Intelligent monitoring method for uniformity of magnetron sputtering coating film of solar heat collecting tube

Abstract

This invention relates to the field of magnetron sputtering coating monitoring and control technology, and more specifically, to an intelligent monitoring method for the uniformity of magnetron sputtering coatings on solar collector tubes. The method includes: constructing a transmission-type acousto-optic caustic monitoring environment; acquiring caustic images and inverting them to obtain transient acoustic pressure distribution; performing singularity analysis on the pressure distribution using the box-counting method to calculate the microscopic inhomogeneity index of the coating layer; generating airflow pulse commands by combining the real-time acoustic resonance frequency of the inter-tube cavity; and driving a standing wave gas knife to inject pulsed gas into the inter-tube gap, using the wave interference principle to block the parasitic hollow cathode effect. This invention, through acousto-optic inversion and fractal topology filtering, achieves precise quantification and online closed-loop suppression of invisible micro-arc discharge lattice damage, significantly improving coating uniformity.

Description

Technical Field

[0001] This invention relates to the field of magnetron sputtering coating monitoring and control technology. More specifically, this invention relates to an intelligent monitoring method for the uniformity of magnetron sputtering coatings on solar collector tubes. Background Technology

[0002] In the solar thermal power generation industry chain, the quality of the heat-absorbing film layer of the collector tube directly determines the photothermal conversion efficiency. Currently, magnetron sputtering is the mainstream technology for preparing high-performance heat-absorbing films. In continuous production processes, glass tubes are usually placed on rotating conveyor rollers for coating, and wedge-shaped gaps inevitably form between adjacent glass tubes. When a high-energy plasma environment is present, these wedge-shaped gaps can easily induce parasitic hollow cathode effects, leading to abnormally high local plasma density and the generation of micro-arc discharges. The high-energy particle shock waves generated by these micro-arc discharges bombard the growing film layer, causing local lattice distortion and a decrease in density, ultimately severely affecting the optical performance and thermal stability of the collector tube.

[0003] However, existing membrane quality monitoring technologies have significant limitations in practical applications, mainly in the following two aspects:

[0004] On the one hand, traditional ellipsometry or colorimeters can only acquire film thickness information at local points, ignoring the fact that the coating process is a continuous, temporal evolution. This static detection method cannot capture the transient defect propagation characteristics caused by micro-arc discharge. Furthermore, existing monitoring methods cannot effectively distinguish between periodic vibrations caused by mechanical transmission and random lattice damage caused by plasma discharge. On high-speed conveyor lines, slight roller vibrations can cause fluctuations in the detection data. If feedback adjustments are made indiscriminately, it can easily lead to malfunctions in the control system, ultimately reducing the overall uniformity of the film.

[0005] On the other hand, traditional methods often rely on simple data fitting or threshold judgments without considering the specific production process mechanisms. For example, existing vacuum coating monitoring systems struggle to detect the acoustic resonance state within the inter-tube cavities, making it impossible to establish a physical correlation between process parameters and the parasitic hollow cathode effect. When a decrease in uniformity is detected, traditional methods typically only adjust the target power or transmission speed, failing to specifically block the physical root cause of the defects—the parasitic discharge in the inter-tube gaps. This lack of physical constraints leads to lag and blindness in coating quality control, making it difficult to meet the production requirements of high-precision solar collector tubes.

[0006] Therefore, there is an urgent need to develop an intelligent monitoring method that can accurately separate mechanical vibrations from lattice defects and actively suppress parasitic discharges based on physical processes. Summary of the Invention

[0007] To address the limitations of existing technologies in accurately distinguishing between mechanical vibrations and lattice defects, and in effectively suppressing inter-tube parasitic discharges that lead to poor film uniformity, this invention proposes an intelligent monitoring method for the uniformity of magnetron sputtering coatings on solar collector tubes. This method includes the following steps:

[0008] A transmission-type acoustic-optical caustic light field monitoring environment was constructed, and caustic images of solar collector tubes on the coating transmission line were acquired in real time using a line light source and an industrial camera.

[0009] The transient acoustic pressure distribution sequence was obtained by inverting the caustic image using the acoustic refractive index distortion measurement model.

[0010] The singular point skeleton analysis of the transient acoustic pressure distribution sequence was performed using the box dimension method, and the micro-inhomogeneity index of the film layer was calculated. The micro-inhomogeneity index of the film layer is used to characterize the degree of lattice defects in the film layer caused by micro-arc discharge.

[0011] The real-time acoustic resonance frequency of the inter-tube cavity formed between adjacent solar collector tubes is obtained. Based on the micro-inhomogeneity index of the film layer and the real-time acoustic resonance frequency of the inter-tube cavity, an airflow pulse control command for suppressing parasitic discharge is generated.

