comparison with calculated thrust.71 Reproduced with permission from Polk et al., “In-flight performance of the NSTAR ion propulsion system on the Deep Space One mission,” in IEEE Aerospace Conference Proceedings (Cat. No. 00TH8484), Big Sky, MT (IEEE, 2000), Vol. 4, pp. 123–148. Copyright 2000 IEEE.

Close modal

The differences in the functional dependance of the ground test thrust deviations with increasing thruster power in comparison to the flight test data suggests a facility effect may be at play, but the direct cause is not yet known. The fact that the calculated thrust did not agree with the measured suggests a higher doubly charged ion content, a higher beam divergence at exit plane, or a difference in the ambient plasma potential locally, resulting in a change in coupling voltage. As discussed and shown previously in Fig. 4, the energy of accelerated ions is related to the potential difference between the screen voltage and the ambient plasma potential at approximately the neutralization plane. A poorly neutralized beam results in a reduced effective beam voltage and thus a reduced measured thrust. In space, one would expect the thruster-enhanced plasma density surrounding the spacecraft to increase linearly with thruster power. The deviation in thrust during spaceflight is relatively linear in nature, indicating CEX plasma is playing a role—resulting in increasing neutralizer coupling to spacecraft ground potential surfaces. Because the plasma screen was grounded in the facility test, the neutralizer coupling voltage may behave in a nonlinear fashion since the neutralizer can couple to both the beam and ground potential surfaces. Charge exchange plasma density will increase with increasing beam current, increasing plasma density in the neutralizer-to-beam plasma bridge, and increasing background pressure, leading to reduced neutralizer coupling impedance to the beam. As a result, one would expect the deviations between calculated and measured thrust on the ground to decrease with power, which was observed.71 In conclusion, limiting ambient thruster produced CEX plasma density to space-like levels in ground tests is critical to decrease the thrust deviation as seen in DS1.

In general, in ground tests, increased accelerator grid current in the presence of elevated background pressure is typically observed. The elevated accelerator grid flux is due to increased charge exchange plasma density which accompanies higher background pressure in the facility relative to the that experienced in space. Figure 23 illustrates a comparison between ground test accelerator grid current and that observed in space for the DS1 engine. As can be seen here, while accelerator grid current is lowest in space, as expected, when compared to ground test facilities the magnitude of the difference is relatively small. Flight data after 432 h. and ground test at GRC though are in good agreement. The accelerator grid current difference between ground test and space impacts overall service life predictions.72 It is unknown whether the spatial distribution of neutrals, which determines in part the CEX driven spatial erosion pattern of the accelerator grid, in the ground test is like that occurring in space.

FIG. 23.

Accelerator grid current variations in space and on ground.71 Reproduced with permission from Polk et al., “In-flight performance of the NSTAR ion propulsion system on the Deep Space One mission,” in IEEE Aerospace Conference Proceedings (Cat. No. 00TH8484), Big Sky, MT (IEEE, 2000), Vol. 4, pp. 123–148. Copyright 2000 IEEE.

FIG. 23.

Accelerator grid current variations in space and on ground.71 Reproduced with permission from Polk et al., “In-flight performance of the NSTAR ion propulsion system on the Deep Space One mission,” in IEEE Aerospace Conference Proceedings (Cat. No. 00TH8484), Big Sky, MT (IEEE, 2000), Vol. 4, pp. 123–148. Copyright 2000 IEEE.

Close modal

The neutralizer as discussed earlier is also sensitive to test conditions such as background pressure, CEX plasma density, and proximity to grounded surfaces. As can be seen in Fig. 24, neutralizer keeper voltage in space was of order 2 V lower than that on the ground. The required keeper voltage tended to decrease over time as well. This drift has been attributed to cathode conditioning and was observed on the ground test as well. As to why the keeper voltage was lower in space is not well understood. It is possible that this may be an electron temperature effect. Electrons are confined in the beam via a potential well and thus if the beam is better neutralized in space, then this potential is shallower. As a result, for neutralization of the beam, lower energy electrons are therefore required.15 This is consistent with the reduced neutralizer potential relative to spacecraft ground as seen in DS1 data of Fig. 25. Here clearly the inferred coupling voltage is larger for the ground test cases, suggesting extra voltage is needed to get the electrons into the beam—likely due to competition with other facility grounded structures. Additionally, CEX plasma Debye shielding effects present in ground test facilities can reduce the actual coupling field allowing for greater competition for electrons between grounded facility surfaces and the beam. A larger coupling voltage may naturally develop under those conditions along with the need for a larger keeper voltage.

FIG. 24.

Neutralizer keeper voltage variations in space and in ground test vs throttle level.71 Reproduced with permission from Polk et al., “In-flight performance of the NSTAR ion propulsion system on the Deep Space One mission,” in IEEE Aerospace Conference Proceedings (Cat. No. 00TH8484), Big Sky, MT (IEEE, 2000), Vol. 4, pp. 123–148. Copyright 2000 IEEE.

FIG. 24.

Neutralizer keeper voltage variations in space and in ground test vs throttle level.71 Reproduced with permission from Polk et al., “In-flight performance of the NSTAR ion propulsion system on the Deep Space One mission,” in IEEE Aerospace Conference Proceedings (Cat. No. 00TH8484), Big Sky, MT (IEEE, 2000), Vol. 4, pp. 123–148. Copyright 2000 IEEE.

Close modal
FIG. 25.

Neutralizer potential relative to ground comparison with flight.71 Reproduced with permission from Polk et al., “In-flight performance of the NSTAR ion propulsion system on the Deep Space One mission,” in IEEE Aerospace Conference Proceedings (Cat. No. 00TH8484), Big Sky, MT (IEEE, 2000), Vol. 4, pp. 123–148. Copyright 2000 IEEE.

