Plasma arc management system, method, electronic device, and storage medium

Abstract

The application relates to the field of semiconductor manufacturing equipment, and discloses a plasma arc management system, a method, an electronic device and a storage medium, the system comprises a sensor module, a signal processing module, a control module and an inhibition module; the sensor module comprises a plurality of sensing units arranged at different spatial positions around a plasma processing chamber, is used for continuously receiving an electromagnetic wave signal and receiving a light signal when being activated; the signal processing module is used for determining a coarse positioning position according to the electromagnetic wave signal and activating the sensing unit of the corresponding region, determining a fine positioning position according to the light signal and generating an arc event; the control module is used for generating an inhibition instruction according to the arc event; and the inhibition module is used for inhibiting the arc according to the inhibition instruction. The application solves the problem that in the related art, the arc spatial position cannot be accurately positioned in the chamber, and realizes millimeter-level positioning accuracy.

Description

Technical Field

[0001] This invention relates to the field of semiconductor manufacturing technology, and more specifically to plasma arc management systems, methods, electronic devices, and storage media. Background Technology

[0002] In the semiconductor manufacturing industry, micro-arc discharge phenomena in plasma processing pose a serious threat to wafer yield. Micro-arcs can cause sputtering particles from the target material, damage to the chamber walls, and destruction of wafer surface patterns, even resulting in the scrapping of an entire batch of wafers. Therefore, rapidly and accurately detecting and suppressing arcs, while simultaneously locating the arc's origin, is crucial for improving process yield and equipment reliability.

[0003] Currently, the existing technologies related to this invention mainly include the following solutions: First, the electrical parameter monitoring method, which uses clamp-on sensors or direct connection to a DC power supply to monitor transient changes in voltage, current, or power at a sampling rate of up to 250kHz. This method has a fast response, but cannot locate the spatial position of the arc and is easily affected by process noise, leading to false alarms. Second, the dual-channel method of digital polling and analog acquisition, which reads the arc count from the internal register of the DC power supply and acquires voltage / power waveforms at high speed, can achieve full coverage of micro-arcs and hard arcs, but still cannot provide spatial positioning information. Third, the optical monitoring method, which uses photomultiplier tubes (PMTs) or fiber optic probes to monitor changes in plasma luminescence intensity, but the optical window is easily contaminated by deposits and has blind spots, making it difficult to achieve full chamber coverage and accurate positioning. Fourth, the wavelet analysis signal processing method, which improves the signal-to-noise ratio and reduces false alarms by combining high-speed sampling with discrete wavelet transform, but this method focuses on data processing algorithms and does not involve hardware integration and spatial positioning. Fifth is the power-off suppression method, which quickly shuts off the radio frequency (RF) output after detecting an arc. Although this can effectively extinguish the arc, it will cause the plasma to be completely deionized, disrupting the process continuity, and it is impossible to adjust the suppression strategy according to the physical location of the arc.

[0004] In summary, the relevant technologies generally suffer from the following insurmountable drawbacks: the inability to determine the three-dimensional spatial position of the arc within the chamber; difficulty in balancing response speed and spatial information (e.g., blind spots due to optical contamination, limited electrical information); a single and lagging suppression method, lacking refined energy control and spatial context linkage, leading to a high false alarm rate; and the disruption of process continuity caused by power-off suppression. Therefore, there is an urgent need for an integrated arc management system capable of simultaneously achieving rapid detection, precise spatial positioning, differentiated decision-making, and continuous plasma process protection. Summary of the Invention

[0005] This invention provides a plasma arc management system, method, electronic device, and storage medium to solve the problem of inaccurate spatial positioning of the arc in plasma processing.

[0006] In a first aspect, the present invention provides a plasma arc management system, comprising: a sensor module including multiple sensing units disposed at different spatial locations around a plasma processing chamber, the sensing units being configured to continuously receive electromagnetic wave signals generated by the arc and, when activated, receive optical signals generated by the arc; a signal processing module, communicatively connected to the sensor module, configured to determine a coarse positioning position of the arc based on the electromagnetic wave signals and activate the sensing units in the corresponding areas to determine a fine positioning position of the arc based on the optical signals, and then generate an arc event based on the coarse positioning position, the fine positioning position, the electromagnetic wave signals, and the optical signals, the arc event including the position, energy, and duration of the arc; and a control module, communicatively connected to the signal processing module. The system is configured to acquire the effective device region layout map of the current wafer, map the arc position in the arc event to the wafer coordinate system, determine whether the arc position falls within the effective device region of the effective device region layout map, and generate suppression commands with different suppression intensities corresponding to different risk levels based on the determination result; the suppression module is connected to the power supply of the control module and the plasma processing chamber, and is configured to acquire the real-time output waveform of the power supply of the plasma processing chamber according to the suppression command; compare the real-time output waveform with a preset expected waveform to obtain the waveform error; generate a compensation pulse with opposite phase and matching amplitude to the waveform error according to the suppression intensity and the waveform error, and inject the compensation pulse into the output terminal of the power supply to suppress the arc.

[0007] Beneficial Effects: The plasma arc management system provided by this invention continuously receives electromagnetic wave signals through multiple sensing units located at different spatial positions around the chamber, and receives optical signals for activated sensing units, achieving multimodal sensing of electromagnetic waves and optical signals. The signal processing module first uses the electromagnetic wave signals to quickly obtain a coarse positioning position, then activates the sensing units in the corresponding area for fine positioning using optical signals, and finally fuses the signals to generate an arc event containing position, energy, and duration. This allows the system to achieve millimeter-level high-precision positioning results while maintaining nanosecond-level rapid response of electromagnetic waves. Compared to traditional electrical detection schemes that can only determine the presence of an arc but not its specific location, this invention can accurately obtain the spatial position of the arc within the chamber, providing a reliable spatial information basis for subsequent differentiated suppression intensity and process optimization.

[0008] In one optional implementation, the signal processing module includes an electromagnetic wave positioning unit, which is configured to: perform bandpass filtering on the electromagnetic wave signals received by each sensing unit; detect the arrival time of each filtered electromagnetic wave signal; calculate the arrival time difference between other sensing units and the reference sensing unit, using the arrival time of one of the sensing units as a reference; establish a hyperbolic equation system about the coordinates of the arc source based on the spatial coordinates of each sensing unit and the arrival time difference; and linearize the hyperbolic equation system and then iteratively solve it using the least squares method to obtain the coarse positioning position of the arc.

[0009] Beneficial effects: The signal processing module performs bandpass filtering, detects arrival time, calculates arrival time difference, and establishes a hyperbolic equation system based on the spatial coordinates of each sensor unit by combining the electromagnetic wave signals received by each sensor unit and solving it linearly. This enables rapid coarse positioning of electromagnetic waves based on the solution of the hyperbolic equation system. This process can be completed at the nanosecond level, providing spatial range constraints for subsequent fine positioning, avoiding blind scanning of the entire chamber, and significantly reducing the computational complexity of fine positioning.

[0010] In an optional implementation, the signal processing module further includes an optical tomography unit, which is configured to: activate a subset of sensing units closest to the coarse positioning position based on the coarse positioning position; perform smooth filtering on the optical signal received by the subset of sensing units to extract the light intensity amplitude; establish a spatial luminous intensity distribution model of the arc luminous intensity based on the light intensity amplitude; perform tomography reconstruction on the spatial luminous intensity distribution model based on multi-angle projection data and combining the principles of three-dimensional spatial projection and back projection reconstruction to obtain the three-dimensional luminous intensity distribution of the arc; and take the position of the maximum intensity in the three-dimensional luminous intensity distribution as the fine positioning position.

[0011] Beneficial effects: The signal processing module activates the nearest subset of sensor units based on the coarse positioning location. It then smooths and filters the optical signals from the activated sensor units, extracts the light intensity amplitude, establishes a spatial luminous intensity distribution model, and performs tomographic reconstruction. The location with the maximum intensity is taken as the fine positioning location. This scheme utilizes the coarse positioning results to constrain the optical solution domain, significantly reducing the amount of reconstruction computation and enabling real-time millimeter-level precise positioning. Simultaneously, the subset activation method reduces unnecessary data processing overhead.

[0012] In an optional implementation, the signal processing module further includes a data fusion unit, which is configured to: take the coarse positioning position and the fine positioning position as observations, perform weighted fusion of the coarse positioning position and the fine positioning position based on state estimation and error covariance matrix, and output the arc position; estimate the energy and duration of the arc based on the intensity integral of the optical signal and the energy of the electromagnetic wave signal; and generate the arc event based on the arc position, the energy, and the duration.

