A PCB copper electroplating uniformity dynamic compensation control method and system

Abstract

This invention relates to the field of circuit board manufacturing technology, specifically a method and system for dynamic compensation control of copper layer uniformity in PCB electroplating. The method includes: constructing a topological mapping relationship between the anode block spatial layout and the spatial frequency response range; identifying low-frequency contour compensation groups and high-frequency detail compensation groups; simultaneously applying multi-frequency micro-perturbation electrochemical impedance spectroscopy measurements to calculate the polarization correlation factor; using the polarization correlation factor to correct the current-deposition transfer gain and combining it with spatial frequency domain deviation spectrum inversion to retrieve dynamic compensation current waveform commands; monitoring the characteristic frequency amplitude attenuation slope and feeding back the time-varying parameters of the polarization correlation factor to complete online collaborative self-correction. This invention achieves targeted collaborative compensation for large-area contour deviations and local detail deviations, and possesses the ability to adaptively offset process state drift during long-cycle operation, improving the control accuracy of plating uniformity throughout the entire electroplating cycle.

Description

Technical Field

[0001] This invention relates to the field of circuit board manufacturing technology, and in particular to a method and system for dynamic compensation control of the uniformity of copper plating in PCBs. Background Technology

[0002] In PCB copper plating, the uniformity of the cathode plating thickness is a key indicator determining the conductivity and reliability of the printed circuit board. Current technologies typically control plating uniformity by adjusting the total output current of the plating bath, adjusting the anode-cathode spacing, or adding auxiliary baffles. These methods are mostly static or open-loop macroscopic compensation approaches. As printed circuit boards become increasingly high-density and have finer lines, plating thickness deviations exhibit complex spatial distributions, with both local undulations and overall tilt. Traditional methods lack effective analysis and targeted suppression of the spatial frequency components of plating thickness deviations, making it difficult to simultaneously compensate for both low-frequency large-area contour deviations and high-frequency local detail deviations. Some existing solutions attempt to introduce electrochemical impedance spectroscopy (EIS) measurements to sense changes in the plating solution state. However, impedance spectral characteristics are usually only used as a reference for adjusting macroscopic process parameters and fail to establish a quantitative correlation with the spatial frequency distribution of plating thickness deviations. Furthermore, they cannot couple the dynamic changes in the plating solution polarization state to the compensation current distribution of the anode blocks in real time. Furthermore, existing compensation approaches based on spatial frequency decomposition often employ fixed frequency band divisions and static anode mapping relationships, which cannot adapt to the drift of plating solution impedance characteristics and anode response capability over time during the electroplating process. This leads to accuracy degradation in long-cycle continuous production. Therefore, to address the shortcomings of the existing technology, this invention proposes a dynamic compensation control method and system for the uniformity of copper plating layers in PCB electroplating. Summary of the Invention

[0003] This invention overcomes the shortcomings of the prior art and provides a method and system for dynamic compensation control of the uniformity of PCB electroplated copper layer.

[0004] To achieve the above objectives, the technical solution adopted by the present invention is as follows: The first aspect of this invention discloses a dynamic compensation control method for the uniformity of copper plating in PCBs, comprising: S1. Real-time acquisition of coating thickness distribution on cathode plate surface, conversion of coating thickness distribution into spatial frequency domain deviation spectrum through two-dimensional spatial frequency domain decomposition, and extraction of the characteristic frequency component to be suppressed that constitutes the dominant deviation of uniformity from the spatial frequency domain deviation spectrum. S2. Based on the spatial layout of each anode block, construct the topological mapping relationship between each anode block and different spatial frequency response ranges, and dynamically identify the anode blocks as low-frequency contour compensation group and high-frequency detail compensation group according to the extracted characteristic frequency components to be suppressed. S3. Simultaneously apply multi-frequency micro-perturbation electrochemical impedance spectroscopy measurement during the electroplating process, extract the real-time impedance spectrum characteristic parameters of the plating solution, and calculate the polarization correlation factor related to the spatial frequency based on the impedance spectrum characteristic parameters. S4. Using the polarization correlation factor, the current-deposition transfer gain of the anode block to be suppressed characteristic frequency components is corrected, and combined with the spatial frequency domain deviation spectrum inversion, dynamic compensation current waveform commands are generated for the low-frequency contour compensation group and the high-frequency detail compensation group respectively. S5. Continuously monitor the attenuation slope of the characteristic frequency amplitude in the spatial frequency domain deviation spectrum, and adjust the time-varying parameter in the polarization correlation factor according to the feedback of the attenuation slope to complete the online collaborative self-correction of impedance spectrum characteristics and spatial frequency topology mapping.

[0005] Preferably, S1 specifically comprises: Real-time acquisition of instantaneous measurements of coating thickness at discrete sampling points on the cathode plate surface to construct the original thickness distribution matrix in a two-dimensional spatial domain; The original thickness distribution matrix is ​​subjected to two-dimensional high-frequency texture envelope filtering processing, namely, morphological opening operation is performed along the transfer direction of the plated part and the transverse direction of the plate surface to extract the inherent non-defect fine texture envelope surface of the plate surface, and the original thickness distribution matrix is ​​subtracted from the texture envelope surface to obtain the substrate thickness deviation distribution field after removing the substrate texture interference. A two-dimensional spatial Fourier transform is performed on the substrate thickness deviation distribution field to map the thickness deviation information in the spatial location domain to the spatial frequency domain, generating a global spatial frequency domain deviation spectrum characterized by the amplitude and phase of different spatial frequency components. The standard spatial frequency amplitude envelope sequence stored in the historical database is retrieved, and statistical generalization processing is performed on each standard spatial frequency amplitude envelope sequence to fit and generate a historical stationary generalization boundary benchmark that characterizes the allowable range of normal sedimentary fluctuations. The real-time amplitude of each spatial frequency component in the global spatial frequency domain deviation spectrum is compared with the historical stationary generalization boundary benchmark frequency by frequency. Spatial frequency components whose real-time amplitude continuously exceeds the historical stationary generalization boundary benchmark are identified and marked as abnormal frequency domain deviation feature clusters. The frequency component with the largest amplitude proportion is selected from the abnormal frequency domain deviation feature clusters as the characteristic frequency component to be suppressed that constitutes the dominant deviation of uniformity.

