Remote plasma source air outlet water pipe fault diagnosis method

Abstract

The invention discloses an air outlet water pipe fault diagnosis method of a remote plasma source in the field of remote plasma sources, which is characterized by comprising the following steps of: 1, monitoring flow resistance pressure drop delta P in real time through a pressure difference sensor, and acquiring local temperature distribution data delta T through an infrared thermal imager, meanwhile, the environment humidity change delta R is monitored through a capacitive humidity sensor; 2, fusing the data of delta P, delta T and delta R based on a Kalman filtering algorithm to obtain state estimation values of pressure drop, temperature and humidity; 3, constructing a scaling / blocking objective function M and a leakage objective function N; fourthly, when M exceeds a first threshold value M1, scaling is judged, and when M exceeds a second threshold value M2, blockage is judged; and when N exceeds the leakage threshold N1, micro leakage is judged. According to the gas outlet pipeline fault diagnosis method, accurate identification of scaling, blockage and leakage is achieved through multi-sensor cooperative monitoring (differential pressure, temperature and humidity) and Kalman filtering data fusion.

Description

Technical Field

[0001] The present invention relates to the technical field of remote plasma sources, and particularly to a method for diagnosing faults in the water pipe at the gas outlet of a remote plasma source. Background Art

[0002] A remote plasma source (RPS) is an advanced plasma generation device, also known as a remote high-density plasma generator, and is commonly used for the cleaning and etching processes in the process chambers during integrated circuit manufacturing. Its core function is to efficiently and stably ionize the gas used for cleaning (NF3) or the gas process gas (NH3, O2), thereby greatly enhancing the activity of the reaction gas and the consistency of the process. Compared with traditional plasma sources, there is a physical separation between the plasma generation region and the processing region in a remote plasma source. After the plasma is generated, it is transmitted to the processing region, and the active particles (such as free radicals, ions, and neutral particles) diffuse during the transmission process. This design enables the plasma to act on the surface of the material to be processed more uniformly and efficiently.

[0003] The dissociation rate is one of the important parameters characterizing the reaction efficiency of a remote plasma source. It refers to the proportion of the reaction gas dissociated into active species (such as free radicals, single-core ions, multi-core ions, and neutral atoms) in a strong electric field environment. A higher and more stable dissociation rate means that at the same input power, a higher plasma density can be obtained, thereby reducing the cleaning time of the process chamber, improving the processing efficiency and quality. Therefore, the measurement of the dissociation rate and its stability is crucial. As the integrated circuit process advances to nodes below 3nm, the control accuracy requirements for the dissociation rate of the reaction gas in the plasma process become increasingly stringent. Therefore, a complete set of devices is needed to achieve reliable control of gas flow and pressure, so as to complete the performance test of the RPS dissociation rate.

[0004] There are certain defects in the traditional design of the cooling water pipe at the gas outlet of the remote plasma device. The traditional design lacks an intelligent monitoring module and cannot sense fouling, blockage, or micro-leakage in real time. Fouling and blockage can reduce the heat dissipation efficiency by 60%, and water micro-leakage can cause production safety accidents; the traditional design of the cooling water pipe often has difficulty adapting to this change and cannot meet the heat dissipation requirements of higher-level intelligent monitoring. Summary of the Invention

[0005] The present application provides a method for diagnosing faults in the water pipe at the gas outlet of a remote plasma source, which solves the problem of the lack of fault monitoring for the cooling water pipe at the gas outlet in the prior art, thereby realizing the accurate identification of fouling, blockage, and leakage faults.

[0006] The embodiments of the present application provide a method for diagnosing faults in the water pipe at the gas outlet of a remote plasma source, including the following steps: Step 1: The differential pressure sensor installed at the inlet and outlet of the outlet water pipe monitors the flow resistance pressure drop ΔP in real time, and the infrared thermal imager outside the pipe wall of the outlet water pipe collects the local temperature distribution data ΔT. At the same time, a capacitive humidity sensor is arranged near the outlet water pipe to monitor the environmental humidity change ΔR; Step 2: Based on the Kalman filtering algorithm, the data of ΔP, ΔT, and ΔR are fused to obtain the state estimation values of pressure drop, temperature, and humidity, ΔP(k), T(k), and R(k); Step 3: Construct the fouling / clogging objective function M and the leakage objective function N, ; ; where t(k) is the current moment, t(k - 1) is the previous moment, ΔP(k), v(k), T(k), and R(k) are the state estimation values of differential pressure, flow rate, temperature, and humidity at the current moment, and ΔP(k - 1), v(k - 1), T(k - 1), and R(k - 1) are the state estimation values of differential pressure, flow rate, temperature, and humidity at the previous moment. λ 1, λ 2, λ 3 are the weight ratios of differential pressure, flow rate, and temperature respectively, and λ4 is the weight ratio of humidity, where λ4 > λ1, λ2; Step 4: When M exceeds the first threshold M1, it is determined as fouling, and when it exceeds the second threshold M2, it is determined as clogging; when N exceeds the leakage threshold N1, it is determined as micro-leakage, and a hierarchical alarm and maintenance instruction are triggered.

[0007] The beneficial effects of the above embodiments are as follows: This fault diagnosis method for the outlet pipeline realizes accurate identification of fouling, clogging, and leakage (accuracy rate ≥ 95%) through multi-sensor collaborative monitoring (differential pressure, temperature, humidity) and Kalman filtering data fusion, and the false alarm rate is reduced to less than 5%; the fault risk is dynamically quantified based on the objective function, and the hierarchical alarm mechanism improves the maintenance response efficiency by 60%, avoiding false alarms or misjudgments caused by traditional single thresholds.

[0008] Based on the above embodiments, the present application can be further improved as follows: In one embodiment of the present application, the accuracy of the differential pressure sensor is ±0.1% FS. When ΔP continuously rises to the first differential pressure threshold P1, it is determined that the flow resistance is abnormal. Through high-precision differential pressure monitoring, early warning of fouling or clogging is achieved, avoiding a decrease in heat dissipation efficiency.

[0009] In one embodiment of the present application, when the infrared thermal imager detects that the local temperature ΔT > 10 °C, it is determined as the fouling area, and the high-temperature point coordinate information is generated. The heat conduction abnormal area caused by water scale is quickly located, improving the maintenance efficiency.

[0010] In one embodiment of the present application, non-invasive flow velocity measurement is combined, and the water flow rate v is calculated through the time difference of acoustic wave propagation. When the water flow velocity drops abnormally to the flow rate threshold v1, it is determined that the flow rate is abnormal. When an alarm command appears, stop the machine for inspection in time to avoid further losses.

[0011] In one embodiment of the present application, the capacitive humidity sensor triggers a leakage alarm when the ambient humidity suddenly rises to the humidity threshold R 1, and double verification is carried out in combination with the differential pressure decay rate. Leakage is judged jointly by humidity and pressure drop, and the false alarm rate is reduced to less than 3%, preventing safety accidents caused by coolant leakage.

[0012] In one embodiment of the present application, the system state equation of the Kalman filter algorithm is x(k)=Ax(k - 1)+Bu(k)+w(k), and the observation equation is y(k)=Hx(k)+v(k). The Kalman gain K(k) is iteratively updated to optimize the state estimate value. Dynamically correct the deviation of sensor data (such as temperature drift), reduce the estimation errors of pressure drop, temperature and humidity, and improve the reliability of fault diagnosis.

[0013] In one embodiment of the present application, the weight coefficients λ1~λ4 in the objective function are obtained by training based on historical fault data and are dynamically adjusted according to the service life of the pipeline. Adaptive optimization of the fault determination threshold, adapting to pipeline aging or working condition changes, makes the diagnostic accuracy stable in the long term. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for the description of the specific embodiments or the prior art. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn to actual scale.

