Arc detection processing method and apparatus for radio frequency plasma processing systems

Abstract

This application provides an arc detection and processing method and apparatus for a radio frequency plasma processing system, belonging to the field of semiconductor equipment technology. The method includes: acquiring analog voltage and current signals from the matching unit side as analog electrical signals; acquiring image data within the reaction chamber; fusing the analog electrical signals and image data to obtain arc level information and spatial location information; executing real-time arc extinguishing control based on the level information; and executing a location-based processing strategy based on the spatial location information. This application achieves accurate arc detection and location through the fusion of electrical signals and image information, efficiently realizes arc extinguishing control based on a graded arc extinguishing strategy, and guides root cause management through a location-based processing strategy. This solves the problem of high arc misjudgment rate in existing technologies and helps to eliminate the causes of arc occurrence at their source.

Description

Technical Field

[0001] This application relates to the field of semiconductor equipment technology, specifically to an arc detection and processing method and apparatus for a radio frequency plasma processing system. Background Technology

[0002] In semiconductor manufacturing, radio frequency plasma processing systems, consisting of radio frequency power supplies, matching circuits, and reaction chambers, are widely used in coating, etching, and cleaning / removing processes. These systems are the fundamental devices for plasma excitation. During plasma excitation, arcing discharge often occurs due to design defects in the reaction chamber or product tooling, electrode insulation design defects, reactant deposition, aging or damage to insulating materials, etc. Severe arcing can generate instantaneous temperatures exceeding 10,000 degrees Celsius. This can not only cause burnout or damage to insulating components in the product, reaction chamber, or equipment, but also lead to drastic changes in the impedance of the reaction chamber, thereby damaging the radio frequency power supply or shortening its lifespan.

[0003] Based on this, in response to the application requirements of such systems, related technologies for arc detection and processing have emerged, such as those involved in prior patents such as CN200680022298.X - "Arc Detection and Processing in Radio Frequency Power Applications", CN202180014708.0 - "System and Method for Arc Detection Using Bias Radio Frequency Generator Signals", CN202410026781.5 - "A Method, Device, Equipment and Storage Medium for Arc Detection Applied to Radio Frequency Systems", and CN202510206823.8 - "A Method and System for Rapid Detection of Radio Frequency Power Arcs". The core idea of ​​these technologies is to sample forward and reverse radio frequency power through a coupler, or to collect voltage and current information, and to process the collected data. By comparing it with corresponding thresholds, the arc and the number of arcs can be monitored and information can be obtained. Based on this, relevant processing can be performed. In addition, in patent CN202410481606.5- "A Method for Arc Detection and Processing of a Plasma Generating Radio Frequency Power Supply", in addition to sampling radio frequency power, a photosensitive device is added to realize arc detection. Its main purpose is to assist in determining whether an arc has been generated.

[0004] In developing this invention, the inventors discovered that existing technologies primarily detect electric arcs by sampling radio frequency (RF) power or voltage and current signals along the RF power transmission path. The basic principle is that an electric arc causes a significant change in the impedance of the load or transmission circuit; therefore, by monitoring these electrical signals and extracting abrupt changes in the data, the generation of an electric arc can be determined. However, in actual manufacturing processes, the tuning of the matching circuit and the ignition of gas within the chamber can cause drastic changes in circuit and chamber impedance. These normal impedance fluctuations can also lead to changes in parameters such as reflected power, voltage, and current. If these normal changes cannot be distinguished from the abnormal abrupt changes caused by an electric arc, misjudgments are highly likely. To avoid this problem, existing solutions typically rely on setting a reasonable threshold to identify electric arcs: only when the signal change exceeds this threshold is it considered an electric arc event.

[0005] However, this threshold detection method has significant limitations. Any signal exceeding the threshold is counted as an arc, failing to distinguish between arc strength levels or pinpoint the exact location of the arc. Therefore, setting the threshold becomes extremely critical yet difficult to optimize: if the threshold is set too low, it will not only cause misjudgments but also lead to frequent activation of the arc extinguishing function, interfering with normal processes; if the threshold is set too high, it will be difficult to effectively monitor and handle weak arcs, leaving potential process hazards. Furthermore, due to the lack of arc location capabilities, existing systems can only attempt to extinguish the arc by simply reducing power, unable to take more targeted measures against the root cause of the arc, making it difficult to eradicate the arc.

[0006] Furthermore, while some existing arc detection technologies have incorporated photosensitive devices to assist in monitoring luminescence within the reaction chamber, their design purpose remains limited to assisting in electrical signal detection to reduce false alarm rates. They do not endow the system with the ability to spatially locate the arc's position. Therefore, existing solutions can only determine whether an arc has occurred, but cannot establish a correspondence between the arc event and the physical location of the chamber. This lack of direction for subsequent process improvements makes it difficult to eliminate the root cause of arcing.

[0007] On the other hand, existing detection architectures also have significant shortcomings in hardware deployment. Accurate arc capture relies on high-precision sampling of analog quantities such as voltage and current, but the sampling accuracy at the output of conventional RF power supplies is often insufficient to meet the requirements of real-time, weak arc detection. To address this issue, existing technologies typically employ the addition of a dedicated sampling module and integration of the arc detection function within the RF power supply. This approach not only increases hardware costs and system complexity, but more importantly, because the detection point is far from the reaction chamber, impedance changes in the matching unit during normal tuning are easily coupled to the detection circuit at the power supply end through the transmission link, leading to a significantly increased arc misjudgment rate and affecting the reliability and process stability of the arc extinguishing system. Summary of the Invention

[0008] To at least partially overcome the problems existing in related technologies, this application proposes an arc detection and processing method and apparatus for a radio frequency plasma processing system. The method employs a fusion of electrical signals and image information to detect the arc level and location, and achieves precise arc extinguishing of different arc levels through a cyclic power reduction control strategy.