[0013] Beneficial effects: The signal processing module uses coarse and fine positioning positions as observations and performs weighted fusion based on state estimation and error covariance matrix to output the final arc position. It also estimates the arc energy and duration based on the integral of the optical signal intensity and the electromagnetic wave signal energy, generating a complete arc event. This fusion processing improves the robustness of the positioning; even if one mode is disturbed, another mode can still provide effective information. The estimation of energy and duration provides the control module with a complete decision-making basis for risk level determination and differentiated suppression intensity.

[0014] In one optional implementation, the control module includes a sensing unit configured to: receive arc events output by the signal processing module in real time; obtain the wafer identifier and process step identifier of the current wafer, and obtain the corresponding chip layout diagram based on the wafer identifier; parse the chip layout diagram, extract the effective device region coordinates of the wafer, generate a wafer mapping table based on the effective device region coordinates, and transform the wafer mapping table from the wafer coordinate system to the global coordinate system of the plasma processing chamber.

[0015] Beneficial effects: The control module receives arc events in real time and obtains the wafer identifier and process step identifier of the current wafer. Based on the wafer identifier, it obtains the corresponding chip layout diagram, parses and extracts the coordinates of the effective device areas, generates a wafer mapping table, and accurately maps the wafer mapping table to the global coordinate system of the plasma processing chamber through coordinate system transformation. This solution introduces process context information into arc management, enabling subsequent risk level determination to be based on the actual wafer layout. This avoids the defect in traditional solutions where fluctuations in non-critical areas are misjudged as high-risk events due to the lack of spatial context.

[0016] In an optional implementation, the control module further includes a decision-making and optimization unit, configured to: obtain the arc position based on the arc event; project the three-dimensional coordinates of the arc position onto the wafer plane; determine the corresponding position of the arc position in the wafer mapping table by judging the spatial relationship between the projection point and the boundary of the effective device region; determine whether the arc position falls within the effective device region of the wafer mapping table; if the arc position falls within the effective device region, it is determined to be of a high-risk level; if the arc position does not fall within the effective device region, it is determined to be of a low-risk level; and generate the suppression instruction based on the risk level.

[0017] Beneficial effects: The control module projects the three-dimensional coordinates of the arc location onto the plane of the wafer, and performs spatial mapping by judging the spatial relationship between the projection point and the boundary of the effective device area. Based on this, it determines the risk level and generates corresponding suppression commands. This scheme achieves spatially differentiated judgment of arc risk—falling into the effective device area is judged as high risk, and not falling into it is judged as low risk. This allows the suppression strategy to be accurately matched according to the actual process impact of the arc, avoiding false alarms or overreactions caused by traditional single suppression strategies, and significantly reducing the false alarm rate.

[0018] In an optional implementation, the decision-making and optimization unit is further configured to: generate a suppression instruction and alarm information with a first suppression strength when the risk level is high risk; and generate a suppression instruction with a second suppression strength when the risk level is low risk; wherein the first suppression strength is greater than the second suppression strength.

[0019] Beneficial effects: The control module generates differentiated suppression commands based on risk level: high risk corresponds to the first suppression intensity and an alarm is issued, low risk corresponds to the second suppression intensity, and the suppression intensity for high risk is greater than that for low risk. This solution achieves refined management of arc energy, with high-risk events receiving strong suppression and alarm notifications, while low-risk events require only weak compensation, avoiding disruption to process continuity. Simultaneously, the alarm information provides engineers with precise fault location information.

[0020] In one optional implementation, the suppression module includes a power unit configured to: acquire the output voltage or output current waveform of the power supply in real time as the real-time output waveform; and a reverse compensation pulse injection unit configured to: compare the real-time output waveform with the preset desired waveform point by point, and if the absolute value of the error at a certain point exceeds a preset threshold, determine it as a waveform error; generate a compensation pulse with an amplitude matching the waveform error and a phase opposite to the waveform error based on the amplitude and phase of the waveform error and the suppression intensity in the suppression command, and adjust the injection timing of the compensation pulse based on the system delay compensation amount; and inject the compensation pulse into the output terminal of the power supply in the next switching cycle.

[0021] Beneficial effects: The suppression module acquires the real-time output waveform of the power supply and compares it with the preset desired waveform to obtain the waveform error. Based on the waveform error and the suppression intensity in the suppression command, it generates a compensation pulse with opposite phase and matching amplitude. The injection timing of the compensation pulse is adjusted in conjunction with the system delay compensation amount and injected into the power supply output terminal in the next switching cycle. This scheme achieves high-speed reverse pulse injection suppression based on GaN devices. The compensation pulse, superimposed with the arc distortion, restores the net output waveform to the desired value. Arc energy can be offset without shutting down the power supply, maintaining the plasma bulk without deionization and ensuring process continuity.

[0022] In an optional implementation, the reverse compensation pulse injection unit is further configured to: continuously monitor the waveform error after injecting the compensation pulse; if the waveform error still exists, then perform the generation and injection of the compensation pulse again until the waveform error disappears; if the waveform error still does not disappear after repeated multi-cycle continuous compensation, then it is determined to be a hard arc, and the power supply is turned off for a preset time.

[0023] Beneficial effects: After injecting the compensation pulse, the suppression module continuously monitors the waveform error. If the error still exists, it repeats the generation and injection of the compensation pulse until the error disappears. If the error persists after repeated execution, it is determined to be a hard arc, and the power is turned off for a preset duration. This scheme achieves multi-cycle continuous compensation and hard arc fallback processing—multi-cycle compensation ensures effective suppression of continuous arcs, while a brief power outage assists in extinguishing the arc for the very few hard arcs where multi-cycle compensation is ineffective. This ensures rapid suppression of conventional arcs while providing a safety fallback mechanism for extreme cases.

[0024] In one optional implementation, the control module further includes a recording unit configured to: record the location information, energy, time, risk level, and suppression results of each arc event and store them in a database; periodically calculate the arc occurrence frequency of each process step and each spatial location; update the arc heat map based on the arc occurrence frequency, wherein the arc heat map uses a three-dimensional model of the chamber as a base and color depth to represent arc density; and generate an early warning message or send parameter optimization suggestions to the process controller when the arc frequency at the same spatial location in the same process step exceeds a preset threshold within a predetermined time.

[0025] Beneficial effects: The control module records complete information on each arc event and stores it in the database. It periodically calculates the arc occurrence frequency at each process step and spatial location and updates the arc heat map. When the arc frequency at the same spatial location in the same process step exceeds a threshold within a predetermined time, it generates an early warning or parameter optimization suggestion. This solution achieves proactive prevention based on big data—the heat map intuitively displays arc accumulation areas, and when the frequency exceeds the threshold, it predicts the risk of hard arcs in advance and triggers optimization, transforming arc management from a passive response to proactive prevention, reducing arc occurrences at the source.

[0026] In one optional implementation, the control module further includes a yield correlation analysis unit, which is configured to: acquire wafer electrical test yield data; acquire arc positions based on multiple arc events and map the arc positions to wafer chip layout; overlay the arc positions with the yield data to generate an arc influence distribution map; and, based on the arc influence distribution map and the spatial correspondence between arc positions and chip failure positions, identify key arc types and positions that cause yield loss.

[0027] Beneficial Effects: The control module acquires wafer electrical test yield data, maps arc locations to the wafer chip layout, and overlays this data with the yield data to generate an arc impact distribution map. Based on the spatial correspondence between arc locations and chip failure locations, it identifies the key arc types and locations that cause yield losses. This solution achieves accurate attribution of arc events to yield—traditional yield analysis cannot pinpoint whether failed chips are related to arcs, while this solution, through spatial mapping and spatial correspondence analysis, accurately identifies which arc events actually caused yield losses, providing direct data support for equipment maintenance and process optimization.

