Electric propulsion systems are becoming increasingly important for all satellite classes, since they require considerably less fuel than chemical alternatives, thereby making a variety of missions of interplanetary and interstellar flights (DeepSpace, BepiColombo, LISA, Euclid, etc.) and complex satellite constellations or formations (OneWeb, Starlink, Kuiper, UN:IO etc.) feasible. In parallel, small satellites, especially micro- and nanosatellites, are gaining in importance and their demand is growing rapidly. Megaconstellations of small satellites for Earth observation, telecommunication, IoT as well as servicing and removal of space objects require a significant number of efficient, robust and cost-efficient propulsion systems.

In the framework of a joint project funded by the Deutsches Zentrum für Luft- und Raumfahrt (DLR), an Engineering Model (EM) of the IonJet electric propulsion system based on a gridded ion thruster (GIT) with a maximum thrust level of up to 1 mN for small satellites has been developed and tested [1, 2]. The EM consists of a GIT with an inductively coupled plasma discharge including a plasma bridge neutralizer (CTU - Coaxial Thruster Unit), the Power Processing Unit (PPS), a Propellant Storage Tank (PST), a Fill and Drain Unit (FDU), a dedicated Flow Control Unit (FCU) and the Microflow Control Unit (MCU). The entire EM ist controlled by an adapted Propulsion System Control Unit (PSCU). The overall concept for the EM and particular responsibilities of involved project partners are shown in Fig. 1.

Fig. 1
figure 1

Overview of the IonJet EM Main Assemblies (IOM in blue, Aerospace Innovation in green)

In a first step, the electrical power of the system required to achieve a defined thrust and specific impulse has been estimated for xenon gas as propellant based on the power balance equation between power absorbed by the plasma and total energy loss of interaction processes between neutrals, ions, electrons and the walls. The assessed power consumption also includes the power for ion extraction, electron generation in the neutralizer and the operation of the control units.

On the basis of these calculations as well as the aim to achieve a compact size of the propulsion system and to make use of low-cost commercial-off-the-shelf (COTS) components if possible, the requirements on thruster and neutraliser, power supply system and flow control unit were established. Afterwards, the components were designed, manufactured, integrated and tested on component, subsystem and system level regarding functionality and performance.

For both thruster and neutralizer it is planned to use xenon as propellant, however, the use of other gases like argon or krypton is possible as well. First tests using krypton have been performed and the performance compared to xenon.

Results and Discussion

Calculation of the total power consumption of the Thruster subsystem

The total power \(P_t\) available on a satellite platform and the choice of the thruster technology determines the maximum achievable thrust and specific impulse for the electric propulsion system. For the IonJet EM in this work, a gridded ion thruster will be assumed. The total power is required by the thruster unit for beam extraction \(P_b\), plasma generation \(P_{rf}\), neutralization of the beam \(P_n\) and control of the system \(P_c\) as shown in Eq. 1:

$$\begin{aligned} P_t = P_b + P_{rf} + P_n + P_c \end{aligned}$$

The power supplies in the individual assemblies have power efficiencies \(\eta\) that are less than 1, so that power losses will occur. These efficiencies are not taken into account in the calculations below.

In the following, the achievable thrust is estimated under the condition that only limited electrical power is available. For gridded ion thrusters, the thrust \(T_i\) generated by ions of mass \(m_i\) can be easily calculated from the electrical parameters beam voltage \(U_b\) and ion current \(I_b\), because the energy of the extracted ions is only shifted by the plasma potential \(U_p\) and has a narrow energy distribution (Eq. 2 with \(q_0\) as elementary charge). The plasma potential can be determined from the measurement of the ion energy distribution of the extracted ions using energy-selective mass spectrometers (ESMS) or retarding potential analyzers (RPA). These measurements show that the plasma potential is typically less than 25 V and can therefore be neglected in the power estimation.

$$\begin{aligned} T_i = v_i \cdot \frac{d m_i}{dt} = I_b \sqrt{\frac{2 m_i (U_b + U_p)}{q_0}} \approx P_b \sqrt{\frac{2 m_i}{U_b q_0}} \end{aligned}$$