[0009] First aspect This application provides an arc detection and processing method for a radio frequency plasma processing system, the arc detection and processing method comprising: The analog voltage and analog current on the matching side are collected as analog electrical signals, and image data inside the reaction chamber is acquired. The analog electrical signal and the image data are fused together to obtain the arc's level information and spatial location information. Real-time arc extinguishing control is performed based on the level information, and a location-based processing strategy is executed based on the spatial location information.

[0010] Second aspect This application provides an arc detection and processing apparatus for a radio frequency plasma processing system, used to implement the steps of the method described in any of the implementations of the first aspect above, the arc detection and processing apparatus comprising: An electrical signal acquisition unit, located inside the matching unit, is used to acquire analog voltage and analog current signals on the matching unit side. An image acquisition unit is installed in the reaction chamber and is used to acquire image data within the reaction chamber; The fusion processing unit is connected to the electrical signal acquisition unit and the image acquisition unit respectively, and is used to fuse the analog electrical signal and the image data to obtain the level information and spatial location information of the electric arc. A control execution unit, connected to the fusion processing unit, is used to perform real-time arc extinguishing control according to the level information and to perform a location-based processing strategy according to the spatial location information. The image acquisition unit, the fusion processing unit, and the control execution unit are all electrically connected to the power supply module of the matcher.

[0011] The technical solution provided in this application, by collecting and fusing analog voltage and current quantities from the matching device side and image data from the reaction chamber, can simultaneously obtain the arc's level and spatial location information. Compared to existing technologies that rely solely on a single electrical signal threshold for judgment, this solution achieves dual verification of electrical signals and image information, effectively distinguishing normal process fluctuations such as matching device tuning and gas ignition from real arc events, significantly reducing the false judgment rate. By obtaining the arc's level information, it solves the problem that existing technologies cannot distinguish between arc strength levels, providing a foundation for subsequent differentiated real-time arc extinguishing control for arcs of different levels. Spatial location information overcomes the limitations of existing technologies that can only determine whether an electric arc has occurred but cannot pinpoint its location. This makes location-based processing strategies (such as guiding physical modifications to specific areas of the reaction chamber) possible, thus providing a directional basis for eliminating the root causes of electric arcs. Based on this, real-time arc extinguishing control is executed according to the level information, enabling timely handling of currently occurring electric arcs and preventing damage to products and equipment. At the same time, executing location-based processing strategies based on spatial location information can accumulate electric arc frequency data to guide process improvements. This achieves rapid response to electric arcs and provides data support for suppressing the recurrence of electric arcs at the source. Attached Figure Description

[0012] Figure 1 A schematic flowchart of an arc detection and processing method for a radio frequency plasma processing system provided in one embodiment of this application; Figure 2 This is a schematic diagram illustrating the arc extinguishing process in an arc detection and processing method for a radio frequency plasma processing system according to an embodiment of this application. Figure 3 This is a block diagram of an arc detection and processing device for a radio frequency plasma processing system provided in one embodiment of this application; Figure 4 A schematic block diagram of an arc detection and processing device for a radio frequency plasma processing system provided in an embodiment of this application; Figure 5 for Figure 4 The illustrated embodiment is a schematic diagram illustrating the arc extinguishing process. Detailed Implementation

[0013] To make the purpose, technical solution and advantages of this application clearer, the technical solution of this application will be described in detail below.

[0014] As described in the background section, in realizing this invention, the inventors discovered that existing technologies primarily detect electric arcs by sampling radio frequency power or voltage and current signals along the radio frequency power transmission path. The basic principle is that an electric arc causes a significant change in the impedance of the load or transmission circuit; therefore, by monitoring these electrical signals and extracting abrupt changes in the data, the generation of an electric arc can be determined. However, in actual manufacturing processes, the tuning of the matching device and the ignition of gas within the chamber can cause drastic changes in circuit impedance and chamber impedance. These normal impedance fluctuations can also lead to changes in parameters such as reflected power, voltage, and current. If such normal changes cannot be distinguished from the abnormal abrupt changes caused by an electric arc, misjudgments are highly likely. To avoid this problem, existing solutions typically rely on setting a reasonable threshold to identify electric arcs: only when the signal change exceeds this threshold is it considered an electric arc event.

[0015] However, this threshold detection method has significant limitations. Any signal exceeding the threshold is counted as an arc, failing to distinguish between arc strength levels or pinpoint the exact location of the arc. Therefore, setting the threshold becomes extremely critical yet difficult to optimize: if the threshold is set too low, it will not only cause misjudgments but also lead to frequent activation of the arc extinguishing function, interfering with normal processes; if the threshold is set too high, it will be difficult to effectively monitor and handle weak arcs, leaving potential process hazards. Furthermore, due to the lack of arc location capabilities, existing systems can only attempt to extinguish the arc by simply reducing power, unable to take more targeted measures against the root cause of the arc, making it difficult to eradicate the arc.

[0016] Furthermore, while some existing arc detection technologies have incorporated photosensitive devices to assist in monitoring luminescence within the reaction chamber, their design purpose remains limited to assisting in electrical signal detection to reduce false alarm rates. They do not endow the system with the ability to spatially locate the arc's position. Therefore, existing solutions can only determine whether an arc has occurred, but cannot establish a correspondence between the arc event and the physical location of the chamber. This lack of direction for subsequent process improvements makes it difficult to eliminate the root cause of arcing.

[0017] On the other hand, existing detection architectures also have significant shortcomings in hardware deployment. Accurate arc capture relies on high-precision sampling of analog quantities such as voltage and current, but the sampling accuracy at the output of conventional RF power supplies is often insufficient to meet the requirements of real-time, weak arc detection. To address this issue, existing technologies typically employ the addition of a dedicated sampling module and integration of the arc detection function within the RF power supply. This approach not only increases hardware costs and system complexity, but more importantly, because the detection point is far from the reaction chamber, impedance changes in the matching unit during normal tuning are easily coupled to the detection circuit at the power supply end through the transmission link, leading to a significantly increased arc misjudgment rate and affecting the reliability and process stability of the arc extinguishing system.

