An ultrahigh vacuum material outgassing rate testing system and method

Abstract

The application belongs to the technical field of vacuum testing, and discloses a kind of ultra-high vacuum material outgassing rate test system and method, the system includes vacuum chamber, vacuum pumping unit, temperature control unit, sensor unit, intelligent control and processing unit, the application can actively learn and compensate background drift under variable temperature condition, have multi-source data fusion diagnostic ability and realize accurate measurement in wide temperature range, solve the technical problems that measurement accuracy of existing test system is susceptible to temperature interference, cannot deduct dynamic background in real time and lack system state self-diagnosis.

Description

Technical Field

[0001] This invention belongs to the field of vacuum testing technology, specifically relating to a system and method for testing the outgassing rate of ultra-high vacuum materials. Background Technology

[0002] Material outgassing rate is a key indicator for evaluating the suitability of materials in vacuum environments. Cutting-edge fields such as high-energy particle accelerators, aerospace devices, semiconductor manufacturing, and controlled nuclear fusion require ultra-high and extremely high vacuum environments, with many applications demanding system ultimate pressures of 10... -8 Pa even 10 -9 The fabrication of these ultra-high vacuum chambers, operating on the Pa scale, requires materials with extremely low outgassing rates, such as 316L stainless steel, oxygen-free copper, and NEG getter films. Measuring these outgassing rates necessitates the use of an ultra-high vacuum material outgassing rate testing system. When the material outgassing rate is as low as 10... -12 Pa·m 3 / s·cm 2 At this scale, existing testing systems face the following technical bottlenecks: (1) The nonlinear drift interference of background gas release under variable temperature conditions is difficult to eliminate; In traditional symmetrical dual-chamber systems, the effectiveness of background subtraction relies on the complete thermodynamic consistency of the two chambers. When the system undergoes temperature changes (such as cooling after baking or thermal cycling in simulated space), even with perfectly symmetrical geometry, transient temperature differences inevitably exist between the two chambers due to factors such as power differences in heating mantles and thermostatic chambers, environmental airflow disturbances, slight differences in material heat capacity, and complex heating or cooling test procedures. Studies have shown that the material outgassing rate of stainless steel vacuum chambers is highly sensitive to temperature—a 1°C temperature difference near room temperature can lead to a 20%–50% change in the material outgassing rate. This difference in background outgassing caused by transient temperature differences cannot be eliminated by simple structural symmetry, becoming a core obstacle restricting measurement accuracy under variable temperature conditions.

[0003] (2) It is difficult to extract signals with ultra-low background. When the material outgassing rate is as low as 10 -12 At the order of magnitude, the signal is extremely weak, producing a pressure change of only 10. -9 ~10 -8 Pa level. Single-channel measurement cannot effectively distinguish between "material venting" and "system background change", and the traditional "no-load-load" two-measurement method cannot deduct real-time background drift due to the time difference.

[0004] (3) Lack of proactive ability to identify interference in the measurement environment; In ultra-low material outgassing rate testing, factors such as molecular pump vibration, environmental electromagnetic interference, and vacuum gauge filament emission current drift can all introduce measurement noise. Test results are highly dependent on operator experience, lacking online monitoring and diagnostic capabilities for the test system's own health status, such as vacuum gauge drift and minor leaks. Furthermore, prolonged measurement under conditions of operator fatigue can lead to increased cumulative errors in the measurement results. Existing systems lack real-time monitoring and automatic labeling of these interference sources, resulting in the inability to effectively identify and eliminate suspicious data, thus reducing the reliability of the test results.

[0005] (4) Insufficient level of intelligence, relying on the experience of operators; Existing systems often employ simple difference or averaging algorithms for data processing, failing to dynamically adjust compensation strategies based on system status. The reliability of test results heavily relies on operator experience and lacks system-level self-diagnostic capabilities for health status. Summary of the Invention

[0006] The purpose of this invention is to provide a test system and method for the outgassing rate of ultra-high vacuum materials, which can actively learn and compensate for background drift under variable temperature conditions, has multi-source data fusion diagnostic capabilities, and achieves accurate measurement over a wide temperature range. This solves the technical problems of existing test systems, such as measurement accuracy being easily affected by temperature, inability to deduct dynamic background in real time, and lack of system state self-diagnosis.

[0007] The technical solution adopted in this invention is an ultra-high vacuum material outgassing rate testing system, including a vacuum chamber, a vacuum pumping unit, a temperature control unit, a sensor unit, and an intelligent control and processing unit, wherein: The vacuum chamber includes a sample chamber, a first reference chamber, a downstream measurement chamber, and a second reference chamber. The sample chamber is used to hold the material sample to be tested. The first reference chamber is used for dynamic testing and monitoring during the orifice conductance method and the dual-channel gas path conversion method, serving as a real-time background reference for the sample chamber. The second reference chamber is used to continuously monitor the long-term background drift of the system when the sample chamber is subjected to static pressure boosting testing, serving as an auxiliary reference. The downstream measurement chamber is used to measure the material outgassing rate under ultra-high vacuum conditions. The second reference chamber is independently isolated. The sample chamber, the first reference chamber, and the downstream measurement chamber are coaxially connected in series along the same axis. The sample chamber is connected to the downstream measurement chamber through a fifth valve and a conventional flow conductance orifice. The first reference chamber is connected to the downstream measurement chamber through a third valve and a high flow conductance orifice. The second measurement chamber is connected to the downstream measurement chamber through a seventh valve. The vacuum pumping unit is connected to the downstream measuring chamber via a valve, enabling the downstream measuring chamber to obtain and maintain an ultimate pressure better than 5 × 10⁻⁶. -8 Pa ultra-high vacuum environment; The temperature control unit is installed in the sample chamber, the first reference chamber, the downstream measurement chamber, and the second reference chamber, and is used to independently control the temperature in each chamber and acquire spatially distributed temperature field data. The sensor unit is installed on the vacuum chamber and is used to acquire data from the full-range vacuum gauge, micro-current sensor, quadrupole mass spectrometer and vibration sensor in real time. The intelligent control and processing unit is used to control the vacuum chamber, vacuum pumping unit, temperature control unit, and sensor unit to achieve multi-sensor fusion and self-learning temperature compensation, complete the ultra-high vacuum material outgassing rate test, and diagnose and output prompt information for at least one of the following: the working status of the full-range vacuum gauge, the system leakage status, and the performance degradation status of the auxiliary pumping equipment.