The extracted ion current is given by the plasma density \(n_p\) in the plasma sheath, the electrical potentials at the grids of the grid system and the geometry of the grid system itself. The plasma density can be calculated on the basis of the energy balance Eq. [3]. The absorbed RF power \(P_{abs}\) is equal to the total power loss (Eq. 3) which includes excitation, ionisation and elastic collision processes between ions, electrons and neutrals, and the losses by interaction with the chamber wall.

$$\begin{aligned} P_{abs} = q_0 n_p u_B A_{eff} E_T \end{aligned}$$

\(u_B\) is the Bohm velocity, \(A_{eff}\) the effective loss area for the excitation processes in the plasma and \(E_T\) the total energy loss per electron-ion pair. This method was already used for the calculation of the plasma bridge neutralizer based on an inductively coupled discharge [4]. The plasma density calculated with this method is used to calculate the extraction of ions through a grid system. The trajectory software IGUN was used to calculate the ion trajectories and to optimize the geometry of the grids. Figure 2 shows the maximum achievable thrust for three different RF powers (6 W, 10 W and 14 W) as a function of the total input power, i.e. the sum of RF power \(P_{rf}\) and beam power \(P_b=I_b \cdot U_b\).

Fig. 2
figure 2

Calculated thrust as a function of total input power for different RF powers based on the power balance equation as theoretical model

BIROS resistojet propulsion system MICROJET 2000

The MICROJET 2000 is a modularly designed “green” propulsion system developed by Aerospace Innovation GmbH (AI) [5, 6] in cooperation with DLR for nano-, micro- and mini-satellites based on the gas-resistojet concept and was launched 2016 on the BIROS satellite. MICROJET 2000 consists of a propellant storage tank for nitrogen which is filled or drained, respectively, through a fill and drain unit, a flow control unit responsible for the control of correct propellant mass flow, as well as redundant Thruster Units (THU) (see Fig. 3). Each of these THUs contains a plenum, a pulse valve and a nozzle for the actual thrust generation.

Fig. 3
figure 3

MICROJET 2000 propulsion system architecture (left) and implementation on BIROS satellite (right)

The BIROS mini-satellite architecture is based on DLRs space proven technology, developed within the frame of the very successful BIRD and TET-1 missions [7]. BIROS is equipped with a modular MICROJET 2000 resistojet propulsion system (see Fig. 4) configured into two independent lines comprising all the functional elements. After the successful completion of the AVANTI experiments in 2016 MICROJET 2000 is currently being used for station keeping and regular collision avoidance maneuvers.

Fig. 4
figure 4

BIROS satellite with the propulsion system and payload segment (left) and BIROS flight model (right)

Development of the IonJet EM

The IonJet EM consists of the thruster subsystem EM developed by IOM and the flow control subsystem EM developed by AI. At an early stage of the IonJet project it was decided to build as much as possible on already available space qualified parts and subsystems or COTS components. Regarding the flow control subsystem, AI has successfully demonstrated the operation of the MICROJET 2000 propulsion system on the BIROS satellite since 2016. Therefore, several space qualified subsystems were available for the flow control subsystem of the IonJet EM and thus considerable cost and time savings were possible. In addition, the modular design of the propulsion system enables a wide range of applications without the need for costly new developments.

For the thruster subsystem EM, the market was screened for COTS components for the DC power supplies needed to bias the grids of the ion extraction system as well as for an RF generator needed for plasma generation in the discharge chamber of the thruster. In both cases, no suitable components were identified that could be used as they were. Therefore, the power processing unit of IonJet EM is based on commercially available components that were either modified for in-vacuum use in case of the DC power supplies or served as building blocks in the RF generator developed in-house.