[0018] Based on this, this application proposes an arc detection and processing method for a radio frequency plasma processing system. The method uses the fusion of electrical signals and image information to detect the arc level and location, and achieves precise arc extinguishing for arcs of different levels through a cyclic power reduction control strategy.

[0019] like Figure 1 As shown, in one embodiment, the arc detection processing method for a radio frequency plasma processing system in this application includes: Step S110: Collect analog voltage and analog current signals from the matching unit side as analog electrical signals, and collect image data inside the reaction chamber. Specifically, an electrical signal acquisition unit located inside the matching unit acquires analog voltage and current signals in real time, which are used to detect the electric arc. Simultaneously, an image acquisition unit located in the reaction chamber—for example, an observation window can be installed on the side of the ion reaction chamber—uses a camera to acquire real-time image data of the interior of the reaction chamber through the observation window. This image data is used to capture changes in light radiation generated by the electric arc discharge within the chamber. After acquisition, the analog electrical signals and image data are transmitted to a subsequent fusion processing unit for further fusion processing and analysis.

[0020] Based on step S110, such as Figure 1 As shown, in step S120, the analog electrical signal and image data are fused to obtain the arc's level information and spatial location information. Specifically, in this step, the fusion processing unit simultaneously receives analog voltage and current signals from the electrical signal acquisition unit and image data from the image acquisition unit, and performs the following processing: On the one hand, the analog electrical signal is processed to obtain the arc level information. In some embodiments, the process of obtaining the arc level information includes: The analog voltage and analog current signals are converted into digital signals and then subjected to time differentiation processing to obtain voltage change rate data and current change rate data. The voltage change rate data is compared with a preset zero voltage threshold, and the current change rate data is compared with a preset zero current threshold. The zero voltage threshold and the zero current threshold are set according to the voltage change rate amplitude and the current change rate amplitude caused by the matching process, respectively. Based on the comparison results of the voltage change rate data and the zero voltage threshold, the comparison results of the current change rate data and the zero current threshold, and the mutual verification results between the voltage change rate and the current change rate, it is determined whether an electric arc has occurred. When an electric arc is detected, the absolute value of the voltage change rate data is compared with a preset first voltage threshold to an Nth voltage threshold, and simultaneously the absolute value of the current change rate data is compared with a preset first current threshold to an Nth current threshold. This yields a first-level judgment result corresponding to the voltage change rate and a second-level judgment result corresponding to the current change rate. The first voltage threshold to the Nth voltage threshold correspond to different arc intensity levels, and the first current threshold to the Nth current threshold correspond to different arc intensity levels, where N≥5. Based on the first-level judgment result and the second-level judgment result, the intensity level of the current arc is determined, and the number of times the arc of that level occurs within the current detection cycle is recorded.

[0021] In this embodiment, both voltage and current sampling are compared to determine whether an arc has occurred. This is mainly to prevent false alarms caused by interference. If both samples are simultaneously greater than the zero voltage / current threshold, an arc is determined to have occurred. However, in some practical engineering scenarios, based on actual engineering experience and requirements, the voltage change rate can be used as the primary criterion (a larger voltage change amplitude results in a more sensitive response) and the current change rate as a secondary criterion when determining whether an arc has occurred. In other words, the voltage change rate is prioritized in determining whether an arc has occurred. In this embodiment, the intensity level of the electric arc is determined based on the change in current during the comparison and determination process. Specifically, the intensity level of the electric arc is determined based on the first level judgment result and the second level judgment result. When the first level judgment result is consistent with the second level judgment result, the consistent result shall prevail. When the two are inconsistent, the second level judgment result shall prevail.

[0022] In practical implementation, the received analog voltage and current signals can be converted into 16-bit digital signals and processed by time differentiation to obtain voltage change rate data (dv / dt) and current change rate data (di / dt). These change rate data are then compared with corresponding preset zero voltage / current thresholds. These zero voltage / current thresholds are set based on the amplitude of voltage and current change rates caused by the normal tuning process of the matching device, and are used to filter out normal signal fluctuations generated by the matching device's tuning operation. When both voltage and current change rate data exceed the zero voltage / current threshold, an arc is confirmed and recorded. After confirming the arc, the absolute value of the corresponding voltage or current change rate data is compared with preset first, second, ... fifth voltage / current thresholds, which correspond to different arc intensity levels from low to high. Based on the comparison results (using the current change comparison result as the standard), the intensity level of the current arc is determined, and the number of times an arc of that level occurs within the current detection cycle is recorded, thus obtaining the arc level information.

[0023] It should be noted that the intensity level of the electric arc is determined by comparing the absolute values ​​of the voltage change rate and / or current change rate with several preset thresholds. The specific threshold values ​​are related to the actual process conditions and the impact of the electric arc on the process. For example, the normal process fluctuation range and arc characteristics vary for different process types (such as coating, etching, and cleaning), different process formulations (such as gas type, pressure, and power), and different equipment configurations (such as chamber size and electrode structure). Therefore, in practical applications, the quantified values ​​of the zeroth voltage / current threshold to the fifth voltage / current threshold are not fixed but need to be calibrated and adjusted according to the actual process environment. For example, based on historical process data or experimental test results, arcs with an absolute value of voltage change rate in the range of 0-50V / μs can be defined as the first level (first-level micro-arc), those in the range of 50-100V / μs as the second level (second-level micro-arc), those in the range of 100-200V / μs as the third level (soft arc), those in the range of 200-400V / μs as the fourth level (soft arc), and those greater than 400V / μs as the fifth level (hard arc), etc.