[0012] The airflow pulse control command is sent to the execution controller of the standing wave air knife, which drives the standing wave air knife to spray pulsed gas into the gap between the tubes in the tube cavity. The pulsed gas is used to block the parasitic hollow cathode effect in the gap between the tubes in real time, so as to realize the online closed-loop control of the uniformity of the magnetron sputtering coating.

[0013] This invention transforms invisible refractive index perturbations into visible sound pressure distributions through acousto-optic inversion, solving the problem of detection blind zones; it achieves accurate description of lattice defects through physical field inversion and topological analysis; and finally, it fundamentally suppresses parasitic discharges by disrupting standing wave conditions through physical interference, thereby significantly improving the uniformity of the film.

[0014] Preferably, the construction of the transmission-type acousto-optic caustic light field monitoring environment includes:

[0015] A parallel line laser source is set on the backlight side of the solar collector tube, and a high-speed industrial camera is set on the transmission side of the solar collector tube.

[0016] The photosensitive surface of the high-speed industrial camera is set at a preset monitoring distance from the axis of the solar collector tube, so that the focal plane of the solar collector tube, which acts as a cylindrical lens, is located in front of the photosensitive surface, thereby forming a caustic pattern containing acoustic refractive index modulation information on the photosensitive surface.

[0017] This invention utilizes the cylindrical geometry of the solar collector tube as a nonlinear lens, employing the principle of defocus caustics to optically amplify minute refractive index changes within the glass tube wall. A preset monitoring distance range ensures the highest contrast of the caustic fringes generated by acoustic modulation, enabling micron-level lattice stress changes to be clearly captured by an industrial camera.

[0018] Preferably, the preset monitoring distance ranges from 150 mm to 300 mm.

[0019] Preferably, obtaining the transient acoustic pressure distribution sequence includes:

[0020]

[0021] In the formula, Indicates the first The objective function of the residuals at time step; Represents the sampling time; A sequence index representing pixels in a caustic image; Represents the total number of pixels in a caustic image; For the first The caustic image acquired at time 1 The brightness value of the caustic spot at each pixel; For the first The average brightness value of the caustic image acquired at any given time; For the Laplace operator; The inversion obtained the first Transient acoustic pressure values ​​at the locations corresponding to each pixel; The preset reference sound pressure constant, The optical-acoustic coupling coefficient is obtained through pre-calibration;

[0022] Minimize the residual objective function to obtain the transient acoustic pressure value at the corresponding position of each pixel, and denote the sequence of transient acoustic pressure values ​​of all pixels as the transient acoustic pressure distribution sequence.

[0023] This invention utilizes the second derivative of light intensity to reflect the curvature change of the light beam, which is caused by the refractive index gradient induced by sound pressure. By solving this equation, the influence of uneven ambient lighting can be eliminated, revealing the true physical quantities of the sound field and providing an accurate data foundation for subsequent defect assessment.

[0024] Preferably, the formula for calculating the micro-inhomogeneity index of the film layer is:

[0025]

[0026] In the formula, Indicates time The microscopic non-uniformity index of the film layer; This represents the grid scale in box dimension calculation; Indicates at time The grid scale is The number of grids required to cover the sound pressure singularities in a transient acoustic pressure distribution sequence; This represents the reference fractal dimension caused by purely mechanical transmission; This represents the sono-induced lattice defect conversion coefficient.

[0027] This invention accurately separates the roughness component caused solely by micro-arc discharge from a strong mechanical noise background by subtracting the baseline dimension, thus greatly reducing the false alarm rate.

[0028] Preferably, the method for obtaining the sono-induced lattice defect transformation coefficient includes:

[0029] A set of glass tube samples that had undergone parasitic discharge of different intensities were obtained;

[0030] Calculate the difference between the calculated fractal dimension of the glass tube sample and the reference fractal dimension caused by the pure mechanical transmission;

[0031] The density of microscopic lattice defects per unit area of ​​the film layer in the glass tube sample was statistically analyzed using a scanning electron microscope.

[0032] Establish a curve showing the relationship between the microscopic lattice defect density and the difference, and use the slope of the curve as the acoustic lattice defect conversion coefficient.

[0033] Preferably, the method for obtaining the reference fractal dimension caused by the purely mechanical transmission includes:

[0034] Under no-load conditions with the magnetron sputtering target power supply off and the transmission system running normally, a sequence of reference caustic images of the solar collector tube was acquired.

[0035] The reference caustic image sequence is inverted to obtain the reference sound pressure distribution sequence; the fractal dimension of each frame of data in the reference sound pressure distribution sequence is calculated using the box dimension method.

[0036] Calculate the arithmetic mean of the fractal dimensions of all frame data, and use the arithmetic mean as the reference fractal dimension caused by the purely mechanical transmission.

