Surface Microdischarge Detection for High-Voltage Power Supplies in Electrostatic Chucks: Technical Principles and Application Value

In the manufacturing of advanced semiconductor processes, the electrostatic chuck (ESC), as a core component for precise wafer clamping and temperature control, has its performance stability directly determining the yield of wafer processing. The clamping force of the ESC relies on the strong electric field provided by a high-voltage power supply (HVPS). However, under high-voltage operating conditions, surface microdischarge (SMD) is prone to occur at the insulating interface between the power supply and the chuck. Although this local, low-energy discharge phenomenon has a negligible single impact, long-term accumulation can lead to aging of insulating media, imbalance in surface charge distribution, not only causing fluctuations in clamping force that result in wafer displacement or scratches, but also interfering with precision processes such as photolithography and etching through electric field coupling. It may even shorten the service life of the HVPS. Therefore, conducting surface microdischarge detection for HVPS in electrostatic chucks has become a key link in the reliability management of semiconductor equipment.
The occurrence of surface microdischarge in the HVPS of electrostatic chucks is essentially the synergistic effect of electric field distortion and charge accumulation at the insulating interface. When the HVPS outputs DC or pulsed high voltage to the ESC electrodes, the surface of insulating media (such as aluminum oxide and aluminum nitride ceramics) is prone to charge accumulation due to the charge trapping effect. If the medium has micro-defects (such as pores and scratches), the electric field intensity at the defect sites will be much higher than the average electric field. When the local electric field exceeds the breakdown threshold of the medium, micro-scale gas ionization discharge will occur. Different from macro-breakdown, surface microdischarge has the characteristics of locality, intermittency, and accumulation: the discharge area is usually limited to micro-scale defect points, the discharge energy is only at the nanojoule to microjoule level, and it presents periodic pulses with the repeated accumulation and release of charges. Long-term action will gradually erode the surface of the insulating medium and damage its dielectric properties.
The current mainstream surface microdischarge detection technologies are built around discharge signal capture and feature recognition, and the core technical paths can be divided into three categories. The first is the Pulse Current (PC) Method, which captures the nanoampere-level pulse current signals generated by microdischarge by connecting a high-precision current sensor in the high-voltage circuit, and then extracts characteristic parameters such as discharge frequency and amplitude through signal amplification and filtering. This method has high sensitivity (capable of detecting 10⁻¹²A-level current), but is easily affected by power grid noise and equipment electromagnetic interference, requiring shielding and anti-interference design. The second is the Ultra-High Frequency (UHF) Detection Method, which uses the UHF band electromagnetic radiation (300MHz-3GHz) accompanied by microdischarge, receives the radiation signal through an antenna array and conducts spectrum analysis. This method has strong anti-interference ability and can realize non-contact detection, but it needs to accurately locate the discharge source to avoid signal attenuation. The third is the Optical Detection Method, which is based on the weak photon emission generated by gas ionization during microdischarge. It uses a high-sensitivity CCD or photomultiplier tube to capture the discharge spot, and can intuitively present the discharge position and intensity distribution through image analysis. However, limited by the closed structure inside the chuck, it is necessary to reserve an optical observation window during the equipment design stage.
From the perspective of application scenarios, surface microdischarge detection technology has been deeply integrated into the full-life cycle management of semiconductor equipment. During the commissioning stage before equipment delivery, by simulating the microdischarge characteristics under different voltage, temperature, and vacuum environments, the safe operating threshold of the HVPS can be determined. During the operation and maintenance stage of the mass production line, the online detection system (such as the HVPS module integrated with UHF sensors) can be used to monitor the change of microdischarge signals in real time. When the discharge frequency or amplitude exceeds the preset threshold, an early warning is triggered to avoid sudden failures. Especially in the manufacturing of 3D NAND and advanced logic chips below 7nm, the expansion of wafer size (12 inches and above) and the improvement of process precision have higher requirements for the electric field stability of the ESC. Surface microdischarge detection can increase the wafer processing yield by 3%-5%, and at the same time extend the maintenance cycle of the HVPS by 2-3 times, significantly reducing the equipment operation cost.
Currently, surface microdischarge detection for HVPS in electrostatic chucks still faces two major challenges: first, the high vacuum and high temperature environment of the semiconductor process will interfere with the detection signal, requiring the development of sensors that can withstand extreme environments; second, it is difficult to distinguish weak microdischarge signals from background noise, requiring the optimization of signal feature extraction models combined with machine learning algorithms. In the future, as detection technology develops towards integration and onlineization, the integrated design of HVPS and microdischarge detection modules will be gradually realized, and the high-voltage output parameters will be dynamically adjusted through real-time data feedback, providing more reliable technical support for the stable operation of advanced semiconductor processes.