[0006] Preferably, S2 specifically comprises: The spatial geometric center coordinates and boundary contour dimensions of each independent controllable anode block in the electroplating tank are obtained. Based on the boundary contour dimensions of each anode block, electric field edge diffusion envelope modeling is adopted to generate the current density diffusion envelope field of each anode block on the cathode plate. The current density diffusion envelope field characterizes the spatial coverage of the current excitation of each position on the cathode plate by the anode block. Two-dimensional spatial Fourier transforms are performed on the current density diffusion envelope field of each anode block to extract the frequency response bandwidth and main response band of each anode block in the spatial frequency domain, and a spatial frequency excitation response coupling scale between each anode block and different spatial frequency response ranges is established. Based on the frequency values ​​of the extracted characteristic frequency components to be suppressed, frequency matching search is performed in the spatial frequency excitation response coupling scale to screen out the anode blocks that generate response coupling with the characteristic frequency components to be suppressed, forming a set of candidate compensation anode blocks. Calculate the low-frequency-high-frequency dominant deviation vector of the slow-changing deviation characteristic frequency and the fluctuating deviation characteristic frequency in the characteristic frequency component to be suppressed for each anode block in the candidate compensation anode block set; Based on the response dominance trend characterized by the low-frequency to high-frequency response dominance deviation vector, the anode blocks are dynamically identified as a low-frequency profile compensation group and a high-frequency detail compensation group.

[0007] Specifically, the low-frequency to high-frequency response dominant deviation vector of the slow-changing deviation characteristic frequency and the fluctuating deviation characteristic frequency in the characteristic frequency components to be suppressed is calculated for each anode block in the candidate compensation anode block set. From the characteristic frequency components to be suppressed, the slow-change deviation characteristic frequency, which characterizes the overall thickness tilt trend of the coating, and the fluctuation deviation characteristic frequency, which characterizes the local coating undulation, are separated according to the spatial frequency value. The current density diffusion envelope field of each anode block in the candidate compensation anode block set is retrieved. Low-frequency driving induction excitation corresponding to the slow-changing deviation characteristic frequency and high-frequency driving induction excitation corresponding to the fluctuation deviation characteristic frequency are applied to the current density diffusion envelope field of each anode block respectively. The first driving induction coupled electromagnetic field harmonic response spectrum of each anode block under low-frequency driving induction excitation and the second driving induction coupled electromagnetic field harmonic response spectrum under high-frequency driving induction excitation are obtained. The main peak of the anode response amplitude and the anode response phase shift under low-frequency induction excitation are extracted from the harmonic response spectrum of the first induction coupled electromagnetic field, and the secondary peak of the anode response amplitude and the anode response phase shift under high-frequency induction excitation are extracted from the harmonic response spectrum of the second induction coupled electromagnetic field. The amplitude ratio of the main peak and the secondary peak of the anode response amplitude of the same anode block is normalized to generate the driving-induction dominant amplitude ratio of the corresponding anode block. The driving-induction dominant amplitude ratio is then calculated with the anode response phase offset to generate the frequency-phase vector value that characterizes the dominant bias of the anode block's response to the slow-changing deviation characteristic frequency and the fluctuation deviation characteristic frequency. The frequency phase vector values ​​of each candidate compensation anode block are normalized by performing deviation calculation to generate a numerical low-frequency to high-frequency response dominant deviation vector for each anode block, so as to characterize the degree of response bias of the anode block to slow-changing deviation or fluctuating deviation.

[0008] Preferably, S3 specifically comprises: During the cathode deposition phase of the electroplating process, a multi-frequency perturbation excitation signal composed of multiple sinusoidal perturbation currents of different frequencies is simultaneously injected into the plating solution through anode blocks, and the potential response signal of each sensing area on the cathode plate is simultaneously acquired. The potential response signal and the multi-frequency perturbation excitation signal are decoupled in the frequency domain to establish a multi-frequency domain impedance complex spectrum characterizing the impedance amplitude and phase relationship of the plating solution system at multiple frequencies. The cathode plate is divided into multiple local micro-elements of electrode surface reaction according to the spatial resolution that matches the spatial frequency value in the characteristic frequency component to be suppressed. The impedance amplitude and phase of each frequency point in the multi-frequency domain impedance complex spectrum are mapped and distributed to each local micro-element of electrode surface reaction according to the spatial distribution of the potential response signal of each sensing area of ​​the cathode plate, thus forming the local relaxation characteristic parameter spectrum of each local micro-element of electrode surface reaction at multiple frequencies. Spatial frequency-relaxation feature correlation analysis is performed on the local relaxation feature parameter spectrum of the local reaction local micro-element on each electrode surface. That is, the local relaxation feature parameter spectrum of the local reaction local micro-element on each electrode surface is arranged in a spatial sequence along the spatial direction of the cathode plate surface, and the fluctuation features of the local relaxation feature parameter as the spatial position changes are extracted to generate the spatial frequency-relaxation feature correlation density distribution. From the spatial frequency-relaxation feature correlation density distribution, query the relaxation feature density value that matches the spatial frequency value in the feature frequency component to be suppressed. Perform bias normalization operation on the relaxation feature density value and the base impedance reference value of the plating solution to generate the polarization correlation factor associated with the spatial frequency.