[0008] Furthermore, the first and second reference chambers have the same structure, size, and material as the sample chamber, and are always unloaded for real-time background subtraction. The two reference chambers verify each other to eliminate system deviations.

[0009] Furthermore, the outer walls of the sample chamber, the first reference chamber, the downstream measurement chamber, and the second reference chamber are equipped with full-scale vacuum gauges for measuring the full-scale vacuum level from the atmospheric side to the ultra-high vacuum side within the vacuum chamber. A micro-current sensor is connected in series in the filament circuit of the full-scale vacuum gauge to monitor minute changes in the filament emission current of the ionization gauge in real time, thereby determining the health status of the vacuum gauge itself. Quadrupole mass spectrometers are installed on the outer walls of the sample chamber and the downstream measurement chamber via eleventh and tenth valves, respectively, to collect the mass spectral peak intensities of characteristic gases H2, H2O, CO, and CO2. Vibration sensors are installed on the outer walls of the sample chamber and the first reference chamber to monitor the interference of mechanical vibration on the ultra-low pressure measurement signal in real time.

[0010] Furthermore, the vacuum pumping unit includes a main pumping device and an auxiliary pumping device; the main pumping device includes a first turbomolecular pump, a second turbomolecular pump, and a dry vortex mechanical pump, the dry vortex mechanical pump is connected to the first turbomolecular pump through a ninth valve, the first turbomolecular pump is connected to the second turbomolecular pump through a first valve, the second turbomolecular pump is connected to the downstream measurement chamber through a second valve, and the auxiliary pumping device is connected to the downstream measurement chamber and the first turbomolecular pump through an eighth valve.

[0011] Furthermore, the temperature control unit includes an independently controllable heating and cooling device and a multi-point distributed fiber optic temperature sensor array; the heating and cooling device includes an independent heating jacket, a radiation heater, and a liquid nitrogen-cooled finger tube. The outer walls of the sample chamber, the first reference chamber, the downstream measurement chamber, and the second reference chamber are respectively wrapped with independent heating jackets. A high-low temperature sample stage is provided in the sample chamber, and a radiation heater and a liquid nitrogen-cooled finger tube are provided on the high-low temperature sample stage. The fiber optic temperature sensor is spirally attached along the inner walls of the sample chamber, the first reference chamber, the downstream measurement chamber, and the second reference chamber.

[0012] Furthermore, the special gas inlet is located on the right side of the sample chamber and is connected to the twelfth valve and the fine-tuning valve.

[0013] A method for testing the outgassing rate of ultra-high vacuum materials, based on the aforementioned ultra-high vacuum material outgassing rate testing system, specifically includes the following steps: S1, Data acquisition for training the self-learning temperature compensation model; When no sample of the material to be tested is placed in the sample chamber, the intelligent control and processing unit controls the temperature control unit to execute a preset variable temperature scanning program, so that the system goes through at least one heating stage and one cooling stage, and during the variable temperature scanning process, input feature data and target output data are collected simultaneously. The input feature data includes multiple spatially distributed temperature measurement point values, temperature change rate, and real-time temperature difference between the sample chamber and the reference chamber, all obtained from a multi-point distributed fiber optic temperature sensor array. The target output data is the real-time pressure difference between the sample chamber and the reference chamber, which, under no-load conditions, fully characterizes the system background gas drift caused by temperature changes. S2, Self-learning temperature compensation model training; Using the input feature data as input and the target output data as output, a self-learning temperature compensation model is trained, and a nonlinear mapping relationship between the temperature field and the system background gas drift is constructed based on a deep neural network model. S3, the formal test, includes the following steps: S31, Test mode selection; The system of this invention automatically switches valves, allowing staff to select any one of three testing modes—orifice flow conduction method, dual-channel gas path conversion method, or static pressure boosting method—based on the expected outgassing rate of the material sample and testing requirements. S32, Calculation of material venting volume; Regardless of the test mode used, during the measurement process, the intelligent control and processing unit will input the real-time collected temperature field characteristics into the self-learning temperature compensation model to obtain the background gas release compensation term ΔP_model(t); (1) Orifice flow conduction method Open the small orifice passage between the sample chamber and the downstream measurement chamber, so that the gas released from the sample of the material to be tested flows into the downstream measurement chamber through the conventional flow conduction orifice / high flow conduction orifice, while the downstream measurement chamber is continuously evacuated by the main evacuation equipment; After the vacuum chamber reaches dynamic equilibrium, record the sample chamber pressure P1 and the downstream measurement chamber pressure P0. The material venting volume Q is calculated using the formula: Q = C × (P1 - P0) In the formula, C is the molecular conductivity of the flow-conducting orifice, which is corrected according to the gas type and temperature; (2) Dual-channel gas path conversion method By alternately opening the fourth and sixth valves, the measurement channel can be quickly switched between the reference chamber and the sample chamber; At any sampling time t, the system simultaneously acquires the sample chamber pressure P1(t) and the reference chamber pressure P2(t); The material venting capacity Q(t) is calculated using the formula: Q(t)=C×[P1(t)-P2(t)-ΔP_model(t)] In the formula, ΔP_model(t) is the background gas release compensation term output in real time by the self-learning temperature compensation model; (3) Static boosting method Close all valves between the sample chamber and the downstream measurement chamber, and simultaneously close the connection between the downstream measurement chamber and the vacuum pumping unit, so that the sample chamber becomes a closed volume V, and record the rate of increase of the pressure in the sample chamber over time, dP / dt. The material venting volume Q is calculated using the formula: Q = V × (dP / dt) S33, Calculation of material outgassing rate; In the small-hole flow conductance test mode, the material outgassing amount Q is divided by the nominal geometric surface area A of the material sample to be tested to obtain the material outgassing rate per unit area. In the static pressure boosting test mode, the material outgassing amount Q is divided by the nominal geometric surface area A of the material sample to be tested, which gives the material outgassing rate per unit area. In the dual-channel gas path conversion method, the material release rate q(t) is calculated in real time according to the formula: q(t)=C×[P1(t)-P2(t)-ΔP_model(t)] / A In the formula, q(t) is the outgassing rate of the material at time t, with units of Pa·m. 3 / (s·cm 2P1(t) is the pressure value in the sample chamber at time t, in Pa; P2(t) is the pressure value in the reference chamber at time t, in Pa; ΔP_model(t) is the background gas release compensation term output by the self-learning temperature compensation model, in Pa; A is the nominal geometric surface area of ​​the sample material, in cm². 2 ; S34, Data Quality Marker; During the formal testing process, the intelligent control and processing unit controls the sensor unit to synchronously collect auxiliary sensing data and performs real-time analysis on the auxiliary sensing data: when the vibration amplitude exceeds the threshold or the full-range vacuum gauge filament current fluctuation exceeds the preset range, the material outgassing rate calculation value at the corresponding moment is marked as "low confidence" for reference or rejection in subsequent data analysis. The auxiliary sensing data includes vibration sensor data installed on the outer wall of the vacuum chamber and microcurrent sensor data installed in the filament circuit of the full-range vacuum gauge. S4, Post-test data processing; After the test, the system automatically generates a test report, which includes the curve of material outgassing rate changing over time, the average and standard deviation of material outgassing rate, statistics of data points marked as low confidence, the main outgassing gas components and their partial pressures measured by quadrupole mass spectrometer, material outgassing rate, and system health status diagnostic information.