[0037] Preferably, the airflow pulse control command generated to suppress parasitic discharge satisfies the following relationship:

[0038]

[0039] In the formula, Indicates time The airflow pulse control command; Indicates the flow gain coefficient; This represents the total number of steps in the historical backtracking process; Representing historical moments The microscopic non-uniformity index of the film layer at that location; This represents the reference fractal dimension caused by purely mechanical transmission; Indicates time The real-time acoustic resonant frequency of the inter-tube cavity; This indicates the phase angle.

[0040] This invention utilizes the accumulation term in the relation to extract the accumulated defects of the memory history, eliminate steady-state errors, and ensure strong suppression of continuous discharge; it also utilizes the sinusoidal term in the relation to ensure that the output airflow has a specific frequency and phase, thereby physically forming an antiphase wave to achieve destructive interference of the wave.

[0041] Preferably, the method for obtaining the real-time acoustic resonance frequency of the inter-tube cavity is as follows:

[0042]

[0043] In the formula, Indicates time The real-time acoustic resonant frequency of the inter-tube cavity; Indicates the cavity mode correction factor; Indicates the adiabatic index of a gas; Represents the gas constant; Indicates time The real-time absolute temperature inside the vacuum chamber; This indicates the distance between adjacent solar collector tubes.

[0044] Preferably, the method for obtaining the cavity modal correction factor includes:

[0045] In a cold environment in a vacuum chamber, the standing wave air knife is driven by a frequency sweep signal to eject the gas flow.

[0046] The sound pressure response was collected at the gap between the tubes using a microphone, and the physical frequency at which the sound pressure response reached its peak was identified.

[0047] The cavity mode correction factor is calculated based on the physical frequency, the tube spacing, and the sound velocity at the initial room temperature.

[0048] The present invention has the following beneficial effects:

[0049] This invention successfully transforms invisible micro-arc discharge lattice damage into visualized sound pressure distribution data by constructing a transmission-based acousto-optic caustics monitoring and inversion system. Utilizing fractal dimension as a topological tool, it effectively removes mechanical vibration noise, achieving a high signal-to-noise ratio defect description. By establishing an acoustic resonance model with temperature correction and an integral feedback control law, it achieves frequency locking and adaptive intensity suppression of parasitic standing wave fields. Finally, by utilizing the wave interference principle of pulsed gas, it precisely blocks the parasitic hollow cathode effect with minimal energy cost. This significantly improves the uniformity and yield of magnetron sputtering coatings for solar collector tubes. Attached Figure Description

[0050] Figure 1 This is a flowchart of the steps of the intelligent monitoring method for the uniformity of magnetron sputtering coating of solar collector tubes provided in this embodiment of the invention.

[0051] Figure 2 This is a graph illustrating the effect of the film micro-uniformity index provided in an embodiment of the present invention;

[0052] Figure 3 This is a schematic diagram of the uniform distribution of the axial film layer in a solar collector tube provided in an embodiment of the present invention. Detailed Implementation

[0053] Please see Figure 1 The diagram illustrates a flowchart of the intelligent monitoring method for the uniformity of the magnetron sputtering coating layer of a solar collector tube provided in Example 1. The method includes the following steps:

[0054] S1: Construct a transmission-type acousto-optic caustic light field monitoring environment, and use a line light source and industrial camera to collect caustic images of solar collector tubes on the coating transmission line in real time.

[0055] It should be noted that during the magnetron sputtering coating process of high borosilicate glass heat collectors, the microscopic uniformity of the film layer is highly susceptible to disruption by the parasitic hollow cathode effect. When micro-arc discharge occurs, the resulting shock wave is essentially ultrasonic waves within the glass medium, causing minute periodic oscillations in the glass's refractive index. These oscillations have extremely low amplitudes and cannot be captured by conventional focusing imaging. Using conventional methods, this can easily lead to invisible lattice damage and missed detections. Therefore, the core objective of this step is to utilize the principle of defocus caustics to establish an optical magnification mechanism, converting invisible phase changes into visible amplitude changes.

[0056] Preferably, as an example, a transmission-type acousto-optic caustic light field monitoring environment is constructed, utilizing a line light source and an industrial camera to acquire caustic images of the solar collector tubes on the coated transmission line in real time, including:

[0057] First, a parallel laser source arranged on the back side of the solar collector tube emits monochromatic parallel light with a constant wavelength to illuminate the solar collector tube.

[0058] Next, the cylindrical geometry of the solar collector tube itself is used as a cylindrical lens to nonlinearly focus the transmitted light passing through the tube wall.

[0059] Subsequently, using a high-speed industrial camera positioned on the transmission side, the photosensitive surface was adjusted to a distance from the axis of the glass tube. This position places the photosensitive surface behind the cylindrical lens.

[0060] It should be noted that the preset monitoring distance The optimal range of values ​​is Within this range, caustic fringes produced by acoustic refractive index modulation exhibit the highest contrast.