[0028] Secondly, the present invention provides a plasma arc management method, comprising: continuously receiving electromagnetic wave signals generated by an arc through multiple sensing units disposed at different spatial locations around a plasma processing chamber; receiving optical signals generated by the arc through the activated sensing units when the sensing units are activated; determining a coarse positioning position of the arc based on the electromagnetic wave signals; activating sensing units in the corresponding area based on the coarse positioning position and receiving optical signals generated by the arc through the activated sensing units; determining a fine positioning position of the arc based on the optical signals; and generating an arc event based on the coarse positioning position, the fine positioning position, the electromagnetic wave signals, and the optical signals, wherein the arc event includes... The system calculates the location, energy, and duration of the electric arc; obtains the effective device area layout of the current wafer and maps the arc position in the arc event to the wafer coordinate system; determines whether the arc position falls within the effective device area of ​​the effective device area layout and generates suppression commands with different suppression intensities corresponding to different risk levels based on the determination result; obtains the real-time output waveform of the power supply of the plasma processing chamber based on the suppression command; compares the real-time output waveform with a preset expected waveform to obtain the waveform error; generates a compensation pulse with opposite phase and matching amplitude to the waveform error based on the suppression intensity and the waveform error, and injects the compensation pulse into the output terminal of the power supply to suppress the arc.

[0029] Thirdly, the present invention provides an electronic device comprising: a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the computer instructions to perform the plasma arc management method of the first aspect or any corresponding embodiment described above.

[0030] Fourthly, the present invention provides a computer-readable storage medium storing computer instructions for causing a computer to execute the plasma arc management method of the first aspect or any corresponding embodiment described above.

[0031] This invention achieves multimodal sensing by continuously receiving electromagnetic wave signals through multiple sensing units of the sensor module and activating the receiving optical signals as needed. The signal processing module uses electromagnetic waves for rapid coarse positioning and activates the corresponding area for fine optical positioning, generating an arc event containing position, energy, and duration. The control module maps the arc position to the wafer coordinate system based on the wafer layout diagram and performs risk assessment based on the spatial relationship between the spatial projection and the boundary of the effective device area, generating suppression commands with different suppression intensities corresponding to the risk level. The suppression module generates compensation pulses with opposite phase and matching amplitude based on the suppression commands and real-time waveform errors and injects them into the power output terminal to suppress the arc. This solves the problem of accurately locating the arc spatial position in the cavity in related technologies, achieving millimeter-level positioning accuracy and providing a reliable spatial information basis for subsequent differentiated suppression intensity and process optimization. Attached Figure Description

[0032] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0033] Figure 1 This is a schematic diagram of the modules of the plasma arc management system according to an embodiment of the present invention; Figure 2 This is an overall architecture diagram of the plasma arc management system according to an embodiment of the present invention; Figure 3 This is a schematic diagram of the layout of a sensor module according to an embodiment of the present invention; Figure 4 This is a data processing flowchart of the signal processing module according to an embodiment of the present invention; Figure 5 This is a flowchart of the context awareness and decision-making process in the control module according to an embodiment of the present invention; Figure 6 This is an arc thermogram according to an embodiment of the present invention; Figure 7 This is a distribution diagram of the electric arc influence according to an embodiment of the present invention; Figure 8 This is a flowchart of power supply and reverse pulse injection according to an embodiment of the present invention; Figure 9 This is a system linkage timing diagram according to an embodiment of the present invention; Figure 10 This is a comparison chart of arc positioning accuracy according to an embodiment of the present invention; Figure 11This is a diagram illustrating the reverse pulse injection suppression effect according to an embodiment of the present invention; Figure 12 This is a schematic flowchart of a plasma arc management method according to an embodiment of the present invention; Figure 13 This is a schematic diagram of the hardware structure of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0034] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0035] It is understood that before using the technical solutions disclosed in the various embodiments of the present invention, users should be informed of the types, scope of use, and usage scenarios of the personal information involved in the present invention and their authorization should be obtained in accordance with relevant laws and regulations through appropriate means.

[0036] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0037] This invention provides a plasma arc management system and method. Addressing the problems in related technologies, such as the inability of electrical monitoring to determine the specific location of the arc within the chamber, the limited coverage of the entire chamber by optical monitoring due to line-of-sight constraints, and the single suppression method lacking differentiated processing, this invention achieves multimodal sensing of electromagnetic waves and optical signals, a two-level positioning architecture consisting of coarse positioning based on the time difference of electromagnetic wave arrival and fine positioning based on optical tomography, differentiated suppression based on positioning results and wafer layout diagrams, and complete closed-loop management of arc events. This results in rapid detection, accurate positioning, and intelligent suppression.

[0038] Reference Figure 1 As shown, this application provides a plasma arc management system, including: a sensor module, a signal processing module, a control module, and a suppression module.

[0039] Specifically, refer to Figure 2As shown, the sensor module includes multiple sensing units positioned at different locations around the plasma processing chamber. Each sensing unit is configured to continuously receive electromagnetic wave signals generated by the electric arc and, when activated, to receive optical signals generated by the electric arc. Specifically, the number of integrated sensing units can be set to N according to actual needs. Each integrated sensing unit integrates an ultra-wideband antenna and a fiber optic probe. The integrated sensing unit continuously receives electromagnetic wave signals generated by the electric arc through the ultra-wideband antenna and receives optical signals generated by the electric arc through the fiber optic probe when activated.

[0040] The signal processing module communicates with the sensor module and is configured to determine the coarse location of the arc based on electromagnetic wave signals and activate the sensing units in the corresponding area. It then determines the fine location of the arc based on optical signals and generates an arc event based on the coarse location, fine location, electromagnetic wave signals, and optical signals. The control module communicates with the signal processing module and is configured to acquire the current wafer's effective device area layout, map the arc position to the wafer coordinate system, determine whether the arc position falls within the effective device area, and generate suppression commands with different suppression intensities corresponding to different risk levels based on the determination result. The suppression module is connected to the control module and the power supply of the plasma processing chamber. It is configured to acquire the real-time output waveform of the plasma processing chamber's power supply based on the suppression commands; compare the real-time output waveform with a preset desired waveform to obtain the waveform error; and generate a compensation pulse with the opposite phase and matching amplitude to the waveform error based on the suppression intensity and waveform error. This compensation pulse is then injected into the power supply's output to suppress the arc.

[0041] The aforementioned plasma arc management system continuously receives electromagnetic wave signals and optical signals from multiple sensing units located at different spatial positions around the chamber, achieving multimodal sensing of electromagnetic and optical signals. The signal processing module first uses the electromagnetic wave signals to quickly obtain a coarse location, then activates the corresponding sensing units for fine optical signal positioning. Finally, it fuses these signals to generate an arc event containing location, energy, and duration, enabling the system to achieve millimeter-level high-precision positioning while maintaining nanosecond-level rapid electromagnetic wave response. Compared to traditional electrical detection schemes that can only determine the presence of an arc but not its specific location, this invention can accurately acquire the spatial position of the arc within the chamber, providing a reliable spatial information basis for subsequent differentiated suppression intensity and process optimization.

[0042] Reference Figure 3 As shown, specifically, multiple sensing units are distributed three-dimensionally around the plasma processing chamber. A typical layout consists of at least eight units, including three dielectric windows arranged at the top with a 120° interval, four viewing windows arranged in the middle at different azimuth angles, and one observation port arranged at the bottom. The three-dimensional coordinates of each sensing unit are pre-calibrated using a laser tracker and stored in the system.

[0043] Specifically, during the initial system deployment phase, the precise three-dimensional coordinates of each integrated sensing unit in the global chamber coordinate system are measured and recorded using a laser tracker. This coordinate matrix is ​​then written into the configuration file of the signal processing module, serving as the known reference point coordinates in the Time Difference of Arrival (TDOA) positioning algorithm. Furthermore, the relative positional deviations of the antennas and fiber optic probes within each sensing unit must be calibrated to eliminate minor positional errors caused by physical packaging, ensuring precise matching between the TDOA and the projection angle of the optical signal.

[0044] The aforementioned three-dimensional layout, by covering the chamber space from multiple angles and heights, provides sufficient spatial diversity and redundancy for subsequent electromagnetic wave arrival time difference positioning and optical tomography.

