ZHANG Haofeng ¹, LIU Zhuo ², LUO Bing ¹, XIAO Wei ¹, ZHONG Zheng ¹, WANG Jiaxi ², WANG Xilin ²

(1. Southern Power Grid Research Institute Co., Ltd., Guangzhou 510670, Guangdong, China; 2. Tsinghua Shenzhen International Graduate School, Shenzhen 518055, Guangdong, China)

Extended Abstract:[Background and purpose] The accumulation of surface charges can easily lead to surface flashover accidents. Regarding the issue of charge accumulation on the surface of basin-type insulators for DC-GIS and DC-GIL under DC voltage, as well as post insulators for ultra-high voltage DC wall bushings, existing research has mainly focused on the gas-solid interface charge accumulation laws of polymers such as epoxy resin and polytetrafluoroethylene at DC voltages. There is relatively few research on the charge accumulation and dissipation characteristics of ceramics and a lack of comparative studies on different ceramic materials, which limits the application of ceramic insulation materials. To clarify the mechanism of charge accumulation and dissipation on ceramic surfaces and to establish a model for charge accumulation and dissipation of ceramic materials under direct current electric fields, charges were injected into the surface of the samples with a needle-plate electrode. The surface charge density, of three ceramic materials, Al₂O₃, AlN and Si₃N₄, was tested using the electrostatic probe method, while their trap distribution characteristics were analyzed using the isothermal surface potential decay method.[Methods] Before the experiments, the humidity within the sealed chamber was controlled and stabilized at ±3% of the target value. The samples were cleaned with an ultrasonic cleaning machine and then dried in an oven for 12 h to eliminate any pre-existing surface charges. Subsequently, the samples were placed under the electrostatic probe, where the measured potential value was close to 0, indicating that the original charges had been mostly eliminated. The samples were then moved beneath the needle electrode, aligning the needle electrode with the center of the sample. The grounding switch of the displacement platform was closed and the power supply was turned on to apply a +10 kV voltage between the needle electrode and the platform in a stepwise manner for 3 min, injecting charges onto the surface of the sample. Then, the power supply was turned off, while simultaneously disconnecting the grounding switch of the displacement platform to maintain the floating potential. The samples were then moved to a position 3 mm below the electrostatic probe to begin measuring the surface potential. When measuring the central charge density, the electrostatic probe was positioned above the center of the sample to record the central potential values over time. To measure the surface charge distribution, the displacement platform was controlled to move the probe in a square area with a side length of 60 mm centered on the sample. The scanning path started at the center of the area, moved to the lower-left corner and then scanned the entire area along an S-shaped path to the upper-right corner, finally returning to the center of the area. The scanning process was symmetric in time and space. The scanning speed was set at 20 mm·s⁻¹ and the entire scanning process took approximately 60 s. Considering that the potential of the sample surface may decay during this time, the measured potential values were corrected.[Results] At a relative humidity of 32%, the initial central potential after removing the voltage of the three samples was measured and the central charge density was calculated. Si₃N₄ had the highest charge density, followed by Al₂O₃, while AlN had the lowest value. All three samples accumulated positive charges of the same polarity as the applied voltage. The potential was highest at the center of charge injection and decreased radially. The area outside the sample on the displacement platform exhibited almost no charge distribution. At 60% and 32% humidities, the surface potential values of the three samples were measured over time after the voltage was removed. Subsequently, a double-exponential function was used to fit the changes in surface potential. Comparatively, at 32% humidity, the initial surface potential values of the samples were slightly higher than those at 60% humidity, while the potential decay rate was significantly lower than that at 60% humidity. In terms of initial charge density at 60% humidity, the order was: Al₂O₃>Si₃N₄>AlN, and at 32% humidity, it was Si₃N₄ > Al₂O₃>AlN. Regarding the charge dissipation rates, regardless of low or high humidity, the order was AlN>Si₃N₄>Al₂O₃. At 32% humidity, using the moment of voltage removal as the initial time, the charge distribution morphology of the three samples was measured at the initial moment, 5 min, 15 min and 60 min. All three samples conformed to have the characteristic of "no significant change in charge distribution morphology", when bulk dissipation was the main dissipation pathway, while no pronounced volcanic-like distribution or expansion of charge distribution range was observed. In Al₂O₃ and Si₃N₄, a phenomenon was observed where the potential at a distance of 4 mm from the center first increased and then decreased after the voltage was removed, but remained lower than the potential at the center. This was ascribed to the charge received from the center direction, which exceeded the charge lost from bulk dissipation at that location, thereby resulting in an increase in charge density and a rise in potential. This phenomenon indicates that there exists a tangential electric field induced by surface charges or a charge dissipation process driven by charge density differences at different locations along the surface. In AlN, the aforementioned phenomenon was not observed, as its bulk dissipation rate was relatively fast, meaning that the charge lost from non-central areas due to bulk dissipation was always greater than the charge gained from high charge density areas due to surface dissipation, resulting in a continuous decrease in potential in non-central areas.[Conclusion] All three ceramic surfaces accumulated charges of the same polarity as the applied voltage, with charge density decreasing radially from the injection point. At low humidity, the central accumulated charge density was Si₃N₄>Al₂O₃>AlN. At high humidity, the order was Al₂O₃>Si₃N₄>AlN. The dissipation of surface charge after removing the applied voltage is the result of the combined action of surface dissipation and bulk dissipation, with dissipation rate of AlN>Si₃N₄>Al₂O₃. The surface conductivity of materials is higher at high humidity, leading to a faster dissipation of charges along the surface. According to the analysis of the isothermal surface potential decay model, shallow traps are the main types of traps in AlN, while deep traps are the main types of traps inAl₂O₃and Si₃N₄.

Key words: Al₂O₃ ceramics; AlN ceramics; Si₃N₄ ceramics; surface charge