[0009] Preferably, S4 specifically comprises: Retrieve the generated polarization correlation factor, and perform gain bias calibration on the polarization correlation factor and the basic current-deposition transfer gain coefficient of the anode block in the unpolarized state. That is, use the polarization correlation factor as a gain bias correction term to compensate the bias of the basic current-deposition transfer gain coefficient, and generate the dynamic current-deposition transfer gain coefficient of the anode block after polarization correction. From the spatial frequency domain deviation spectrum, based on the frequency values ​​in the characteristic frequency components to be suppressed, the deviation amplitude and deviation phase at the deviation frequency are extracted, and the deviation amplitude and deviation phase are used to construct a two-dimensional component matrix of the spatial spectrum at the deviation frequency. The deviation amplitude and deviation phase corresponding to the slowly changing deviation characteristic frequency in the two-dimensional component matrix of the spatial spectrum at the deviation frequency are used as the low-frequency contour compensation amplitude table and the low-frequency contour compensation phase reference. Through frequency-guided signal waveform synthesis, a low-frequency contour envelope wave with the low-frequency contour compensation amplitude table as the amplitude envelope and the low-frequency contour compensation phase reference as the phase reference is generated. Meanwhile, the deviation amplitude and deviation phase corresponding to the wave-like deviation characteristic frequency in the two-dimensional component matrix of the spatial spectrum at the deviation frequency are decomposed into frequency bands with the main response frequency band of the corresponding anode block in the spatial frequency excitation response coupling scale as the center frequency band. The wavelet components of each frequency band are then synthesized into high-frequency detail ripple primitives. The low-frequency contour envelope wave and the high-frequency detail ripple primitive are waveform-shaped and superimposed, and the superimposed waveform is amplitude-distributed by the dynamic current-deposition transfer gain coefficient to generate dynamic compensation current waveform commands that act on the low-frequency contour compensation group and the high-frequency detail compensation group respectively.

[0010] Preferably, S5 specifically comprises: Within multiple consecutive monitoring time windows after the execution of the dynamic compensation current waveform command, the coating thickness distribution on the cathode plate is successively acquired and converted into a spatial frequency domain deviation spectrum. The characteristic frequency amplitude corresponding to the characteristic frequency component to be suppressed is extracted from the spatial frequency domain deviation spectrum of each conversion, and the evolution trend curve of the characteristic frequency amplitude with the attenuation slope of the monitoring time window is generated. The decay slope evolution trend curve is dynamically time-warped and matched with the multi-scale historical decay slope trend primitive library pre-stored in the database to identify time-varying perturbation features in the decay slope evolution trend curve that deviate from the historical decay slope trend primitive library. Based on the identified time-varying perturbation features, the spatial frequency-relaxation feature correlation density distribution is successively time-delayed tracked to extract the relaxation feature density drift trend associated with the time-varying perturbation features in the relaxation feature density distribution. The relaxation characteristic density drift direction is decoupled from the polarization correlation factor by time-varying parameters, and the time-varying relaxation correction term affected by the gradual change of the polarization state of the plating solution is extracted from the polarization correlation factor. Using the attenuation slope change amplitude of each monitoring time window in the attenuation slope evolution trend curve as the adjustment step size benchmark, the amplitude of the time-varying relaxation correction term is updated by recursion, and the updated time-varying relaxation correction term is fed back to the generation stage of spatial frequency-relaxation feature correlation density distribution to complete the coordinated time-delay adaptive adjustment of impedance spectrum features and spatial frequency excitation response coupling scale.

[0011] The second aspect of the present invention discloses a dynamic compensation control system for the uniformity of copper plating in PCB, comprising a memory and a processor. The memory stores a program for a dynamic compensation control method for the uniformity of copper plating in PCB. When the program for the dynamic compensation control method for the uniformity of copper plating in PCB is executed by the processor, the steps of the method described in any one of the present invention are implemented.

[0012] This invention addresses the technical deficiencies in the prior art and has the following beneficial effects: This invention organically integrates spatial frequency domain decomposition of coating thickness deviation, multi-frequency micro-perturbation electrochemical impedance spectroscopy measurement, anode block spatial frequency response mapping, and time-varying parameter adaptive adjustment into a closed-loop control system. This fundamentally solves the problem of the disconnect between plating solution polarization state sensing and spatial selective compensation in existing technologies. It can accurately identify the characteristic frequency components that currently dominate the uniformity deviation from the spatial distribution of coating thickness, and dynamically classify compensation groups based on the response capabilities of each anode block to different spatial frequencies, thereby enabling targeted suppression of large-area contour deviations and local detail deviations. Simultaneously, the polar... The introduction of the polarization correlation factor quantitatively couples the real-time polarization state of the plating solution to the correction link of the current-deposition transfer gain, enabling the generation of the compensation current waveform command to adaptively adjust with the changes in the electrochemical state of the plating solution, avoiding the accuracy decay of the fixed mapping relationship during long-term operation; the online collaborative self-correction mechanism uses the characteristic frequency amplitude attenuation slope as feedback basis to continuously adjust the correlation between the polarization correlation factor and the spatial frequency topological mapping, so that the system can maintain stable compensation accuracy under the conditions of plating solution composition evolution and anode state drift, and finally achieves continuous and precise control of coating uniformity throughout the entire electroplating cycle. Attached Figure Description

[0013] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other embodiments can be obtained from these drawings without creative effort.