[0024] By performing time differentiation on the analog voltage and current quantities and comparing them with the preset zero voltage / current threshold, interference caused by normal process fluctuations such as matching device tuning can be effectively filtered out. After confirming the occurrence of an arc, the absolute value of the rate of change data is compared with at least five preset level thresholds to obtain the arc intensity level and the number of occurrences of each level. This enables the quantitative differentiation of the arc strength and provides a refined data foundation for subsequent implementation of differentiated graded arc extinguishing control.

[0025] Furthermore, since the essence of arc discharge is a sudden and dramatic change in local impedance, its physical characteristics manifest as a sudden drop in voltage and a sudden rise in current, meaning that the rates of change of voltage and current must exhibit strictly opposite polarities. While normal process procedures such as matching circuit tuning and gas ignition may cause fluctuations in electrical signal amplitude, they do not possess this instantaneous reverse mutation characteristic. Therefore, this application also introduces "polarity reversal judgment" as a verification condition, adding a logical judgment dimension at the physical level on top of threshold comparison, thereby effectively filtering out signal fluctuations caused by non-arc factors such as matching circuit tuning. In some embodiments, the determination of whether an electric arc has occurred specifically involves: cross-checking the voltage change rate data and current change rate data at the same moment; when the voltage change rate data is greater than the zero voltage threshold, the current change rate data is greater than the zero current threshold, and the polarities of the voltage change rate and the current change rate are opposite, an electric arc is confirmed to have occurred and recorded.

[0026] In step S120, specifically, the process of processing image data to obtain the spatial location information of the electric arc includes: pre-acquiring images of the reaction chamber at different focal lengths to establish a mapping relationship between pixel positions and the spatial position of the chamber; acquiring image data within the reaction chamber in real time through at least two image acquisition modules, each responsible for covering different areas within the chamber and providing near-focal-length and far-focal-length images of the corresponding areas; processing the image data, comparing the current frame image with the previous frame image, extracting areas of brightness change, and obtaining the pixel position of the electric arc in the image; and calculating the actual spatial coordinates of the electric arc within the reaction chamber based on the difference in pixel position between the near-focal-length and far-focal-length images provided by the image acquisition module corresponding to the area where the electric arc is located, combined with the mapping relationship, as the location information of the electric arc occurrence.

[0027] In the specific implementation process, the first step is to use the configured image acquisition modules (e.g., at least two image acquisition modules to acquire different areas of the chamber) to pre-capture images of the reaction chamber at different focal lengths, establishing a mapping relationship between image pixel positions and the actual spatial position of the chamber, providing a reference for subsequent coordinate transformation. Then, the image acquisition modules are used to acquire image data within the reaction chamber in real time. Each module is responsible for covering different areas within the chamber and simultaneously provides near-focal-length and far-focal-length images of the corresponding areas. When processing the acquired image data, the differences between the current frame and the previous frame are compared to extract areas where brightness changes, thereby obtaining the pixel position of the electric arc in the image. Finally, based on the difference in pixel position of the electric arc in the near-focal-length and far-focal-length images provided by the image acquisition modules corresponding to the arc's location, combined with the pre-established pixel-space mapping relationship, the actual spatial coordinates of the electric arc within the reaction chamber are calculated using the parallax principle, serving as the location information of the electric arc's occurrence. By combining multi-focal-length image acquisition with parallax calculation, precise three-dimensional spatial positioning of the arc occurrence location is achieved, avoiding the limitation of existing technologies that can only determine the presence or absence of an arc but cannot locate it. This provides accurate coordinate basis for subsequent position-based processing strategies.

[0028] Furthermore, in some embodiments, processing the image data further includes: determining the intensity information of the electric arc based on the amount of brightness change in the brightness change area, wherein the intensity information is proportional to the amount of brightness change and is used to assist in verifying the level information of the electric arc.

[0029] Specifically, in obtaining the spatial location information of the electric arc, after extracting the brightness change area, the brightness change amount of that area can be calculated (e.g., a sudden change from a grayscale value of 50 to 200 results in a change of 150), which is proportional to the arc intensity. Simultaneously, a correspondence between brightness change and arc level is pre-established (e.g., brightness change of 0-30 corresponds to a first-level micro-arc, 30-90 to a second-level micro-arc, and so on). Thus, if the image processing module detects a brightness change of 80 in a certain area, it initially determines that the arc may belong to a second-level arc. Subsequently, this image judgment result is compared with the second-level arc information obtained from the aforementioned electrical signal processing: if they match, the arc is confirmed to be a second-level arc; if they do not match (e.g., the electrical signal determines it to be second-level but the image brightness displays as third-level), a prompt or re-verification message can be issued, thereby achieving dual verification of the arc level information and further improving the accuracy of the level determination.

[0030] Continue back Figure 1 Based on step S120, step S130 is performed, which involves executing real-time arc extinguishing control based on the level information and executing a location-based processing strategy based on the spatial location information.

[0031] In some embodiments, step S130, executing a location-based processing strategy based on spatial location information includes: calculating the arc frequency in each spatial region based on the location information of the arc occurrence; and issuing a regional alarm prompt when the arc frequency in a certain region exceeds a preset threshold to guide physical rectification of the region.

[0032] In the specific implementation process, the arc occurrence location coordinates obtained in step S120 can be used to divide the reaction chamber into different spatial regions, and the arc occurrence frequency in each region within a certain time period (e.g., per unit process time or per unit product quantity) can be statistically analyzed. When the arc frequency in a specific region exceeds a preset threshold, an area alarm prompt is automatically generated. This prompt may include the region identifier, arc frequency statistics, and historical trend information for reference by process engineers or equipment maintenance personnel. Based on this alarm prompt, operators can carry out targeted physical rectification of the region, such as checking for defects in the electrode insulation design, abnormal deposits, or structural aging, thereby eliminating the root cause of frequent arc occurrences in the region. This effectively avoids blind investigation and ineffective rectification, significantly improving the efficiency of arc problem resolution and the long-term stability of equipment operation.