[0045] As one specific implementation of the sensing unit, each sensing unit integrates an ultra-wideband antenna and a fiber optic probe. The ultra-wideband antenna can be a planar inverted-F antenna (PIFA) or a Vivaldi antenna, or other ultra-wideband antenna forms, with a frequency coverage range of 100MHz to 6GHz to accommodate the radiation spectrum of arcs of different sizes. The antenna is encapsulated in a high-temperature resistant, low-loss dielectric material (such as polytetrafluoroethylene) and attached to the outside of the dielectric window of the chamber or embedded in the insulating layer of the chamber wall. The fiber optic probe uses radiation-resistant multimode fiber (200μm core diameter, 0.22 numerical aperture), with a lens group at the front end to focus plasma emission. A narrowband filter is located at the front end of the fiber optic probe, with the center wavelength corresponding to the characteristic spectral lines of the plasma processing gas (such as argon Ar I 811.5nm) to suppress background light interference. The antenna output is connected to the data acquisition card via a semi-rigid coaxial cable, and the fiber optic output is connected to a photodetector module to convert the optical signal into a voltage signal. The center wavelength of the narrowband filter can also be adjusted according to the characteristic spectral lines of the specific process gas. For example, for the nitrogen process, the characteristic spectral lines of the second positive band system of nitrogen can be selected.

[0046] The aforementioned sensing unit structure achieves simultaneous single-point acquisition of electromagnetic waves and optical signals through the physical integration of antenna and optical fiber, eliminating the need for multiple independent probes, reducing the number of cavity openings, and lowering installation complexity. The electromagnetic waves and optical signals are naturally synchronized in time, providing a precise time reference for subsequent fusion. Narrowband filters effectively remove background light interference, ensuring the signal-to-noise ratio of the optical signal.

[0047] As a modified implementation of the sensing unit, refer to Figure 4As shown, in one specific implementation of the signal processing module, the signal processing module uses a multi-channel high-speed data acquisition card to synchronously acquire electromagnetic wave signals and optical signals from all sensing units. The sampling rate of the electromagnetic wave channel is greater than or equal to 1 GHz, and the sampling rate of the optical fiber channel is greater than or equal to 250 kHz. The acquisition card has deep onboard memory to support continuous recording. In some specific embodiments, the multi-channel high-speed data acquisition card can be implemented using an acquisition card with a single-channel maximum sampling rate of 12.5 GS / s, such as the NI PXIe-5186 acquisition card.

[0048] Specifically, the signal processing module includes an electromagnetic wave positioning unit, an optical tomography unit, and a data fusion unit.

[0049] The electromagnetic wave positioning unit is configured to perform bandpass filtering (100MHz to 6GHz) on the electromagnetic wave signals received by each sensing unit to remove power frequency interference. The arrival time of each filtered electromagnetic wave signal is detected using the short-time energy method or wavelet transform modulus maxima method. Using the arrival time of one sensing unit as a reference, the arrival time difference between the other sensing units and the reference sensing unit is calculated. Based on the spatial coordinates of each sensing unit and the arrival time difference, a hyperbolic equation system about the arc source coordinates is established. After linearizing the hyperbolic equation system, iteratively solving it using the least squares method yields the coarse positioning position (x0, y0, z0) of the arc. In this embodiment, the establishment and iterative solution of the hyperbolic equation system can be implemented using the Chan algorithm or Taylor series method. This process is completed by a Field-Programmable Gate Array (FPGA) hardware, with a total time consumption of less than 1μs and centimeter-level accuracy.

[0050] The aforementioned electromagnetic wave positioning method removes power frequency interference through bandpass filtering, detects the arrival time using the short-time energy method or wavelet transform modulus maxima method, and performs hyperbolic positioning by combining the chamber geometric model and sensor unit coordinates. This achieves rapid coarse positioning at the nanosecond level, providing a solution domain constraint for subsequent optical tomography and significantly reducing the computational load.

[0051] In the coarse localization problem solved using the Chan algorithm or Taylor series method, a hyperbolic equation system regarding the arc source coordinates is first established based on the chamber geometry model. This nonlinear equation system is then linearized and solved iteratively using the least squares method to obtain sub-microsecond coarse localization coordinates. During optical tomography reconstruction, a constraint matrix is ​​constructed based on the coarse localization results. Only the spatial grid within a preset radius around the coarse localization coordinates is used as an unknown variable in the Algebraic Reconstruction Technique (ART) iteration, rather than a full-chamber grid scan. This significantly reduces the computational load, enabling millimeter-level precision localization to be completed within microseconds.

[0052] As a further configuration of the signal processing module, the optical tomography unit is configured as follows: Based on the coarse positioning position, activate the subset of sensor units closest to the coarse positioning position, for example, selecting the four with the strongest signals from eight sensor units. Smoothly filter the optical signal received by the subset of sensor units to extract the light intensity amplitude. Establish a spatial luminous intensity distribution model of the electric arc based on the light intensity amplitude. Based on multi-angle projection data, and combining the principles of three-dimensional spatial projection and back-projection reconstruction, perform tomography reconstruction on the spatial luminous intensity distribution model to obtain the three-dimensional luminous intensity distribution of the electric arc. Take the position of the maximum intensity in the reconstruction result as the fine positioning position (x). f ,y f ,z f In this embodiment, the spatial luminescence intensity distribution model can specifically adopt a Gaussian distribution model, and the back-projection reconstruction principle can be implemented using algebraic reconstruction technology or a filtered back-projection algorithm, with an accuracy of millimeters and a time consumption of approximately 10μs.

[0053] The aforementioned optical tomography fine-tuning method utilizes coarse-tuning results to constrain the optical solution domain, avoiding the enormous computational overhead of full-chamber scanning and enabling real-time millimeter-level precise positioning. Simultaneously, it further reduces the data volume by selecting the subset with the strongest signal. This scheme significantly reduces the amount of reconstruction computation by establishing a spatial luminescence intensity distribution model and combining the principles of three-dimensional spatial projection and back-projection reconstruction.

[0054] As an improved implementation of the signal processing module, the data fusion unit is configured to: use the coarse and fine positioning positions as observations, and perform weighted fusion of the coarse and fine positioning positions based on state estimation and error covariance matrix to output the final arc position (x, y, z). In this embodiment, this fusion method can specifically be implemented using Kalman filtering fusion. Simultaneously, the arc energy E and duration t are estimated based on the intensity integral of the optical signal and the energy of the electromagnetic wave signal, and an arc event (x, y, z, E, t) is generated based on the final arc position, energy, and duration. The fusion process takes less than 5 μs.

[0055] The aforementioned data fusion method improves positioning robustness by fusing multimodal observations, ensuring that even if one mode is disturbed, another mode can still provide valid information. The estimation of energy and duration provides the control module with a complete decision-making basis for risk level determination and differentiated suppression intensity.

[0056] Reference Figure 5 As shown, in one specific implementation of the control module, the control module communicates with the factory host (Manufacturing Execution System (MES), Equipment Automation Program (EAP)), Fault Detection and Classification (FDC) system, and process controller via industrial Ethernet. Specifically, the control module includes a sensing unit, a decision-making and optimization unit, a recording unit, and a yield correlation analysis unit.

[0057] The sensing unit is configured to: receive the arc event output by the signal processing module in real time, obtain the wafer identifier and process step identifier of the current wafer, obtain the corresponding chip layout diagram based on the wafer identifier, parse the chip layout diagram (such as CAD file parsing) to extract the effective device area coordinates of the wafer, and generate a wafer mapping table based on the effective device area coordinates.

[0058] It should be noted that when the control module parses the chip layout diagram, it first extracts the chip die outline and boundary coordinates from the CAD file, and then transforms the chip coordinate system to the global coordinate system of the process chamber based on the wafer center and pre-alignment marks (Notch or Flat). Subsequently, the three-dimensional coordinates of the arc event are projected onto the wafer plane, and the corresponding position of the arc in the wafer mapping table is determined by judging the spatial relationship between the projection point and the boundary of the effective device area. In this embodiment, the determination of the projection and position relationship can specifically employ a geometric inclusion algorithm (such as the Point-in-Polygon algorithm), thereby determining the specific chip index where the arc falls and its spatial position on the wafer within microseconds.

[0059] The above-mentioned wafer mapping table generation method introduces process context information (wafer identifier, process step identifier, chip layout diagram) into arc management, enabling subsequent risk level determination to be based on the actual wafer layout. Furthermore, by transforming the coordinate system, the wafer mapping table is accurately mapped to the global coordinate system of the plasma processing chamber, avoiding the defect in traditional solutions where fluctuations in non-critical areas are misjudged as high-risk events due to the lack of spatial context.

[0060] As a variation of the chip layout diagram, the chip layout diagram is a CAD file, and the control module extracts the coordinates of the valid chip area by parsing the CAD file; in other embodiments, the chip layout diagram can also be wafer layout data stored in other formats.