Thruster subsystem EM

Conventional GITs consist of the ion thruster to generate thrust and a separate neutralizer to compensate the space charge of the ion beam. This requires power supplies for both the operation of the ion thruster and the neutralizer as well as the necessary gas supplies. Within this project, a new design for a compact GIT with integrated neutralizer was developed (patent pending) that combines the plasma generation for the ions and for the electrons in a specially designed discharge chamber where two electrically isolated plasmas are generated (see Fig. 5). In addition to the plasma chamber, the grid system has also been re-designed. It has a ring-shaped structure with a hexagonal arrangement of grid holes in the outer area for extracting the ions whereas the inner region is designed to extract electrons. This allows simultaneous extraction of both ions and electrons from the discharge chambers. Due to this design, only one RF generator including a matching network and one gas supply unit are required for plasma generation. Furthermore, a separate electrical supply for the neutraliser could be spared by combining it with the power supply for the thruster. This saves space and reduces power losses as well as overall complexity of the propulsion system. The electrical wiring of the power supply always ensures that the extracted ion current is compensated by the extracted electron current.

A significant advantage of using an RF discharge for electron generation is the relaxed demand on propellant purity compared to hollow cathode neutralizers. This allows using less pure and therefore cheaper xenon gas.

Fig. 5
figure 5

Schematic of the gridded ion thruster with integrated neutralizer

The thruster unit consists of the gridded ion thruster with the integrated neutralizer, the module for igniting the plasma, the RF generator and the DC power supplies for thrust generation (Fig. 6).

Fig. 6
figure 6

Gridded ion thruster (top) with power processing unit (bottom)

For the selection of the materials and individual components, especially for the electronic assemblies, care was taken to use COTS components whenever possible, at least as the basis for the power processing unit. The PPU consists of two parts. One is providing the bias voltages for the ion extraction and is based on customized commercial high-voltage modules. The other one is a compact RF generator developed at the IOM for powering the plasma discharge. It comprises a commercially available DC-DC converter as the main power supply and a single board design containing the actual RF generator module as well as supporting modules like smaller DC-DC converters, a module to determine output and reflected power and further modules to measure two temperature values including the creation of several status indications like an over-temperature warning. The actual RF generator consists of three parts, the signal generation by a common quartz oscillator, an integrated circuit as the gate driver and finally a compact RF amplifier formed around a single LDMOS-transistor.

All modules of the RF generator have been tested separately and as a whole unit including tests in vacuum and with the described thruster as a real load. RF output powers larger than 20 W at 13.56 MHz and efficiencies larger than 74% could be achieved.

The individual assemblies were examined with regard to their vacuum suitability and thermal stability during operation in a vacuum. During the tests, the temperature was measured as a function of time using a thermocamera in order to identify individual hot spots (see Fig. 7). Based on the tests, critical components have been replaced by more suitable ones in order to achieve stable operation in vacuum.

Fig. 7
figure 7

False color thermal image of the RF generator assembly under vacuum operation. The hotspots in yellow are coils for RF filtering. Their temperatures remain below 100\(^\circ\)C which is considered uncritical

Flow Control subsystem EM

As already mentioned above, it was decided for IonJet EM to build as much as possible on already available space qualified parts and subsystems. For this reason, the IonJet flow control subsystem was designed in consideration of the MICROJET main assemblies FCU and FDU, which were in the meantime further developed in close cooperation with smallsat manufacturers, DLR and TU Berlin. As part of the development process the design requirements and interfaces of all main assemblies were iterated in close cooperation between AI and IOM. A dedicated mathematical simulation model was then developed for the entire IonJet flow control subsystem. During the design process of the PSCU care was taken to ensure that it is capable of controlling the entire IonJet EM. In parallel, all main assemblies were designed, manufactured and integrated. After performing functional and leak tests the main assemblies were integrated into the IonJet flow control subsystem.

In the first step, the IonJet flow control subsystem was pressurized with nitrogen and helium and first functional, leakage and flow tests were performed at 10\(^{-6}\) mbar in the high vacuum chambers of AI and IOM (see Fig. 8). The desired flow was set at FCU and MCU, afterwards the flow control subsystem was filled with xenon gas of purity 4.8 and the flow was fine-tuned in the vacuum chamber at IOM. After completion of these initial tests the flow control subsystem was integrated with the thruster subsystem to IonJet EM and successfully tested in the vacuum chamber of IOM. Initial EM-level tests showed that the flow control subsystem performed as desired and provided the required xenon flow.