[0014] The beneficial effects of this invention are as follows: (1) By using a self-learning temperature compensation model, it actively learns and compensates for the nonlinear drift of background gas release caused by transient temperature difference, breaking through the limitation of traditional symmetrical structures that can only eliminate steady-state common-mode interference.

[0015] (2) Combining multi-point temperature field sensing and neural network nonlinear fitting capabilities, it can accurately measure the material outgassing rate in a wide temperature range from -196℃ to +400℃, meeting the testing needs of aerospace materials, superconducting materials and other extreme environments.

[0016] (3) For the first time, auxiliary sensing data such as vibration and microcurrent were incorporated into the fusion analysis framework of the material outgassing rate test system, realizing the active identification and marking of measurement noise and online diagnosis of vacuum gauge health, system leakage, saturation of auxiliary pumping equipment, etc., which greatly improved the reliability of test results.

[0017] (4) When the system state changes (such as the cavity background increases after long-term operation), the model can perform incremental learning by performing rapid temperature scanning to maintain the compensation accuracy and has good adaptability to working conditions. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the overall structure of the system of the present invention.

[0019] In the diagram, the following are the markings: VC1 - Sample chamber; VC2 - First reference chamber; VC3 - Downstream measurement chamber; VC4 - Second reference chamber; F1 - First turbomolecular pump; F2 - Second turbomolecular pump; W1 - Dry vortex mechanical pump; B1 - Auxiliary pumping equipment; Y - Special gas inlet; V1 - First valve; V2 - Second valve; V3 - Third valve; V4 - Fourth valve; V5 - Fifth valve; V6 - Sixth valve; V7 - Seventh valve; V8 - Eighth valve; V9 - Ninth valve; V10 - Tenth valve; V11 - Eleventh valve; V12 - Twelfth valve; V13 - Fine-tuning valve; C1 - Conventional flow conduction orifice; C2 - High flow conduction orifice; G - Full-range vacuum gauge; QMS - Quadrupole mass spectrometer. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this invention clearer and more understandable, the technical solutions of this invention will be clearly and completely described below in conjunction with the accompanying drawings.

[0021] Background drift: The slow, persistent deviation of the background signal (such as electrical signal, fluorescence intensity, pressure reading, etc.) output by the instrument over time when there is no sample input or an ideal reference state.

[0022] Background gas outgassing drift: The phenomenon in which the system background pressure / background mass spectrometry signal undergoes a slow unidirectional shift due to the continuous and time-decrease-driven gas outgassing effect of the cavity and components.

[0023] like Figure 1 As shown, an ultra-high vacuum material outgassing rate testing system includes a vacuum chamber, a vacuum pumping unit, a temperature control unit, a sensor unit, and an intelligent control and processing unit.

[0024] The vacuum chamber includes a sample chamber VC1, a first reference chamber VC2, a downstream measurement chamber VC3, and a second reference chamber VC4. The sample chamber VC1 is used to place the sample of the material to be tested. The first reference chamber VC2 is used for dynamic testing and monitoring in the small-hole conductance method and the dual-channel gas path conversion method, serving as a real-time background reference for the sample chamber VC1. The second reference chamber VC4 is used to continuously monitor the long-term background drift of the system when the sample chamber VC1 is subjected to static pressure boosting testing, serving as an auxiliary reference. The downstream measurement chamber VC3 is used to measure the material outgassing rate under ultra-high vacuum conditions. The second reference chamber VC4 is set up independently and isolated. The sample chamber VC1, the first reference chamber VC2, and the downstream measurement chamber VC3 are set up coaxially in series along the same axis. The sample chamber VC1 is connected to the downstream measurement chamber VC3 through the fifth valve V5 and the conventional flow conductance small hole C1. The first reference chamber VC2 is connected to the downstream measurement chamber VC3 through the third valve V3 and the high flow conductance small hole C2. The second reference chamber VC4 is connected to the downstream measurement chamber VC3 through the seventh valve V7.