[0061] Finally, using a high-speed industrial camera and its image acquisition card, the light field modulated by acoustic refractive index is photoelectrically converted to obtain a caustic image containing acoustic refractive index modulation information.

[0062] S2: The caustic image is inverted using the acoustic refractive index distortion metric model to obtain the transient acoustic pressure distribution sequence.

[0063] It should be noted that although caustic images amplify the signal, they are only a superficial representation of light intensity distribution and are easily affected by factors such as ambient light and uneven glass transmittance. To accurately assess the intensity of micro-arc discharge, it is necessary to reconstruct the physical essence causing the changes in light intensity, namely the acoustic pressure field within the glass tube wall.

[0064] Preferably, as an example, the caustic image is inverted using an acoustically induced refractive index distortion metric model to obtain a transient acoustic pressure distribution sequence, including:

[0065] First, construct the residual objective function, which satisfies the following relation:

[0066]

[0067] In the formula, Represents the sampling time; A sequence index representing pixels in a caustic image; Represents the total number of pixels in a caustic image; For the first The caustic image acquired at time 1 The brightness value of the caustic spot at each pixel; For the first The average brightness value of the caustic image acquired at any given time; For the Laplace operator; The first one to be solved Transient acoustic pressure values ​​at the locations corresponding to each pixel; For example, a preset reference sound pressure constant. Take 100 ; For the pre-calibrated optical-acoustic coupling coefficient, For the first The residual objective function at time step 1. This is a model for measuring acoustic refractive index distortion.

[0068] Understandably, the strong second derivative term The image brightness's concavity or curvature was extracted. Caustics are formed due to the intense convergence or divergence of light rays in a localized area. When a section of the pipe wall becomes denser due to sound pressure, it acts as a miniature convex lens, causing light rays to converge and thus producing extremely high second-order derivative values ​​in the image. Therefore, It is the best operator for capturing the lensing effect. Physical constraint terms. This represents the expected value of optical distortion caused by sound pressure. By minimizing... Find a sound pressure distribution This causes the optical distortion that should occur due to this sound pressure distribution to differ from the optical distortion actually observed. Infinitely close. This mathematically enforces the conservation of light and sound.

[0069] Then, by minimizing the residual objective function, the transient acoustic pressure value at the corresponding position of each pixel is obtained, and the sequence of transient acoustic pressure values ​​of all pixels is denoted as the transient acoustic pressure distribution sequence.

[0070] It should be pointed out that, The method for obtaining parameters includes: during the offline calibration stage, using a standard ultrasonic transducer attached to a stationary glass tube wall, applying a standard sound pressure signal of known amplitude. Simultaneously acquire caustic images and calculate their Laplace response intensity. ; calculated through linear regression fitting .

[0071] S3: The singular point skeleton analysis of the transient acoustic pressure distribution sequence is performed using the box dimension method to calculate the micro-inhomogeneity index of the film layer. The micro-inhomogeneity index of the film layer is used to characterize the degree of lattice defects in the film layer caused by micro-arc discharge.

[0072] It should be noted that directly analyzing sound pressure data still presents challenges: the mechanical vibration of the conveyor rollers also generates sound pressure fluctuations. However, mechanical vibration is a low-frequency, large-scale overall motion, and its sound pressure distribution curve is smooth and continuous, topologically close to a straight line. In contrast, micro-arc discharge is a high-frequency, microscopic explosive impact, and the resulting sound pressure field is full of peaks, fractures, and rough structures. The core logic of this step lies in using fractal dimension as a topological filter to filter out large-amplitude smooth fluctuations and keenly capture tiny but complex rough structures, thereby achieving precise signal separation.

[0073] Preferably, as an example, the box-dimensional method is used to perform singular point skeleton analysis on the transient acoustic pressure distribution sequence to calculate the micro-inhomogeneity index of the film layer, including:

[0074] First, using the threshold extraction method, high-pressure singularities in the transient acoustic pressure distribution sequence are identified, and an acoustic pressure ridge skeleton is constructed.

[0075] Next, using a series of grids of different scales, the acoustic pressure ridge skeleton is covered statistically to obtain the number of grids required to cover the acoustic pressure singularities in the transient acoustic pressure distribution sequence.

[0076] It should be noted that the preset range of grid scale values ​​is [range missing]. Data points.

[0077] Finally, the micro-inhomogeneity index of the film layer is calculated, which satisfies the following relationship:

[0078]

[0079] In the formula, Indicates time The microscopic non-uniformity index of the film layer; This represents the grid scale in box dimension calculation; Indicates at time The grid scale is The number of grids required to cover the sound pressure singularities in a transient acoustic pressure distribution sequence; This represents the reference fractal dimension caused by purely mechanical transmission; This represents the sono-induced lattice defect conversion coefficient.