[0015] Figure 1 It is a front view structural schematic diagram of a dissociation rate test device for a remote plasma source in an embodiment of the present application; Figure 2 It is a three-dimensional structural schematic diagram of a dissociation rate test device for a remote plasma source in an embodiment of the present application; Figure 3 It is a schematic diagram of the relationship between components in the process of obtaining the dissociation rate in an embodiment of the present application; Figure 4 It is a schematic diagram of the function diagram of the RPS control test bench display screen in an embodiment of the present application; Figure 5 It is a method flow chart for obtaining the dissociation rate in an embodiment of the present application; Figure 6It is a schematic structural diagram of the tail gas treatment equipment in the embodiment of the present application; Figure 7 It is a connection diagram of the control module of the tail gas treatment equipment in the embodiment of the present application; Figure 8 It is a working flow chart of the tail gas treatment device in the embodiment of the present application; Figure 9 It is a schematic diagram of the traditional structure of the water-cooled pipe at the gas outlet, where 9(a) is the internal structure diagram and 9(b) is the external structure diagram; Figure 10 It is a schematic diagram of the structure of the water-cooled pipe at the gas outlet in the embodiment of the present application, where 10(a) is the internal structure diagram and 10(b) is the external structure diagram; Figure 11 It is a step flow chart of a method for diagnosing faults in the water outlet pipe of a remote plasma source in the embodiment of the present application; Figure 12 It is a flow chart for evaluating the importance degree relationship between scaling, blockage, micro-leakage and process parameters; Figure 13 It is a flow chart for constructing a decision tree for evaluating the importance degree relationship; Figure 14 It is a schematic diagram of the score ranking of the importance degree of scaling, blockage, micro-leakage and process parameters.

[0016] Among them, 100 - remote plasma source; 101 - air inlet; 102 - equipment status display screen; 103 - water-cooled pipe; 1031 - cooling water channel; 104 - etching test bench; 105 - etching observation window; 106 - RPS inlet and outlet; 107 - gas pressure gauge; 108 - sampling gas pressure gauge; 200 - gas ionization detection device; 201 - Fourier transform infrared spectrometer; 202 - ionization gas sample collection pipeline; 203 - spectrometer tail gas output pipeline; 300 - RPS control test bench; 301 - test bench display screen; 302 - test bench status operation indicator light; 400 - host computer system; 500 - vacuum pump; 501 - tail gas transmission pipeline; 600 - tail gas treatment equipment; 601 - tail gas treatment equipment air inlet; 602 - tail gas treatment equipment air outlet. Specific embodiments

[0017] The following further clarifies the present invention in conjunction with specific embodiments. It should be understood that these embodiments are only used to illustrate the present invention and not to limit the scope of the present invention. After reading the present invention, various equivalent modifications made by those skilled in the art to the present invention all fall within the scope defined by the appended claims of the present application.

[0018] It should be noted that similar reference numerals and letters denote similar items in the following figures. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. In the description of the present invention, it should also be noted that unless otherwise clearly specified and limited, the terms "arranged", "installed", "connected", and "coupled" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium, and it can be the communication inside two components. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0019] By providing a dissociation rate test device for a remote plasma source in Embodiment 1 of the present application, the problem of inconvenient testing of the dissociation rate of RPS in the prior art is solved, so as to conveniently and accurately test the dissociation rate of RPS.

[0020] By providing an electrothermal adsorption type tail gas treatment device for RPS dissociation rate test in Embodiment 2 of the present application, the problem of inconvenient tail gas treatment after the RPS dissociation rate test in the prior art is solved, so as to efficiently treat the remaining gas after the RPS dissociation rate test.

[0021] By providing an outlet water pipe structure for heat dissipation of a remote plasma source in Embodiment 3 of the present application, the problems that the uneven temperature of the outlet water pipe for heat dissipation in the prior art affects the accuracy of the dissociation rate test and the heat dissipation efficiency of the water pipe for heat dissipation is low are solved, so as to improve the temperature balance and heat dissipation efficiency of the water pipe for heat dissipation.

[0022] By providing a method for diagnosing faults in the outlet water pipe of a remote plasma source in Embodiment 4 of the present application, the problem of lack of fault monitoring for the outlet water pipe for heat dissipation in the prior art is solved, so as to accurately identify the faults of scaling, blockage, and leakage.

[0023] Embodiment 1: As Figures 1-3 shown, a dissociation rate test device for a remote plasma source includes a remote plasma source 100, a gas ionization detection device 200, a Fourier transform infrared spectrometer 201, an RPS control test bench 300, a host computer system 400, and a tail gas treatment device 600 ( Figure 2 omitted in The remote plasma source 100 is used to dissociate process gases and perform processes such as thin film deposition and cleaning under vacuum conditions. The remote plasma source 100 includes a dissociation chamber, an inlet 101, an outlet, and an RPS water inlet / outlet 106. The outlet is connected to the etching stage 104 through an outlet water cooling pipe 103. The etching stage 104 is provided with an etching observation window 105 and a gas pressure gauge 107. The RPS water inlet / outlet 106 is controlled by the RPS control stage 300 to control the water flow rate of the cooling water. The etching stage 104 is used to perform thin film etching on materials such as silicon wafers and wafers. The gas pressure gauge 107 is used to monitor the internal pressure of the etching stage 104 and is regulated by the RPS stage. The remote plasma source 100 is also provided with an equipment status display screen 102, which is used to observe and set the operating status, such as information on power, voltage, and current.

[0024] The gas ionization detection device 200 includes a gas dissociation circuit control device (maintenance gas input control module) for maintaining gas input, which is connected to the dissociation chamber of the remote plasma source 100 and is used to control the amount of maintenance gas input into the dissociation chamber. The gas ionization detection device 200 is also provided with a sampling gas pressure gauge 108 to ensure an ultra-low pressure test environment of 100 mTorr and guarantee the accuracy of the remaining gas amount collection. The gas ionization detection device 200 also includes an output detection unit, which is connected to the dissociation chamber through an ionized gas sample collection pipeline 202 and is used to detect the amount of remaining maintenance gas in the gas of the dissociation chamber. The gas ionization detection device 200 also includes an output calculation unit, which is connected to the maintenance gas input gas dissociation circuit control device and the output detection unit and is used to calculate the ionization rate based on the amount of maintenance gas input into the dissociation chamber and the amount of remaining maintenance gas in the gas output from the dissociation chamber.

[0025] The gas ionization detection device 200 respectively detects the amount M of the remaining gas in the gas of the dissociation chamber and the amount N of the input maintenance gas, then the dissociation rate is 1 - M / N. During the remote plasma process, when nitrogen trifluoride NF3 is used as a process gas and is ionized by plasma, it decomposes into nitrogen and fluorine atoms, and the fluorine atoms perform thin film etching on the silicon wafer. When the amount M of the remaining NF3 gas is less than the residual gas threshold, it means that the NF3 waste gas is sufficiently small, and NF3 is toxic. At this time, the etching stage 104 can be opened to take out the silicon wafer, and the heating furnace of the tail gas treatment device also stops working at this time to increase the power application efficiency. When the amount M of the remaining NF3 gas is greater than or equal to the residual gas threshold, it means that the ionization process of the process gas has not ended, and the silicon wafer is continuously etched, and the heating furnace of the tail gas treatment device also needs to work continuously.

[0026] The Fourier transform infrared spectrometer 201 is used to collect the infrared spectral data of the remaining maintenance gas. The spectrometer accesses the spectral data of the remaining maintenance gas in the sampling dissociation chamber of the etching test bench 104 through the ionized gas sample collection pipeline 202; it is connected to the exhaust gas transmission pipeline through the spectrometer exhaust gas output pipeline 203 and is output to the exhaust gas treatment device by the vacuum pump 500.