[0025] The standard conductivity orifice C1 has a conductivity rating of 6.0 L / s (20℃) for N2 molecules and 22.4 L / s for H2 molecules, suitable for outgassing tests of conventional materials. The high conductivity orifice C2 has a conductivity rating of 24.0 L / s (20℃) for N2 molecules and 90 L / s for H2 molecules, suitable for high-sensitivity tests such as NEG thin film materials. Both orifices are circular, thin-walled structures made of oxygen-free copper with electrolytically polished inner surfaces; the thickness-to-diameter ratio is <0.1. Conductivity values ​​were calibrated in situ using the standard gas injection method.

[0026] The first reference chamber VC2 and the second reference chamber VC4 have the same structure, dimensions, and materials as the sample chamber VC1. They are always unloaded and used for real-time background subtraction. The two reference chambers cross-calibrate to eliminate system bias. The sample chamber VC1, the first reference chamber VC2, the downstream measurement chamber VC3, and the second reference chamber VC4 are all made of SUS316L stainless steel. The inner surfaces are electrolytically polished to a roughness Ra=0.18μm and degassed at high temperature in a vacuum furnace at 900℃ for 4 hours.

[0027] The outer walls of the sample chamber VC1, the first reference chamber VC2, the downstream measurement chamber VC3, and the second reference chamber VC4 are equipped with a full-range vacuum gauge G (measuring range 10). -9 ~10 -2 The full-range vacuum gauge G is used to measure the full-scale vacuum from the atmospheric side to the ultra-high vacuum side in the vacuum chamber. A micro-current sensor (accuracy ±0.1%) is connected in series in the filament circuit of the full-range vacuum gauge G to monitor the minute changes in the filament emission current of the ionization gauge in the full-range vacuum gauge G in real time, so as to judge the health status of the vacuum gauge itself and ensure the long-term reliability of the ultra-low outgassing rate measurement data. The outer walls of the sample chamber VC1 and the downstream measurement chamber VC3 are equipped with quadrupole mass spectrometers (mass number 1~200 amu) through the eleventh valve V11 and the tenth valve V10. They are used to collect the mass spectral peak intensities of characteristic gases H2, H2O, CO, and CO2 for outgassing component analysis. Vibration sensors (accelerometers, ±10g range) are installed on the outer walls of the sample chamber VC1 and the first reference chamber VC2. They are used to monitor the interference of mechanical vibrations such as molecular pump vibration, ground vibration, and personnel walking on the ultra-low pressure measurement signal in real time, and to identify and eliminate false noise signals introduced by environmental vibration.

[0028] The sensor unit is installed on the vacuum chamber to acquire data in real time from the full-range vacuum gauge G, microcurrent sensor, quadrupole mass spectrometer QMS, and vibration sensor.

[0029] The vacuum pumping unit is connected to the downstream measuring chamber VC3 via a valve, enabling the downstream measuring chamber VC3 to obtain and maintain an ultimate pressure better than 5×10⁻⁶. -8 Ultra-high vacuum environment of Pa.

[0030] The vacuum pumping unit includes a main pumping device and an auxiliary pumping device B1. The main pumping device includes a first turbomolecular pump F1, a second turbomolecular pump F2 (pumping speed of 1000 L / s for N2), and a dry vortex mechanical pump W1 (pumping speed of 12 L / s). The dry vortex mechanical pump W1 is connected to the first turbomolecular pump F1 through the ninth valve V9. The first turbomolecular pump F1 is connected to the second turbomolecular pump F2 through the first valve V1. The second turbomolecular pump F2 is connected to the downstream measurement chamber VC3 through the second valve V2. The auxiliary pumping device is connected to the downstream measurement chamber VC3 and the first turbomolecular pump F1 through the eighth valve V8. The auxiliary pumping device B1 includes a cryogenic pump (pumping speed of 500 L / s for H2) and an ion pump (400 L / s). During the formal measurement, the eighth valve V8 remains closed.

[0031] The temperature control unit is installed in the sample chamber VC1, the first reference chamber VC2, the downstream measurement chamber VC3, and the second reference chamber VC4. It is used to independently control the temperature in each chamber and acquire spatially distributed temperature field data.

[0032] The temperature control unit includes independently controllable heating and cooling devices and a multi-point distributed fiber optic temperature sensor array (16 measuring points per chamber, spatial resolution 5mm, temperature measurement accuracy ±0.3℃). The heating and cooling devices include independent heating jackets (2kW / chamber), radiant heaters (1.5kW, maximum 400℃), and liquid nitrogen-cooled finger tubes (minimum -196℃). Independent heating jackets are wrapped around the outer walls of sample chamber VC1, the first reference chamber VC2, the downstream measurement chamber VC3, and the second reference chamber VC4. A high- and low-temperature sample stage is installed inside sample chamber VC1, with radiant heaters and liquid nitrogen-cooled finger tubes mounted on it. Fiber optic temperature sensors are spirally attached to the inner walls of sample chamber VC1, the first reference chamber VC2, the downstream measurement chamber VC3, and the second reference chamber VC4 to acquire spatially distributed temperature field data in real time. The fiber optic temperature sensor array enables millimeter-level spatial resolution temperature field measurement, with a single fiber replacing dozens of thermocouples, reducing the sources of gas release within the vacuum chamber.