[0033] In some embodiments, step S130, performing real-time arc extinguishing control based on the level information includes: activating the arc detection function and determining whether an arc exists. If the fusion processing is an independent process, the detection here is performed after completing the arc detection and quantity statistics to determine whether an arc exists. If no arc exists, the arc detection is performed cyclically. If an arc exists, the number of arcs at each level is counted as the initial baseline, and the corresponding processing strategy is executed based on the arc level distribution obtained from the detection.

[0034] Specifically, in some embodiments, based on specific threshold settings and the impact of corresponding levels of arcs, micro-arcs can generally be eliminated through active arc extinguishing processes (power parameter reduction) without affecting the normal operation of the process; while soft arcs require a larger power reduction than micro-arcs, and their continued occurrence will affect the process; hard arcs cannot be eliminated through active arc extinguishing processes, and their continued occurrence will damage the equipment or workpiece, requiring shutdown for physical intervention before they can be eliminated.

[0035] Based on this, the above-mentioned processing strategy based on the detected arc level distribution includes: pre-setting multiple processing methods corresponding to arc levels, wherein: For arcs of the first and second levels, they are defined as micro-arcs, and the arc extinguishing process is executed without stopping the output of the radio frequency power supply; for arcs of the third and fourth levels, they are defined as soft arcs, and an alarm signal is issued while the arc extinguishing process is executed; for arcs of the fifth level, they are defined as hard arcs, and an alarm signal is issued while the output of the radio frequency power supply is stopped.

[0036] By classifying electric arcs into three categories—micro-arcs, soft arcs, and hard arcs—and implementing differentiated processing strategies, the system can quickly suppress minor arcs without shutting down the machine, effectively ensuring process continuity and production efficiency. At the same time, it can promptly alarm and cut off the output for severe arcs, avoiding irreversible damage to equipment and products caused by hard arcs, thus balancing process stability and equipment safety.

[0037] As a specific implementation method, in some embodiments, such as Figure 2 As shown, the arc extinguishing process includes the following steps: Step S1: Restart the arc detection function; Step S2: Determine if there is an electric arc; Step S3: If it is determined that there is no electric arc, return to step S1; otherwise, count the number of electric arcs at each level and initialize the number of arc extinguishing times N=0. Step S4: Send an arc extinguishing command to reduce the output power of the RF power supply by a single power reduction value (e.g., reduce the output power by 2.5%), and update N=N+1; Step S5: Enter the preset detection time to perform arc detection and quantity statistics (e.g., the arc detection cycle is 8ms to 10ms) to obtain the current number of arcs of each level and their corresponding spatial location information. Step S6: Determine if an electric arc exists. If no electric arc exists, record the arc extinguishing power as N × the single power reduction value, and return to step S1. Step S7: If an electric arc exists, compare the current electric arc with the electric arc counted in step S3 to determine whether an electric arc of the corresponding level has been eliminated. Step S8: If yes, record the arc extinguishing power required for that level of arc as N × the single power reduction value, and return to step S4. Step S9: If not, determine whether N is less than or equal to the preset maximum allowed number of arc extinguishing times; Step S10: If N is less than or equal to the maximum allowed number of arc extinguishing attempts, then return to step S4; Step S11: If N is greater than the maximum allowed number of arc extinguishing attempts, stop arc extinguishing and issue an alarm signal, reset the number of arc extinguishing attempts N to 0, and return to step S1.

[0038] It should be noted that in step S7 above, the elimination of the electric arc refers to the effective suppression of the electric arc events (initial records) detected and marked with a level (e.g., first-level micro-arc, second-level micro-arc, third-level soft arc, etc.) in the first detection cycle at the start of this arc extinguishing process, after several power reduction operations. If the remaining number has been reduced to below a preset proportion of the initial number (corresponding to the initial record), this preset proportion can be configured according to specific process requirements, for example, set to 10%. That is, when the number of electric arcs of a certain initial level is reduced to less than 10% of the initial number, it can be considered that the electric arc of that level has been basically eliminated and effectively suppressed; or if no new electric arc is detected in the historical occurrence area of ​​the electric arc of that level for at least two consecutive detection cycles, it can also be said that it has been effectively suppressed.

[0039] In this application's technical solution, the initial record is used as the basis for judgment because during the power reduction process, the arc intensity may weaken as the power decreases, causing it to be reclassified as a lower-level arc in subsequent detection cycles, thus rendering it meaningless. Therefore, it is necessary to track the initial situation of each arc event to accurately determine whether an arc of a specific intensity level has truly been suppressed, providing a reliable data foundation for subsequently recording the arc extinguishing power required for that level of arc.

[0040] Furthermore, based on this, in practice, based on the data accumulation during the arc extinguishing process, self-learning optimized arc extinguishing control can also be achieved. That is, according to the recorded arc extinguishing power required for each level of arc, a mapping relationship between different levels of arc and arc extinguishing power is established; based on the mapping relationship, the arc extinguishing parameters are reset so that when the corresponding level of arc is detected later, a one-time power reduction operation is directly executed to extinguish the arc.