[0014] Figure 1 This is a flowchart of the dynamic compensation control method for the uniformity of the copper plating layer in this PCB. Figure 2 This is a diagram of the architecture of the dynamic compensation control system for the uniformity of the copper plating layer on this PCB. Detailed Implementation

[0015] To better understand the above-mentioned objectives, features, and advantages of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in these embodiments can be combined with each other.

[0016] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and therefore the scope of protection of the invention is not limited to the specific embodiments disclosed below.

[0017] like Figure 1As shown, the first aspect of this invention discloses a dynamic compensation control method for the uniformity of copper plating in PCBs, comprising: S1. Real-time acquisition of coating thickness distribution on cathode plate surface, conversion of coating thickness distribution into spatial frequency domain deviation spectrum through two-dimensional spatial frequency domain decomposition, and extraction of the characteristic frequency component to be suppressed that constitutes the dominant deviation of uniformity from the spatial frequency domain deviation spectrum. S2. Based on the spatial layout of each anode block, construct the topological mapping relationship between each anode block and different spatial frequency response ranges, and dynamically identify the anode blocks as low-frequency contour compensation group and high-frequency detail compensation group according to the extracted characteristic frequency components to be suppressed. S3. Simultaneously apply multi-frequency micro-perturbation electrochemical impedance spectroscopy measurement during the electroplating process, extract the real-time impedance spectrum characteristic parameters of the plating solution, and calculate the polarization correlation factor related to the spatial frequency based on the impedance spectrum characteristic parameters. S4. Using the polarization correlation factor, the current-deposition transfer gain of the anode block to be suppressed characteristic frequency components is corrected, and combined with the spatial frequency domain deviation spectrum inversion, dynamic compensation current waveform commands are generated for the low-frequency contour compensation group and the high-frequency detail compensation group respectively. S5. Continuously monitor the attenuation slope of the characteristic frequency amplitude in the spatial frequency domain deviation spectrum, and adjust the time-varying parameter in the polarization correlation factor according to the feedback of the attenuation slope to complete the online collaborative self-correction of impedance spectrum characteristics and spatial frequency topology mapping.

[0018] Preferably, S1 specifically comprises: Real-time acquisition of instantaneous measurements of coating thickness at discrete sampling points on the cathode plate surface to construct the original thickness distribution matrix in a two-dimensional spatial domain; The original thickness distribution matrix is ​​subjected to two-dimensional high-frequency texture envelope filtering processing, namely, morphological opening operation is performed along the transfer direction of the plated part and the transverse direction of the plate surface to extract the inherent non-defect fine texture envelope surface of the plate surface, and the original thickness distribution matrix is ​​subtracted from the texture envelope surface to obtain the substrate thickness deviation distribution field after removing the substrate texture interference. A two-dimensional spatial Fourier transform is performed on the substrate thickness deviation distribution field to map the thickness deviation information in the spatial location domain to the spatial frequency domain, generating a global spatial frequency domain deviation spectrum characterized by the amplitude and phase of different spatial frequency components. The standard spatial frequency amplitude envelope sequence stored in the historical database is retrieved, and statistical generalization processing is performed on each standard spatial frequency amplitude envelope sequence to fit and generate a historical stationary generalization boundary benchmark that characterizes the allowable range of normal sedimentary fluctuations. The real-time amplitude of each spatial frequency component in the global spatial frequency domain deviation spectrum is compared with the historical stationary generalization boundary benchmark frequency by frequency. Spatial frequency components whose real-time amplitude continuously exceeds the historical stationary generalization boundary benchmark are identified and marked as abnormal frequency domain deviation feature clusters. The frequency component with the largest amplitude proportion is selected from the abnormal frequency domain deviation feature clusters as the characteristic frequency component to be suppressed that constitutes the dominant deviation of uniformity.

[0019] It should be noted that the PCB substrate surface itself has subtle texture undulations unrelated to the electroplating process. If subsequent analysis is performed directly based on the original measurements, these inherent textures will be superimposed on the actual unevenness of the plating layer, leading to inaccurate extraction of deviation features. Therefore, this invention constructs a two-dimensional high-frequency texture envelope by performing morphological opening operations along the plating transfer direction and the orthogonal lateral direction of the board surface. This envelope is essentially a lower envelope formed by tracing local minima in the original thickness distribution matrix, which can approximate the inherent non-defective undulations of the substrate itself. Subtracting the original thickness distribution matrix from the texture envelope yields a substrate thickness deviation distribution field that effectively removes the substrate texture components; its numerical changes primarily reflect the unevenness of the plating layer's deposition.

[0020] Performing a two-dimensional spatial Fourier transform on the substrate thickness deviation distribution field physically maps the thickness deviation distribution, originally with two-dimensional spatial coordinates as independent variables, from the spatial location domain to the spatial frequency domain. In the resulting global spatial frequency domain deviation spectrum, each spatial frequency component contains two parameters: amplitude and phase. The amplitude characterizes the strength of the fluctuation feature corresponding to that spatial frequency in the coating deviation, while the phase characterizes the spatial offset position of that fluctuation feature on the board surface. Thus, periodic deviation patterns that were difficult to directly discern in the spatial domain can be clearly presented in the frequency domain. The standard spatial frequency amplitude envelope sequence in the historical database refers to the set of envelope curves formed by the amplitudes of each spatial frequency component in the spatial frequency domain deviation spectrum of each batch of coated parts, collected and stored under normal deposition processes where coating uniformity has been repeatedly confirmed. Each envelope curve represents the amplitude distribution profile of a normal batch on the spatial frequency axis, and they collectively reflect the statistical range of allowable amplitude fluctuations at each spatial frequency under normal process conditions. Statistical generalization processing is performed on the amplitude envelope sequences of each standard spatial frequency. This involves statistically fitting the amplitudes of multiple normal batches at the same spatial frequency point to obtain a historical stable generalization boundary benchmark characterizing the upper and lower limits of normal deposition fluctuations. The amplitudes of each frequency component in the currently obtained global spatial frequency domain deviation spectrum are compared one by one with this boundary benchmark. Frequency components whose amplitudes significantly and persistently exceed the boundary benchmark are considered abnormal frequency domain deviation feature clusters that deviate significantly from normal process conditions. The frequency component with the largest amplitude proportion is selected as the dominant deviation in coating uniformity at the current moment and needs to be suppressed, thus providing a clear frequency domain suppression target for targeted compensation of anode blocks.