[0027] The RPS control test bench 300 is equipped with a gas flow meter, a pressure gauge, a water flow meter, a test bench display screen 301, and a test bench status operation indicator light 302. The gas flow meter, pressure gauge, and water flow meter are used to detect the intake gas flow rate, the chamber pressure, and the cooling water flow rate of the remote plasma source 100. The RPS control test bench 300 is used to control the process gas flow rate, vacuum pressure, water cooling system, etc. of the remote plasma source 100, including the display of parameters such as the operating state, power, voltage, and current of the equipment, as well as the recording of fault information and the modification of process parameters. Among them, the gas flow meter controls the gas flow rate entering the RPS and cooperates with the vacuum pump 500 to achieve the output of different powers of the RPS, simulating the actual operating conditions.

[0028] Among them, the function diagram on the RPS test bench display screen 301 is as Figure 4 shown. The test bench control module includes a process gas module, a vacuum pressure module, a water cooling module, a parameter display module, a fault information recording module, and a process parameter module; the process gas module can control and real-time read the flow rate of the input gas, and cooperate with the vacuum pump 500 to achieve the output of different powers of the RPS, simulating the actual operating conditions; the vacuum pressure module can control and real-time read the pressure in the RPS cavity and the Fourier transform infrared spectrometer 201 by setting the pressure value; the water cooling module provides water cooling for the plasma source, can control the inlet and outlet water flow rates and real-time read them; The parameter display module can read the actual power value, bus bar current and voltage in the RPS, as well as the operating state of the equipment; first, the power switch is started, after introducing the process gas and setting the power value, it is determined whether the "Initialization State (Ready)" and "Power Input (AC OK)" signal lamp indicators are always on. "Initialization State (Ready)" represents that the equipment initialization state is normal; "Power Input (AC OK)" indicates that the power input is within the allowable range; then press the ignition switch. "Ignition State (Plasma ON)" indicates that the equipment ignition is successful or the equipment is processing the process gas, otherwise it means that the equipment is not ignited or the ignition fails during operation; "Equipment Alarm (Fault)" always on indicates that the equipment has an alarm, otherwise there is no alarm for the equipment; The fault information recording module can be clicked to view the faults sent by the device, including the time when the fault occurred, the type of fault that occurred, and the number of times the current fault type has occurred; the fault types include: over-temperature alarm (OT), over-power alarm (OP), abnormal input voltage (ACV), output current drop (LC), output current exceeding the limit (OC), water leakage alarm (Water Leak); the process parameter module can select the ignition process or the burn-in aging process to test the RPS. The detection device is also designed with a real-time monitoring function, which can immediately feedback the dissociation state of the plasma. Through real-time monitoring, the working conditions of the plasma source can be adjusted in time to ensure the stability and consistency of the production process.

[0029] The host computer system 400 is used to perform dissociation rate detection step operations such as preprocessing, feature extraction, and pattern recognition on the Fourier transform infrared spectroscopy data.

[0030] The exhaust gas treatment device 600 is a system for treating various process exhaust gases on site, provided with an exhaust gas treatment device inlet 601 and an exhaust gas treatment device inlet 602. The exhaust gas treatment device inlet 601 is connected to the outlet of the vacuum pump 500 through a pipeline. The inlet of the vacuum pump 500 is connected to the dissociation chamber through an exhaust gas transmission pipeline 501. The vacuum pump 500 ensures a low-pressure environment in the reaction chamber. At low pressure, the collisions between gas molecules are reduced, enabling more effective excitation of gas molecules during the plasma treatment process; it also promotes the stable generation and maintenance of the plasma, and the exhaust gas pollutants are introduced into the exhaust gas treatment device 600.

[0031] Among them, as Figure 5 shown, in this embodiment, the dissociation rate detection method adopts a traditional method, such as S1. The Fourier transform infrared spectrometer 201 collects the near-infrared spectra of different dissociation rates of the remote plasma source 100; the gas ionization detection device 200 respectively detects the amount M of the remaining gas and the amount N of the input maintenance gas in the gas of the dissociation chamber, and the dissociation rate is 1 - M / N. S2. Establish an ionization rate data set of the remote plasma source 100 using the near-infrared spectrum data of different dissociation rates. S3. Adopt the same near-infrared spectrum collection conditions, sample the near-infrared spectrum of the plasma source to be detected, compare the measured characteristic spectral peak values with the ionization data set, and judge the current concentration of the remaining gas (NF3), then the dissociation rate is 1 - the concentration of the remaining gas (NF3).

[0032] This device uses infrared mass spectrometry analysis and detection technology, which can achieve high-precision and high-sensitivity measurement of various active particles in the plasma, thereby accurately calculating the dissociation rate and providing reliable data for the performance evaluation of the plasma source. By applying near-infrared spectral data with different dissociation rates at different pressures, flow rates, and temperatures, an ionization rate data set is established. By identifying the characteristic spectral peaks based on the detected amount of residual NF3 gas and comparing them with the ionization rate data set, the gas dissociation rate at different pressures, flow rates, and temperatures can be detected in real time.

[0033] This test device integrates an automated control system, an integrated infrared mass spectrometry analysis, and an RPS tail gas treatment system. By coordinating the work of each functional module through the automated control system, it can automatically complete the detection of relevant technical indicators of the RPS. This device realizes reliable testing of the RPS dissociation rate and also simulates the actual operating conditions of the RPS, eliminating the need to detect the RPS on the actual integrated circuit production terminal equipment, with low cost and short cycle. This test device realizes the full-process monitoring and closed-loop control of the operating state of the plasma source, improving the detection accuracy of the dissociation rate and process stability.

[0034] Example 2: As Figures 6-7 shown, an electrothermal adsorption type tail gas treatment device for a remote plasma source dissociation rate test device includes a control system, a gas content detection system, and a heating system, a steam injection system, a cooling system, and an adsorption system (neutralization tower) that are connected in sequence. A particle collection tank is also provided at the outlet of the steam injection system to collect solid particles (such as CaF2 precipitate) generated during the reaction, avoiding pipeline blockage and extending the service life of the equipment. A heat exchanger is provided between the outlet of the steam injection system and the inlet of the heating system to recover the waste heat of the tail gas to preheat the inlet gas, raising the inlet temperature to 300 °C through waste heat recovery and reducing the energy consumption of the heating system. The inlet of the heating system is connected to the vacuum pump of the remote plasma source dissociation rate test device.

[0035] Among them, the heating system uses a three-stage coaxial nested nickel-based alloy reaction tube (Inconel 600), and the length ratio of each section is 1:2:1 (corresponding to the preheating section, the main decomposition section, and the deep cracking section), and the temperature gradient is 500 °C (preheating section) → 800 °C (main decomposition section) → 1000 °C (deep cracking section). The inner wall is sprayed with a 200 μm thick Al2O3 ceramic layer (purity ≥ 99.5%) to withstand fluorine corrosion. High-frequency induction coils (copper tube diameter 8 mm, turn spacing 5 mm) are wound around the outer wall of each section, and the output power (0~20 kW) is controlled by an independent variable frequency power supply (frequency 50 - 100 kHz) to achieve precise adjustment of the temperature gradient.

[0036] Adopt high-frequency induction heating technology: Utilize the electromagnetic eddy current effect to directly heat the reaction tube wall, with a heating rate of 50 °C / s, saving 30% energy compared with traditional resistance heating.