[0061] As a further configuration of the control module, the decision and optimization unit is configured as follows: It obtains the arc position based on the arc event, projects the three-dimensional coordinates of the arc position onto the plane where the wafer is located, and determines the corresponding position of the arc position in the wafer mapping table by judging the spatial relationship between the projection point and the boundary of the effective device area; it judges whether the arc position falls within the effective device area of ​​the wafer mapping table; if the arc position falls within the effective device area, it is determined to be of a high-risk level; if the arc position does not fall within the effective device area, it is determined to be of a low-risk level; and it generates a suppression instruction based on the risk level.

[0062] The aforementioned spatially differentiated risk assessment method projects the three-dimensional coordinates of the arc location onto the plane of the wafer, and performs spatial mapping by judging the spatial relationship between the projection point and the boundary of the effective device area. Based on this, the risk level is determined and a corresponding suppression command is generated. If the arc falls into the effective device area, it is judged as high risk and strong suppression is triggered; if it does not fall into the effective device area, it is judged as low risk and weak compensation is used. This avoids false alarms or overreactions caused by traditional single suppression strategies and significantly reduces the false alarm rate.

[0063] As a variation of the control module, for arcs located at the edge of the effective device region of the wafer, an intermediate risk level can be set according to the distance between the arc and the boundary of the effective device region, and a corresponding medium suppression intensity can be adopted.

[0064] As a further configuration of the decision-making and optimization unit, when the risk level is high, a suppression command with a first suppression strength (e.g., compensation coefficient K=1.0) and alarm information are generated; when the risk level is low, a suppression command with a second suppression strength (e.g., compensation coefficient K=0.3) is generated, with the first suppression strength being greater than the second suppression strength. The alarm information includes an arc location tag and can be sent to the manufacturing execution system or fault detection and classification system via the factory automation system.

[0065] The control module generates differentiated suppression commands based on the risk level: high risk corresponds to the first suppression intensity and an alarm is issued, while low risk corresponds to the second suppression intensity, with the suppression intensity for high risk being greater than that for low risk. This solution achieves refined management of arc energy, providing strong suppression and alarm notifications for high-risk events, while requiring only weak compensation for low-risk events, thus avoiding disruption to process continuity. Simultaneously, the alarm information provides engineers with precise fault location information.

[0066] As a variation of alarm information implementation, alarm information can be sent from the factory automation system to the manufacturing execution system, fault detection and classification system, or process controller to achieve linkage with the factory-level automation system.

[0067] As a further configuration of the control module, the recording unit is configured to: record the location information, energy, time, risk level and suppression results of each arc event and store them in the database; periodically count the arc occurrence frequency of each process step and each spatial location after each batch of wafers is processed, and update the arc thermal map according to the arc occurrence frequency.

[0068] Reference Figure 6 As shown, the arc thermal map uses a three-dimensional model of the chamber as its base, with the arc density represented by the depth of color.

[0069] The aforementioned heatmap update method statistically analyzes the location information of historical electric arc events according to process steps and spatial locations. Using a 3D model of the chamber as the base and color depth to represent arc density, it enables engineers to intuitively identify areas where electric arcs accumulate and process steps with high frequency, providing visual data support for equipment maintenance and process optimization.

[0070] As a variation of arc thermal mapping, arc thermal mapping can be filtered and displayed by wafer batch, process step, and time period, so that engineers can analyze the arc distribution characteristics under different dimensions.

[0071] As a further configuration of the recording unit, when the arc frequency at the same spatial location in the same process step exceeds a preset threshold (e.g., more than 5 times) within a predetermined time period (e.g., per hour), an early warning message is generated or parameter optimization suggestions are sent to the process controller. Parameter optimization suggestions include at least one of the following: reducing the power ramp-up slope, adjusting the gas flow rate, cleaning the target material, replacing components, or performing a cleaning procedure.

[0072] The aforementioned preventative optimization method, when statistical analysis reveals frequent micro-arcs occurring within the same process step and area over a short period, predicts the risk of hard arcs in advance and triggers optimization, transforming arc management from a passive response to proactive prevention, thus reducing arc occurrences at the source. This solution visually displays arc accumulation areas using heat maps; when the frequency exceeds a threshold, it predicts the risk of hard arcs in advance and triggers optimization, shifting arc management from a passive response to proactive prevention, and reducing arc occurrences at the source.

[0073] As a variation of the parameter optimization suggestion, the parameter optimization suggestion includes at least one of reducing the power ramp rate, adjusting the gas flow rate, or performing a cleaning procedure; the warning information can be notified to the engineer through a display interface, SMS or email.

[0074] Reference Figure 7As shown, in another configuration of the control module, the yield correlation analysis unit is configured to: acquire wafer electrical test yield data (pass / fail for each chip); obtain arc locations based on multiple arc events and map them to the wafer chip layout; overlay the arc locations with the yield data to generate an arc influence distribution map; and, based on the spatial correspondence between arc locations and chip failure locations, identify the key arc types and locations that cause yield loss. In this embodiment, this spatial correspondence analysis can be implemented using statistical correlation analysis (such as chi-square test) or machine learning methods.

[0075] The aforementioned yield traceability method generates an arc impact distribution map by spatially overlaying the location of arc events with the yield data of each chip on the wafer. Based on the spatial correspondence between arc locations and chip failure locations, it identifies the key arc types and locations that lead to yield losses. This solution achieves accurate attribution of arc events to yield—traditional yield analysis cannot pinpoint whether failed chips are related to arcs, while this solution, through spatial mapping and spatial correspondence analysis, accurately identifies which arc events actually caused yield losses, providing direct data support for equipment maintenance and process optimization.

[0076] Reference Figure 8 As shown, in one specific implementation of the suppression module, the suppression module includes a power unit and a reverse compensation pulse injection unit. The power unit uses gallium nitride (GaN) high electron mobility transistors (HEMTs) to construct an RF / DC power output stage with a switching frequency of tens of MHz and a response time of less than 1 ns. The suppression module integrates a high-speed analog-to-digital converter (sampling rate greater than or equal to 1 GS / s) and a field-programmable gate array (FPGA). The FPGA is used to perform real-time waveform monitoring, error detection, and compensation pulse calculation.

[0077] During the process development phase, output voltage / current waveforms under normal, arc-free conditions are collected for different formulation steps (ignition, steady state, power switching) to train a desired waveform model (either an average template or an ARMA model). This model is pre-stored in a field-programmable gate array (FPGA) lookup table. When acquiring the desired waveform model, for the same process formulation step, real-time output waveforms of multiple batches of wafers under normal process operation are collected, and the mean and standard deviation at each moment are calculated to form a dynamic desired template. When the waveform exceeds the mean ± 3σ range at a certain moment, it is considered a waveform error, and the FPGA immediately triggers the reverse compensation pulse generation logic. This statistical threshold method further reduces false triggering caused by normal process disturbances and improves the robustness of arc detection.

[0078] During real-time operation, the output waveform is continuously acquired. , and the desired waveform Comparison calculation of instantaneous error If the absolute value of the error exceeds the preset threshold, it is determined to be a waveform error.

[0079] The reverse compensation pulse injection unit is configured to generate a compensation pulse whose amplitude matches the waveform error and whose phase is opposite to the waveform error, based on the amplitude and phase of the waveform error and the suppression intensity in the suppression command, and adjust the injection timing of the compensation pulse based on the system delay compensation amount.

[0080] In this embodiment, the preset threshold for the error is specifically 5% of the rated voltage; the compensation pulse can be expressed as... ,in The compensation coefficient determined according to the suppression command (e.g., under high risk) Under low risk ), This is for system delay compensation (including detection, calculation, and switching delays, determined through calibration). The pulse is injected into the power supply output during the next switching cycle, superimposed on the arc distortion to bring the net output closer to the desired value.

[0081] The aforementioned reverse pulse injection method, through real-time waveform monitoring and error detection, generates a compensation pulse with opposite phase and matched amplitude based on the waveform error and the suppression intensity in the suppression command. The injection timing of the compensation pulse is adjusted in conjunction with the system delay compensation, and injected into the power output terminal in the next switching cycle. This restores the net output waveform to the desired value, eliminating the need to shut down the power supply to counteract arc energy, maintaining the plasma bulk without deionization, and ensuring process continuity. The switching speed of gallium nitride power devices allows for a compensation pulse delay of less than 5 ns, achieving sub-microsecond suppression.