[0041] Regarding the timing of resetting the arc extinguishing parameters, in practical applications, it can be flexibly configured according to the equipment operating mode and process requirements. For example, it can be set to automatically update at a preset cycle. For instance, after every 24 hours of continuous operation or after processing 1000 wafers, the system automatically calculates the arc extinguishing power required for each level of arc recorded within that cycle, calculates its average or weighted value, and updates the arc extinguishing parameters, allowing the parameters to be dynamically optimized as the equipment ages or the process drifts. Another example is that parameter updates can be triggered immediately when the number of samples for a certain level of arc reaches a preset threshold (e.g., accumulating 10 valid arc extinguishing data records), setting the arc extinguishing power of that level of arc to the sample average, ensuring that subsequent similar arcs can directly use this parameter to perform a one-time power reduction arc extinguishing. Furthermore, for different process formulations (such as gas type, pressure, power, etc.), a set of arc extinguishing parameter mapping relationships can be maintained independently for each formulation. When the operator switches process formulations, the historical arc extinguishing parameters corresponding to that formulation are automatically loaded; if there is no historical data, the default parameters are temporarily used, and specific data is accumulated for that formulation after executing the active arc extinguishing process. After cleaning the reaction chamber, replacing electrodes, or performing other physical modifications, operators can manually trigger the "relearn" mode via the human-machine interface to clear the chamber's historical arc-extinguishing data and restart the accumulation process, thus establishing an arc-extinguishing parameter mapping relationship adapted to the new state. Through these diverse parameter reset timings, the technical solution of this application can adaptively maintain optimized arc-extinguishing performance, achieving rapid, one-time arc extinguishing for different levels of arcs, further improving process efficiency and equipment reliability.

[0042] Figure 3 This is a block diagram of an arc detection and processing device for a radio frequency plasma processing system provided in one embodiment of this application, as shown below. Figure 3 As shown, the arc detection and processing device 200 for a radio frequency plasma processing system includes: The electrical signal acquisition unit 201 is located inside the matching unit and is used to acquire analog voltage and analog current on the matching unit side. The image acquisition unit 202 is installed in the reaction chamber and is used to acquire image data within the reaction chamber. The fusion processing unit 203 is connected to the electrical signal acquisition unit and the image acquisition unit respectively, and is used to fuse the analog electrical signal and image data to obtain the level information and spatial location information of the electric arc. The control execution unit 204 is connected to the fusion processing unit and is used to execute real-time arc extinguishing control according to the level information and execute position-based processing strategies according to the spatial position information. The image acquisition unit 202, the fusion processing unit 203, and the control execution unit 204 are all electrically connected to the power supply module of the matcher.

[0043] Regarding the arc detection and processing device 200 for the radio frequency plasma processing system in the above-mentioned embodiments, the specific manner in which each module performs its operation has been described in detail in the embodiments of the relevant method, and will not be elaborated here.

[0044] In this embodiment, the arc detection and processing device 200 for the radio frequency plasma processing system, by placing the electrical signal acquisition unit 201 inside the matching unit, can directly acquire the analog electrical signal quantity using the voltage and current acquisition module built into the matching unit itself, without the need for an additional dedicated sampling circuit, thereby significantly reducing hardware costs while ensuring sampling accuracy; by placing the image acquisition unit 202 in the reaction chamber, it can acquire image data inside the chamber in real time, providing raw data for subsequent spatial positioning; the fusion processing unit 203 is connected to the electrical signal acquisition unit 201 and the image acquisition unit 202 respectively, and can simultaneously receive the analog electrical signal quantity and image data and perform fusion processing. The system obtains arc intensity level information through electrical signal analysis and arc spatial location information through image analysis, enabling the perception of arc intensity classification and occurrence location. This avoids the limitation of existing technologies that can only determine the presence or absence of an arc but cannot distinguish its intensity or location. The control execution unit 204 is connected to the fusion processing unit 203, and can perform real-time arc extinguishing control based on the obtained intensity information (e.g., cyclic power reduction arc extinguishing for micro-arcs and soft arcs, and direct shutdown for hard arcs). It also executes location-based processing strategies based on spatial location information (e.g., statistically analyzing the arc frequency in each area and issuing area alarms), realizing a comprehensive solution that combines real-time arc handling with root cause management. Furthermore, the image acquisition unit, fusion processing unit, and control execution unit are all electrically connected to the power supply module of the matching unit. By reusing the existing power supply resources inside the matching unit, there is no need to equip the newly added functional modules with independent power supplies, which further reduces the system complexity and manufacturing cost. At the same time, since the detection point is located on the matching unit side rather than the RF power supply side, the signal transmission path is shortened, reducing the interference of impedance changes during the matching unit tuning process coupled to the detection circuit through the transmission link, effectively improving the reliability and process stability of arc detection.

[0045] like Figure 4 The diagram shown is a schematic block diagram of the actual application of the arc detection and processing device of the present application for the radio frequency plasma processing system in one embodiment.

[0046] like Figure 4As shown, the radio frequency plasma processing system includes a radio frequency power supply, a matching unit, and an ion reaction chamber. The radio frequency power supply serves as the system's radio frequency energy source, and its energy enters the matching unit through radio frequency cables. In the matching unit, the voltage and current acquisition modules transmit the acquired analog electrical signals to the matching unit control module to achieve normal system process control. The arc detection and processing device of this application reuses the voltage and current acquisition modules of the matching unit to realize the function of the electrical signal acquisition unit, and independently sets up an image acquisition unit (corresponding to...) at the ion reaction chamber. Figure 4 The system includes a 3D image acquisition module and an ARC detection and control board within the matcher to function as a fusion processing unit and control execution unit. The ARC detection and control board includes a data storage and processing module, a filtering module, and an algorithm control module. The analog electrical signal data acquired by the voltage and current acquisition modules of the matcher first enters the data storage and processing module, which converts the acquired analog data into 16-bit digital signals. This digital signal then enters the filtering module for level classification. After classification, the digital signal enters the algorithm control module. After processing by the ARC algorithm, the number of arcs at different intensity levels and the corresponding cavity coordinates of each arc level are obtained, enabling real-time arc extinguishing control and position-based processing strategies.