FIG. 25.

Neutralizer potential relative to ground comparison with flight.71 Reproduced with permission from Polk et al., “In-flight performance of the NSTAR ion propulsion system on the Deep Space One mission,” in IEEE Aerospace Conference Proceedings (Cat. No. 00TH8484), Big Sky, MT (IEEE, 2000), Vol. 4, pp. 123–148. Copyright 2000 IEEE.

Close modal

The DS1 spacecraft contained a suite of electrostatic probes to examine the local thruster-generated plasma environment. The so-called Ion Propulsion Diagnostic Subsystem (IDS) was located radially about a meter off axis of the thruster centerline roughly in the exit plane. The subsystem included a retarding potential analyzer, a planar Langmuir probe, and a spherical Langmuir probe. The module also housed two flux gate magnetometers. This package allowed for unprecedented analysis of the plasma-plume environment generated by an ion thruster operating in interplanetary space and thus specifies the target environment for ground tests. High quality data were obtained during the so-called S-Peak operating period of the mission where the engine was throttled sequentially through three operating points up to a peak power of 1.94 kW. Both electron temperature and local plasma potential varied with thruster operating condition. The local CEX plasma density at the IDS was around ∼106/cc. It was also observed that at the highest power throttle point, the CEX ion production rate tended to go down owing to the higher utilization. This is the opposite of what is observed during ground testing. At higher power throttle conditions, the facility struggles to keep up with throughput and results in an increasing background neutral pressure. As with any plasma in a bound volume, there are confining sheaths at the walls whose voltage drop is dependent on the electron temperature. The sheath adjusts to assure equal electron and ion loss rates to the wall. In this respect, the sheaths act to confine the lower energy electrons and thus CEX plasma density in the chamber can build up with increasing throttle power. This contrasts with space flight, where the bounding loss surface and associated confining sheaths are absent. The absence of natural expansion in ground test facilities prevents the beam from developing the potential profile that prevails in space. These processes affect not only beam neutralization but also electric fields that control CEX plasma transport.

The DS1 NSTAR engine flight data has since been studied and compared to models and the findings are discussed here. Local plasma properties as measured from the IDS are shown in Fig. 26 where ML represents mission throttle level and is typically increasing in throttle and power.48 The engine essentially flies in a cloud of self-produced CEX plasma. Backflowing CEX ion energy and charge exchange plasma environmental conditions produced by the engine are also shown. Note the marked difference between plasma density in the beam in the near field (calculated at the exit plane) and the surrounding charge exchange plasma density (measured ∼1 m off axis). The expansion of the higher density plume into the low density CEX plasma gives rise to a drop off in plasma potential. The steepness of this potential gradient depends on the relative differences in the plasma density. This suggests that an artificially high CEX density in a ground test facility will affect the plume potential profile and thus neutralization dynamics along with CEX ion transport. It also implies that the sensitivity of plume potential profile will depend on power level due to the CEX background plasma density created. Additionally, during space flight it was found that the electron temperature increased with increasing beam current, suggesting that the heating is due to the higher electric field associated with the potential gradient. Neutralization dynamics are dependent on the ratio of electron thermal velocity to ion beam velocity so in some respects, this is expected to be different in ground tests owing to the smaller gradient in density and the corresponding potential gradient.

FIG. 26.

Plasma conditions near engine at different throttle conditions.48 Reproduced with permission from Wang et al., J. Spacecr. Rockets 37(5), 545–555 (2000). Copyright 2000 American Institute of Aeronautics and Astronautics.

FIG. 26.

Plasma conditions near engine at different throttle conditions.48 Reproduced with permission from Wang et al., J. Spacecr. Rockets 37(5), 545–555 (2000). Copyright 2000 American Institute of Aeronautics and Astronautics.

Close modal

The difference in dynamics of CEX ions in space flight compared to ground tests is also not well understood. Interestingly, it was found that CEX ions could make it all the way to the plasma sensor called the plasma experiment for planetary exploration (PEPE) located on the opposite side of the engine (opposite side of the beam) when running at high power during the DS1 space flight tests. It was postulated by Wang and colleagues that one can model this mechanism as follows: the sheath around the spacecraft is associated with ambient space plasma yielding a Debye length of ∼10 m, which is fairly large. This invariably overlaps with the layer of CEX ions repelled out of the ion beam. The CEX ions can therefore be channeled to the opposite side of the spacecraft by fields in the space plasma sheath. This effect is difficult to simulate in ground tests owing to the large length scales over which the transport takes place and the fact that the weaker space plasma halo does play a role as well. In this case, we see that a plume–spacecraft interaction process would be difficult to assess if one does not match the local plasma environment or at least be able to test sensitivity to changes in this environment (altering artificially the plasma density via quenching or electron cooling/heating locally).

While the engine determines in large part the charge transport from the vehicle, if the vehicle contains structures such as biased solar arrays, then collection here also becomes part of the circuit and charge balance. It is likely necessary that subscale tests such as that which was carried out by Kuninaka and colleague to understand neutralization processes, will be ultimately necessary to truly understand system level differences between facility testing and space operation.72 The importance of addressing facility effects from a system rather than thruster-centric perspective is critically important if one is to access an electric propulsion spacecraft's performance in space. Indeed, charging effects on the Hayabusa 2 mission led to significant spacecraft sputtering owing to operational problems with the neutralizer. In this case, the potential of the spacecraft was driven sufficiently negative allowing CEX plasma to bombard the spacecraft.73 If one is to capture or study such events, it is critical that facility effects include aspects of the system that will be affected by engine operation, which is clearly different on the ground relative to space.