[0037] The steam injection system is a steam injection ring (made of 316L stainless steel with a pore diameter of 0.5 mm) installed at the outlet of the deep cracking section. The injection amount of ultrapure steam is controlled by a mass flow meter (MFC) and dynamically matched with the NF3 molar ratio (range 0.8:1 - 1.2:1). Dynamically matching the molar ratio of steam to NF3 (0.8:1 - 1.2:1) ensures the full conversion of F2 into HF and reduces the risk of secondary pollution by by-products. The decomposition of NF3 requires high-temperature cracking or catalytic hydrolysis reactions, and its chemical equation is usually: 2NF3 + 3H2O → 6HF + NO + NO2; steam acts as a hydrolysis medium in this reaction, providing hydroxyl -OH to accelerate the breaking of the N-F bond. If the steam is insufficient, i.e., the molar ratio is lower than 0.8:1, NF3 may not be completely decomposed, and the remaining NF3 will be emitted in the form of toxic gas; if the steam is excessive, higher than 1.2:1, it may dilute the reaction system, reduce the reaction efficiency, and even inhibit the decomposition reaction due to a local temperature drop.

[0038] The cooling system is used to cool the mixed gas, and the cooling is determined by the pipe length and water flow rate, generally cooling to 500 ± 50 °C.

[0039] The adsorption system is a neutralization absorption tower, designed with two - stage series connection. The first - stage tower is filled with Ca(OH)2 particles (particle size 3 - 5 mm) for preliminary adsorption of HF; the second - stage tower is equipped with a circulating spray system with pH value feedback (Ca(OH)2 solution concentration 10 wt%), and the spray rate is 5 L / min. By circulating the spray of the Ca(OH)2 solution, HF is fixed as CaF2 precipitate, achieving a tail - gas fluorine content < 1 ppm.

[0040] The gas content detection system includes several NF3 detectors + electrochemical HF sensors, and the control system controls the above - mentioned systems according to the set program. By real - time monitoring the NF3 and HF concentrations, the heating power and steam flow rate are dynamically adjusted to ensure that the tail - gas treatment meets the standards (NF3 decomposition rate > 99%, HF < 1 ppm) and adapts to high - concentration shock - load working conditions.

[0041] The tail - gas treatment equipment in this Example 1 uses the above - mentioned electrothermal adsorption type tail - gas treatment equipment, including the processes of thermal oxidative decomposition and reagent adsorption. The waste gas generated by the process is decomposed at high temperature through the heating system and then discharged to the plant utility treatment system through dry - type adsorption treatment. A high - temperature system (400 - 1200 °C) generated by a heating furnace is used. It decomposes and oxidizes the NF3 gas that was not dissociated by the previous - stage equipment at high temperature, treating the toxic gas into non - harmful reactants. After cooling by the cooling system, an activated chemical adsorption system is used for activated chemical adsorption or reaction to generate deposited substances, achieving the complete absorption of toxic and corrosive gases. The harmless gas after treatment will then enter the plant exhaust system.

[0042] As shown Figure 8 in the figure, this embodiment also provides a working flowchart of an electrothermal adsorption type tail gas treatment device, and the main process is as follows: S1. Heating system: A nickel-based alloy reaction tube with three-stage independent temperature control (lined with ALO, and a ceramic coating to prevent oxidation of fluorine atoms F and reduce the service life) is adopted, and the temperature gradient is 500 °C (preheating section) → 800 °C (main decomposition section) → 1000 °C (deep cracking section), ensuring that NF is gradually decomposed into NF3 and F2 to improve the decomposition efficiency of NF3.

[0043] Among them, NF3 detectors are equipped for all three-stage independent temperature controls. In the preheating stage, the concentration of NF3 in the input gas is detected as n vol%, and the flow rate is l m 3 / h. If the concentration n is less than the first concentration threshold N1 and the flow rate l is less than the flow rate threshold L, the heating furnace temperature in the preheating section is set to 500 °C (heating rate 30 °C / s, power 8 kW). (2) Main decomposition section: 800 °C (constant temperature accuracy ±5 °C, power 15 kW). When the NF3 concentration is less than the second concentration threshold N2, it enters the deep cracking stage; otherwise, continue heating. (3) Deep cracking section: 1000 °C (power 18 kW, residence time ≥2 s). When the NF3 concentration is less than the third concentration threshold N3, it enters the steam injection system; When the concentration of NF3 in the input gas is detected as n vol% and the flow rate is l m 3 / h in the preheating stage, if the concentration n is greater than or equal to the first concentration threshold N1, that is, the system automatically adjusts when the concentration increase is detected within 0.5 s by the heating system, and raises the temperature of the main decomposition section to 850 °C and increases the power to 20 kW. And when the NF3 gas flow rate l is greater than or equal to the flow rate threshold L, the inlet gas valve is used to reduce the inlet gas flow rate by 15% and extend the residence time in the preheating stage; when the NF3 concentration is less than the second concentration threshold N2, it enters the deep cracking stage; otherwise, continue heating. Deep cracking section: 1000 °C (power 18 kW, residence time ≥2 s). When the NF3 concentration is less than the third concentration threshold N3, it enters the steam injection system; and synchronously increases the steam flow rate to k m 3 / h (molar ratio 1.05:1).

[0044] S2. In-situ neutralization system: When the concentration of NF3 in the input gas is detected as n vol% and the flow rate is l km 3 / h in the preheating stage, if the concentration n is greater than or equal to the first concentration threshold N1 or when the NF3 gas flow rate 1 is greater than or equal to the flow rate threshold L, the steam injection system needs to synchronously increase the steam flow rate by k m 3 / h.

[0045] Steam injection system: Ultra-pure steam (flow rate to NF3 molar ratio 1:1) is injected at the end of the reactor to convert F2 into HF gas. Reaction formula: 2F2 + 2H2O → 4HF + O2.

[0046] Multi-stage alkaline adsorption system: The neutralization absorption tower adopts a two-stage series design. The first-stage tower is filled with Ca(OH)2 particles (particle size 3 - 5 mm) for preliminary adsorption of HF; the second-stage tower is equipped with a circulating spray system with pH feedback (Ca(OH)2 solution concentration 10 wt%), and the spray rate is 5 L / min. The pH value of the spray liquid is monitored in real time and the concentration of the Ca(OH)2 solution (10 wt%) is adjusted to ensure the HF neutralization efficiency (tail gas HF concentration < 1 ppm), and at the same time prevent fouling of the adsorption tower. The Ca(OH)2 solution is circulated and sprayed to fix HF as CaF2 precipitate, achieving a tail gas fluorine content < 1 ppm. When the NF3 concentration monitored by the NF3 detector is less than the fourth concentration threshold N4 after passing through the adsorption system, it is discharged into the plant utility system, otherwise it is re-introduced into the multi-stage alkaline adsorption system by the air pump. Since NO and NO# can be absorbed by the calcium hydroxide solution when mixed in a 1:1 ratio to form calcium nitrite and water, and nitrogen dioxide can directly react with calcium hydroxide to form calcium nitrate and calcium nitrite, the nitrogen oxides in the tail gas can be treated together by the Ca(OH)2 solution.

[0047] S3, Waste heat recovery and intelligent control: The heat exchanger conducts countercurrent heat exchange between the 1000°C tail gas and the inlet gas, raising the temperature of the NF3 inlet gas from 25°C to 300 ± 10°C. The waste heat of the 1000°C tail gas is recovered through the heat exchanger to preheat the inlet to 300°C, which can reduce energy consumption.

[0048] Dynamic regulation is carried out using gas sensors (NF3 detector + electrochemical HF sensor): The NF3 concentration is monitored in real time through the gas sensors, and the heating power and the inlet gas flow rate are adaptively adjusted to ensure a stable decomposition rate.

[0049] Example 1 of the working scenario of the above electro-adsorption type tail gas treatment equipment: Standard working condition treatment: A. Operating parameters: Input gas: NF3 concentration 10 vol%, flow rate 12 m 3 / h, with O2 as the carrier gas (accounting for 90%).