[0033] The special gas inlet Y is located on the right side of the sample chamber VC1 and is connected to the twelfth valve V12 and the fine-tuning valve V13 (the test system needs to be equipped with an inlet structure for nitrogen cleaning test system, special gas introduction test, etc.).

[0034] The sample chamber VC1 is connected to the downstream measurement chamber VC3 via a sixth valve V6, and the downstream measurement chambers VC3 are also connected via a fourth valve V4.

[0035] The intelligent control and processing unit is used to control the above modules (vacuum chamber, vacuum pumping unit, temperature control unit, and sensor unit) to achieve multi-sensor fusion and self-learning temperature compensation, complete the ultra-high vacuum material outgassing rate test, and diagnose and output prompt information for at least one of the following: the working status of the full-range vacuum gauge G, the system leakage status, and the performance degradation status of the auxiliary pumping equipment B1.

[0036] After system assembly, leak detection using a helium mass spectrometer showed an overall leak rate better than 1.5 × 10⁻⁶. -11 Pa·m 3 / s.

[0037] System debugging: Heat the entire system to 250℃ at a heating rate of 1.5℃ / min, maintain this temperature for 48 hours, and then allow it to cool naturally to 25℃; start the main exhaust system, and after 24 hours, the system's ultimate pressure reaches 4.8×10⁻⁶. -8 Pa; Auxiliary pumping equipment B1 is turned on (eighth valve V8 is opened) and pumping continues for 12 hours, reducing the background pressure to 2.1 × 10⁻⁶ Pa. -8 Pa, then the eighth valve V8 was closed, and the pressure stabilized at 3.5 × 10 Pa. -8 Pa, meets the design specifications.

[0038] A method for testing the outgassing rate of ultra-high vacuum materials, based on the aforementioned ultra-high vacuum material outgassing rate testing system, specifically includes the following steps: S1, Data acquisition for training the self-learning temperature compensation model; With no sample of the material to be tested placed in sample chamber VC1, the intelligent control and processing unit controls the temperature control unit to execute a preset variable temperature scanning program, causing the vacuum chamber to undergo at least one heating stage and one cooling stage. During the variable temperature scanning process, the following two types of data are collected simultaneously: Input feature data: at least includes multiple spatially distributed temperature measurement point values, temperature change rate, and real-time temperature difference between sample chamber VC1 and reference chamber obtained by a multi-point distributed fiber optic temperature sensor array; Target output data: Real-time pressure difference between the sample chamber VC1 and the reference chamber, which fully characterizes the system background gas drift caused by temperature changes under no-load conditions.

[0039] S2, Self-learning temperature compensation model training; Using input feature data as input and target output data as output, a self-learning temperature compensation model is trained, and a nonlinear mapping relationship between the temperature field and the system's background gas outburst drift is constructed based on a deep neural network model.

[0040] Specifically, in one embodiment, the model input features include: The fiber optic temperature sensor array collects n temperature measurement points T1~Tn (n=16~32); The real-time pressure difference between the sample chamber VC1 and the reference chamber is ΔP = P1 - P2; The average temperature change rate dT / dt of the downstream measuring chamber VC3; The real-time temperature difference ΔT_sr between the sample chamber VC1 and the reference chamber; Mass spectral peak intensities of characteristic gases H2, H2O, CO, and CO2 collected by a quadrupole mass spectrometer; Historical background pressure time series (first 10 time steps); Approximately 30,000 sets of valid data were collected, and vibration amplitudes exceeding 0.1 m / s were removed. 2 The remaining data points are Z-score standardized and then divided into training and validation sets in an 8:2 ratio.

[0041] The deep neural network model structure is as follows: Input layer: 40-dimensional feature vector (16 temperature measurement points + temperature change rate + temperature difference + partial pressure of 4 gases + historical pressure value); Hidden layer 1: Fully connected layer (128 nodes) + BatchNorm + ReLU + Dropout (0.2); Hidden layer 2: Fully connected layer (256 nodes) + BatchNorm + ReLU + Dropout (0.2); Hidden layer 3: Fully connected layer (128 nodes) + BatchNorm + ReLU + Dropout (0.2); Hidden layer 4: Fully connected layer (64 nodes) + BatchNorm + ReLU; Output layer: Fully connected layer (1 node) + Linear activation, outputting the predicted background gas exhalation compensation term ΔP_model(t).

[0042] A genetic algorithm was used to optimize the initial weights and thresholds of the network (population 50, 100 generations), followed by fine-tuning using the Adam optimizer (learning rate 0.001, batch size 32, early stopping mechanism). After training, the model's coefficient of determination R on the validation set was calculated. 2 =0.987, the root mean square error of prediction is 1.6×10 -12 Pa meets the measurement requirements. The training strategy employs a two-stage approach combining genetic algorithm optimization of the initial weights of the BP neural network with error backpropagation fine-tuning to avoid getting trapped in local optima and improve the model's generalization ability.

[0043] S3, the formal test, includes the following steps: S31, Test mode selection; The system of this invention allows operators to select any one of three testing modes—orifice flow conduction method, dual-channel gas path conversion method, or static pressure boosting method—based on the expected outgassing rate of the material sample and testing requirements through automatic valve switching.

[0044] S32, Calculation of material venting volume; Regardless of the test mode used, during the measurement process, the intelligent control and processing unit will input the real-time collected temperature field characteristics into the self-learning temperature compensation model to obtain the background venting compensation term ΔP_model(t).