Fig. 8
figure 8

IonJet flow control subsystem testing: (a) in AI HVC Vacuum Chamber, (b,c) testing at IOM

Experimental characterization of the IonJet EM

In order to characterize the operation and performance of the newly designed ion thruster, it was installed together with the PPU in a vacuum test chamber at IOM. The Xe gas was provided by a conventional mass flow controller in this test campaign. The total volumetric flow for ion thruster and neutralizer ranged between 1.29 sccm and 1.58 sccm. Various ignition procedures were investigated to determine the optimum setting for the gas flow and gas peak. Figure 9 shows the thruster in operation with the plume of the neutralizer visible in the center and the ion beam forming concentrically.

Fig. 9
figure 9

The gridded ion thruster with integrated neutralizer in operation

Various performance tests of the thruster subsystem were carried out using two grid configurations: (i) an unrestricted grid system with 237 grid holes and (ii) an restricted grid system with only 114 grid holes. Figure 10 shows the extracted beam current \(I_b\) as function of RF power for a beam voltage of \(U_b=1000\) V and Xe volumetric flow rate of 1.29 sccm for the restricted grid system with 114 grid holes. The experimental values are in good agreement with the simulations based on the power balance equation.

Fig. 10
figure 10

Extracted ion beam current as function of the RF power in comparison to simulated values

Tables 1, 2 and 3 summarize thruster performance parameters (beam current, thrust, beam power, specific impulse) in dependence on the beam voltage and RF power for the unrestricted grid system with 237 grid holes. Thrusts well in excess of 1 mN have been achieved indicating potential for miniaturization of the thruster if 1 mN maximum thrust is envisaged. This would also allow to reduce the number of grid holes, the gas consumption and increase the specific impulse \(I_{sp}\). With the present configuration of the engineering model, \(I_{sp}\) values up to \(\sim 1100\) s have been demonstrated. The mass efficiencies \(\eta _m\) ranged between 0.15 and 0.22 within the RF power range of 6 W to 14 W and already include the additional gas consumption of the neutralizer. Neglecting power losses in the power supplies, the thrust to power ratio values were in the range \((0.030 - 0.036)\) mN/W.

Table 1 Measured performance parameters in dependence on the beam voltage for an RF power of 6 W and a volumetric Xe flow rate of 1.58 sccm. The thrust and specific impulse \(I_{sp}\) were calculated from the measured values using Eq. 2 and Eq. 4, respectively
Table 2 Measured performance parameters in dependence on the beam voltage for an RF power of 10 W and a volumetric Xe flow rate of 1.50 sccm. The thrust and specific impulse \(I_{sp}\) were calculated from the measured values using Eq. 2 and Eq. 4, respectively
Table 3 Measured performance parameters in dependence on the beam voltage for an RF power of 14 W and a volumetric Xe flow rate of 1.50 sccm. The thrust and specific impulse \(I_{sp}\) were calculated from the measured values using Eq. 2 and Eq. 4, respectively

As the last step of testing the IonJet EM, the flow control subsystem EM and the thruster subsystem EM were integrated to the IonJet EM (see Fig. 11). The dimensions of IonJet EM in centimeters are \(\text{ length }\times \text{ width }\times \text{ height }=33\times 33\times 20\). The weight is approx. 9 kg. The whole system was mounted in a vacuum chamber at IOM and the Xe flow rates were optimized for thruster operation and gas peak. Afterwards, the thruster was successfully ignited several times and stable operation of the IonJet EM was demonstrated, including a refilling of the plenum by the FCU under thruster operation.

Fig. 11
figure 11

The IonJet EM: (a) front-side view, (b) top view, (c) front view after mounting in the vacuum chamber, (d) in operation. The vertical box left from the thruster is an additional device to measure the reflected RF power and not integral part of the EM

Performance comparison between krypton and xenon

IonJet was also tested with krypton as an alternative propellant using the restricted grid system. No changes or optimizations were made to the thruster with regard to using krypton before the test. The first ignition tests showed that the gas flow of krypton had to be increased by about 50% compared to xenon to ensure stable operation. Figure 12 shows the extracted ion current for krypton and xenon as a function of the RF power at a beam voltage of 1200 V.