[0080] Understandably, the limiting term This is the standard value for the box-dimensionality. It measures the self-similarity filling ability of a signal at different scales. If the sound pressure level curve is very smooth, it will look like a line no matter how much it is magnified. The grid number is inversely proportional to the scale, and the ratio approaches 1 / 2. If the sound pressure level curve is full of spikes, more details will be revealed when magnified, including the number of grid lines. The growth rate is much faster than the scaling rate, and the ratio is significantly greater than that of the scale reduction rate. Difference Term It is the core topology filtering logic. This represents the inherent mechanical background dimension of the production line. By subtracting... This eliminates the contribution of mechanical vibration to signal complexity, retaining only the additional roughness caused by the discharge. This is more effective than traditional frequency domain filtering because micro-arc discharge is broadband noise that is difficult to separate in the frequency domain.

[0081] Thus, the above process utilizes the differences in topological properties to accurately eliminate mechanical interference and effectively extract information on the degree of lattice defects in the film layer caused by micro-arc discharge.

[0082] Figure 2 The graph shows the effect of the film micro-inhomogeneity index. The thin solid line curve represents the transient acoustic pressure distribution sequence, which includes large-amplitude low-frequency fluctuations and local high-frequency spikes superimposed on it. The thick dashed line curve represents the background noise baseline identified by the mechanical vibration of the conveyor rollers; this curve exhibits a smooth, continuous fluctuation. The thick dotted-dashed line curve represents the extracted film micro-inhomogeneity index, which remains low for most of the time period, with sharp spikes only appearing at specific moments. The thin dotted straight line represents the quality control warning line set by the process, which is distributed horizontally.

[0083] Comparing the thin solid line curve and the thick dashed line curve reveals that significant fluctuations in the original signal are effectively filtered out and not reflected in the thick dashed line curve; while minute high-frequency spikes in the original signal are accurately extracted and amplified into sudden peaks on the thick dashed line. This demonstrates that the proposed scheme can accurately separate mechanical interference and discharge signals based on differences in the topological dimension of the signal. The peak value of the thick dashed line curve directly corresponds to the timing and intensity of micro-arc discharge, proving that the microscopic inhomogeneity index of the film constructed by this scheme can serve as a reliable basis for quantitatively evaluating the lattice quality of the film.

[0084] It should be pointed out that, Methods for obtaining parameters include:

[0085] Under no-load conditions with the magnetron sputtering target power supply off and the transmission system running normally, a sequence of reference caustic images of the solar collector tube was acquired.

[0086] The reference caustic image sequence is inverted to obtain the reference sound pressure distribution sequence; the fractal dimension of each frame of data in the reference sound pressure distribution sequence is calculated using the box dimension method.

[0087] Calculate the arithmetic mean of the fractal dimensions of all frame data, and use the arithmetic mean as the reference fractal dimension caused by the purely mechanical transmission.

[0088] It should also be noted that the acoustic lattice defect transformation coefficient Methods for obtaining [the information] include:

[0089] A set of glass tube samples that had undergone parasitic discharge of different intensities were obtained;

[0090] Calculate the difference between the calculated fractal dimension of the glass tube sample and the reference fractal dimension caused by the pure mechanical transmission;

[0091] The density of microscopic lattice defects per unit area of ​​the film layer in the glass tube sample was statistically analyzed using a scanning electron microscope.

[0092] Establish a curve showing the relationship between the microscopic lattice defect density and the difference, and use the slope of the curve as the acoustic lattice defect conversion coefficient.

[0093] S4: Obtain the real-time acoustic resonance frequency of the inter-tube cavity formed between adjacent solar collector tubes, and generate airflow pulse control commands to suppress parasitic discharge based on the micro-inhomogeneity index of the film layer and the real-time acoustic resonance frequency of the inter-tube cavity.

[0094] It should be noted that the wedge-shaped gaps formed between adjacent heat collectors acoustically constitute a resonant cavity, which provides crucial standing wave maintenance conditions for the occurrence of parasitic hollow cathode effects. According to the ideal gas velocity formula, the velocity of sound is positively correlated with the square root of thermodynamic temperature, thus the inherent resonant frequency is temperature-dependent. During the magnetron sputtering process, as the ambient temperature inside the vacuum chamber increases, the resonant frequency will inevitably experience thermal drift. If a fixed-frequency interference strategy is adopted, it will lead to a frequency mismatch between the control signal and the physical resonance point, thus failing to effectively disrupt the standing wave conditions. Furthermore, the formation of lattice defects in the film exhibits significant temporal accumulation characteristics. Although the transient damage caused by a weak discharge at a single moment is small, continuous discharge accumulation will lead to irreversible damage to the microstructure of the film. Therefore, the core of this step lies in constructing a dynamic control strategy that integrates real-time temperature compensation and defect integral feedback to ensure that the control command can follow changes in the physical environment in real time to eliminate accumulated errors.