[0045] (1) Orifice conduction method (dynamic flow method) Open the small orifice passage between the sample chamber VC1 and the downstream measurement chamber VC3 (e.g., by opening the fifth valve V5) so that the gas released from the sample of the material to be tested flows into the downstream measurement chamber VC3 through the conventional flow conduction orifice C1 / high flow conduction orifice C2. At the same time, the downstream measurement chamber VC3 is continuously evacuated by the main evacuation equipment. After the vacuum chamber reaches dynamic equilibrium, record the pressure P1 in the sample chamber VC1 and the pressure P0 in the downstream measurement chamber VC3. The material venting volume Q is calculated according to Formula 1: Q = C × (P1 – P0) Formula 1 In the formula, C is the molecular conductivity of the flow-conducting orifice (conventional flow-conducting orifice C1 / high flow-conducting orifice C2), which is corrected according to the gas type and temperature.

[0046] (2) Dual-channel gas path conversion method By alternately opening the fourth valve V4 (connecting the first reference chamber VC2 and the downstream measurement chamber VC3) and the sixth valve V6 (connecting the sample chamber VC1 and the downstream measurement chamber VC3), the measurement channel can be quickly switched between the reference chamber and the sample chamber VC1; At any sampling time t, the system simultaneously acquires the pressure P1(t) of the sample chamber VC1 and the pressure P2(t) of the reference chamber; The material venting capacity Q(t) is calculated according to Formula 2: Q(t) = C × [P1(t) - P2(t) - ΔP_model(t)] Formula 2 In the formula, ΔP_model(t) is the background outgassing compensation term output in real time by the self-learning temperature compensation model. This method effectively eliminates the influence of system background drift through real-time differential and temperature compensation, and is especially suitable for the accurate measurement of material samples with extremely low material outgassing rates.

[0047] (3) Static boost method (cumulative method) Close all valves (such as V5 and V6) between the sample chamber VC1 and the downstream measurement chamber VC3, and at the same time close the connection between the downstream measurement chamber VC3 and the vacuum pumping unit, so that the sample chamber VC1 becomes a closed volume V, and record the rate of increase of the pressure in the sample chamber VC1 over time dP / dt. The material venting volume Q is calculated according to formula 3: Q = V × (dP / dt) Formula 3 This method does not rely on orifice conductivity calibration and can be used as a verification tool for other methods.

[0048] S33, Calculation of material outgassing rate; In the small-hole flow conductance test mode, the material outgassing amount Q is divided by the nominal geometric surface area A of the material sample to be tested, which gives the material outgassing rate per unit area.

[0049] In the static pressure boosting test mode, the material outgassing amount Q is divided by the nominal geometric surface area A of the material sample to be tested, which gives the material outgassing rate per unit area.

[0050] The material release rate q(t) is calculated in real time according to Formula 4: q(t) = C × [P1(t) - P2(t) - ΔP_model(t)] / A (Formula 4) In the formula: q(t) is the outgassing rate of the material at time t, with units of Pa·m 3 / (s·cm 2 P1(t) is the pressure value in sample chamber VC1 at time t, in Pa; P2(t) is the pressure value in reference chamber at time t, in Pa; ΔP_model(t) is the background gas release compensation term output by the self-learning temperature compensation model, in Pa; A is the nominal geometric surface area of ​​the sample material under test, in cm². 2 .

[0051] Wherein, ΔP_model(t) = f(T(t), dT / dt, ΔTsr(t)), that is, ΔP_model(t) = f(temperature field(t), temperature change rate(t), temperature difference(t)). The temperature change self-learning background gas release compensation term ΔP_model(t) is an AI-driven dynamic zero-point corrector. The system can distinguish between real material gas release and temperature-induced background drift through this compensation term.

[0052] S34, Data Quality Marker; During the formal testing process, the intelligent control and processing unit controls the sensor unit to synchronously collect auxiliary sensing data and performs real-time analysis on the auxiliary sensing data: when the vibration amplitude exceeds the threshold or the filament current fluctuation of the full-range vacuum gauge G exceeds the preset range, the calculated material outgassing rate at the corresponding moment is marked as "low confidence" for reference or rejection in subsequent data analysis.

[0053] The auxiliary sensing data includes at least the vibration sensor data installed on the outer wall of the vacuum chamber and the micro-current sensor data installed in the filament circuit of the full-range vacuum gauge G.

[0054] S4, Post-test data processing; After the test, the system automatically generates a test report, which includes the material outgassing rate change curve over time, the average and standard deviation of the material outgassing rate, statistics of data points marked as low confidence, the main outgassing gas components and their partial pressure material outgassing rate measured by quadrupole mass spectrometer (the partial pressure material outgassing rate is calculated by replacing the total pressure in the orifice conductance formula with the partial pressure value of a specific mass number, and the conductance value adopts the molecular conductance of the corresponding gas), and system health status diagnostic information (such as the full-range vacuum gauge G status, leakage warning, auxiliary pumping equipment performance prompts, etc.).

[0055] This invention utilizes a self-learning temperature compensation model to actively learn and compensate for the nonlinear drift of background gas release caused by transient temperature differences. This overcomes the limitation of traditional symmetrical structures, which can only eliminate steady-state common-mode interference. Experiments show that at a temperature change rate of 1℃ / min, the background compensation accuracy is improved by more than 80% compared to the traditional differential method, which helps to extend the limits of the material gas release rate testing system. Combining multi-point temperature field sensing with the nonlinear fitting capability of neural networks, it can achieve accurate measurement of material gas release rate in a wide temperature range from -196℃ to +400℃, meeting the testing needs of aerospace materials, superconducting materials, and other extreme environments. For the first time, it incorporates auxiliary sensing data such as vibration and microcurrent into the fusion analysis framework of the material gas release rate testing system, realizing the active identification and marking of measurement noise and online diagnosis of vacuum gauge health, system leakage, and auxiliary pumping equipment B1 saturation, significantly improving the reliability of test results. When the system state changes (such as the cavity background increasing after long-term operation), the model can perform incremental learning by executing rapid temperature change scanning to continuously maintain compensation accuracy and has good adaptability to operating conditions.