[0021] Preferably, S2 specifically comprises: The spatial geometric center coordinates and boundary contour dimensions of each independent controllable anode block in the electroplating tank are obtained. Based on the boundary contour dimensions of each anode block, electric field edge diffusion envelope modeling is adopted to generate the current density diffusion envelope field of each anode block on the cathode plate. The current density diffusion envelope field characterizes the spatial coverage of the current excitation of each position on the cathode plate by the anode block. Two-dimensional spatial Fourier transforms are performed on the current density diffusion envelope field of each anode block to extract the frequency response bandwidth and main response band of each anode block in the spatial frequency domain, and a spatial frequency excitation response coupling scale between each anode block and different spatial frequency response ranges is established. Based on the frequency values ​​of the extracted characteristic frequency components to be suppressed, frequency matching search is performed in the spatial frequency excitation response coupling scale to screen out the anode blocks that generate response coupling with the characteristic frequency components to be suppressed, forming a set of candidate compensation anode blocks. Calculate the low-frequency-high-frequency dominant deviation vector of the slow-changing deviation characteristic frequency and the fluctuating deviation characteristic frequency in the characteristic frequency component to be suppressed for each anode block in the candidate compensation anode block set; Based on the response dominance trend characterized by the low-frequency to high-frequency response dominance deviation vector, the anode blocks are dynamically identified as a low-frequency profile compensation group and a high-frequency detail compensation group.

[0022] It should be noted that for each independently controllable anode block in the electroplating tank, the current excitation does not only act on the area corresponding to the orthographic projection of the cathode plate, but extends in a diffuse manner from the geometric boundary of the anode block to the surrounding area, and the diffusion range is closely related to the boundary contour size of the anode block. Therefore, this invention uses the boundary contour size of each anode block as an input parameter and employs electric field edge diffusion envelope modeling. Specifically, it uses the geometric contour of the anode block as the boundary of the electric field emission source, and combines the edge diffusion attenuation characteristics of the electrolyte current field in the plating medium to calculate the current density distribution formed by the diffusion of each anode block to various positions on the cathode plate along the cathode plate direction. This generates a current density diffusion envelope field on the cathode plate surface for each anode block. This envelope field depicts the current excitation coverage intensity of the anode block at various coordinate points on the cathode plate surface, with the highest density in the anode geometric center projection area, gradually decreasing towards the edge.

[0023] After obtaining the current density diffusion envelope field of each anode block, a two-dimensional spatial Fourier transform is performed on each envelope field. The physical meaning of the transform is to map the spatial current excitation distribution of the anode block from the spatial domain to the spatial frequency domain, obtaining the spatial frequency range and intensity distribution of the influence that the anode block can exert. Two key parameters are extracted from the transform results: one is the frequency response bandwidth, which is the span of the spatial frequency range in which the anode block can generate an effective current excitation response; the other is the main response band, which is the spatial frequency segment where the peak response intensity is located. After calibrating the frequency response bandwidth and main response band of each anode block to their spatial location, a spatial frequency excitation response coupling scale between each anode block and different spatial frequency response ranges is established. This scale clarifies the response coverage range of each anode block on the spatial frequency axis.

[0024] Subsequently, in the group identification stage, based on the specific frequency values ​​of the identified characteristic frequency components to be suppressed, frequency matching retrieval is performed on the spatial frequency excitation response coupling scale to screen out anode blocks with effective response coupling to that frequency component, forming a candidate compensation anode block set. For each anode block in the candidate set, the low-frequency-high-frequency response dominant deviation vector of the two different types of deviation frequencies in the characteristic frequency components to be suppressed (i.e., the slowly changing deviation characteristic frequency reflecting the overall tilt trend of the coating and the fluctuating deviation characteristic frequency reflecting local undulations) is further calculated. The magnitude and direction of the deviation vector intuitively characterize the degree of bias between the low-frequency compensation capability and the high-frequency compensation capability of the anode block. For example, if the main response frequency band of a certain anode block is concentrated in the low-frequency range and the dominant deviation vector of the low-frequency-high-frequency response points significantly towards the low-frequency direction, it indicates that the anode block is more suitable for undertaking large-area contour compensation tasks and should be classified into the low-frequency contour compensation group. Conversely, if the main response frequency band of another anode block covers the higher-frequency range and the deviation vector points towards the high-frequency direction, it indicates that it is more suitable for compensating for local fine undulations and should be classified into the high-frequency detail compensation group.

[0025] In this way, each anode block can be dynamically allocated to different compensation groups in each control cycle according to its response matching with the current frequency component of the deviation to be suppressed, thus realizing the adaptive spatial frequency division of compensation resources.