[0095] Preferably, as an example, the real-time acoustic resonance frequency of the inter-tube cavity formed between adjacent solar collector tubes is obtained. Based on the micro-inhomogeneity index of the film layer and the real-time acoustic resonance frequency of the inter-tube cavity, an airflow pulse control command for suppressing parasitic discharge is generated, including:

[0096] First, the real-time absolute temperature at various times is collected using a temperature sensor installed inside the vacuum chamber.

[0097] Next, the real-time acoustic resonance frequency of the inter-tube cavity is calculated based on the real-time absolute temperature. The real-time acoustic resonance frequency satisfies the following relationship:

[0098]

[0099] In the formula, Indicates time The real-time acoustic resonant frequency of the inter-tube cavity; Indicates the cavity mode correction factor; Indicates the adiabatic index of a gas; Represents the gas constant; Indicates time The real-time absolute temperature inside the vacuum chamber; This indicates the distance between adjacent solar collector tubes.

[0100] It is understandable that in the relational expression The formula for the speed of sound of an ideal gas. (Denominator) This represents the half-wavelength condition for a standing wave. Therefore, this relationship, based on the first principles of thermodynamics, establishes the correlation between temperature, speed of sound, and frequency. It ensures that the calculated frequency remains consistent regardless of changes in the vacuum chamber temperature. Always precisely locked at the moment At the acoustic resonance point of the cavity, control failure caused by frequency thermal drift is avoided.

[0101] It should be noted that the methods for obtaining the cavity modal correction factor include:

[0102] In a cold environment in a vacuum chamber, the standing wave air knife is driven by a frequency sweep signal to eject the gas flow.

[0103] The sound pressure response was collected at the gap between the tubes using a microphone, and the physical frequency at which the sound pressure response reached its peak was identified.

[0104] Based on the physical frequency, the tube spacing, and the sound velocity at the initial room temperature, the cavity modal correction factor is calculated. The calculation method for the cavity modal correction factor is as follows:

[0105]

[0106] In the formula, This is the cavity mode correction factor; The physical frequency corresponding to the peak sound pressure response measured by a microphone frequency sweep experiment in a vacuum chamber under cold conditions. The distance between adjacent solar collector tubes; The speed of sound at the initial room temperature.

[0107] Finally, using an integral feedback control algorithm, the micro-inhomogeneity index of the membrane layer at historical moments and the real-time acoustic resonance frequency are modulated to obtain the airflow pulse control command, which specifically satisfies the following relationship:

[0108]

[0109] In the formula, Indicates time The airflow pulse control command; This represents the flow gain coefficient, for example. Take 0.6; This represents the total number of steps in the historical backtracking, for example. Take 30; Representing historical moments The microscopic non-uniformity index of the film layer at that location; This represents the reference fractal dimension caused by purely mechanical transmission; Indicates time The real-time acoustic resonant frequency of the inter-tube cavity; The phase angle is a preset parameter used to adjust the phase difference between the pulsed airflow and the standing wave to achieve destructive interference. Preferably, in this embodiment, the optimal preset value of the phase angle is... At this angle, the airflow pulse and the standing wave can form an antiphase cancellation relationship.

[0110] Understandably, the summation term As an integral controller in control theory, it focuses not only on the current defect but also on the cumulative total of defects over a period of time. This means that even if the discharge at the current moment is very weak, if there have been weak discharges over the past few seconds, the accumulated damage will be significant, and the integral term will affect the control output. Gradually increase until the defect is completely eliminated. This eliminates the steady-state error. Sine modulation term. It is an active interference wave, and its frequency is... By strictly following the aforementioned resonant frequency, the airflow emitted by the air knife is no longer a blind, random blowing, but a sound wave with a specific frequency and phase, thus causing destructive interference with the standing wave in the gap.

[0111] S5: The airflow pulse control command is sent to the execution controller of the standing wave air knife to drive the standing wave air knife to spray pulsed gas into the gap between the tubes in the tube cavity. The pulsed gas is used to block the parasitic hollow cathode effect in the gap between the tubes in real time, so as to realize the online closed-loop control of the uniformity of the magnetron sputtering coating.

[0112] It should be noted that the maintenance of the parasitic hollow cathode effect depends on the formation of a stable plasma standing wave field within the slit. Continuous direct current airflow purging can only dilute the gas macroscopically and is insufficient to disrupt the microscopic standing wave structure, resulting in extremely low efficiency. The core of this step lies in the precise wave-on-wave strike. A pulsed airflow with the same frequency as the standing wave is generated using a fluid bistable oscillator, and its phase is adjusted to make it an antiphase wave. When the pulsed airflow enters the slit, utilizing the principle of destructive wave interference, the high-pressure peak of the airflow precisely fills the low-pressure trough of the standing wave, instantly smoothing out the pressure gradient. This causes the electrons to lose their acceleration mechanism, and the plasma is subsequently extinguished.