Additionally, it was pointed out that on the Hayabusa 2 mission the neutralizer coupling performance degraded over time at a rate ten times that which was observed during standalone ground testing. The neutralizer on this asteroid sample return mission was configured in a manner like that of SERT II where it was biased relative to spacecraft ground. With the neutralizer supply set in constant current mode, the coupling voltage shifts to whichever value is necessary to neutralize the beam. This means that the actual emission current does not need to be equal to the beam current. Under these conditions the electrons can also couple to any surface at a higher potential, this could include any nearby spacecraft ground potential surfaces. To understand the origin of the increase in coupling voltage over time, a series of ground tests were carried out to investigate the effect of neutralizer current going to surrounding grounded surfaces. It was found that the presence of surrounding ground potential surfaces reduced the coupling voltage. The neutralizer is more efficient if there are nearby ground potential surfaces to collect some of the emitted electron current—which means that the neutralizer emission current is higher than the beam current. It was concluded that future missions should incorporate sputter resistant, conductive ground potential surfaces around the thruster to assure lower coupling voltages. It was conjectured that during Hayasbusa 2 mission, the ground potential surfaces near the engine, which consisted of a thin conductive film coating, was sputtered away over time leading to the more rapid upward drift in the absolute value of the coupling voltage, which also would tend to hasten the erosion rate.74 This example illustrates the sensitivity of neutralizer coupling to overall ground test configuration.

While the DS1 mission was a technology demonstration that was complete with diagnostics suite, the Dawn mission, which also featured the same gridded ion thrusters, was not outfitted with a diagnostics suite as the engine system was only integrated as an off the shelf propulsion system. In this regard, only engine telemetry data are available for the engine's operation in space. Based on these data, the engine operation very closely matched the ground tested thruster used in the NSTAR ELT. There were two exceptions: (1) observed roll torque—which truly can only be assessed in space and (2) neutralizer common error, were the neutralizer potential exceeds the clamping voltage of Zener diode between the neutralizer common and spacecraft ground.

The roll torque about the Dawn thruster axis was observed at a magnitude of ∼10 μN m. This roll torque was most likely due to grid clocking error or the interaction of ions with stray magnetic fields leaking out of the discharge chamber.75 The neutralizer common fault occurred because of the interaction between the spacecraft and the CEX plasma while no beam extraction was occurring. Since the vehicle is floating, it will be driven to negative voltage when immersed in a plasma of a few eV. Neutralizer common on the other hand acts like a plasma contactor and clamps to a potential near space plasma ground. If a large solar array is attached to the spacecraft, then its collection of electrons can drive the potential of the spacecraft ground very negative, which means neutralizer common will be above spacecraft ground on the order of the voltage of the array and thus can exceed the Zener diode isolation of 40 V relative to ground. In this respect, this effect would not necessarily be observed in a ground test facility as it also involved interaction with the solar array, but it does point to the care that must be taken in isolating experiments where charging can give rise to unexpected results and mask operation in space as previously discussed.76 This result also suggests the need for mockups that mimic the entire spacecraft to account for on some level other sources of charge pumps and collectors associated with nominal vehicles operation (e.g., solar arrays).

NASA's Double Asteroid Redirection Test (DART) used a low-cost impactor spacecraft targeted for the asteroid Dimorphos that utilized a NEXT-C gridded ion engine, the commercial version of the NEXT engine. The NEXT-C inaugural space flight on the DART mission began operation as expected but was cut short due to an anomaly in the spacecraft's power system electronics (PSE).77 It was found that PSE's Interface-Controller-Low-voltage-power-supply (ICL) component clamped due to electromagnetic-interference (EMI) that was produced during a recycle during engine operation. Recycles are high voltage arcs that can occur between the grids, from the grids to spacecraft ground, or grid to anode. Recycle events are typically fast and can create line noise. The magnitude of ground test background plasma density and neutral gas density influences recycle frequency, which will be different from space operation at higher-power operation, thus making this a facility effect in addition to a system engineering problem. As a result, out of precaution, the engine was prematurely turned-off. Later, it was determined that the issue could be replicated in ground test. Due to cost and time restraints, an integrated thruster on the spacecraft system was not properly tested and deemed an acceptable risk. From this event, it was realized that the thruster and the spacecraft system should be ground tested to mimic spaceflight as ground testing cannot be only thruster-centric but must also include the whole spacecraft system as events which occurred on the Hayabusa and DART missions have shown. Furthermore, characterization of gridded ion engine recycles in ground test compared to space flight also should be performed. Differences in magnitude, frequency, and type of recycle is important to understanding if damage to the spacecraft system may occur and how it may differ as functions of background pressure or electrical ground test configuration.

In addition to NEXT-C flight data, the NEXT array ground test, which featured a diagnostic suite, provides some insight into ground test challenges of testing high power engine systems. The magnitude of impact of the facility effect associated with background pressure on CEX plasma density as mentioned previously and can be seen here in Fig. 27.78,79 Here the ground test of 3-NEXT engines along with a dormant thruster was tested in an array at NASA VF 6 With increasing power/throughput, there is a monotonic increase in facility pressure (Fig. 27, left). The denotation of “EM1” refers to engine number one, and “FP” refers to full power. The concomitant increase in CEX plasma density is observed as well. The CEX plasma density in space is also expected to increase with power but likely not with the same slope as ground tests owing to the ability of the space CEX plasma to expand, not being limited by grounded wall surfaces.

FIG. 27.

The facility effect—background pressure of VF 6 vs flow and the response, the local plasma density at the multi-thruster array (FP nominally 7 kW per thruster-NEXT).78,79 Adapted with permission from McEwen et al., AIAA Paper No. AIAA 2006-5183, 2006. Copyright 2006 American Institute of Aeronautics and Astronautics and Foster et al., Report No. NASA TM-2006-214402, 2006. Copyright 2006 Work of the US Gov. Public use permitted.