[0026] Specifically, the low-frequency to high-frequency response dominant deviation vector of the slow-changing deviation characteristic frequency and the fluctuating deviation characteristic frequency in the characteristic frequency components to be suppressed is calculated for each anode block in the candidate compensation anode block set. From the characteristic frequency components to be suppressed, the slow-change deviation characteristic frequency, which characterizes the overall thickness tilt trend of the coating, and the fluctuation deviation characteristic frequency, which characterizes the local coating undulation, are separated according to the spatial frequency value. The current density diffusion envelope field of each anode block in the candidate compensation anode block set is retrieved. Low-frequency driving induction excitation corresponding to the slow-changing deviation characteristic frequency and high-frequency driving induction excitation corresponding to the fluctuation deviation characteristic frequency are applied to the current density diffusion envelope field of each anode block respectively. The first driving induction coupled electromagnetic field harmonic response spectrum of each anode block under low-frequency driving induction excitation and the second driving induction coupled electromagnetic field harmonic response spectrum under high-frequency driving induction excitation are obtained. The main peak of the anode response amplitude and the anode response phase shift under low-frequency induction excitation are extracted from the harmonic response spectrum of the first induction coupled electromagnetic field, and the secondary peak of the anode response amplitude and the anode response phase shift under high-frequency induction excitation are extracted from the harmonic response spectrum of the second induction coupled electromagnetic field. The amplitude ratio of the main peak and the secondary peak of the anode response amplitude of the same anode block is normalized to generate the driving-induction dominant amplitude ratio of the corresponding anode block. The driving-induction dominant amplitude ratio is then calculated with the anode response phase offset to generate the frequency-phase vector value that characterizes the dominant bias of the anode block's response to the slow-changing deviation characteristic frequency and the fluctuation deviation characteristic frequency. The frequency phase vector values ​​of each candidate compensation anode block are normalized by performing deviation calculation to generate a numerical low-frequency to high-frequency response dominant deviation vector for each anode block, so as to characterize the degree of response bias of the anode block to slow-changing deviation or fluctuating deviation.

[0027] It should be noted that the technical meaning of the slow-change deviation characteristic frequency is as follows: In the spatial frequency domain deviation spectrum of the coating thickness, the lower spatial frequency components correspond to a thickness fluctuation pattern with a large plate surface scale and gradual changes. This is manifested as a gradual thickening or thinning trend of the coating from one side to the other across the entire plate, hence the term "slow-change deviation characteristic frequency." Conversely, the higher spatial frequency components correspond to a rapidly fluctuating thickness fluctuation pattern in a localized area of ​​the plate surface, exhibiting a densely distributed peak-valley deviation, i.e., a fluctuating deviation characteristic frequency. Classification is based on spatial frequency thresholds. For example, frequency components with spatial frequency values ​​below 0.05 cycles per millimeter can be classified as slow-change deviation characteristic frequencies, while those above this value can be classified as fluctuating deviation characteristic frequencies. The specific threshold is determined comprehensively based on the plate size and historical deviation statistics. After obtaining the current density diffusion envelope field of each anode block in the candidate compensation anode block set, two types of driving excitations were applied to this envelope field: one was a quasi-steady-state low-frequency driving excitation with the slowly varying deviation characteristic frequency as the center frequency and the excitation amplitude bandwidth covering the range adjacent to this frequency, aiming to stimulate the electromagnetic response characteristics of the anode block in low-frequency compensation mode; the other was a pulsed high-frequency driving excitation with the fluctuating deviation characteristic frequency as the center frequency and the excitation amplitude bandwidth covering this frequency and its harmonic range, aiming to stimulate the electromagnetic response characteristics of the anode block in high-frequency compensation mode. The first driving-coupled electromagnetic field harmonic response spectrum and the second driving-coupled electromagnetic field harmonic response spectrum obtained after applying the excitation recorded the change curves of the response amplitude of the anode block with frequency and the corresponding phase shift under low-frequency and high-frequency excitation conditions, respectively.

[0028] After extracting the main peak and secondary peak of the anode response amplitude, as well as their respective anode response phase offsets, from the two sets of harmonic response spectra, a frequency-phase vector calculation is performed: the ratio of the dominant amplitude of the induction and the anode response phase offset of the same anode block are respectively regarded as the horizontal and vertical coordinate components in a two-dimensional evaluation plane. Together, they form a two-dimensional vector pointing from the origin to this coordinate point. The length of the vector reflects the significance of the response dominance, while the direction of the vector reflects its tendency to favor low-frequency or high-frequency responses. Subsequently, a deviation standardization operation is performed on the frequency-phase vector values ​​of each candidate compensation anode block: the maximum and minimum values ​​of the frequency-phase vector values ​​of all candidate blocks are taken, and the difference between the frequency-phase vector value and the minimum value of each block is divided by the difference between the maximum and minimum values. This maps the response bias of each block to a unified numerical range, thereby eliminating the influence of absolute value differences on grouping judgment. The numerical result generated after deviation standardization is the low-frequency-high-frequency response dominant deviation vector. If its value approaches one end of the interval, it indicates that the anode block tends to compensate for slow-change deviation, while if it approaches the other end, it indicates that it tends to compensate for volatile deviation.