[0050] Temperature control: (1) Preheating section: 500°C (heating rate 30°C / s, power 8 kW). (2) Main decomposition section: 800°C (constant temperature accuracy ±5°C, power 15 kW). (3) Deep cracking section: 1000°C (power 18 kW, residence time ≥ 2 s).

[0051] B. Neutralization system: Steam flow rate 6 m 3 / h, the spraying rate of Ca(OH)2 solution is 8 L / min.

[0052] C. Performance results: (1) Decomposition efficiency: The decomposition rate of NF3 is 99.7% (the residual NF3 concentration is <50 ppm analyzed by FTIR spectroscopy). (2) By-product control: The conversion rate of F2 > 99.9%, and the HF concentration in the tail gas <0.5 ppm (meeting the GB16297-1996 standard). (3) The purity of CaF2 precipitate at the outlet of the secondary absorption tower ≥98%, which can be directly recycled as industrial raw materials. (4) Energy consumption index: The comprehensive power consumption is 1.2 kWh / m 3 NF3, saving 52% energy compared with the traditional incineration method.

[0053] Example 2 of the working scenario of the above electro-adsorption type tail gas treatment equipment: Treatment of high-concentration NF3 shock load: A. Operating parameters: Input gas: The NF3 concentration is 30 vol%, and the flow rate suddenly increases to 20 m 3 / h (simulating abnormal process conditions).

[0054] B. Dynamic response: The heating system detects the increase in concentration within 0.5 s and automatically adjusts: The temperature of the main decomposition section rises to 850 °C, and the power is increased to 20 kW. The steam flow rate synchronously increases to 10 m 3 / h (molar ratio 1.05:1). The inlet gas flow rate is reduced by 15%, and the residence time is extended to 3 s.

[0055] C. Performance results: (1) Decomposition efficiency: The decomposition rate of NF3 is 99.6%, and there is no F2 leakage (the sensor alarm threshold is 0.1 ppm). (2) System stability: The temperature fluctuation < ±10 °C, and the pH value of the neutralization tower is maintained at 8.5 - 9.0 (to avoid scaling of Ca(OH)2).

[0056] The key process verification data of the above working scenarios Examples 1 and 2 are shown in Table 1 below:

[0057] Handling of special situations: This electro-thermal adsorption type tail gas treatment equipment also includes an emergency pressure relief valve arranged between the inlet of the heating system and the inlet of the adsorption system. When the inlet pressure increases, the emergency pressure relief valve will open, and the tail gas will pass through the bypass pipeline to the adsorption tank to prevent back-end pressure buildup.

[0058] This tail gas treatment device realizes the step-by-step decomposition of NF3 through segmented heating, combines a multi-stage adsorption system to efficiently neutralize harmful gases (such as HF), can effectively remove harmful substances in the tail gas, and reduce environmental pollution. The principle is as follows: Preheating stage: gradually raise the temperature of the waste gas to the initial reaction temperature to avoid uneven reaction or partial undecomposition of the gas due to sudden high temperature. Preheating can also reduce the damage of thermal stress to the equipment.

[0059] Main decomposition stage: The temperature rises to the optimal decomposition temperature of 800°C, at which the chemical bonds of NF3, i.e., NF bonds, are more easily broken, generating N2 and F2 through thermal decomposition or catalytic reaction, or further converting into stable products such as HF. This stage ensures that most NF3 is decomposed quickly.

[0060] Deep cracking stage: Maintain or slightly exceed the temperature of the main reaction zone at 1000°C, extend the residence time to completely decompose the residual NF3, prevent the by-products from recombination, and inhibit the reverse reaction of F2.

[0061] NF3 decomposition is an endothermic reaction, and high temperature is conducive to the thermodynamic equilibrium moving toward the product direction. However, maintaining high temperature throughout the process will significantly increase energy consumption. And a single high temperature may lead to side reactions, such as the generation of NO X or other fluorides. Controlling the temperature in stages can optimize the reaction path, ensuring that NF3 is preferentially decomposed into target products, such as N2, F2 / HF, and reducing harmful byproducts. Through three-stage temperature gradient control, the heating furnace achieves fine regulation of the NF3 decomposition process, taking into account the reaction rate, thoroughness and energy efficiency, and ultimately increases the NF3 decomposition rate to more than 99.5%, effectively reducing the emission of this potent greenhouse gas.

[0062] Embodiment 3: Figure 9 The conventional structure diagram of the air outlet water cooling pipe is shown. Figure 9 Where L is the length of the water pipe, l is the inner diameter of the cooling water channel, d is the inner diameter of the water pipe, and D is the outer diameter of the water pipe; an increase in the length of the heat dissipation water pipe may mean a longer cooling time, resulting in a lower temperature in the plasma area. If the temperature is too low, it may be detrimental to the dissociation reaction, and the dissociation rate may drop by 20%-30%, because the gas dissociation process requires high temperature with sufficient energy. On the contrary, if the water pipe is too short, the cooling is insufficient, or the temperature is too high, it may cause equipment damage or reaction runaway, interfering with the stability of the dissociation reaction.

[0063] A water pipe with a larger diameter can allow for a larger water flow rate, improve cooling efficiency, and remove heat faster to reduce temperature. However, a too large diameter may reduce flow rate, affect turbulence, and thus reduce heat transfer efficiency. On the contrary, a smaller diameter may increase flow rate, but may cause excessive pressure drop, increase water pump load, and even cause poor cooling effect due to insufficient flow.

[0064] In the remote plasma test system, the design of the outlet water cooling pipe directly affects the stability of the plasma temperature and free radical distribution, and thus affects the measurement accuracy of the dissociation rate. A pipe that is too short will lead to insufficient heat dissipation, resulting in an increase in the outlet temperature, accelerated free radical recombination, and a low dissociation rate measurement value; a pipe that is too long will increase the flow resistance pressure drop ΔP, requiring higher pump power, and may cause local boiling due to excessive coolant temperature rise, destroying temperature uniformity.

[0065] Figure 10 The optimized structure diagram of the outlet water cooling pipe 103 of the dissociation rate test device of a remote plasma source provided in this embodiment is shown; the outlet water cooling pipe structure is improved and divided into multiple sections with gradually decreasing diameters, each section having its own pipe length L, pipe diameter D and inner diameter l of the cooling water channel 1031, specifically: S1: Determine the heat dissipation requirements of the water cooling pipes through thermodynamic calculations to avoid interference of local temperature fluctuations on the plasma dissociation state.

[0066] Calculate the heat load at the plasma outlet based on the plasma source power, gas flow rate and gas type (e.g. NF3 / Ar mixture) Q . Heat load formula: Q = P · η ,in P is the input power, η is the energy conversion efficiency. Temperature change It can be further calculated from the gas parameters: ; Where m is the mass flow rate of gas flow rate, C m It is the constant pressure specific heat capacity of the mixed gas, which is determined by the gas type and the mixing ratio. Select deionized water or ethylene glycol solution as the coolant and set the flow rate V (L / min) and cooling water temperature rise T in , usually control the temperature rise of the water outlet and water inlet between 10~20℃ to avoid vaporization or affect the heat dissipation efficiency; ensure that the coolant temperature rise Δ T Controlled within a reasonable range, where Δ T ≤5℃.

[0067] The flow rate can be expressed as ,in ρ is the coolant density, where C p is the specific heat capacity of water (about 4186 J / (kg·℃)).

[0068] S2: By controlling the length of the pipeline, the heat dissipation efficiency and flow resistance are balanced to avoid the interference of temperature gradient on the dissociation rate measurement.