FIG. 27.

The facility effect—background pressure of VF 6 vs flow and the response, the local plasma density at the multi-thruster array (FP nominally 7 kW per thruster-NEXT).78,79 Adapted with permission from McEwen et al., AIAA Paper No. AIAA 2006-5183, 2006. Copyright 2006 American Institute of Aeronautics and Astronautics and Foster et al., Report No. NASA TM-2006-214402, 2006. Copyright 2006 Work of the US Gov. Public use permitted.

Close modal

A range of processes germane to ion engine ground testing exists that can confound the interpretation of how the engine will operate space. It is clear that the key issues associated with interpreting the impact of facility effects that require attention include, but not limited to, are: (1) beam neutralization, (2) engine lifetimes due to facility backsputtered material that deposits on the engine, (3) thruster-plume interactions from gas phase and CEX plasma effects associated with elevated background pressure. Gridded ion engine performance in space appears to faithfully represent performance on the ground at least for low to medium power operation. Operation at higher powers exacerbates the aforementioned issues. For higher power engine testing, these four focus areas must be carefully examined. This can also be done in combination with simulation.

Simulations can predict, for example, the multitude of neutralization and circulating current pathways in a vacuum facility. Models that account for these unique elements of ground tests are needed such that the prediction of thruster performance in space can be made in the absence of these elements. The problem is that the model required must be multiscale. This is to say, that tracking the current pathways is not enough. Current flow pathways have associated with them impedance considerations and thus ultimately must be self-consistent with the prevailing neutralizer operating conditions and the CEX environment. So, performance modification to the neutralizer must be predictable. Thus, modelers and experimentalist must collaborate to validate simulations via experimental data to understand first, the neutralization pathways, then use that knowledge to create experimental standards and procedures to create test environments that mimic space impedance conditions for neutralization. Key needed areas include, realistic modeling of neutralizer coupling bridge and electron–ion beam plasma mixing.

Deposition of backsputtered material (which is protective) influences the wear rate of the grids. Additionally, if the wear rate is modified over time during a life test, then the neutral losses may also be modified (aperture enlargement rate). As result, the discharge performance and the spatial distribution of the neutral losses are modified, leading to changes in the CEX spatial distribution near the engine. So, none of the effects are truly isolated. It is the job of the facility test engineer to determine those key parameters that quantify the system and the key facility effects. These are the inputs for a model. Again, working with the modelers, experimentalists can determine best practices for ground testing, and using corrections for the deposited material where needed from the models. This too is still an active area of research.

The way the thruster interacts with the gas/plasma environment it operates in is of the upmost importance when preparing correction strategies for ground test facility effects. Regarding backsputtered material, it is important to map out the spatial distribution of the flux toward the engine to assess spatial variations in arrival rates and the flux magnitude. What is the ionization fraction of the sputtered efflux? It may be possible to generate such maps using laser absorption spectroscopy. Mapping is necessary primarily in the near field. For the neutralizer it is critical that the engine is isolated from ground surfaces or at least a quantification of the distribution of ground potential surfaces to which the neutralizer can couple. As was seen with the Hayabusa 2 mission, this collection area critically determines the magnitude of the coupling voltage. In the near field, neutralizer plasma and CEX plasma electrons diffuse to the walls. How does one separate neutralizer plasmas electrons from the background CEX plasma electrons? Is it possible to map out the current loops? It may be possible to use diagnostic like collisional laser collisional induced fluorescence (LCIF) to create electron maps in the near field of the neutralizer to track charge flow.80 Tracking the current at the beam target is also important. The ideal situation would be to tag the exiting ion beam—using a laser to place them in a long-lived excited state and then read the flux at the target. This tagging method has been used successfully in the plumes of chemical rockets.81 One can then compare actual beam current with the combination of CEX plasma ions and the beam. This may be possible if a small engine is used with flat grids. Additionally, the role of secondary electrons has not been well explored at the beam target or chamber walls. The secondary electron emission yields from beam ions are typically ignored because the emission coefficient is small (<0.1), but this applies only to clean metal surfaces. It has been shown that the secondary electron coefficient for ions with typical ion thruster energies (>0.5 keV) on “dirty” (oxides) wall surfaces exceeds unity suggesting they may play an important role determination of CEX plasma and neutralization processes.82 These effects therefore must be accounted for realistic simulations and interpretation of ground test performance. Figure 28 shown here from reference 88 illustrates the large yield observed on oxidized surfaces for argon ions. Spatially mapping and modeling the charge exchange plasma in the near field is also critical as it influences not only neutralizer coupling but also the spatial distribution of ions that erode the downstream surface of the accelerator grid. As idealized test facilities are elusive, we are left with models.

FIG. 28.

Secondary electron emission yield for beam relevant energies on dirty and clean metal surfaces.82 Reproduced with permission from Phelps and Petrovic, Plasma Sources Sci. Technol. 8, R21 (1999). Copyright 1999 IOP.

FIG. 28.

Secondary electron emission yield for beam relevant energies on dirty and clean metal surfaces.82 Reproduced with permission from Phelps and Petrovic, Plasma Sources Sci. Technol. 8, R21 (1999). Copyright 1999 IOP.