[0056] Any content not described in detail in this specification belongs to the prior art in this technical field.

Claims

1. A system for testing the outgassing rate of ultra-high vacuum materials, characterized in that, It includes a vacuum chamber, a vacuum pumping unit, a temperature control unit, a sensor unit, and an intelligent control and processing unit, among which: The vacuum chamber includes a sample chamber (VC1), a first reference chamber (VC2), a downstream measurement chamber (VC3), and a second reference chamber (VC4). The sample chamber (VC1) is used to hold the sample material to be tested. The first reference chamber (VC2) is used for dynamic testing and monitoring during the orifice conductance method and the dual-channel gas path conversion method, serving as a real-time background reference for the sample chamber (VC1). The second reference chamber (VC4) is used to continuously monitor the long-term background drift of the system when the sample chamber (VC1) undergoes static pressure boosting testing, serving as an auxiliary reference. The downstream measurement chamber (VC3) is used to measure ultra-high vacuum... Material outgassing rate in an air environment; the second reference chamber (VC4) is independently isolated; the sample chamber (VC1), the first reference chamber (VC2), and the downstream measurement chamber (VC3) are coaxially connected in series along the same axis; the sample chamber (VC1) is connected to the downstream measurement chamber (VC3) through the fifth valve (V5) and a conventional flow guide orifice (C1); the first reference chamber (VC2) is connected to the downstream measurement chamber (VC3) through the third valve (V3) and a high flow guide orifice (C2); and the second reference chamber (VC4) is connected to the downstream measurement chamber (VC3) through the seventh valve (V7). The vacuum pumping unit is connected to the downstream measuring chamber (VC3) via a valve, enabling the downstream measuring chamber (VC3) to obtain and maintain an ultimate pressure better than 5 × 10⁻⁶. -8 Pa ultra-high vacuum environment; The temperature control unit is installed in the sample chamber (VC1), the first reference chamber (VC2), the downstream measurement chamber (VC3), and the second reference chamber (VC4) to independently control the temperature in each chamber and acquire spatially distributed temperature field data. The sensor unit is installed on the vacuum chamber and is used to acquire data from the full-range vacuum gauge (G), microcurrent sensor, quadrupole mass spectrometer (QMS), and vibration sensor in real time. The intelligent control and processing unit is used to control the vacuum chamber, vacuum pumping unit, temperature control unit, and sensor unit to achieve multi-sensor fusion and self-learning temperature compensation, complete the ultra-high vacuum material outgassing rate test, and diagnose and output prompt information for at least one of the following: the working status of the full-range vacuum gauge (G), the system leakage status, and the performance degradation status of the auxiliary pumping equipment (B1).

2. The ultra-high vacuum material outgassing rate testing system according to claim 1, characterized in that, The first reference chamber (VC2) and the second reference chamber (VC4) have the same structure, size and material as the sample chamber (VC1), and are always unloaded. They are used for real-time background subtraction, and the two reference chambers verify each other to eliminate system deviation.

3. The ultra-high vacuum material outgassing rate testing system according to claim 1, characterized in that, The outer walls of the sample chamber (VC1), the first reference chamber (VC2), the downstream measurement chamber (VC3), and the second reference chamber (VC4) are equipped with full-scale vacuum gauges (G) for measuring the full-scale vacuum level from the atmospheric side to the ultra-high vacuum side within the vacuum chamber. A micro-current sensor is connected in series with the filament circuit of the full-scale vacuum gauge (G) to monitor minute changes in the filament emission current of the ionization gauge in the full-scale vacuum gauge (G) in real time, thereby determining the health status of the vacuum gauge itself. Quadrupole mass spectrometers (QMS) are installed on the outer walls of the sample chamber (VC1) and the downstream measurement chamber (VC3) via eleventh valves (V11) and tenth valves (V10) to collect the mass spectral peak intensities of characteristic gases H2, H2O, CO, and CO2. Vibration sensors are installed on the outer walls of the sample chamber (VC1) and the first reference chamber (VC2) to monitor the interference of mechanical vibration on the ultra-low pressure measurement signal in real time.

4. The ultra-high vacuum material outgassing rate testing system according to claim 1, characterized in that, The vacuum pumping unit includes a main pumping device and an auxiliary pumping device (B1). The main pumping device includes a first turbomolecular pump (F1), a second turbomolecular pump (F2), and a dry vortex mechanical pump (W1). The dry vortex mechanical pump (W1) is connected to the first turbomolecular pump (F1) through a ninth valve (V9). The first turbomolecular pump (F1) is connected to the second turbomolecular pump (F2) through a first valve (V1). The second turbomolecular pump (F2) is connected to the downstream measuring chamber (VC3) through a second valve (V2). The auxiliary pumping device is connected to the downstream measuring chamber (VC3) and the first turbomolecular pump (F1) through an eighth valve (V8).

5. The ultra-high vacuum material outgassing rate testing system according to claim 1, characterized in that, The temperature control unit includes an independently controllable heating and cooling device and a multi-point distributed fiber optic temperature sensor array. The heating and cooling device includes an independent heating jacket, a radiation heater, and a liquid nitrogen-cooled finger tube. The outer walls of the sample chamber (VC1), the first reference chamber (VC2), the downstream measurement chamber (VC3), and the second reference chamber (VC4) are respectively wrapped with independent heating jackets. A high-low temperature sample stage is provided inside the sample chamber (VC1). A radiation heater and a liquid nitrogen-cooled finger tube are provided on the high-low temperature sample stage. The fiber optic temperature sensor is spirally attached along the inner walls of the sample chamber (VC1), the first reference chamber (VC2), the downstream measurement chamber (VC3), and the second reference chamber (VC4).