[0069] Pipe lengthL , k is the thermal conductivity of the pipe material, K is the proportionality coefficient, and K is an empirical constant, usually taking a value of 0.8 - 1.2. Then the length of the outlet cooling pipe can be expressed as: (1); Among them, the length L of the outlet cooling pipe can be represented by the lengths L1, L2, L of the first, second, and nth sections of the water pipe, that is: n That is: (2); S3: Within the allowable ΔP range, optimize the design by adjusting the length L or the diameter D. If the pressure loss is too high, it is necessary to increase the pipe diameter or shorten the length; Control the target water flow velocity v between 1 and 3 m / s to avoid excessive turbulence leading to high pressure loss or laminar flow causing poor heat dissipation of the water pipe. From the flow rate formula , initially derive the pipe diameter: (3); Among them, D1, D2, D n represent the diameters of the first, second, and nth sections of the water pipe; Q1, Q2, Q n represent the heat loads of the plasma outlets in the first, second, and nth sections of the water pipe; The relationship between the water pipe length L and the pressure drop ΔP can be used to optimize the calculation of the length of each section of the water pipe. According to the Darcy - Weisbach formula, there is: (4); Among them, f is the friction coefficient, f is related to the Reynolds number Re and the pipe wall roughness. When water is at 20°C, the relationship is .

[0070] If there is a relationship of Re < 2000 in laminar flow, then f = 64 / Re; The inner diameter of the ith section of the water channel can be expressed as: (5); If P is greater than the first pressure drop threshold P1 of the set flow resistance, that is, the pressure loss ΔP is too high at this time, it is necessary to increase the pipe diameter or shorten the length to optimize the current ith section of the water pipe; From equation (4), the following formula can be derived: (6); Then, the pipe diameter and length after optimized design are expressed as equations (7) and (8): (7); (8); Precisely match the parameters of each section of the pipeline based on thermodynamics and fluid mechanics formulas to ensure that the coolant temperature rise ΔT ≤ 5°C, maintain the stability of the plasma dissociation temperature, and reduce the measurement error of the dissociation rate.

[0071] The temperature at the remote plasma outlet is the highest, and the temperature away from the outlet decreases sequentially. After design, when the water pipe approaches the outlet, the water pipe diameter D is increased and the inner diameter l of the water channel is decreased to increase the water channel density to solve the problem of excessive temperature at the outlet; Figure 9 Compared with the traditional water pipe design scheme, this embodiment adopts the method of water-cooled temperature gradient adjustment, and designs suitable lengths L, diameters D, and inner diameters l of the water channels according to different water pipe sections, improving the heat dissipation efficiency of the remote plasma water cooling.

[0072] The heat dissipation water pipe structure at the outlet balances the heat dissipation efficiency and flow resistance through variable diameter design, ensures the matching of heat dissipation requirements and flow pressure drop, avoids local boiling or insufficient cooling, and maintains the temperature stability of the plasma dissociation reaction. The heat dissipation water pipe structure at the outlet solves the problem of free radical recombination caused by excessive local temperature at the outlet, improves the measurement accuracy of the dissociation rate, and improves the heat dissipation efficiency.

[0073] For the heat dissipation efficiency verification requirements of the remote plasma source cooling system design, it is necessary to quantify the influence of different water channel parameters on temperature gradient control through multiple groups of control experiments; Example 1 is a three-section cooling water pipe design, and the variable parameters are set as follows: the inner diameter of the cooling water pipe is 4.2 cm, and the total length of the cooling water pipe L is 20.0 cm; the diameter of the outlet section D 1 is 12.3 cm, the length L 1 = 8.2 cm, and the inner diameter of the first section of the water channel l 1 is 1.5 cm; the diameter of the middle section D 2 is 8.3 cm, the length L 2 = 6.7 cm, and the inner diameter of the second section of the water channel l 2 is 2.0 cm; the diameter of the last section D 3 is 6.0 cm, the length L 3 = 5.1 cm, and the inner diameter of the third section of the water channel l 3 is 2.5 cm; Example 2 is a five-section cooling water pipe design, and the variable parameters are set as follows: the inner diameter of the cooling water pipe is 4.2 cm, and the total length of the cooling water pipe L is 20 cm; the diameter of the outlet section D 1 is 15.2 cm, the length L 1 = 6.0 cm, and the inner diameter of the first section of the water channel l 1 is 1.0 cm; the diameter of the second section of the water pipe D 2 is 12.1 cm, the length LThe inner diameter of the second water channel is 5.0 cm, and the inner diameter of the second waterway is l 1.2 cm; the middle section diameter D The third section is 10.0 cm in diameter and the length L The third water channel is 4.0 cm, and the inner diameter of the third waterway l is 1.5 cm; the diameter of the fourth water pipe D The fourth section is 8.1 cm in diameter and the length L The fourth water channel is 3.0 cm, and the inner diameter of the fourth waterway l is 1.7 cm; the diameter of the last section D The fifth section is 6.0 cm in diameter and the length L The fifth water channel is 2.0 cm, and the inner diameter of the fifth waterway l is 2.0 cm; The control group is a traditional cooling water pipe, and its variable parameters are: the inner diameter of the cooling water pipe d is 4.2 cm, and the outer diameter D is 6.1 cm, and the total length of the cooling water pipe L is 20.0 cm, and the inner diameter l of the cooling water channel is 2.5 cm; The remote plasma source is preheated to a steady state at full power. After maintaining for 10 minutes, it is ignited and turned off. The cooling water circulation is started, and the water flow rate is maintained at 5 L / min. After 30 s, an infrared thermal imager is used to record the temperature at the gas outlet of the remote plasma source, the middle section temperature, and the temperature at the outlet of the cooling water end respectively. The experiment is repeated 3 times and the average value is taken, as shown in Table 2 below.

[0074]

[0075] Compared with Example 1, the 5-segment gradient water channel has better temperature control at the gas outlet, middle section, and end through refined segmented design; adding 2 more segments of high-density water channels at the gas outlet reduces the local thermal resistance and has a significant temperature drop effect; the refined segmentation of the 5-segment water pipe gradient cooling structure makes the temperature gradient in the middle section smoother. At the end of the water channel, the 5-segment water pipe gradient cooling structure can reduce the accumulation of residual heat through a longer cooling section and has a higher cooling efficiency; Comparing the water pipe gradient cooling structure with the traditional water pipe structure, the temperature of the highest temperature zone at the gas outlet is reduced from 525 °C to 384 °C, meeting the requirements of semiconductor processes for heat load control; by adjusting the water channel parameters D , L , l , the temperature drop rate from the gas outlet to the end is increased by 40%, verifying the effectiveness of the temperature gradient design.

[0076] Example 4: In a remote plasma system, the scaling, blockage, or micro-leakage of the outlet water pipe will directly affect the heat dissipation efficiency, the stability of gas dissociation, and even cause safety accidents. A fault diagnosis method for the outlet water pipe of a remote plasma source in this embodiment combines multi-physical field sensing technology, data analysis, and intelligent algorithms to achieve real-time perception and early warning. As Figure 11 shown, it includes the following steps: S1: Real-time monitor the flow resistance pressure drop ΔP through the differential pressure sensors installed at the inlet and outlet of the outlet water pipe, collect the local temperature distribution data ΔT through the infrared thermal imager outside the wall of the outlet water pipe, and at the same time arrange a capacitive humidity sensor near the outlet water pipe to monitor the environmental humidity change ΔR; S2: Based on the Kalman filter algorithm, fuse the ΔP, ΔT, and ΔR data to eliminate noise interference and obtain the state estimation values ΔP(k), T(k), and R(k) of the pressure drop, temperature, and humidity; S3: Construct a scaling / blockage objective function M and a leakage objective function N; Use the prediction mean of all decision trees as the final output to respectively fit the scaling, blockage, and micro-leakage formed by each combination of process parameters, and evaluate the importance degree relationship between scaling, blockage, and micro-leakage and each process parameter; S4: Assign the importance degree of each process parameter to the weight ratios of the differential pressure, flow rate, temperature, and humidity in the scaling, blockage, and micro-leakage objective functions M and N; when M exceeds the first threshold M1, it is determined as scaling, and when it exceeds the second threshold M2, it is determined as blockage; when N exceeds the leakage threshold N1, it is determined as micro-leakage. M1 / M2 / N1 are all set values and are adjusted according to the actual situation.