[0029] Preferably, S3 specifically comprises: During the cathode deposition phase of the electroplating process, a multi-frequency perturbation excitation signal composed of multiple sinusoidal perturbation currents of different frequencies is simultaneously injected into the plating solution through anode blocks, and the potential response signal of each sensing area on the cathode plate is simultaneously acquired. The potential response signal and the multi-frequency perturbation excitation signal are decoupled in the frequency domain to establish a multi-frequency domain impedance complex spectrum characterizing the impedance amplitude and phase relationship of the plating solution system at multiple frequencies. The cathode plate is divided into multiple local micro-elements of electrode surface reaction according to the spatial resolution that matches the spatial frequency value in the characteristic frequency component to be suppressed. The impedance amplitude and phase of each frequency point in the multi-frequency domain impedance complex spectrum are mapped and distributed to each local micro-element of electrode surface reaction according to the spatial distribution of the potential response signal of each sensing area of ​​the cathode plate, thus forming the local relaxation characteristic parameter spectrum of each local micro-element of electrode surface reaction at multiple frequencies. Spatial frequency-relaxation feature correlation analysis is performed on the local relaxation feature parameter spectrum of the local reaction local micro-element on each electrode surface. That is, the local relaxation feature parameter spectrum of the local reaction local micro-element on each electrode surface is arranged in a spatial sequence along the spatial direction of the cathode plate surface, and the fluctuation features of the local relaxation feature parameter as the spatial position changes are extracted to generate the spatial frequency-relaxation feature correlation density distribution. From the spatial frequency-relaxation feature correlation density distribution, query the relaxation feature density value that matches the spatial frequency value in the feature frequency component to be suppressed. Perform bias normalization operation on the relaxation feature density value and the base impedance reference value of the plating solution to generate the polarization correlation factor associated with the spatial frequency.

[0030] It should be noted that existing electrochemical impedance spectroscopy measurements are mostly used for macroscopic process monitoring. The impedance information obtained only reflects the overall average state of the plating solution system and cannot analyze the differences in electrochemical reactions at different spatial locations on the cathode plate, nor can it be linked to subsequent spatial selective compensation. This invention, during the cathode deposition period, synchronously injects a multi-frequency perturbation excitation signal, consisting of several superimposed sinusoidal perturbation currents of different frequencies, into the plating solution in blocks on the anode. Compared to frequency-by-frequency scanning, synchronous injection can complete full-frequency excitation within a single measurement cycle, avoiding measurement inconsistencies caused by time-varying plating solution states. After synchronously acquiring the potential response signals of each sensing area on the cathode plate, the excitation signal and response signal are decoupled in the frequency domain to obtain the impedance amplitude and phase information corresponding to each frequency point. This information is then used to construct a multi-frequency domain impedance complex spectrum to comprehensively characterize the impedance response of the plating solution system at multiple frequencies. To further correlate macroscopic impedance information with spatial location, this invention uses the spatial resolution corresponding to the spatial frequency value in the characteristic frequency component to be suppressed as a basis to divide the cathode plate surface into fine-grained local micro-elements of the electrode surface reaction. The spatial resolution ensures that the scale of each local micro-element matches the spatial variation scale of the deviation to be compensated. After mapping the impedance amplitude and phase of each frequency point in the multi-frequency domain impedance complex spectrum to each local micro-element according to the spatial distribution of the cathode plate surface potential response signal, each micro-element obtains its local relaxation characteristic parameter spectrum at multiple frequencies. The relaxation characteristic parameter characterizes the dynamic response characteristics of the electrode surface reaction at that micro-element to perturbation excitation. Then, the local relaxation characteristic parameter spectra of each micro-element are spatially sequenced along the spatial direction of the cathode plate surface to extract the law of variation of local relaxation characteristic parameters with spatial location fluctuations, reflecting the non-uniform distribution of the plating solution impedance characteristics at different spatial scales. This generates a spatial frequency-relaxation characteristic correlation density distribution, establishing a direct correspondence between the spatial differences in the dynamic behavior of the plating solution and the spatial frequency, thereby generating a polarization correlation factor. The polarization correlation factor numerically represents the additional plating solution resistance effect that needs to be considered for a specific spatial frequency deviation under the current plating solution polarization state.

[0031] Preferably, S4 specifically comprises: Retrieve the generated polarization correlation factor, and perform gain bias calibration on the polarization correlation factor and the basic current-deposition transfer gain coefficient of the anode block in the unpolarized state. That is, use the polarization correlation factor as a gain bias correction term to compensate the bias of the basic current-deposition transfer gain coefficient, and generate the dynamic current-deposition transfer gain coefficient of the anode block after polarization correction. From the spatial frequency domain deviation spectrum, based on the frequency values ​​in the characteristic frequency components to be suppressed, the deviation amplitude and deviation phase at the deviation frequency are extracted, and the deviation amplitude and deviation phase are used to construct a two-dimensional component matrix of the spatial spectrum at the deviation frequency. The deviation amplitude and deviation phase corresponding to the slowly changing deviation characteristic frequency in the two-dimensional component matrix of the spatial spectrum at the deviation frequency are used as the low-frequency contour compensation amplitude table and the low-frequency contour compensation phase reference. Through frequency-guided signal waveform synthesis, a low-frequency contour envelope wave with the low-frequency contour compensation amplitude table as the amplitude envelope and the low-frequency contour compensation phase reference as the phase reference is generated. Meanwhile, the deviation amplitude and deviation phase corresponding to the wave-like deviation characteristic frequency in the two-dimensional component matrix of the spatial spectrum at the deviation frequency are decomposed into frequency bands with the main response frequency band of the corresponding anode block in the spatial frequency excitation response coupling scale as the center frequency band. The wavelet components of each frequency band are then synthesized into high-frequency detail ripple primitives. The low-frequency contour envelope wave and the high-frequency detail ripple primitive are waveform-shaped and superimposed, and the superimposed waveform is amplitude-distributed by the dynamic current-deposition transfer gain coefficient to generate dynamic compensation current waveform commands that act on the low-frequency contour compensation group and the high-frequency detail compensation group respectively.