[0113] Preferably, as an example, the airflow pulse control command is sent to the execution controller of the standing wave air knife to drive the standing wave air knife to inject pulsed gas into the inter-tube gap at the inter-tube cavity. The pulsed gas is used to block the parasitic hollow cathode effect in the inter-tube gap in real time, so as to achieve online closed-loop control of the uniformity of the magnetron sputtering coating, including:

[0114] First, the introduced process argon gas is subjected to hydrodynamic modulation. The received gas flow pulse control command is converted from digital to analog using an execution controller to obtain a voltage control signal. The voltage control signal is then used to drive a proportional valve to introduce the process argon gas source.

[0115] Next, a fluid bistable oscillator was used to modulate the introduced process argon gas using fluid dynamics. Utilizing the Coanda effect, the main jet was automatically switched at high frequency between the two outlets, resulting in a frequency of [frequency value missing]. Pulsed gas.

[0116] Subsequently, the pulsed gas is injected into the inter-tube cavity using a nozzle aligned with the gap between the tubes, thereby disrupting the gas density standing wave distribution within the gap by utilizing the principle of wave interference.

[0117] Finally, the pulsed gas is used to force the accumulated high-density plasma to diffuse in all directions, thereby eliminating the parasitic hollow cathode effect and achieving online closed-loop control of the uniformity of the magnetron sputtering coating layer of the solar collector tube.

[0118] It is understandable that the frequency resonance characteristics of the pulsed airflow are used to specifically disrupt the standing wave maintenance conditions of the cavity; and the ionization path of gas discharge is interrupted by using fluid dynamics.

[0119] In this way, the parasitic hollow cathode effect was precisely blocked by physical interference, and online closed-loop control of the film uniformity was effectively achieved.

[0120] Figure 3This diagram illustrates the axial uniformity distribution of the film layer in a solar collector tube. The horizontal axis represents the axial position of the solar collector tube, and the vertical axis represents the sequence of the production control cycle. Darker areas in the diagram represent higher film layer micro-uniformity indices, while lighter areas represent lower film layer micro-uniformity indices.

[0121] The thick dashed horizontal line in the diagram marks the time boundary for control strategy intervention. Above the thick dashed line, multiple fixed-position vertical dark stripes are displayed, with light-colored areas between the stripes; below the thick dashed line, the dark stripes rapidly fade and eventually disappear, transforming the area into a uniform light-colored region.

[0122] Before the control strategy was implemented, the distribution of longitudinal dark stripes was observed, which directly reflected that the parasitic hollow cathode effect was maintained by the acoustic standing wave within the inter-tube cavity, with defects concentrated at the antinodes of the standing wave. After the control strategy was implemented, the originally stubborn dark standing wave stripes rapidly faded within a short period of time, transforming into a uniform light color. This strongly demonstrates that the pulsed airflow generated by this scheme successfully disrupted the standing wave conditions within the inter-tube cavity, achieving rapid recovery and self-healing of the film's microscopic uniformity.

[0123] This concludes the embodiment.

[0124] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An intelligent monitoring method for the uniformity of magnetron sputtering coating on solar collector tubes, characterized in that, include: A transmission-type acoustic-optical caustic light field monitoring environment was constructed, and caustic images of solar collector tubes on the coating transmission line were acquired in real time using a line light source and an industrial camera. The transient acoustic pressure distribution sequence was obtained by inverting the caustic image using the acoustic refractive index distortion measurement model. The singular point skeleton analysis of the transient acoustic pressure distribution sequence was performed using the box dimension method, and the micro-inhomogeneity index of the film layer was calculated. The micro-inhomogeneity index of the film layer is used to characterize the degree of lattice defects in the film layer caused by micro-arc discharge. The real-time acoustic resonance frequency of the inter-tube cavity formed between adjacent solar collector tubes is obtained. Based on the micro-inhomogeneity index of the film layer and the real-time acoustic resonance frequency of the inter-tube cavity, an airflow pulse control command for suppressing parasitic discharge is generated. The airflow pulse control command is sent to the execution controller of the standing wave air knife, which drives the standing wave air knife to spray pulsed gas into the gap between the tubes in the tube cavity. The pulsed gas is used to block the parasitic hollow cathode effect in the gap between the tubes in real time, so as to realize the online closed-loop control of the uniformity of the magnetron sputtering coating.

2. The intelligent monitoring method for the uniformity of the magnetron sputtering coating layer of a solar collector tube according to claim 1, characterized in that, The construction of the transmission-type acousto-optic caustic light field monitoring environment includes: A parallel line laser source is set on the backlight side of the solar collector tube, and a high-speed industrial camera is set on the transmission side of the solar collector tube. The photosensitive surface of the high-speed industrial camera is set at a preset monitoring distance from the axis of the solar collector tube, so that the focal plane of the solar collector tube, which acts as a cylindrical lens, is located in front of the photosensitive surface, thereby forming a caustic pattern containing acoustic refractive index modulation information on the photosensitive surface.

