3.1. Low Working Current
Under the condition that the anode current 2 A was fixed, the outlet of the keeper was a bright violet plume, and the position of the port of the keeper showed an obvious contraction of the bright white spot, as shown in
Figure 8. At this time, the anode voltage was also relatively stable, and the voltage fluctuation did not exceed 2 V.
The orifice size is the most critical dimension in the design of a hollow cathode. Compared with other structural dimensions, the cathode orifice size directly affects the working range of current, cathode working temperature, and discharge voltage. In this paper, two types of cathodes with different orifice sizes were calculated. The orificed emitter configuration refers to the structure type, where the size of the cathode orifice is small, and the emitter is completely wrapped in the cathode tube [
20,
21]. In traditional low-current Bao-W and Lab6 cathode designs, the orifice size of the cathode is usually less than 1 mm. The main purpose is to maintain a high plasma density in the cathode orifice and emitter areas to prevent the cathode from working at too low a temperature. For the C12A7 electride, this material, with a low work function, is suitable for working under low-current conditions, while too small a cathode orifice may cause the emission temperature of the C12A7 electride to be too high. The open-end emitter orifice keeper configuration was first designed by Aston et al. [
22]. This type of structure is characterized by a large cathode orifice size or the absence of a cathode top, which causes the emitter to be exposed to the plume region. It is often used in a hollow cathode without a high current. The discharge stability of a hollow cathode under a high discharge current can be improved effectively by increasing the orifice size of the cathode orifice and reducing the working voltage. In this paper, the size of the cathode orifice in the C12A7 hollow cathode was 3 mm. In addition to reducing the working temperature under high currents, exposing the emitter can also realize the direct measurement of the working temperature of the emitter of the C12A7 electride.
The atomic number density reflects the amount of gas in the cathode and affects the frequency of ionization collision.
Figure 9 shows the simulation distribution results of the atom number density of the small-orifice cathode (1 mm) and the open-orifice hollow cathode (3 mm) under the state of steady discharge, where the anode current is 2 A and the xenon flow rate is 10 sccm. As can be seen from the density distribution of the xenon atoms inside and outside the cathode, the density of the xenon atoms inside and outside the cathode differed greatly. For the cathode with small pores, the density of the xenon atoms inside the cathode reached up to 6.7 × 10
22/m
3 due to the flow limiting and pressure boosting effect of the cathode orifices, slowly dropping to 5.9 × 10
22/m
3 in the emitter region. Then, it dropped to 2.5 × 10
22/m
3 at the cathode orifice. Compared with the small-orifice cathode, the inner atomic number density of the open-orifice hollow cathode obviously decreased, and the maximum atomic number density was about 4.3 × 10
22/m
3. However, the atomic number density at the cathode outlet increased slightly, about 2.7 × 10
22/m
3. The enlargement of the cathode orifice size reduced the throttling effect of the cathode orifice, resulting in decreased pressure inside the cathode and increased pressure at the outlet.
As can be seen from the electron number density simulation distribution cloud,
Figure 10 shows that the area near the orifice was the area with the highest electron number density inside the hollow cathode, indicating that this area is the location where the particle collision reaction was the most intense. The number density of electrons in the cathode reached a maximum value of about 3.4 × 10
20/m
3 near the cathode orifice, and there was a second small wave peak of about 1.2 × 10
20/m
3 at the cathode orifice outlet. For the open-orifice hollow cathode, the decrease in the neutral atom density inside the cathode led to a decrease in particle collision frequency and plasma density. However, it was also more evenly distributed. Therefore, the region with the highest electron number density in the cathode is located near the emitter and extends outwards, with a maximum value of about 1.45 × 10
20/m
3.
Figure 11 shows the temperature distribution of the C12A7 electride along the axial direction. It can be seen that the downstream of the emitter near the top of the cathode was the part with the highest temperature of the whole cathode. The open-orifice hollow cathode (3 mm) was 50 K lower than the small-orifice cathode (1 mm). This shows that the enlargement of the cathode orifice size is beneficial to reduce the risk of overheating of the C12A7 electride. The smaller orifice size increased the pressure on the cathode and shifted the peak plasma density closer to the cathode pore, which increased the power deposited on the surface of the plasma bombardment in the region, causing the emitter temperature to rise. With the increase of the discharge current, both the emitter temperature and the overall cathode temperature of C12A7 rose. When the working current was 4 A, the maximum temperature of the C12A7 electride reached 1520 K, which is close to the maximum allowable working temperature of 1548 K for the electron emissions of C12A7. The element analysis in the literature [
12] shows that, at more than 1548 K, the C12A7 electride will decompose into C3A and CA, which is detrimental to the hollow cathode’s ability to work stably for a long time. Therefore, the C12A7 hollow cathode is reasonable under low-current working conditions, and its maximum safe working current is 4 A. If it is necessary to increase the safe working current, it is necessary to use copper material with higher thermal conductivity and optimize the heat dissipation structure for the design of a C12A7 hollow cathode.
In addition to the increased the plasma bombardment heat flow, it is also believed that the ohmic heat, caused by high resistance, is related to the high temperature of the emitter. It is difficult to obtain the enthalpy change parameters of the C12A7 electride in a molten state. It is also unknown under what conditions the absorption of heat by melting must be considered for the C12A7 electride. As a result, the computational model of the C12A7 hollow cathode under high-current conditions showed difficulty in reflecting the experimental results.
3.2. High Working Current
With the increase of the discharge current, the anode voltage demonstrated a transition phenomenon, suddenly increasing from more than 10 volts to about 20 volts, and the shock was serious.