Close modal

The experimental isolation of key processes along with an assessment of their sensitivities is important to not only improve our understanding but also to serve as validation points for the development of advanced models to truly interpret ground test data and extrapolate operation in space. Such models, when fully developed, should prove useful in the establishment of testing standards for ground-based experiments where key measurements are taken to provide adequate inputs. Initial experimental studies that quantify and provide corrections for the effects of increased background pressure on operation have been made for doubles-to-singles ratios and simulation work has shown the effects of CEX plume broadening, limitations of plume expansions, and backflow of CEX ions and sputtered material to spacecraft and thruster. However, even these finding are just the beginning as even being armed with this knowledge, there is much factors such as off-axis doubles being greater than on-axis. As such, much collaboration between experimentalists and modelers is needed to derive models based on experimental ground and test flight data to predict future results.

An idealized ion thruster test facility would be too difficult and costly to build in practice. Such an ideal test facility would include walls that absorb all particles incident-neutral and charged. In this case background CEX plasma converges to space conditions and backsputter processes disappear. This absorbing boundary is still problematic for neutralizer operation. The neutralizer's job is just to evacuate excess electron charge whether that's to the absorbing surface or to the beam, it matter little where the electrons end up after emission from the neutralizer. The neutralizer is forced to couple to primarily the beam, provided the absorption surface (the chamber walls) is less positive than the potential of the beam. To absorb electrons at the same rate as ions, the absorbing surface would effectively be at the floating potential. This would mean that a sheath now exists at the boundary that would not exist in space. This sheath would filter out energetic electrons and reflect lower energy ones.

An absorbing boundary's primary benefit would be the control and removal of neutrals, but neutrals are only part of the problem. Stray fields-both electric and thruster derived magnetic field lines—can guide charge to the loss surface as well. In space, these fields establish charge transport in the plasma cloud generated by the engine. The complexity and difficulty of realizing an idealized absorbing surface is immediately realized with this thought experiment. During space flight, large scale plasma structures evolve that facilitate neutralization and transport of charge to the spacecraft. In ground tests, a rather large facility with walls far removed so that plasma structures can develop is needed such that a spacelike environment can be mimicked. In addition, neutral pressures of similar magnitude are required. Such an idealized facility is difficult to realize in practice, rather we are forced to make measurements to understand the magnitude of the facility effects.

With this realization of key facility effects, one can make a series of recommendations to at least quantify the nature of the facility effects, and if possible, adjust the degree to which that effect is impacting thruster operation. This also involves assessing the sensitivity of thruster operations to a facility effect at various operating conditions, with the end goal providing a means of extrapolation to space conditions. It may be possible to benchmark models by capturing engine operation under scaling conditions that more closely mimic space conditions, for example operating small engines in very large vacuum chambers. Additionally, while understanding how engine operation in a vacuum facility differs from flight operation, the engine is part of a system—the spacecraft.

Both experimental and modeling research need to be performed to first identify the effects contributing to error, then to quantify the magnitude of those effects, and finally develop strategies for mitigation and correction. The work on facility effects has just begun but will require collaboration between different agencies as well for establishing best practices and mitigation strategies such that better isolation and corrections for facility effects can be elucidated. This document reviews a fraction of what is known regarding facility-ion thruster interactions and presents along the way key takeaways from ground tests and space performance. It is not meant to be completely comprehensive but more a less a survey of present knowledge on facility impacts. Many of the processes discussed herein are not germane to gridded ion engines but aspects of this review apply to Hall engines as well. High power electric propulsion systems are mission enabling but their deployment and ultimate success will be dependent on a thorough understanding facility effects and strategies to interpret these effects for accurate prediction of performance in space.

The authors would like to thank NASA for supporting the compilation of this review under the JANUS project (Grant No. 80NSSC21K1118).

The authors have no conflicts to disclose.

John E. Foster: Writing – original draft (lead). Tyler Topham: Writing – original draft (supporting).

Data sharing is not applicable to this article as no new data were created or analyzed in this study.