[0082] As a modified implementation of reverse compensation pulse generation, the compensation coefficient It can be dynamically adjusted, ranging from 0.9 to 1.1, to adaptively suppress the intensity based on the arc energy, avoiding overcompensation or undercompensation.

[0083] As a further configuration of the reverse compensation pulse injection unit, the suppression module continuously monitors the waveform error after injecting the compensation pulse. If the waveform error still exists, the generation and injection of the compensation pulse are performed again until the waveform error disappears. If the waveform error still does not disappear after repeated multi-cycle continuous compensation, it is determined to be a hard arc, and the power is turned off for a preset duration. In some more specific embodiments, the preset duration of the power-off is set to less than 10μs.

[0084] The aforementioned multi-cycle continuous compensation method ensures effective suppression of continuous arcs through multi-cycle compensation; for the very few hard arcs where multi-cycle compensation is ineffective, arc extinguishing is assisted by briefly shutting off the power supply, which not only ensures rapid suppression of conventional arcs, but also provides a safety fallback mechanism for extreme situations.

[0085] Reference Figure 9 As shown, in this embodiment, when an electric arc occurs in the plasma chamber (at t=0), the arc source simultaneously radiates broadband electromagnetic waves and characteristic optical signals. The sensing units of the sensor module synchronously capture both types of signals (t+ several ns to several μs) and send them to the signal processing module. The signal processing module first performs coarse localization based on the time difference of electromagnetic wave arrival, with a total time of less than 1 μs, and outputs a coarse localization position with centimeter-level accuracy; then, based on the coarse localization result, it activates the nearest subset of sensing units to perform fine localization using optical tomography, which takes about 10 μs and outputs a fine localization position with millimeter-level accuracy; finally, it performs data fusion, which takes less than 5 μs, outputs the final arc position, and estimates the arc energy and duration based on the integral of the light intensity and the electromagnetic wave signal energy, generating an arc event (t+ < 15 μs). The control module acquires the current wafer's process context information and determines the risk level (t+ < 50μs) based on the arc event. High risk corresponds to strong compensation (K=1.0) and an alarm is sent; low risk corresponds to weak compensation (K=0.3). The suppression command is sent to the suppression module via a high-speed bus (such as LVDS or fiber optic link) with a delay of less than 1μs (t+ < 55μs). The GaN smart power supply of the suppression module injects a compensation pulse into the power output terminal within 5μs after receiving the command (t+ < 60μs). If the arc continues, compensation is continuously applied in subsequent cycles until the error disappears (t+60μs to t+ several hundredμs). The entire closed-loop process is completed in microseconds, realizing a complete cycle of "detection → location → decision → suppression → recording → optimization". The control module stores all information of this arc event in the database and updates the heat map (t+ several hundredμs to t+ several ms). If the risk is high, an alarm with a location tag is sent to the FDC; if statistics show that micro-arcs frequently occur in the same process step and the same area within a short period of time, preventive optimization is triggered. After the wafer completes electrical testing (t+ several hours to several days), the control module imports the yield data to generate an arc influence distribution map.

[0086] Based on the above system linkage timing, it can be concluded that the time from arc occurrence to the issuance of the suppression command is approximately 55μs, and the time from the suppression command to the injection of the compensation pulse is approximately 5μs, with a total suppression delay of approximately 60μs. This total suppression delay is far less than the millisecond-level response of traditional power-off schemes, and it has the advantage of a complete closed loop of "detection-location-decision-suppression-recording-optimization", which greatly improves the intelligence level of arc management and process yield.

[0087] It should be noted that although the number of sensing units in the above embodiment is 8, in other embodiments, different numbers of sensing units can be set according to the cavity size and positioning accuracy requirements, such as 6, 10 or more. The specific position of each sensing unit can also be adjusted according to the cavity structure, in principle ensuring that a three-dimensional coverage with multiple angles and heights is formed around the cavity.

[0088] In some specific application scenarios of this invention, the aforementioned plasma arc management system can be deployed in various plasma processing chambers, such as physical vapor deposition (PVD), plasma enhanced chemical vapor deposition (PECVD), and etching (Etch) chambers. The following example, a 12-inch physical vapor deposition tantalum nitride (TaN) process chamber, illustrates the specific application process of this invention.

[0089] Step 1: Arc generation and physical signal radiation (t=0 time) In a 12-inch physical vapor deposition tantalum nitride process chamber, deposits peeled off from the target edge. During process operation, these peeled-off deposits entered the plasma region, triggering a micro-arc discharge. The arc source simultaneously radiated broadband electromagnetic waves (100MHz to 6GHz) and characteristic optical signals (argon I 811.5nm spectrum).

[0090] Step 2: Sensor module synchronously captures data (t + nanoseconds to microseconds). Eight sensing units arranged three-dimensionally around the chamber (three at 120° intervals at the top, four at different azimuth angles in the middle, and one at the bottom) synchronously receive the aforementioned signals. The electromagnetic wave signal output by the antenna is transmitted to the acquisition card via a coaxial cable, and the optical signal received by the fiber optic probe is converted into a voltage signal through photoelectric conversion. All analog signals are sent to the signal processing module in real time and synchronously.

[0091] Step 3: Coarse localization based on electromagnetic wave arrival time difference (t+ < 1 μs) The signal processing module performs bandpass filtering on the antenna signals through the electromagnetic wave positioning unit, detects the arrival time using the short-time energy method, calculates the time difference of arrival using a reference antenna, and solves for the initial three-dimensional coordinates of the arc using the Chan algorithm. This process is completed by field-programmable gate array hardware, with a total time of less than 1μs, and outputs a coarse positioning position with centimeter-level accuracy.

[0092] Step 4: Fine-tuning with optical tomography (t+ < 10 μs) Based on the coarse localization results, the signal processing module activates the four fiber optic probes closest to the region via the optical tomography unit. The optical signal is smoothed and filtered, the intensity amplitude is extracted, a Gaussian emission model is established, and algebraic reconstruction technology is used for tomographic reconstruction. The location of the maximum intensity is taken as the fine coordinate, achieving millimeter-level accuracy.

[0093] Reference Figure 10 As shown, after adopting the multimodal fusion positioning scheme of the present invention, the arc positioning error is significantly reduced compared with single electromagnetic wave positioning or single optical positioning, and the positioning accuracy is improved to the millimeter level.

[0094] Step 5: Data fusion (t+ < 15μs) The signal processing module uses a data fusion unit to fuse the coarse and fine positioning positions as observations using Kalman filtering, outputting the final arc position. Simultaneously, it estimates the arc energy and duration based on the light intensity integral and antenna signal energy, generating a complete arc event (position, energy, and time).

[0095] Step 6: The control module obtains the process context and determines the risk (t+ < 50μs). The control module obtains the current wafer ID and process step ID from the manufacturing execution system, retrieves the coordinates of the effective device area from the chip layout database, and generates a wafer mapping table. It maps the arc position to the wafer coordinate system and determines whether the arc position falls within the effective device area. Based on the mapping, the arc is located at the edge of the target material—outside the effective device area of ​​the wafer—and is classified as a low-risk event. Actual production line verification shows that combining the context-aware decision logic of the wafer layout map can reduce the false alarm rate by more than 80%.

[0096] Step 7: The control module sends a suppression command (t+ < 55μs) to the suppression module. The control module determines the risk to be low and sends a weak compensation command (compensation coefficient K=0.3) to the gallium nitride smart power supply via the high-speed bus. It only records the event and does not trigger an alarm.

[0097] Step 8: The suppression module performs a reverse pulse injection (t+ < 60μs). After receiving the command, the field-programmable gate array inside the power unit of the suppression module reads the output waveform in real time, compares it with the expected waveform to calculate the error, and then controls the gallium nitride switch by the reverse compensation pulse injection unit to generate a compensation pulse with opposite phase and matching amplitude, which is injected into the output terminal in the next switching cycle. The delay from receiving the command to the actual injection of the compensation pulse is less than 5μs.

[0098] Reference Figure 11As shown, after adopting the reverse pulse injection suppression scheme of the present invention, the power supply output waveform quickly returns to normal after the arc occurs. The compensation pulse effectively cancels the waveform distortion caused by the arc, and the plasma body is not interrupted due to the suppression operation.