6. The ultra-high vacuum material outgassing rate testing system according to claim 1, characterized in that, The special gas inlet (Y) is located on the right side of the sample chamber (VC1) and is connected through the twelfth valve (V12) and the fine-tuning valve (V13).

7. The ultra-high vacuum material outgassing rate testing system according to claim 1, characterized in that, The sample chamber (VC1) and the downstream measurement chamber (VC3) are also connected by a sixth valve (V6), and the downstream measurement chambers (VC3) are also connected by a fourth valve (V4).

8. A method for testing the outgassing rate of ultra-high vacuum materials, implemented based on the ultra-high vacuum material outgassing rate testing system according to any one of claims 1-7, characterized in that, Specifically, the following steps are included: S1, Data acquisition for training the self-learning temperature compensation model; When no sample of the material to be tested is placed in the sample chamber (VC1), the intelligent control and processing unit controls the temperature control unit to execute a preset variable temperature scanning program, so that the system goes through at least one heating stage and one cooling stage, and during the variable temperature scanning process, input feature data and target output data are collected simultaneously. The input feature data includes multiple spatially distributed temperature measurement point values, temperature change rate, and real-time temperature difference between the sample chamber (VC1) and the reference chamber, obtained by a multi-point distributed fiber optic temperature sensor array. The target output data is the real-time pressure difference between the sample chamber (VC1) and the reference chamber, which, under no-load conditions, fully characterizes the system background gas drift caused by temperature changes. S2, Self-learning temperature compensation model training; Using the input feature data as input and the target output data as output, a self-learning temperature compensation model is trained, and a nonlinear mapping relationship between the temperature field and the system background gas drift is constructed based on a deep neural network model. S3, the formal test, includes the following steps: S31, Test mode selection; The system of this invention automatically switches valves, allowing staff to select any one of three testing modes—orifice flow conduction method, dual-channel gas path conversion method, or static pressure boosting method—based on the expected outgassing rate of the material sample and testing requirements. S32, Calculation of material venting volume; Regardless of the test mode used, during the measurement process, the intelligent control and processing unit will input the real-time collected temperature field characteristics into the self-learning temperature compensation model to obtain the background gas release compensation term ΔP_model(t); (1) Orifice flow conduction method Open the small orifice passage between the sample chamber (VC1) and the downstream measurement chamber (VC3) so that the gas released from the sample of the material to be tested flows into the downstream measurement chamber (VC3) through the conventional flow conduction orifice (C1) / high flow conduction orifice (C2), while the downstream measurement chamber (VC3) is continuously evacuated by the main evacuation equipment; After the vacuum chamber reaches dynamic equilibrium, record the pressure P1 in the sample chamber (VC1) and the pressure P0 in the downstream measurement chamber (VC3). The material venting volume Q is calculated using the formula: Q = C × (P1 - P0) In the formula, C is the molecular conductivity of the flow-conducting orifice, which is corrected according to the gas type and temperature; (2) Dual-channel gas path conversion method By alternately opening the fourth valve (V4) and the sixth valve (V6), the measurement channel can be quickly switched between the reference chamber and the sample chamber (VC1); At any sampling time t, the system simultaneously acquires the pressure P1(t) of the sample chamber (VC1) and the pressure P2(t) of the reference chamber; The material venting capacity Q(t) is calculated using the formula: Q(t)=C×[P1(t)-P2(t)-ΔP_model(t)] In the formula, ΔP_model(t) is the background gas release compensation term output in real time by the self-learning temperature compensation model; (3) Static boosting method Close all valves between the sample chamber (VC1) and the downstream measurement chamber (VC3), and simultaneously close the connection between the downstream measurement chamber (VC3) and the vacuum pumping unit, so that the sample chamber (VC1) becomes a closed volume V, and record the rate of increase of the pressure in the sample chamber (VC1) over time, dP / dt. The material venting volume Q is calculated using the formula: Q = V × (dP / dt) S33, Calculation of material outgassing rate; In the small-hole flow conductance test mode, the material outgassing amount Q is divided by the nominal geometric surface area A of the material sample to be tested to obtain the material outgassing rate per unit area. In the static pressure boosting test mode, the material outgassing amount Q is divided by the nominal geometric surface area A of the material sample to be tested, which gives the material outgassing rate per unit area. In the dual-channel gas path conversion method, the material release rate q(t) is calculated in real time according to the formula: q(t)=C×[P1(t)-P2(t)-ΔP_model(t)] / A In the formula, q(t) is the outgassing rate of the material at time t, with units of Pa·m. 3 / (s·cm 2 P1(t) is the pressure value in the sample chamber (VC1) at time t, in Pa; P2(t) is the pressure value in the reference chamber at time t, in Pa; ΔP_model(t) is the background gas compensation term output by the self-learning temperature compensation model, in Pa; A is the nominal geometric surface area of ​​the sample material under test, in cm². 2 ; S34, Data Quality Marker; During the formal testing process, the intelligent control and processing unit controls the sensor unit to synchronously collect auxiliary sensing data and performs real-time analysis on the auxiliary sensing data: when the vibration amplitude exceeds the threshold or the full-range vacuum gauge (G) filament current fluctuation exceeds the preset range, the material outgassing rate calculation value at the corresponding moment is marked as "low confidence" for reference or rejection in subsequent data analysis. The auxiliary sensing data includes vibration sensor data installed on the outer wall of the vacuum chamber and microcurrent sensor data installed in the filament circuit of the full-range vacuum gauge (G). S4, Post-test data processing; After the test, the system automatically generates a test report, which includes the curve of material outgassing rate changing over time, the average and standard deviation of material outgassing rate, statistics of data points marked as low confidence, the main outgassing gas components and their partial pressures measured by quadrupole mass spectrometer, material outgassing rate, and system health status diagnostic information.

Images