Fig. 12
figure 12

Extracted ion current at a beam voltage of 1200 V for operation with xenon and krypton at different RF powers

For a better comparison of the two operating modes, the measured values were modelled by a linear function and extrapolated beyond the measuring range. The linear approach is justified also due to the linear relationship between plasma density and absorbed power (RF power) according to Eq. 3. The thrust for the two operating modes can be calculated from the ion current and the beam voltage according to the Eq. 2. Figure 13 shows the calculated thrust of the ion thruster as a function of the total power, i.e. the sum of RF power \(P_{rf}\) and beam power \(P_b=I_b \cdot U_b\).

Fig. 13
figure 13

Calculated thrust for xenon and krypton operation as a function of total power based on a linear modelling of the measured data in Fig. 12

The total power required to generate a specific thrust is significantly larger for krypton than for xenon. The quotient of the two thrust-performance curves can be seen in the Fig. 14.

Fig. 14
figure 14

Ratio between calculated thrusts for krypton and xenon as a function of total power based on a linear modelling of the measured data in Fig. 12

The thrust ratio between krypton and xenon in the investigated power range between 10 W and 50 W total power is approx. 0.55-0.68 depending on RF power for the described IonJet EM propulsion system. The thrust ratio increases with increasing power, however, a factor close to or even exceeding 1.0 is not expected, because the ionization potential of krypton (\(U_i=\,\)13.999 V) is larger than that of xenon (\(U_i=\,\) 12.139 V), the collision cross sections for electron impact ionization are smaller for krypton than for xenon [8] and the beam current \(I_b\) must be larger by a factor \(\sqrt{131/84}=1.25\) for achieving the same thrust for a given beam voltage \(U_b\) due to the lower ion mass of krypton compared to xenon (see Eq. 2).

Another important parameter that describes the efficiency of the used propellant is the specific impulse \(I_{sp}\). The specific impulse is defined as the ratio of the change in momentum to the change in propellant mass. For gridded ion thrusters, the specific impulse can be simply described by the beam voltage \(U_b\), the mass utilization \(\eta _m\) and the gravitational acceleration \(g_0 \approx\) 9.81 ms\(^{-2}\) (Eq. 4).

$$\begin{aligned} I_{sp} = \frac{\dot{p}}{\dot{m}} = \frac{\eta _m}{g_0} \sqrt{\frac{2 q_0 U_b}{m_i}} \end{aligned}$$

The ratio between the specific impulse of krypton and xenon in the power range from 10 W to 35 W is between 0.50 and 0.62 (Fig. 15). That is, the utilization of krypton as a propellant is worse than for xenon, although the mass of krypton is less than that of xenon, and thus a higher \(I_{sp}\) was to be expected. This is mainly due to the higher volumetric flow required for krypton to ensure stable operation of the IonJet EM in these first Kr tests.

Fig. 15
figure 15

Ratio of specific impulse between krypton and xenon as function of the RF power

In summary, the use of krypton for operating the IonJet EM thruster system leads to both smaller thrusts and lower \(I_{sp}\) values compared to xenon propellant under the condition that the same electrical power is available.


Within a joint project funded by the DLR, IOM and Aerospace Innovation GmbH have developed an engineering model of an electric propulsion system for small satellites based on a novel gridded ion thruster that inherently incorporates the neutralizer. This approach reduces the amount of components and consequently costs and complexity of the propulsion system. Furthermore, the demands on propellant purity can be relaxed compared to hollow cathode neutralizers which provides further potential for cost savings. In line with the requirements set out for this project, the IonJet EM weighs less than 10 kg and is capable to produce thrusts of 1 mN with a total power consumption of less than 50 W. Specific impulse values larger than 1000 s have been achieved as well.

In the next phase, funded by the DLR and starting in July 2022, an engineering qualification model (EQM) of IonJet called IonJet-Evo will be developed and tested by IOM and Aerospace Innovation GmbH. It is planned to significantly reduce the overall size and weight of the system, compatible with the Cubesat specifications for the 6U-16U-format while maintaining or improving the performance figures.