[0099] Step 9: Post-suppression state detection and multi-cycle processing (t+60μs to t+hundredsμs) The reverse compensation pulse injection unit continues to monitor the output waveform error. The error disappears after a single compensation, and the system records success and exits. The entire suppression process is completed within 60μs, which is much smaller than the millisecond response of traditional power-off schemes, significantly reducing the risk of arc damage to the wafer.

[0100] Step 10: The control module records the event, updates the heatmap, and triggers an alert (t + hundreds of μs to t + several ms). The control module stores all information about the current electric arc event (location, energy, time, risk level, and success of suppression) into the database through the recording unit and updates the electric arc heat map. Engineers can visually identify areas where electric arcs accumulate through the heat map—discovering that most arcs are concentrated at a certain point on the edge of the target material.

[0101] Statistics showed that micro-arcs frequently occurred in the same process step and the same area (target edge) during multiple consecutive wafer processing batches, exceeding 5 times per hour, surpassing the preset threshold. The system automatically generated a warning message and suggested checking the area. Engineers inspected the area and found deposits peeling off at the target edge. After cleaning, the arc frequency decreased significantly. After applying this invention to a certain physical vapor deposition process, arc-related defects were reduced by 65%, and the overall yield improved by 2.3%.

[0102] Step 11: Subsequent yield correlation (t + several hours to several days) After the wafer completes all processes and undergoes electrical testing, the yield data for each chip is imported into the system. The control module overlays the locations of historical arc events with the chip layout to generate an arc impact distribution map. Through correlation analysis (such as chi-square test or machine learning), the key arc types and locations causing yield losses are accurately identified—analysis shows that arcs are mainly concentrated in the edge region of the target material, and the failure rate of chips at the wafer edge is significantly higher than that in the center region. Based on this, engineers can optimize process parameters or adjust chamber conditions to achieve accurate yield attribution.

[0103] This specific application example demonstrates that the system of the present invention achieves complete closed-loop management from arc generation to suppression, recording, early warning, and yield traceability in a real physical vapor deposition process environment, significantly reducing the impact of arcs on process yield. The present invention can also be applied to other plasma processing processes (such as plasma-enhanced chemical vapor deposition, etching, etc.) and process chambers of different wafer sizes (such as 200mm, 300mm) and different material systems. The specific number and layout of sensing units can be adjusted according to the chamber geometry and positioning accuracy requirements, all of which fall within the scope of protection of this invention.

[0104] Reference Figure 12 As shown, this embodiment provides a plasma arc management method, which can be used in any of the plasma arc management systems described above. The method is executed by a sensor module, a signal processing module, a control module, and a suppression module. The process includes the following steps: Step S201: Electromagnetic wave signals generated by the electric arc are continuously received by multiple sensing units located at different spatial positions around the plasma processing chamber. When the sensing units are activated, the light signals generated by the electric arc are received by the activated sensing units.

[0105] Step S202: Determine the coarse location of the arc based on the electromagnetic wave signal; activate the sensing unit in the corresponding area based on the coarse location; receive the optical signal generated by the arc through the activated sensing unit; determine the fine location of the arc based on the optical signal; generate an arc event based on the coarse location, fine location, electromagnetic wave signal, and optical signal; the arc event includes the position, energy, and duration of the arc; obtain the effective device area layout map of the current wafer; map the arc position in the arc event to the wafer coordinate system; determine whether the arc position falls within the effective device area of ​​the effective device area layout map; and generate suppression instructions with different suppression intensities corresponding to different risk levels based on the judgment result.

[0106] Step S203: According to the suppression command, the real-time output waveform of the power supply of the plasma processing chamber is obtained; the real-time output waveform is compared with the preset desired waveform to obtain the waveform error; according to the suppression intensity and the waveform error, a compensation pulse with the opposite phase and the matching amplitude of the waveform error is generated, and the compensation pulse is injected into the output terminal of the power supply to suppress the electric arc.

[0107] The plasma arc management method provided in this embodiment of the invention can be applied to the plasma arc management system of any of the above embodiments. The specific implementation of the above method steps has been described in detail in the sensor module, signal processing module, control module and suppression module of the aforementioned system embodiments. For example, the specific process of coarse positioning, fine positioning and arc event generation in step S202 has been fully recorded in the signal processing module. The specific process of generating suppression commands and suppressing arcs in step S203 has been fully recorded in the control module and suppression module, and will not be repeated here.

[0108] This invention also provides a non-transitory computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the plasma arc management method described above. This non-transitory computer-readable storage medium possesses all the beneficial effects of the plasma arc management method described above, which will not be elaborated further here.

[0109] This invention also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described display system control method and has all the beneficial effects of the above-described plasma arc management method. Further details are omitted here.

[0110] This invention also provides an electronic device, including a memory and a processor. The memory stores a computer program; the processor executes the computer program in the memory to implement the steps of the plasma arc management method described above. This electronic device possesses all the beneficial effects of the plasma arc management method described above, which will not be elaborated further here.

[0111] Reference Figure 13 As shown, the electronic device may include a processing system (e.g., a central processing unit, a graphics processing unit, etc.) 301, which can perform various appropriate actions and processes according to a program stored in read-only memory (ROM) 302 or a program loaded from storage system 308 into random access memory (RAM) 303. The RAM 303 also stores various programs and data required for the operation of the electronic device. The processing system 301, ROM 302, and RAM 303 are interconnected via a bus 304. An input / output (I / O) interface 305 is also connected to the bus 304.

[0112] Typically, the following systems can be connected to I / O interface 305: input systems 306 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output systems 307 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage systems 308 including, for example, magnetic tapes, hard disks, etc.; and communication systems 309. Communication system 309 allows electronic devices to communicate wirelessly or wiredly with other devices to exchange data. Although... Figure 13 Electronic devices with various systems are shown, but it should be understood that it is not required to implement or have all of the systems shown. More or fewer systems may be implemented alternatively. Figure 13 Each box shown can represent a system, or multiple systems as needed.

[0113] In particular, according to some embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, some embodiments of the present invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication system 309, or installed from a storage system 308, or installed from a read-only memory 302. When the computer program is executed by the processing system 301, it performs the functions defined in the methods of some embodiments of the present invention.

[0114] It should be noted that, in some embodiments of the present invention, the computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. The computer-readable storage medium may be, for example, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system or device, or any combination thereof, and the present invention does not specifically limit this. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, a random access memory, a read-only memory, an erasable programmable read-only memory (EPROM) or flash memory, an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof.

[0115] In some embodiments of the present invention, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in connection with an instruction execution system, system, or device. In some embodiments of the present invention, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium, which may transmit, propagate, or transmit a program for use by or in connection with an instruction execution system, system, or device. The program code contained on the computer-readable medium may be transmitted using any suitable medium, including but not limited to: wires, optical fibers, RF (radio frequency), etc., or any suitable combination thereof.

[0116] In some implementations, clients and servers can communicate using any currently known or future-developed network protocol, such as Hypertext Transfer Protocol (HTTP), and can interconnect with digital data communication (e.g., communication networks) of any form or medium. Examples of communication networks include Local Area Networks (LANs), Wide Area Networks (WANs), the Internet (e.g., the Internet), and peer-to-peer networks (e.g., ad-hoc peer-to-peer networks), as well as any currently known or future-developed networks.

[0117] The aforementioned computer-readable medium may be included in the aforementioned electronic device or may exist independently without being assembled into the electronic device. The aforementioned computer-readable medium carries one or more programs.

[0118] Computer program code for performing operations of some embodiments of the present invention can be written in one or more programming languages ​​or a combination thereof. These programming languages ​​include object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network (including local area networks or wide area networks), or it can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0119] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function.

[0120] It should also be noted that in some alternative implementations, the functions marked in the box may occur in a different order than those marked in the attached figures.

[0121] For example, two consecutively represented blocks can actually be executed in substantially parallel order, and sometimes they can be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram and / or flowchart, as well as combinations of blocks in block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified functions or operations, or using a combination of dedicated hardware and computer instructions.

[0122] The units described in some embodiments of the present invention can be implemented in software or in hardware.

[0123] The functions described above in this document can be performed at least in part by one or more hardware logic components. For example, exemplary types of hardware logic components that can be used, without limitation, include: Field-Programmable Gate Array (FPGA), Application Specific Integrated Circuit (ASIC), Application Specific Standard Product (ASSP), System on Chip (SOC), Complex Programmable Logic Device (CPLD), etc.