1.
S.
Nakanishi
and
E. V.
Palik
, “
Experimental investigation of a 1.5-m-diam Kaufman thruster
,”
5
,
801
(
1968
).
2.
M. J.
Patterson
and
J. S.
Sovey
, “
History of electric propulsion at NASA Glenn Research Center: 1956 to present
,”
J. Aerosp. Eng.
26
(
2
),
300
316
(
2013
).
3.
D. R.
Jovel
,
M. L.
Walker
, and
D.
Herman
, “
review of high-power electrostatic and electrothermal electric propulsion
,”
J. Propul. Power
38
(
6
),
1051
1081
(
2022
).
4.
J.
Sovey
,
V.
Rawlin
, and
M.
Patterson
, “
A synopsis of ion propulsion development projects in the United States – SERT I to Deep Space 1
,” in
Proceedings of the 35th Joint Propulsion Conference and Exhibit
(
AIAA
,
1999
), AIAA-99-2270/NASA TM-1999-209439.
5.
J. R.
Brophy
, “
NASA's Deep Space 1 ion engine (plenary)
,”
Rev. Sci. Instrum.
73
(
2
),
1071
1078
(
2002
).
6.
R.
Killinger
,
H.
Bassner
, and
J.
Müller
, “
Development of an high performance RF-ion thruster
,” AIAA Paper No. AIAA 1999-2445,
1999
.
7.
H.
Kuninaka
,
K.
Nishiyama
,
I.
Funaki
,
T.
Yamada
,
Y.
Shimizu
, and
J.
Kawaguchi
, “
Powered flight of electron cyclotron resonance ion engines on Hayabusa explorer
,”
J. Propul. Power
23
,
544
551
(
2007
).
8.
D.
Rafalskyi
and
A.
Aanesland
, “
A neutralizer free gridded ion thruster embedded into a 1U Cubsat module
,” in
Proceedings of the 35th International Electric Propulsion Conference (IEPC
)
2017
.
9.
G.
Sohl
,
V. V.
Fosnight
, and
S. J.
Goldner
, “
Electron bombardment cesium ion engine system
,”
Report No. NASA CR-54711
,
1967
.
10.
P. D.
Reader
,
D. P.
White
, and
G. C.
Isaacson
, “
Argon plasma bridge neutralizer operation with a 10‐cm‐beam‐diameter ion etching source
,”
J. Vac. Sci. Technol.
15
(
3
),
1093
1095
(
1978
).
11.
V. K.
Rawlin
and
E. V.
Pawlik
, “
A Mercury plasma-bridge neutralizer
,”
J. Spacecr. Rockets
5
(
7
),
814
820
(
1968
).
12.
J. W.
Ward
and
H. J.
King
, “
Mercury hollow cathode plasma bridge neutralizers
,”
J. Spacecr. Rockets
5
(
10
),
1161
1164
(
1968
).
13.
W. R.
Kerslake
and
D. S.
Domitz
, AIAA Paper No. AIAA 1979-2064 (
1979
).
14.
C.
Othmer
,
K. H.
Glassmeier
,
U.
Motschmann
,
J.
Schule
, and
Ch.
Frick
,
Phys. Plasmas
7
,
5242
(
2000
).
15.
H.
Usui
,
A.
Hashimoto
, and
Y.
Miyake
,
J. Phys.
454
,
012017
(
2013
).
16.
I. B.
Bernstein
,
J. M.
Greene
, and
M. D.
Kruskal
, “
Exact nonlinear plasma oscillations
,”
Phys. Rev.
108
(
3
),
546
(
1957
).
17.
I. H.
Hutchinson
, “
Electron holes in phase space: What they are and why they matter
,”
Phys. Plasmas
24
(
5
),
055601
(
2017
).
18.
A. W.
Trivelpiece
and
R. W.
Gould
, “
Space charge waves in cylindrical plasma columns
,”
J. Appl. Phys.
30
(
11
),
1784
1793
(
1959
).
19.
N.
Nuwal
,
D. A.
Levin
, and
I. D.
Kaganovich
, “
Excitation of surface waves in 3D ion beam neutralization
,”
Phys. Plasmas
30
(
4
),
042109
(
2023
).
20.
S. R.
Oleson
,
L. M.
Burke
,
L. S.
Mason
,
E. R.
Turnbull
,
S.
McCarty
,
A. J.
Colozza
,
J. E.
Fittje
,
J. T.
Yim
,
M.
Smith
,
T. W.
Packard
, and
B. T.
Klefman
, “
Compass final report: nuclear electric propulsion-chemical vehicle 1.2
,”
Report No. NASA/TM/20230017131
(
2021
).
21.
M. J.
Patterson
and
T. R.
Verhey
, “
5 kW xenon ion thruster lifetest
,” AIAA Paper No. AIAA 1990-2543,
1990
.
22.
J. S.
Sovey
,
R. H.
Vetrone
,
S. P.
Grisnik
,
R. M.
Myers
, and
J. E.
Parkes
, “
Test facilities for high power electric propulsion
,” AIAA Paper No. AIAA 91-3499,
1994
.
23.
J. W.
Dankanich
,
M.
Walker
,
M. W.
Swiatek
, and
J. T.
Yim
, “
Recommended practice for pressure measurement and calculation of effective pumping speed in electric propulsion testing
,”
J. Propul. Power
33
(
3
),
668
(
2017
).
24.
J.
Dankanich
,
M.
Swiatek
, and
J.
Yim
, “
A step towards electric propulsion testing standards: Pressure measurements and effective pumping speeds
,” AIAA Paper No. AIAA 2012-3737,
2012
.
25.
Y.
Zhao
,
J.
Wang
, and
H.
Usui
, “
Simulations of ion thruster beam neutralization using a particle–particle model
,”
J. Propul. Power
34
(
5
),
1109
1115
(
2018
).
26.
Y.
Hu
,
J.
Wang
, and
Q.
Sun
, “
Geometrically self-similar ion acceleration in collisionless plasma beam expansion
,”
Plasma Sources Sci. Technol.
29
(
12
),
125004
(
2020
).
27.
W. R.
Kerslake
,
D. C.
Byers
,
V. K.
Rawlin
,
S. G.
Jones
, and
F. D.
Berkopec
, “
Flight and ground performance of SERT II thruster
,” in
Proceedings of the 8th Electric Propulsion Conference