3. The intelligent monitoring method for the uniformity of the magnetron sputtering coating layer of a solar collector tube according to claim 2, characterized in that, The preset monitoring distance ranges from 150 mm to 300 mm.

4. The intelligent monitoring method for the uniformity of the magnetron sputtering coating layer of a solar collector tube according to claim 2, characterized in that, The acquisition of the transient acoustic pressure distribution sequence includes: In the formula, Indicates the first The objective function of the residuals at time step; Represents the sampling time; Represents the sequence index of pixels in a caustic image; Represents the total number of pixels in a caustic image; For the first The caustic image acquired at time 1 The brightness value of the caustic spot at each pixel; For the first The average brightness value of the caustic image acquired at any given time; For the Laplace operator; The inversion obtained the first Transient acoustic pressure values ​​at the locations corresponding to each pixel; The preset reference sound pressure constant, The optical-acoustic coupling coefficient is obtained through pre-calibration; Minimize the residual objective function to obtain the transient acoustic pressure value at the corresponding position of each pixel, and denote the sequence of transient acoustic pressure values ​​of all pixels as the transient acoustic pressure distribution sequence.

5. The intelligent monitoring method for the uniformity of the magnetron sputtering coating layer of a solar collector tube according to claim 1, characterized in that, The formula for calculating the micro-inhomogeneity index of the film layer is as follows: In the formula, Indicates time The microscopic non-uniformity index of the film layer; This represents the grid scale in box dimension calculation; Indicates at time The grid scale is The number of grids required to cover the sound pressure singularities in a transient acoustic pressure distribution sequence; This represents the reference fractal dimension caused by purely mechanical transmission; This represents the sono-induced lattice defect conversion coefficient.

6. The intelligent monitoring method for the uniformity of the magnetron sputtering coating layer of a solar collector tube according to claim 5, characterized in that, The method for obtaining the acoustic lattice defect transformation coefficient includes: A set of glass tube samples that had undergone parasitic discharge of different intensities were obtained; Calculate the difference between the calculated fractal dimension of the glass tube sample and the reference fractal dimension caused by the pure mechanical transmission; The density of microscopic lattice defects per unit area of ​​the film layer in the glass tube sample was statistically analyzed using a scanning electron microscope. Establish a curve showing the relationship between the microscopic lattice defect density and the difference, and use the slope of the curve as the acoustic lattice defect conversion coefficient.

7. The intelligent monitoring method for the uniformity of the magnetron sputtering coating layer of a solar collector tube according to claim 5, characterized in that, The method for obtaining the reference fractal dimension caused by the purely mechanical transmission includes: Under no-load conditions with the magnetron sputtering target power supply off and the transmission system running normally, a sequence of reference caustic images of the solar collector tube was acquired. The reference caustic image sequence is inverted to obtain the reference sound pressure distribution sequence; the fractal dimension of each frame of data in the reference sound pressure distribution sequence is calculated using the box dimension method. Calculate the arithmetic mean of the fractal dimensions of all frame data, and use the arithmetic mean as the reference fractal dimension caused by the purely mechanical transmission.

8. The intelligent monitoring method for the uniformity of the magnetron sputtering coating layer of a solar collector tube according to claim 1, characterized in that, The generated airflow pulse control command for suppressing parasitic discharge satisfies the following relationship: In the formula, Indicates time The airflow pulse control command; Indicates the flow gain coefficient; This represents the total number of steps in the historical backtracking process; Representing historical moments The microscopic non-uniformity index of the film layer at that location; This represents the reference fractal dimension caused by purely mechanical transmission; Indicates time The real-time acoustic resonant frequency of the inter-tube cavity; This indicates the phase angle.

9. The intelligent monitoring method for the uniformity of the magnetron sputtering coating layer of a solar collector tube according to claim 8, characterized in that, The method for obtaining the real-time acoustic resonance frequency of the inter-tube cavity is as follows: In the formula, Indicates time The real-time acoustic resonant frequency of the inter-tube cavity; Indicates the cavity mode correction factor; Indicates the adiabatic index of a gas; Represents the gas constant; Indicates time The real-time absolute temperature inside the vacuum chamber; This indicates the distance between adjacent solar collector tubes.

10. The intelligent monitoring method for the uniformity of the magnetron sputtering coating layer of a solar collector tube according to claim 9, characterized in that, The method for obtaining the cavity modal correction factor includes: In a cold environment in a vacuum chamber, the standing wave air knife is driven by a frequency sweep signal to eject the gas flow. The sound pressure response was collected at the gap between the tubes using a microphone, and the physical frequency at which the sound pressure response reached its peak was identified. The cavity mode correction factor is calculated based on the physical frequency, the tube spacing, and the sound velocity at the initial room temperature.

Images