[0077] The following are the key steps and technical solutions for the function design: The detection of water pipe scaling and blockage is as follows: The formation of water scale is due to the attachment of minerals in hard water on the inner wall of the water pipe under the action of high temperature and long time. And the blockage is caused by the continuous accumulation of water scale gradually reducing the flow area of the water pipe; Install high-precision differential pressure sensors at the inlet and outlet of the water pipe, with an accuracy of ±0.1% FS. Scaling or blockage will cause an increase in local flow resistance and a significant rise in the differential pressure Δ P . Set the initial differential pressure to P 0 kPa. If Δ P continuously rises to the first differential pressure threshold P 1 kPa, an alarm for excessive flow resistance pressure drop will be triggered. Use non-invasive measurement of the water flow rate and calculate the water flow through the time difference of acoustic wave propagation v , and when the water flow rate drops abnormally to the flow rate threshold v1. If the flow rate decreases from 10 L / min to 7 L / min, a flow rate anomaly alarm will be prompted. An infrared thermal imager is installed outside the pipe wall. Due to poor thermal conductivity, local high-temperature points will form in the fouling area. When Δ T temperature difference > 10°C can be recognized. When the local high-temperature point is greater than the temperature threshold T 1, an over-temperature alarm will be triggered. The detection of micro-leakage of the water pipe is as follows: A capacitive humidity sensor is arranged near the air outlet to detect an abnormal increase in ambient humidity. The initial humidity is R 0RH. If the humidity suddenly rises to the humidity threshold R 1RH, a water pipe humidity anomaly alarm will be triggered; in addition, the flow differential pressure and the attenuation rate of the water flow velocity of the water pipe also reflect to a certain extent whether the water pipe has micro-leakage; after the water pump is turned off, a differential pressure sensor is used to monitor the pressure attenuation rate in the water pipe. If it is normal <0.1 bar / min and >0.5 bar / min during leakage. Similarly, the water flow velocity will also be appropriately attenuated during micro-leakage, and a non-invasive method can be used to measure the flow velocity.

[0078] In order to achieve real-time perception and early warning, this embodiment is based on the Kalman filtering algorithm to fuse multi-source data such as differential pressure, flow rate, temperature, and humidity to eliminate noise interference, and infers the state of the water pipe through the observed data.

[0079] The Kalman filter consists of two parts: prediction and update. It adjusts the weights of the prediction part and the observation part by updating the Kalman gain to make the result closer to the true value. The steps are as follows: S1: System state equation: Establish the state equation of the system, that is, describe how the current state of the system is determined by the previous state and external inputs. Usually, this can be represented in the form of a linear dynamic system: x ( k ) = Ax ( k- 1) + Bu ( k ) + w ( k ), where x ( k ) represents the state of the system at time k ; A is n * n the state transition matrix, B is the input control matrix, u ( k ) is the input at time k , w ( k ) is the system process noise. x (k ) is the data value at the current moment, x ( k- 1) is the data value at the previous moment, and the data value includes multi-source data such as differential pressure, flow rate, temperature, humidity, etc.

[0080] S2: Observation equation: An observation equation is established to describe how to obtain the observed value by measuring the system, which is usually also a linear relationship: y ( k ) = Hx ( k ) + v ( k ), where y ( k ) represents the observed value at time k , H is the observation matrix, v ( k ) is the observation noise.

[0081] S3: Prediction step: In the Kalman filter, the prediction step is first carried out, and the system state equation and prior information are used to estimate the current state and state uncertainty of the system. The uncertainty equation is as follows: (9); (10); In the formula, P - ( k ) is the prior estimate value of the covariance matrix at k time; P - ( k - 1) is the prior estimate value of the covariance matrix at k - 1 time; Q is the covariance matrix of the prediction value error.

[0082] S4: Update step: Then the update step is carried out. The state predicted in the previous step is compared with the actual observed value, and the prediction value is corrected through the Kalman gain to obtain an estimated value closer to the true state.

[0083] The three update equations are as follows: (11); In the formula, R is the covariance matrix of the observed value error; K ( k ) is the Kalman gain at k time; (12); (13); wherein P ( k ) is k the covariance matrix at the moment; I is the identity matrix.

[0084] S5: Continuously iterate: Finally, continuously repeat the prediction and update steps, gradually optimizing the estimation of the system state over time, and finally obtaining state estimation values Δ P ( k )、 v ( k )、 T ( k )、 R ( k ) with high accuracy for differential pressure, flow rate, temperature, and humidity.

[0085] S6: Fault type judgment: First, determine whether the system triggers alarms such as excessive flow resistance pressure drop, abnormal flow rate, over-temperature, or abnormal humidity. When an alarm command appears, the system needs to stop for inspection and perform fault type judgment; Respectively establish objective functions for fouling, blockage M and leakage faults N : The relationship between the importance of fouling and blockage fault characteristics and each process parameter is as follows: 1) Differential pressure (λ1): Fouling or blockage will cause mineral accumulation on the inner wall of the pipeline, increasing the flow resistance, thereby increasing the differential pressure of the system. Since the differential pressure reflects the resistance of gas flow in the system, the increased flow resistance caused by fouling will directly increase the differential pressure; 2) Flow rate (λ2): Fouling or blockage affects the flow path, reducing the effective flow area of the pipeline, thereby causing a decrease in gas flow rate; 3) Temperature (λ3): The reduced flow rate caused by fouling or blockage will lead to poor heat dissipation and local temperature increase; 4) Humidity (λ4): The humidity sensor is installed at the RPS outlet and does not reflect the effects of fouling and blockage; then: (14); The relationship between the importance of micro-leakage fault characteristics and each process parameter is as follows: 1) Differential pressure (λ1): When a micro-leakage occurs in the water pipe, the leakage will cause an increase in the pressure difference before and after the leakage point; 2) Flow rate (λ2): Micro-leakage will cause a decrease in the flow rate in the water pipe, resulting in a decrease in the downstream flow rate; 3) Humidity (λ4): The humidity sensor is installed at the outlet of the RPS and can accurately sense the degree of water pipe leakage; then: (15); In equations (14 / 15), t ( k ) is the current moment, t ( k- 1) is the previous moment, Δ P ( k )、 v ( k )、 T ( k )、 R ( k ) are the state estimation values of the pressure difference, flow rate, temperature, and humidity at the current moment, Δ P ( k- 1)、 v ( k -1)、 T ( k -1)、 R ( k -1) are the state estimation values of the pressure difference, flow rate, temperature, and humidity at the previous moment, λ 1、 λ 2、 λ 3 are the weight ratios of the pressure difference, flow rate, and temperature respectively; λ4 is the weight ratio of humidity, where λ4 > λ1, λ2. The weight coefficients λ1~λ4 are obtained by training based on historical fault data and are dynamically adjusted according to the service life of the pipeline.

[0086] When M is greater than the first fault threshold M 1, it is judged that there is scale deposition in the water pipe at this time; when M is greater than the second fault threshold M 2, where M 2 > M 1, then it is judged that there is a local blockage in the water pipe at this time; When N is greater than the leakage fault threshold N 1, it is judged that there is a micro-leakage phenomenon in the water pipe at this time.