(
AIAA
, 1970), AIAA-70-1125/NASA-TM-X-52848.
28.
H. S.
Ogawa
,
R. K.
Cole
, and
J. M.
Sellen
, Jr.
, “
Measurements of equilibration potential between a plasma “thruster” beam and a dilute “space” plasma
,” AIAA Paper No. AIAA 1969-263,
1969
.
29.
J.
Wang
,
O.
Chang
, and
Y.
Cao
,
IEEE Trans. Plasma Sci.
40
(
2
),
230
(
2012
).
30.
I. C.
Bell
,
B. E.
Gilchrist
, and
J. K.
McTernan
, “
Analyzing miniature electrodynamic tether propulsion capabilities and the interaction with the low earth orbit–plasma environment
,” in
Proceedings of the AIAA Space Conference and Exposition
,
2013
.
31.
D.
Hall
,
R.
Kemp
, and
H.
Shelton
, “
Mercury discharge devices and technology
,” AIAA Paper No. AIAA 1967-669,
1967
.
32.
J. E.
Foster
,
G. J.
Williams
, and
M. J.
Patterson
, “
Characterization of an ion thruster neutralizer
,” AIAA Paper No. AIAA 2005-3881,
2005
.
33.
A.
Sengupta
,
J. R.
Brophy
, and
K. D.
Goodfellow
, “
Status of the extended life test of the Deep Space 1 flight spare ion engine after 30,352 hours of operation
,” AIAA Paper No. AIAA 2003-4558,
2004
.
34.
G.
Williams
,
T.
Hickman
,
T.
Haag
,
J. E.
Foster
, and
M. J.
Patterson
, “
Preliminary wear analysis following a 2000 hour test of the HiPEP ion thruster
,” in
IEPC
,
2005
.
35.
C. E.
Huerta
and
R. E.
Wirz
, “
Ion-induced electron emission reduction via complex surface trapping
,”
AIP Adv.
9
(
12
),
125009
(
2019
).
36.
M. J.
Patterson
and
K.
Mohajeri
,
Report No. NASA-TM-105578/IEPC 91-151
,
1991
.
37.
D. M.
Goebel
,
I.
Katz
, and
I. G.
Mikellides
,
Fundamentals of Electric Propulsion
(
John Wiley & Sons
,
2023
).
38.
W. R.
Kerslake
and
L. R.
Ignaczak
, “
SERTII1979-1981 tests: Plasma thrust and neutralizer measurements
,”
J. Spacecr. Rockets
19
(
3
),
236
240
(
1982
).
39.
W.
Kerslake
and
S.
Domitz
, Report No. NASA TM-79271/AIAA 79-2064,
1979
.
40.
Y.-C.
Feng
and
P. J.
Wilbur
, “
The influence of stray magnetic fields on ion beam neutralization
,” AIAA Paper No. AIAA 82-1945,
1945
.
41.
W.
Kerslake
and
L.
Ignaczak
, “SERT II 1980 extended flight thruster experiments,” in
Proceedings of the 15th International Electric Propulsion Conference
(
AIAA
,
1981
), AIAA-81-0665/NASA TM-81685.
42.
R. J.
Bell
,
N. G.
Davey
,
M.
Martinsen
,
R. T.
Short
,
C. G.
Gill
, and
E. T.
Krogh
,
J. Am. Soc. Mass Spectrom.
26
,
201
(
2015
).
43.
J.
Wang
,
D.
Brinza
, and
M.
Young
, “
Three-dimensional particle simulations of ion propulsion plasma environment for Deep Space 1
,”
J. Spacecr. Rockets
38
(
3
),
433
440
(
2001
).
44.
G. C.
Soulas
, “Design and performance of 40 cm ion optics,” in Proceedings of the 27th International Electric Propulsion Conference (NASA, 2001), IEPC-01-090/NASA TM-2001-211275.
45.
V. H.
Chaplin
,
J. E.
Polk
,
I.
Katz
,
G.
Williams
,
G. C.
Soulas
,
J.
Yim
, and
J. R.
Anderson
, AIAA Paper No. AIAA 2018-4812,
2018
.
46.
J.
Polk
,
V.
Chaplin
,
J.
Anderson
,
J.
Yim
,
G.
Soulas
,
G.
Williams
, and
R.
Shastry
, “
Modeling grid erosion in the NEXT ion thruster using the CEX2D and CEX3D codes
,”
J. Electr. Propul.
2
(
1
),
14
(
2023
).
47.
M. W.
Crofton
and
I. D.
Boyd
, “Origins of accelerator grid current: analysis of T5 grid test results”
J. Propul. Power
17
(1),
203
2011
(
2001
).
48.
J.
Wang
,
D. E.
Brinza
,
D. T.
Young
,
J. E.
Nordholt
,
J. E.
Polk
,
M. D.
Henry
,
R.
Goldstein
,
J. J.
Hanley
,
D. J.
Lawrence
, and
M.
Shappirio
, “
Deep Space One investigations of ion propulsion plasma environment
,”
J. Spacecr. Rockets
37
(
5
),
545
555
(
2000
).
49.
E. A.
Sartsev
,
I. D.
Kaganovich
, and
R. C.
Davidson
,
Nucl. Insrum. Methods Phys. Res., Sect. A
733
,
80
(
2014
).
50.
J. S.
Sovey
,
Report No. NASA-TM-82990
,
1982
.
51.
J. R.
Brophy
, AIAA Paper No. AIAA 90-2655,
1990
.
52.
R.
Shastry
,
R. R.
Hofer
,
B. M.
Reid
, and
A. D.
Gallimore
, “
Method for analyzing E × B probe spectra from Hall thruster plumes
,”
Rev. Sci. Instrum.
80
(
6
),
063502
(
2009
).
53.
R. P.
Doerner
,
D. G.
Whyte
, and
D. M.
Goebel
,
J. Appl. Phys.
93
,
5816
(
2003
).
54.
M. A.
Mantenieks
,
J. E.
Foster
,
P. K.
Ray
,
S. V.
Shutthanandan
, and
T. S.
Thevuthasan
, “Low energy xenon ion sputtering yield measurements,” in
Proceedings of the 27th International Electric Propulsion Conference
(IEPC, 2001), IEPC-01-309.
55.
J. D.
Williiams
,
M. M.
Gardner
,
M. L.
Johnson
, and
P. J.
Wilbur