Figure 12 shows the anode voltage jump during high-current operation. The plasma plume at the cathode outlet changed obviously with the change in anode voltage. The cathode was bright violet at start-up and under low-current conditions, as shown in
Figure 13a. However, with the increase of the discharge current, the plume color gradually changed to a darker yellow, as shown in
Figure 13b. When the voltage transition was high, the cathode plume rapidly changed to a very bright orange color, as shown in
Figure 13c. As shown in
Figure 12, the simultaneous oscillation of the voltage and current usually lasted for several seconds to tens of seconds and then automatically jumped back to the low-voltage level, and the cathode plume then returned to normal. As the anode current continued to increase, the voltage transition became more frequent, occurring every few minutes. By comparing the appearance of the hollow cathode in
Figure 13b,c, it can be seen that the keeper of the C12A7 hollow cathode running at a high voltage shows a red-hot state, indicating that the temperature of the hollow cathode was higher at this time. A long high-voltage discharge duration is obviously detrimental to the performance and lifetime of the C12A7 hollow cathode.
Although the reason for this voltage transition was not clear during the experiment, the discharge state can be actively regulated by the contact keeper power supply. After the anode voltage transition occurred again for about 4200 s, the initially closed contact pole power supply was actively turned on, and the contact pole power supply parameters were about 2 A and 8 V. At this time, the anode voltage did not return to normal, but with the contact pole power supply turned off again, the anode voltage immediately returned to the low-voltage state. At 4800 s and 5100 s, it was confirmed that the switch of keeper power supply could make the C12A7 hollow cathode immediately jump from the high-voltage discharge state to the low-voltage discharge state. When the anode current exceeded 15 A, the phenomenon of automatic voltage transition did not appear again, but the anode voltage value was always high in the discharge process.
After the high-current discharge, the C12A7 hollow cathode was then subjected to repeated ignition experiments to study the effect of a high-current discharge on the starting performance of the C12A7 hollow cathode. The results show that the C12A7 hollow cathode can still start, and the ignition parameters do not change, obviously. Interestingly, the voltage transition occurred again at a 1–4 A low current and the cathode plume was no longer bright violet but dull yellow. This phenomenon implies that the voltage transition problem is not determined by the anode current but is more likely because the discharge behavior at a high current completely changes some key structures of the C12A7 hollow cathode, leading to irreversible changes in its discharge characteristics.
The examination of the C12A7 hollow cathode after the test also confirmed this speculation. As shown in
Figure 14a, the area between the keeper and the cathode tube is filled with a large amount of white ash, which may be the residue after the discharge of the emitter. The top of the cathode adheres to the black molten material, which should be the emission of the C12A7 electride after melting, as shown in
Figure 14b. The electride emitter in the cathode tube has almost been consumed, and only a small amount of graphite sleeve remains, as shown in
Figure 14c. In addition to the increased plasma bombardment heat flux, ohmic heat, caused by its own resistance, may be one of the possible reasons for the serious melting of the electride emitter under high-current conditions. A multimeter was used to measure the resistance values at both ends of the C12A7 electride used in the experiment before discharge. Although the size was the same, the resistance values of different emitters at room temperature were different, and the value was from about tens of ohms to hundreds of ohms. The difference in measured resistance may be related to the surface state of different materials, while the lanthanum hexaboride emitter of the same size was stable at about 0.9Ω. Suppose the phenomenon of the high resistance of C12A7 electride is caused by high body resistivity, such as low electron occupancy in the cage. In that case, the ohmic heat caused by the material’s own resistance will increase exponentially under high-current conditions, which may be the main reason for the large area of melting of the electride emitter under a high current.
After 10 A, the emitter temperature of the C12A7 hollow cathode exceeded the effective range of the colorimetric thermometer. Therefore, the emitter of the C12A7 hollow cathode not only melts at high temperatures under high currents but even flows down to the cathode orifice under the convection of gravity and working medium gas. This explains why the anode voltage switches between the high and low values. Hollow cathode discharge normally occurs between the surface of the emitter and the anode plate in the cathode tube. Once the emitter melts and deforms or even flows out of the cathode orifice, the discharge point of the emitter will be displaced, and the emission area will also change. When the discharge point migrates to the molten body outside the cathode tube (because the effective emission area of this part of the emitter is small), in order to maintain a sufficient amount of thermal emission electrons, the anode voltage will increase, and the higher discharge power will make the emitter temperature higher. Since this mode of decharging occurs outside the cathode tube and between the anode, the plume appears brighter. When a small amount of emitters outside the cathode tube is insufficient to maintain high-current discharge or when the contact pole power is turned on, the discharge point will be transferred to the emitter inside the cathode tube again. The effective emission area of this part of the emitter is larger, and only a lower discharge power is needed to maintain the thermal emission temperature and, thus, the anode voltage is low. Because the discharge point is in the cathode tube, the small orifice of the cathode blocks the discharge plume, and therefore, the color of the plume is relatively dim.
However, if the high resistance of the C12A7 electride is caused by high surface resistivity, then the ohmic heat may not be serious because the high-resistance insulating layer on the surface of the emitter is gradually removed during discharge. Since the resistance values of different C12A7 emitters before and after discharge vary greatly, which is obviously related to the discharge conditions experienced and the material’s surface state, the high-resistance behavior of C12A7 electride is more likely to be caused by the change in the surface layer structure. The total working time of a C12A7 hollow cathode under the conditions of a high current (more than 10 A) is not more than two hours, and the emitter in the cathode tube will almost be consumed; such a short life means that a C12A7 hollow cathode under the condition of a high current is difficult to utilize in practical applications.