[0087] This fault diagnosis method for the outlet pipeline realizes accurate identification of scale, blockage, and leakage (accuracy rate ≥ 95%) through multi-sensor collaborative monitoring (pressure difference, temperature, humidity) and Kalman filter data fusion, and reduces the false alarm rate to less than 5%.

[0088] Regarding the evaluation of the importance relationship of the weight ratios of the pressure difference, flow rate, temperature, and humidity λ 1-4 is further described as follows: As Figures 12-13 shown, the random forest regression algorithm evaluates the importance degree relationship between scaling, blockage, micro-leakage and process parameters. The steps are as follows: Step P1: Bootstrap sampling: Using the linear regression model, the approximate linear relationships between scaling, blockage, micro-leakage and each process parameter can be fitted. T sub-sample sets {D1, D2,..., D T} are drawn from the original data set D with replacement, and the sample size of each subset is N. The probability P(N) of each sample being selected in a single sampling is: (16); Then, the out-of-bag data set OOB composed of the unselected samples is used for error estimation and feature importance analysis.

[0089] Step P2: Decision tree construction: Randomly select features. Let the total number of features be M t , for each subset D t , construct a regression decision tree h t (x). When splitting nodes, randomly select m features (m ≤ M t ).

[0090] The goal of node splitting in the regression tree is to minimize the mean squared error MSE. For feature j and split threshold and R R (j, s), select the optimal split pair (j∗, s∗): (17); Among them, the actual threshold y i of scaling and blockage faults is represented by the objective function M, and the actual threshold y i of micro-leakage faults is represented by the objective function N; R L and R R are the sample sets of the left and right child nodes after splitting, and c L and c R are the predicted values of the left and right child nodes respectively. That is, the target mean of the subset samples: (18); Step P3: Prediction result integration: The mean of the predictions of all decision trees is used as the final output, and ensemble learning is used to reduce the uncertainty of the prediction of a single tree. The final output is as follows: (19); Among them, is the average result; h t (x) is the prediction result of a single decision tree; T is the number of decision trees.

[0091] Step P4: Model performance evaluation metrics: Use the root mean square error (RMSE) metric to quantify the accuracy of predicting scaling, blockage, and micro-leakage respectively: (20); Judge whether the root mean square error of scaling, blockage, and micro-leakage and process parameters meets the error accuracy requirement, and the error accuracy is 0.1; otherwise, return to step P1 to continue iteration until the maximum number of iterations is reached; Step P5: Feature importance evaluation and analysis Measure the feature importance through the change in the out-of-bag (OOB) error after permuting the feature values. After randomly permuting the feature Xj, measure its importance degree through the change in the OOB error. If the error increases significantly after permutation, it indicates that this feature has an important impact on model prediction. The importance degree is calculated as follows: (21); where, MSE t is the OOB error of the t-th tree.

[0092] Accordingly, the score rankings of the importance degrees of scaling, blockage, and micro-leakage and process parameters are as Figure 14 shown. The importance degree relationships between scaling and blockage faults and pressure difference, flow rate, and temperature are 0.51, 0.30, and 0.19 respectively, and are assigned to the weight values of the objective functions M pressure difference, flow rate, and temperature of scaling and blockage; similarly, the importance degree relationships between micro-leakage and humidity, flow rate, and pressure difference are 0.83, 0.10, and 0.07 respectively. Then, the weight values of the objective function M humidity, flow rate, and pressure difference of micro-leakage can be obtained.

[0093] The technical solutions in the embodiments of the present application at least have the following technical effects or advantages: 1. The detection device integrates an advanced automatic control system and can automatically complete the detection process.

[0094] 2. The device adopts infrared mass spectrometry analysis and detection technology, can achieve high-precision and high-sensitivity measurement of various active particles in the plasma, and thus accurately calculate the dissociation rate, providing reliable data for the performance evaluation of the plasma source.

[0095] 3. The device is designed with a real-time monitoring function and can instantly feedback the dissociation state of the plasma. Through real-time monitoring, the working conditions of the plasma source can be adjusted in a timely manner to ensure the stability and consistency of the production process.

[0096] 4. The outlet of the RPS device is designed with a variable cycle heat dissipation device to solve the influence of temperature imbalance problems on the measurement reliability of the dissociation rate.

[0097] 5. The detection device is designed with an exhaust gas treatment device, which can effectively remove harmful substances in the exhaust gas and reduce environmental pollution.

[0098] 6. The evaluation of the dissociation rate of the remote plasma source does not need to be detected in actual integrated circuit production equipment, with low cost and short cycle.

[0099] Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those of ordinary skill in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present invention.

Claims

1. A method for diagnosing faults in a water pipe at an outlet of a remote plasma source, characterized in that: The following steps are involved: Step 1: monitor the flow resistance pressure drop ΔP in real time by means of a pressure difference sensor installed at the inlet and outlet of the air outlet water pipe, collect local temperature distribution data ΔT by means of an infrared thermal imager outside the wall of the air outlet water pipe, and monitor the ambient humidity change ΔR by means of a capacitive humidity sensor arranged near the air outlet water pipe; Step 2: Based on the Kalman filter algorithm, the ΔP, ΔT, and ΔR data are fused to obtain the state estimation values ​​ΔP(k), T(k), and R(k) of the voltage drop, temperature, and humidity; Step 3: Construct the fouling / clogging objective function M and the leakage objective function N. ； ； Where t(k) is the current moment, t(k-1) is the previous moment, ΔP(k), v(k), T(k), and R(k) are the estimated values ​​of the pressure difference, flow rate, temperature, and humidity at the current moment, and ΔP(k-1), v(k-1), T(k-1), and R(k-1) are the estimated values ​​of the pressure difference, flow rate, temperature, and humidity at the previous moment. λ 1. λ 2. λ 3 are the weighted ratios of pressure difference, flow rate and temperature respectively, λ4 is the weighted ratio of humidity, λ4>λ1,λ2; Step 4: When M exceeds the first threshold M1, it is determined to be scaling, and when it exceeds the second threshold M2, it is determined to be clogging; when N exceeds the leakage threshold N1, it is determined to be a micro-leakage, and a graded alarm and maintenance instruction are triggered.

2. The method for diagnosing a fault in a gas outlet water pipe according to claim 1, characterized in that: The accuracy of the differential pressure sensor is ±0.1% FS. When ΔP continues to rise to the first differential pressure threshold value P1, it is determined that the flow resistance is abnormal.

3. The method for diagnosing a fault in a gas outlet water pipe according to claim 1, characterized in that: When the infrared thermal imager detects that the local temperature ΔT>10°C, it determines it as a scaling area and generates high temperature point coordinate information.

4. The method for diagnosing a fault in a gas outlet water pipe according to claim 1, characterized in that: Combined with the non-invasive flow velocity measurement, the water flow v is calculated by the sound wave propagation time difference. When the water flow velocity abnormally drops to the flow threshold v1, the flow is judged to be abnormal.

5. The method for diagnosing a fault in a gas outlet water pipe according to claim 1, characterized in that: The capacitive humidity sensor is activated when the ambient humidity suddenly rises to a humidity threshold R 1, the leakage alarm is triggered and double verification is performed in combination with the pressure difference decay rate.

6. The method for diagnosing a fault in a gas outlet water pipe according to claim 1, characterized in that: The system state equation of the Kalman filter algorithm is x(k)=Ax(k-1)+Bu(k)+w(k), and the observation equation is y(k)=Hx(k)+v(k). The state estimation value is optimized by iteratively updating the Kalman gain K(k).

7. The method for diagnosing a fault in a gas outlet water pipe according to claim 1, characterized in that: The weight coefficients λ1~λ4 in the objective function are obtained through training based on historical fault data and are dynamically adjusted with the use time of the pipeline.

Images