[0124] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A plasma arc management system, characterized in that, include: The sensor module includes multiple sensing units disposed at different spatial locations around the plasma processing chamber. The sensing units are configured to continuously receive electromagnetic wave signals generated by the electric arc and, when activated, receive light signals generated by the electric arc. The signal processing module is communicatively connected to the sensor module and is configured to determine the coarse positioning position of the electric arc based on the electromagnetic wave signal and activate the sensing unit in the corresponding area, determine the fine positioning position of the electric arc based on the optical signal, and then generate an electric arc event based on the coarse positioning position, the fine positioning position, the electromagnetic wave signal and the optical signal. The electric arc event includes the position, energy and duration of the electric arc. The control module is communicatively connected to the signal processing module and is configured to obtain the effective device area layout map of the current wafer, map the arc position in the arc event to the wafer coordinate system, determine whether the arc position falls within the effective device area of ​​the effective device area layout map, and generate suppression instructions with different suppression intensities corresponding to different risk levels based on the judgment result. The suppression module is connected to the power supply of the control module and the plasma processing chamber, and is configured to acquire the real-time output waveform of the power supply according to the suppression command. The real-time output waveform is compared with the preset desired waveform to obtain the waveform error; Based on the suppression intensity and the waveform error, a compensation pulse with the opposite phase and matching amplitude to the waveform error is generated, and the compensation pulse is injected into the output terminal of the power supply to suppress the electric arc.

2. The plasma arc management system according to claim 1, characterized in that, The signal processing module includes an electromagnetic wave positioning unit, which is configured as follows: Bandpass filtering is performed on the electromagnetic wave signals received by each sensing unit; The arrival time of each filtered electromagnetic wave signal is detected; Using one of the sensing units as a reference sensing unit and the arrival time of the reference sensing unit as a reference, calculate the arrival time difference between the other sensing units and the reference sensing unit. Based on the spatial coordinates of each sensing unit and the arrival time difference, a hyperbolic equation system about the coordinates of the arc source is established. After linearizing the hyperbolic equation system, iteratively solving it using the least squares method yields the coarse location of the arc.

3. The plasma arc management system according to claim 2, characterized in that, The signal processing module further includes an optical tomography unit, which is further configured as follows: Based on the coarse positioning position, activate the subset of sensing units closest to the coarse positioning position; The optical signal received by the subset of the sensing units is smoothed and filtered to extract the light intensity amplitude; A spatial luminous intensity distribution model of the electric arc luminous intensity is established based on the light intensity amplitude. Based on multi-angle projection data, and combined with the principles of three-dimensional spatial projection and back projection reconstruction, the spatial luminous intensity distribution model is reconstructed by tomography to obtain the three-dimensional luminous intensity distribution of the electric arc. The location of the maximum intensity in the three-dimensional luminescence intensity distribution is taken as the fine positioning location.

4. The plasma arc management system according to claim 3, characterized in that, The signal processing module further includes a data fusion unit, which is further configured to: Using the coarse and fine positioning positions as observations, the coarse and fine positioning positions are weighted and fused based on state estimation and error covariance matrix to output the arc position; The energy and duration of the electric arc are estimated based on the intensity integral of the optical signal and the energy of the electromagnetic wave signal. The arc event is generated based on the arc location, the energy, and the duration.

5. The plasma arc management system according to claim 1, characterized in that, The control module includes a sensing unit, which is configured as follows: The system receives arc events output by the signal processing module in real time. Obtain the wafer identifier and process step identifier of the current wafer, and obtain the corresponding chip layout diagram based on the wafer identifier; The chip layout diagram is analyzed to extract the coordinates of the effective device area of ​​the wafer. A wafer mapping table is generated based on the coordinates of the effective device area, and the wafer mapping table is transformed from the wafer coordinate system to the global coordinate system of the plasma processing chamber.

6. The plasma arc management system according to claim 5, characterized in that, The control module further includes a decision-making and optimization unit, which is configured as follows: The arc position is obtained based on the arc event, the three-dimensional coordinates of the arc position are projected onto the plane where the wafer is located, and the corresponding position of the arc position in the wafer mapping table is determined by judging the spatial position relationship between the projection point and the boundary of the effective device area. Determine whether the arc position falls within the effective device area of ​​the wafer mapping table. If the arc position falls within the effective device area, it is determined to be of high risk level; if the arc position does not fall within the effective device area, it is determined to be of low risk level. The suppression instruction is generated based on the risk level.

7. The plasma arc management system according to claim 6, characterized in that, The decision-making and optimization unit is further configured as follows: When the risk level is high risk, a suppression command and alarm information of the first suppression strength are generated; When the risk level is low risk, a suppression instruction with a second suppression strength is generated; The first inhibition strength is greater than the second inhibition strength.

8. The plasma arc management system according to claim 1, characterized in that, The suppression module includes: The power unit is configured to acquire the output voltage or output current waveform of the power supply in real time as the real-time output waveform. The reverse compensation pulse injection unit is configured to: compare the real-time output waveform with the preset expected waveform point by point; if the absolute value of the error at a certain point exceeds a preset threshold, it is determined to be a waveform error. Based on the amplitude and phase of the waveform error and the suppression intensity in the suppression command, a compensation pulse with an amplitude matching the waveform error and a phase opposite to the waveform error is generated, and the injection timing of the compensation pulse is adjusted based on the system delay compensation amount. The compensation pulse is injected into the output of the power supply during the next switching cycle.

9. The plasma arc management system according to claim 8, characterized in that, The reverse compensation pulse injection unit is further configured to: After the compensation pulse is injected, the waveform error is continuously monitored; If the waveform error still exists, the generation and injection of the compensation pulse are performed again until the waveform error disappears; If the waveform error still does not disappear after repeated multi-cycle continuous compensation, it is determined to be a hard arc, and the power supply is turned off for a preset time.

10. The plasma arc management system according to claim 1, characterized in that, The control module further includes a recording unit, which is configured as follows: Record the location, energy, time, risk level, and suppression results of each electric arc event and store them in the database; Regularly analyze the frequency of electric arc generation at each process step and spatial location; The arc heat map is updated based on the arc occurrence frequency. The arc heat map uses a three-dimensional model of the chamber as a base and uses color depth to represent arc density. When the arc frequency at the same spatial location in the same process step exceeds a preset threshold within a predetermined time, an early warning message is generated or parameter optimization suggestions are sent to the process controller.

11. The plasma arc management system according to claim 1, characterized in that, The control module further includes a yield correlation analysis unit, which is configured as follows: Obtain wafer electrical test yield data; The arc position is obtained based on multiple arc events, and the arc position is mapped to the wafer chip layout; The arc location is superimposed with the yield data to generate an arc influence distribution map; Based on the arc influence distribution map, and the spatial correspondence between the arc location and the chip failure location, the key arc types and locations that lead to yield loss are identified.

12. A plasma arc management method, applied to the plasma arc management system according to any one of claims 1 to 11, characterized in that, include: The electromagnetic wave signal generated by the electric arc is continuously received by multiple sensing units located at different spatial positions around the plasma processing chamber. When the sensing unit is activated, the light signal generated by the electric arc is received by the activated sensing unit. The coarse location of the electric arc is determined based on the electromagnetic wave signal. The sensing unit in the corresponding area is activated based on the coarse location, and the light signal generated by the electric arc is received through the activated sensing unit. The fine location of the electric arc is determined based on the light signal. An electric arc event is generated based on the coarse location, the fine location, the electromagnetic wave signal, and the light signal. The electric arc event includes the location, energy, and duration of the electric arc. Obtain the effective device area layout map of the current wafer and map the arc position in the arc event to the wafer coordinate system; Determine whether the arc position falls within the effective device area of ​​the effective device area layout diagram, and generate suppression instructions with different suppression intensities corresponding to different risk levels based on the determination result; According to the suppression command, the real-time output waveform of the power supply of the plasma processing chamber is obtained; the real-time output waveform is compared with the preset expected waveform to obtain the waveform error; Based on the suppression intensity and the waveform error, a compensation pulse with the opposite phase and matching amplitude to the waveform error is generated, and the compensation pulse is injected into the output terminal of the power supply to suppress the electric arc.

13. An electronic device, characterized in that, include: The system includes a memory and a processor, which are communicatively connected to each other. The memory stores computer instructions, and the processor executes the computer instructions to perform the plasma arc management method of claim 12.

14. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing a computer to execute the plasma arc management method of claim 12.

Images