The primary focus of this paper is to characterize air ionization using AES. This study also suggests that by modulating the energy of AES and as such forming positive and negative ions, a low-power self-neutralizing SABPT can be achieved. Fig. 1(a) shows a vacuum arc source, a copper anode, cathode, and non-porous alumina ceramic. The cathode spot is formed on the ceramic between cathode and anode. This cathode spot can rotate under the influence of axial magnetic field in J (current density) × B (magnetic field) motion [13].
A complete electron source comprising a vacuum arc source and the electron extraction grid can be observed in Fig. 1(b). The extraction grid is positively biased with respect to the ground. The extracted electron current and energy are the function of applied grid potential. The electron current extracted saturates at some applied voltage. A typical electron source (hollow cathode) would require a vacuum environment, thermionic emitter heating system, and a separate propellant tank for its application. In this research, we have shown that our AES does not require those systems. Most importantly, AES can efficiently operate in the medium pressure range (10-4 – 10-1 Torr). The main objective is to utilize negative ions and positive ions created in the ionization process using AES leading to an extracted neutralized beam at the exit of SABPT. Both negative and positive ions will be accelerated by appropriate electric field arrangement and switching electrodes' polarity [1].
This paper describes the AES using intrusive and non-intrusive plasma diagnostics methods. Fig. 2. shows the schematics for the system design configuration starting with an inductive energy storage circuit [14] to power the AES. The inductance, capacitance, switch, input DC voltage, PWM (pulse width modulation) width, frequency, and the coil magnetic field values were 550 µH, 6800 µF, IXYH50N120C3D1 IGBT, 30 V, 610 µs, 5 Hz, and 0.15 T. During the stage 1, the IGBT gate was closed, and high emf is generated in the inductor as it charges.
The gate was open in the second stage, and the stored inductor energy was discharged into the vacuum arc source. With the addition of a capacitor, the overall discharge current was increased. The radius and length dimensions for the cylindrical cathode, anode, and ceramic are 5.5 and 40, 6.5 and 35, and 12.5 and 30 mm, respectively. These dimensions were chosen tentatively as our primary focus is studying ionization and optimization. The ceramic was thin enough for a good conduction gap and required a thin carbon paint layer [15] (100-200 Ω resistance between cathode and anode) to ignite the arc. The electron extraction grid is an aluminum grid aperture (diameter 4 mm for each opening) placed on a 50 mm diameter 3D printed extraction grid holder. The aperture was 25 mm away from the source to avoid arcing. The chamber was pumped down using a roughing pump to reach a base pressure of 0.05 Torr. An air leak allowed for a small air flow rate into the chamber. A Langmuir probe with a 2 mm wire length was placed in front of the grid for Langmuir probe experiments. A Faraday cup made up of aluminum was used for ion current measurements. For OES, a Stellar Net Inc. [16] OES spectrometer was coupled with an optical probe placed close to the quartz viewing port of the chamber. The SpectraWiz software was used on the PC to monitor the spectrum. Lastly, the magnetic filter (to measure negative ions) was designed for a length and diameter of 50 and 14 mm with a magnetic field of 0.12 T using Alnico permanent magnets (diameter and height as 22 and 8 mm).
2.1 Optimization using Magnetic Field
We designed and added an air-cored coil axial magnetic field of 0.15 T to the vacuum arc source to improve the performance. The magnet system does not require an external circuit, and it can be easily integrated into the source. The addition of a magnetic field causes cathode spot rotation [6,13] and a uniform cathode erosion at the cathode-ceramic interface to improve the performance of the source.
Additionally, plasma bends in the J × B direction leading to improvement in ionization, causing an increment in the plasma density and velocity. The ionization in the source was improved with an increase in the magnetic field. The goal was to obtain the optimum magnetic field condition for AES such that the air ion/electron current was maximized. An air-cored coil magnetic field was simulated in FEMM software [17], as shown in Fig. 3(a). An axisymmetric time-invariant field equations model was simulated with input conditions of the number of turns, wire gauge, input current, coil radius, and height as sixty-eight turns, AWG 24, 36 A, 12.5, and 13 mm. A peak magnetic field of around 0.15 T can be observed in the middle of the core.
Air ion and electron current experiments were performed at different pressures concerning applied magnetic field 0-0.25 T. As observed in Fig. 3(b), 0.15-0.2 T field showed peak performance, but 0.15 T was chosen. Ion and electron currents were measured using an Aluminum Faraday cup (diameter 56 mm) placed 35 mm away from the source by biasing it – (for ions)/+ (for electrons) 40 V. The magnetic field significantly increased the ion and electron/arc current ratio. Beyond 0.2 T, there is a reduction in ion or electron/arc current ratio, this is likely due to electron confinement leading to disruption of the arc current.
2.3 Langmuir Probe and Faraday Cup
The Faraday cup experiments were performed using a cup area of 0.03 m2. The current was calculated using Ohm’s law for a potential drop across a 100 Ω resistor. Additionally, the Langmuir probe was built with a titanium wire for thickness and length of 1, 2 mm exposed to the plasma (the remaining was shielded using a non-porous alumina ceramic). Results have been presented in section 3.1.