[0047] It should also be noted that, Figure 4 The algorithm control module can be implemented in various hardware forms. For example, a digital signal processor (DSP) combined with a field-programmable gate array (FPGA) can be used, where the FPGA is responsible for real-time acquisition and analog-to-digital conversion of voltage / current analog quantities, while the DSP is responsible for executing the classification algorithm and arc detection logic. Alternatively, a heterogeneous system-on-a-chip (SoC) integrating an ARM core and an FPGA architecture can be used, where the FPGA logic performs real-time processing of electrical signals, and the ARM core runs image processing algorithms and executes control strategies. Those skilled in the art should understand that the above hardware implementation methods are merely illustrative, and any programmable logic device, processor, or dedicated circuit capable of executing the algorithm described in this application falls within the protection scope of this application.

[0048] Figure 5 for Figure 4 The illustrated embodiment is a schematic diagram illustrating the arc extinguishing process. (See diagram for example.) Figure 5 As shown, in this embodiment, the algorithm principle for arc extinguishing is the same as in the previous embodiment, but the specific implementation is different. The process is as follows: After entering the arc extinguishing process (determining if there is an arc), first determine if the number of arc extinguishing attempts N=0. If yes, send an arc extinguishing command. If no, increment the number of arc extinguishing attempts to N=N+1, and determine if N is less than or equal to the preset maximum allowed number of arc extinguishing attempts. In this decision branch, if N is greater than the maximum allowable number of arc extinguishing attempts, an alarm signal is issued, and the number of arc extinguishing attempts N is reset, and arc detection is cyclically performed (voltage and current analog signal acquisition and subsequent steps are repeated); if N is less than or equal to the maximum allowable number of arc extinguishing attempts, it is then determined whether a certain level of arc has been extinguished. If not, an arc extinguishing command is sent, reducing the output power of the RF power supply by a single power reduction value. If so, the power used to extinguish this level of arc is recorded, and arc detection is cyclically performed. Figure 5 (Not shown in the image).

[0049] The arc extinguishing command is issued to reduce the power (that is, to reduce the output power of the radio frequency power supply by a single power reduction value), and then the arc detection is repeated.

[0050] In the arc detection process, it is determined whether there is an arc. If no arc is detected, the current power reduction value = N × single power reduction value is recorded, and the number of arc extinguishing times N is reset to 0, and the arc detection cycle is repeated. If there is an arc, the arc extinguishing process is entered, and the cycle is repeated.

[0051] Based on the above embodiments, the technical solution of this application can achieve the following technical effects: In arc detection, by performing time differentiation on analog voltage and current quantities and comparing them with a preset zero threshold, interference from normal process fluctuations such as matcher tuning and gas ignition is effectively filtered out. Furthermore, by utilizing the physical characteristic that the voltage and current change rates have opposite polarities during arc occurrence for mutual verification, high-precision identification of real arc events is achieved, fundamentally solving the misjudgment problem caused by the inability to distinguish between tuning and arcs in existing technologies. After confirming the occurrence of an arc, the absolute value of the change rate data is compared step-by-step with at least five level thresholds to obtain the arc intensity level and the number of occurrences at each level, achieving quantitative differentiation of arc strength and providing a refined data foundation for differentiated arc extinguishing control.

[0052] In terms of arc positioning, by pre-collecting spatial photographs of the reaction chamber at different focal lengths and establishing a pixel-space mapping relationship, and utilizing the difference in arc pixel positions in near-focus and far-focus images provided by at least two image acquisition modules, combined with the parallax principle, the actual spatial coordinates of the arc are calculated, thus achieving precise three-dimensional positioning of the arc occurrence location. This overcomes the limitation of existing technologies that can only determine the presence or absence of an arc but cannot locate it.

[0053] In terms of arc extinguishing control, a tiered approach is implemented based on arc level information: for arcs of levels one and two (defined as micro-arcs) and levels three and four (defined as soft arcs), an active arc extinguishing process is executed, achieving precise suppression through a combination of cyclic power reduction and real-time detection; for level five arcs (defined as hard arcs), an alarm is triggered and output is stopped directly to prevent damage to equipment and products. Based on the power required for extinguishing each arc level recorded during the active arc extinguishing process, a level-power mapping relationship is established and arc extinguishing parameters are optimized, enabling subsequent arcs of the same type to be extinguished quickly in one go, giving the system self-learning and optimization capabilities.

[0054] In addition, the frequency of electric arcs in each area can be counted based on the spatial location information of the electric arcs. When the frequency in a certain area exceeds the threshold, an alarm is issued to guide targeted physical rectification. This combines immediate treatment with root cause management, giving the process improvement a clear and directional basis.

[0055] In terms of hardware architecture, the electrical signal acquisition unit reuses existing modules inside the matcher. The image acquisition unit, fusion processing unit, and control execution unit are all electrically connected to the matcher power supply module, realizing resource reuse and significantly reducing system cost and complexity. At the same time, the detection point is located on the matcher side, which shortens the signal transmission path, reduces the interference of the tuning process on the detection loop, and further improves detection reliability and process stability.

[0056] The above description is merely a preferred embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0057] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0058] Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of this application.

Claims

1. An arc detection and processing method for a radio frequency plasma processing system, characterized in that, include: The analog voltage and analog current on the matching side are collected as analog electrical signals, and image data inside the reaction chamber is acquired. The analog electrical signal and the image data are fused together to obtain the arc's level information and spatial location information. Real-time arc extinguishing control is performed based on the level information, and a location-based processing strategy is executed based on the spatial location information.