, “Xenon sputter yield measurements for ion thruster materials,” in
Proceedings of the 28th International Electric Propulsion Conference
(NASA, 2003), IEPC-03-130/NASA CR-2003-212306.
56.
K. K.
de Groh
,
B. A.
Banks
, and
C. A.
Karniotis
, “NSTAR extended life test discharge chamber flake analyses,” in
Proceedings of the 40th AIAA/ASME/SAE/ASEE Joint Propulsion Conference and Exhibit
(
AIAA
, 2004), AIAA-2004-3612/NASA TM-2005-213195.
57.
R.
Shastry
and
G. C.
Soulas
, “Post-test examination of NASA’s evolutionary xenon thruster long-duration test hardware: Discharge chamber,” in
Proceedings of the 52nd AIAA/SAE/ASEE Joint Propulsion Conference
(
AIAA
, 2016), AIAA-2016-4630/NASA TM-20210009633.
58.
G.
Soulas
and
R.
Shastry
, “Post-test examination of NASA’s evolutionary xenon thruster long-duration test hardware: Ion optics,” in
Proceedings of the 52nd AIAA/SAE/ASEE Joint Propulsion Conference
(
AIAA
, 2016), AIAA-2016-4632/NASA TM 20210009717.
59.
G. C.
Soulas
, “
The impact of back-sputtered carbon on the accelerator grid wear rates of the NEXT and NSTAR ion thrusters
,” in
International Electric Propulsion Conference (No. IEPC-2013-157)
,
2013
.
60.
M. W.
Crofton
,
J. C.
Nocerino
,
W. A.
Cox
,
J. A.
Young
,
K. D.
Diamant
,
E. J.
Beiting
, and
G. C.
Soulas
, “
Next ion engine plume deposition rates: QCM measurements
,” AIAA Paper No. AIAA 2014-0140,
2014
.
61.
J.
Pollard
,
K.
Diamant
,
M.
Crofton
,
M.
Patterson
, and
G.
Soulas
, AIAA Paper No. AIAA 2010-6779,
2010
.
62.
J. S.
Miller
,
S. H.
Pullins
,
D. J.
Levandier
,
Y. H.
Chiu
, and
R. A.
Dressler
,
J. Appl. Phys.
91
,
984
(
2002
).
63.
N. A.
Arthur
and
G. J.
Williams
, “Near field probe measurements in the plume of a NEXT ion thruster,” in
Proceedings of the 36th International Electric Propulsion Conference
(
NASA
, 2019), IEPC-2019-167/GRC-E-DAA-TN72319.
64.
P.
Wilbur
, “
A model for nitrogen chemisorption in on thrusters
,” AIAA Paper No. AIAA 1979-2062,
1979
.
65.
V.
Rawlin
and
M.
Mantenieks
, AIAA Paper No. AIAA 78-665,
1978
.
66.
H.
Gold
,
R.
Rulis
,
F.
Maruna
, and
W.
Hawersaat
, “
Description and operation of spacecraft in SERT I ion thruster flight test
,”
Report No. NASA-TM-X-1077
,
1965
.
67.
R. J.
Cybulski
,
D. M.
Shellhammer
,
R. R.
Lovell
,
E. J.
Domino
,
J. T.
Kotnik
,
J.
Cybulski
, and
R. R.
Loveli
,
Report No. NASA TN D-2718
,
1965
.
68.
W. R.
Kerslake
and
L. R.
Ignaczak
, “Development and flight history of the SERT II spacecraft,”
J. Spacecr. Rockets,
30
(
3
),
258
290
(
1993
).
69.
S. G.
Jones
,
J. V.
Staskus
, and
D. C.
Byers
, Report No. NASA-TM-X-52856, 1970.
70.
W.
Kerslake
and
L.
Ignaczak
, AIAA Paper No. AIAA 91-0665aR,
1991
.
71.
J. E.
Polk
,
R. Y.
Kakuda
,
J. R.
Anderson
,
J. R.
Brophy
,
V. K.
Rawlin
,
J.
Sovey
, and
J.
Hamley
, “
In-flight performance of the NSTAR ion propulsion system on the Deep Space One mission
,”
IEEE Aerospace Conference. Proceedings (Cat. No. 00TH8484), Big Sky, MT
(IEEE,
2000
), Vol.
4
, pp.
123
148
.
72.
H.
Kuninaka
and
P.
Molina-Morales
,
Acta Astronaut.
55
,
27
(
2004
).
73.
T.
Muranaka
,
M.
Kato
,
K.
Ueno
,
S.
Hosoda
, and
K.
Nishiyama
, “Estimation of charging for HAYABUSA2 by analysis of surface erosion in ion thruster operation,” in
Proceedings of the 37th International Electric Propulsion Conference
(IEPC, 2022), IEPC-2022-275.
74.
A.
Nono
,
T.
Morishita
,
S.
Hosoda
,
R.
Tsukizaki
, and
K.
Nishiyama
,
Acta Atronaut.
212
,
130
(
2023
).
75.
N. A.
Arthur
, “
Ion thruster produced roll torque
,” in
Proceedings of the AIAA Propulsion and Energy 2019 Forum
(
AIAA
,
2019
), AIAA 2019-4166/NASA CR 2020-220329.
76.
J. R.
Brophy
et al,
J. Propul. Power
25
,
1189
(
2009
).
77.
J.
John
,
L.
Roufberg
,
G. K.
Ottman
, and
E.
Adams
, “
NEXT-C lessons learned on the DART mission for future integration and test
,” in
IEEE Aerospace Conference
(
IEEE
,
2023
).
78.
H.
McEwen
,
J.
Foster
,
M.
Patterson
,
E.
Pencil
, and
E.
Diaz
, AIAA Paper No. AIAA 2006-5183,
2006
.
79.
J. E.
Foster
,
M.
Patterson
,
E.
Pencil
,
H.
McEwen
,
E.
Diaz
, and
J.
Van Noord
,
Report No. NASA-TM-2006-214402
,
2006
.
80.
B. R.
Weatherford
et al,
Plasma Sources Sci. Technol.
21
,
055030
(
2012
).
81.
L. A.
Ribarov
,
J. A.
Wehrmeyer
,
S.
Hu
, and
R. W.
Pitz
,
Exp. Fluids
37
,
65
(
2004
).
82.
A. V.
Phelps
and
Z. L.
Petrovic
,
Plasma Sources Sci. Technol.
8
,
R21
(
1999
).
83.
M.
Patterson
, private communication (
2022
).