2.3 Optical Emission Spectroscopy (OES) and Plasma Chemistry
The OES technique gave an insight into plasma chemistry by studying the emissions in the visible spectrum. The emission lines are interpreted from the NIST database [18] for atoms and the molecular spectra book [19] for molecules. The results of OES can also assist in understanding plasma chemistry. The electron temperature, electron density, and energy distribution function parameters could be calculated. The natural logarithmic equation (1) is based on two selected line spectrums [20,21]
Here i, I, λ, A, g, kB, Te, and E are spectral line number, spectral line intensity, wavelength (nm), transmission probability, statistical weight, Boltzmann constant (1.38 × 10-23 m2 Kg s-2 K-1), electron temperature (eV) and energy (eV). The equation (1) mentioned above can be curve fitted to obtain electron temperature. When plugged into the ion saturation current equation from the Langmuir probe, the electron temperature can give electron density. Alternately, the Saha equation (Ref. 22 and 21) could also be used to obtain electron density. The above equations are based on the local thermodynamic equilibrium (LTE) condition. The vacuum arc sources have EEDF (electron energy distribution function) in the Maxwellian regime [21] due to LTE and high collision frequency. This condition can be used for our case because the incoming airflow is exceptionally low density. The next step is to obtain the reaction rate coefficient for attachment and ionization reactions. The rate coefficients can determine the rate of positive and negative ions formation based on the plasma chemical reactions and their transport coefficients [23]. The rate coefficient is given by [24,25],
Here, e, and φ, me, and σk (ε) are electron charge, electron energy (eV) and mean electron energy (eV), the mass of the electron (9.1 × 10-31 Kg), and electron impact collisional cross-sections (m2). The formation rate for specific reactions can be obtained by multiplying rate coefficients with their respective reactants' number densities. The above equations 1 and 2 consider LTE. The Boltzmann equation shall be solved in the case of non-Maxwellian distribution (non-thermal plasma) because the Boltzmann transport equation assumes the effect of the inelastic collision. It is computationally expensive. A Boltzmann solver can be used to compute EEDF [23,26] when the EEDF distribution is non-Maxwellian. The GUI of the solver takes an input as a collision cross-section database which can be obtained from the LXcat [27] website. The other inputs are gas temperature, ionization degree, plasma density, mean electron energy, and electric field by number density ratio to compute rate and transport coefficients. Additionally, EEDF and rate coefficients can provide adequate knowledge of plasma chemistry. Hence, we use it to verify the ion formation concerning our parameters, such as extraction grid voltage and pressure.
2.4 Magnetic Filter Design
The negative charge in the air plasma is a blend of electrons and negative ions (O- is dominant [28]). These charges can be differentiated based on their mass, gyro radius, and velocity (distance/delay time). The O- ions are typically formed at lower energies due to the dissociative attachment process. However, the ions get destroyed mainly because of electron impact detachment or mutual neutralization with O+ and O2+ ions [29]. Therefore, researchers have attempted to measure and study negative oxygen ions using the experimental and modeling approach. For instance, McKnight [30] investigated drift velocities and rate constants for negative oxygen (in oxygen plasma) ions experimentally as a function of electric field-neutral density ratio, pressure, and gas temperature, to verify the calculations with their numerical results. In the field of a helicon wave discharge, Mieno et al. [31] experimentally studied negative and positive oxygen ions formation in an oxygen plasma due to rf (radio frequency) power modulation (on-off power) using a time-of-flight mass spectrometer. Additionally, Zhang et al. [32] used a floating harmonic method to investigate negative ion density and electronegativity variation with the radial distance, gas pressure, and power in an inductively coupled plasma. Regarding a mathematical model, the authors [33] obtained a temporal variation of negative ion density by solving a 1D hydrodynamic drift model (motion of charges, ionization, and recombination reactions in a spatial-temporal varying electric field). While most works were conducted for an electronegative gas (O2) plasma, the research lacked experimental negative oxygen ion data for air plasma.
To this end, we propose an approach based on partial magnetization through magnetic field confinement of electrons. The negative ions drifting towards the probe will allow ion current measurement. The magnetic filter length was decided based on the criteria that the dimension (LTube) was selected between the range of Larmor radius for negative ions and electrons. This condition permitted effective electron confinement while ensuring negative ions drift towards the current measuring electrode 2. The measuring tube design system was electrically floating to prevent positive ions from entering electrode 1 (due to the potential difference between electrode 1 and the ground). The length of the tube was inversely related to the applied magnetic field.
The tube design is guided by partial magnetization condition:
where mi, UBattery, ue, ui and B are oxygen ion mass (2.65 × 10-26 Kg), applied voltage, initial electron, initial ion velocity (ratio of the distance between AES-tube and delay time), and the applied magnetic field using Alnico permanent magnet (0.12 T). Based on equation 3, a tube length of 5 cm for an applied, voltage of 80 V was selected for the experiments such that electrons were radially confined, and the negative ions were not magnetized. The applied voltage results in an electric field between the plasma, electrodes 1 and 2 (Fig. 4) extracting the negative charges (electrons and negative ions) as they move from low to high potential, simultaneously repelling the positive charge. Electrons are confined radially using the radial magnetic field. Even though a strong magnetic field was applied, some electrons would still escape the magnetic field due to anomalous electron transport/because of their high mobility. This resulted in the formation of 2 peaks in the time-varying current waveforms. The current was measured using Ohm’s law for a potential drop across a 15 Ω resistor. Later, we discuss the results pertaining to negative ion current measurements in section 3.3.