2. The arc detection and processing method according to claim 1, wherein, The process of obtaining the arc level information includes: The analog voltage and analog current are converted into digital signals and then subjected to time differentiation processing to obtain voltage change rate data and current change rate data. The voltage change rate data is compared with a preset zero voltage threshold, and the current change rate data is compared with a preset zero current threshold. The zero voltage threshold and the zero current threshold are set according to the voltage change rate amplitude and the current change rate amplitude caused by the matching process, respectively. Based on the comparison results of the voltage change rate data and the zero voltage threshold, and the comparison results of the current change rate data and the zero current threshold, it is determined whether an electric arc has occurred. When an electric arc is detected, the absolute value of the voltage change rate data is compared with a preset first voltage threshold to an Nth voltage threshold, and the absolute value of the current change rate data is simultaneously compared with a preset first current threshold to an Nth current threshold to obtain a first-level judgment result corresponding to the voltage change rate and a second-level judgment result corresponding to the current change rate. The first voltage threshold to an Nth voltage threshold correspond to different electric arc intensity levels, and the first current threshold to an Nth current threshold correspond to different electric arc intensity levels, where N≥5; The intensity level of the current electric arc is determined based on the first and second level judgment results, and the number of times the electric arc of that level occurs in the current detection cycle is recorded.

3. The arc detection and processing method according to claim 2, wherein, The determination of whether an electric arc has occurred includes: The voltage change rate data and current change rate data at the same moment are cross-checked. When the voltage change rate data is greater than the zero voltage threshold, the current change rate data is greater than the zero current threshold, and the polarities of the voltage change rate and current change rate are opposite, an electric arc is confirmed to have occurred and is recorded.

4. The arc detection and processing method according to claim 2, wherein, The process of obtaining the spatial location information of the electric arc includes: Pre-acquire images of the reaction chamber at different focal lengths to establish a mapping relationship between pixel positions and chamber spatial positions; Image data within the reaction chamber is acquired in real time using at least two image acquisition modules. Each image acquisition module is responsible for covering different areas within the chamber and provides near-focal and far-focal images of the corresponding areas. The image data is processed by comparing the current frame image with the previous frame image, extracting the brightness change area, and obtaining the pixel position of the electric arc in the image. Based on the pixel position differences of the electric arc in the near-focal distance and far-focal distance images provided by the image acquisition module corresponding to the area where the electric arc is located, and combined with the mapping relationship, the actual spatial coordinates of the electric arc in the reaction chamber are calculated as the location information of the electric arc occurrence.

5. The arc detection and processing method according to claim 4, wherein, Processing the image data further includes: The intensity information of the electric arc is determined based on the amount of brightness change in the brightness change area. The intensity information is proportional to the amount of brightness change and is used to assist in verifying the level information of the electric arc.

6. The arc detection and processing method according to claim 1, wherein, The step of performing real-time arc extinguishing control based on the level information includes: Activate the arc detection function and determine if an arc exists; If no electric arc is found, the electric arc detection is performed repeatedly. If an electric arc is present, the number of electric arcs at each level is counted as an initial baseline, and the corresponding processing strategy is executed based on the distribution of electric arc levels obtained from the detection.

7. The arc detection and processing method according to claim 6, wherein, The step of implementing corresponding processing strategies based on the detected arc level distribution includes: Multiple preset processing methods are provided for different arc levels, including: For the first and second level arcs, they are defined as micro-arcs, and the arc extinguishing process is performed without stopping the output of the radio frequency power supply; For arcs of the third and fourth levels, they are defined as soft arcs, and an alarm signal is issued and an arc extinguishing procedure is executed. For the fifth level of arc, defined as a hard arc, an alarm signal is issued and the output of the radio frequency power supply is stopped.

8. The arc detection and processing method according to claim 7, wherein, The arc extinguishing process includes the following steps: Step S1: Restart the arc detection function; Step S2: Determine if there is an electric arc; Step S3: If it is determined that there is no electric arc, return to step S1; otherwise, count the number of electric arcs at each level and initialize the number of arc extinguishing times N=0. Step S4: Send an arc extinguishing command to reduce the output power of the RF power supply by a single power reduction value and update N=N+1; Step S5: Enter the preset detection time to perform arc detection and quantity statistics, and obtain the current number of arcs of each level and their corresponding spatial location information; Step S6: Determine if an electric arc exists. If no electric arc exists, record the arc extinguishing power as N × the single power reduction value, and return to step S1. Step S7: If an electric arc exists, compare the current electric arc with the electric arc counted in step S3 to determine whether an electric arc of the corresponding level has been eliminated. Step S8: If yes, record the arc extinguishing power required for that level of arc as N × the single power reduction value, and return to step S4. Step S9: If not, determine whether N is less than or equal to the preset maximum allowed number of arc extinguishing times; Step S10: If N is less than or equal to the maximum allowed number of arc extinguishing attempts, then return to step S4; Step S11: If N is greater than the maximum allowed number of arc extinguishing attempts, stop arc extinguishing and issue an alarm signal, reset the number of arc extinguishing attempts N to 0, and return to step S1.

9. The arc detection and processing method according to claim 1, wherein, The execution of the location-based processing strategy based on the spatial location information includes: Based on the location information of the electric arc, the frequency of the electric arc in each spatial region is counted; When the frequency of electric arcs in a certain area exceeds a preset threshold, an area alarm is issued to guide the physical rectification of that area.

10. An arc detection and processing device for a radio frequency plasma processing system, used to implement the method according to any one of claims 1 to 9, characterized in that, include: An electrical signal acquisition unit, located inside the matching unit, is used to acquire analog voltage and analog current signals on the matching unit side. An image acquisition unit is installed in the reaction chamber and is used to acquire image data within the reaction chamber; The fusion processing unit is connected to the electrical signal acquisition unit and the image acquisition unit respectively, and is used to fuse the analog electrical signal and the image data to obtain the level information and spatial location information of the electric arc. A control execution unit, connected to the fusion processing unit, is used to perform real-time arc extinguishing control according to the level information and to perform a location-based processing strategy according to the spatial location information. The image acquisition unit, the fusion processing unit, and the control execution unit are all electrically connected to the power supply module of the matcher.

Images