[0032] It should be noted that the basic current-deposition transfer gain coefficient refers to the conversion efficiency coefficient between the unit current increment output by the anode block and the increase in coating thickness caused by this current increment in the corresponding area of ​​the cathode plate. Its value is jointly determined by factors such as the spatial geometric relationship between the anode block and the cathode plate, the conductivity of the plating solution, and the cathode current efficiency. Essentially, it reflects the linear transfer relationship from current to coating thickness without considering changes in the polarization state of the plating solution. When the polarization state of the plating solution changes during the actual electroplating process, the transfer efficiency will shift. Therefore, this invention introduces a polarization correlation factor as a gain bias correction term to offset and compensate the basic current-deposition transfer gain coefficient. The dynamic current-deposition transfer gain coefficient obtained after correction can follow the changes in the polarization state of the plating solution in real time, making the amplitude allocation of the waveform command more accurate. In the waveform generation stage, the deviation amplitude and deviation phase corresponding to the slowly changing deviation characteristic frequency in the two-dimensional component matrix of the spatial spectrum at the deviation frequency are used as the low-frequency profile compensation amplitude table and the low-frequency profile compensation phase reference, respectively. Frequency-guided signal waveform synthesis refers to using the slow-changing deviation characteristic frequency as the main control frequency for waveform synthesis, the amplitude data of each frequency point in the low-frequency profile compensation amplitude table as the amplitude envelope constraint of each frequency component of the waveform, and the phase data of each frequency point in the low-frequency profile compensation phase reference as the phase calculation reference of each frequency component of the waveform. A continuous low-frequency profile envelope wave is generated by the weighted superposition of multiple frequency sine waves. For example, if the slow-changing deviation characteristic frequency contains two main spatial frequency components with amplitudes A1 and A2 and phases ψ1 and ψ2, then during synthesis, a sine wave with frequency f1, amplitude A1, and initial phase ψ1 is superimposed with a sine wave with frequency f2, amplitude A2, and initial phase ψ2. After superposition, the waveform in the time domain exhibits an amplitude envelope shape constrained by both A1 and A2, and the overall phase is determined by both ψ1 and ψ2, thus forming a compensation waveform that matches the spatial distribution of the current large-area tilt deviation of the coating. Simultaneously, the deviation amplitude and phase corresponding to the characteristic frequency of the fluctuation deviation are decomposed into frequency bands centered on the main response frequency band of the anode block in the spatial frequency excitation response coupling scale. Several narrow-band wavelet components are then synthesized into a high-frequency detail ripple element. This ripple element exhibits a dense, small-amplitude vibration pattern on the time axis, used to compensate for localized fine deviations on the board surface. Finally, the low-frequency profile envelope wave and the high-frequency detail ripple element are waveform-shaped and superimposed. The amplitude of the superimposed waveform is weighted using a dynamic current-deposition transfer gain coefficient to ensure that the actual deposition compensation amount meets expectations under the same current increment. This generates dynamic compensation current waveform commands that are output to the low-frequency profile compensation group and the high-frequency detail compensation group, respectively.

[0033] Preferably, S5 specifically comprises: Within multiple consecutive monitoring time windows after the execution of the dynamic compensation current waveform command, the coating thickness distribution on the cathode plate is successively acquired and converted into a spatial frequency domain deviation spectrum. The characteristic frequency amplitude corresponding to the characteristic frequency component to be suppressed is extracted from the spatial frequency domain deviation spectrum of each conversion, and the evolution trend curve of the characteristic frequency amplitude with the attenuation slope of the monitoring time window is generated. The decay slope evolution trend curve is dynamically time-warped and matched with the multi-scale historical decay slope trend primitive library pre-stored in the database to identify time-varying perturbation features in the decay slope evolution trend curve that deviate from the historical decay slope trend primitive library. Among them, the multi-scale historical attenuation slope trend primitive library refers to the set of typical trend segments that are pre-recorded and extracted in multiple historical normal process operations for different plating solution life stages and different plate specifications, with the attenuation slope change trend at each time scale as the basic unit. Each primitive represents the normal evolution form of the characteristic frequency amplitude attenuation slope after compensation takes effect under specific process conditions.

[0034] Based on the identified time-varying perturbation features, the spatial frequency-relaxation feature correlation density distribution is successively time-delayed tracked to extract the relaxation feature density drift trend associated with the time-varying perturbation features in the relaxation feature density distribution. The relaxation characteristic density drift direction is decoupled from the polarization correlation factor by time-varying parameters, and the time-varying relaxation correction term affected by the gradual change of the polarization state of the plating solution is extracted from the polarization correlation factor. Using the attenuation slope change amplitude of each monitoring time window in the attenuation slope evolution trend curve as the adjustment step size benchmark, the amplitude of the time-varying relaxation correction term is updated by recursion, and the updated time-varying relaxation correction term is fed back to the generation stage of spatial frequency-relaxation feature correlation density distribution to complete the coordinated time-delay adaptive adjustment of impedance spectrum features and spatial frequency excitation response coupling scale.

[0035] It should be noted that during the electroplating process, the composition of the plating solution changes gradually with consumption and replenishment, and the anode surface state evolves due to passivation or consumption. These factors all cause the previously established polarization correlation factor and spatial frequency topological mapping relationship to drift over time. Without adaptive correction, the compensation accuracy will gradually decline. To address this, this invention periodically acquires the cathode plate coating thickness distribution and converts it to a spatial frequency domain deviation spectrum within multiple consecutive monitoring time windows after the execution of the dynamic compensation current waveform command. It continuously tracks the changing trend of the amplitude of the characteristic frequency component to be suppressed, generating an attenuation slope evolution trend curve with the monitoring time window as the horizontal axis and the characteristic frequency amplitude as the vertical axis. The direction and magnitude of the curve's slope change directly reflect the dynamic changes in the effectiveness of the current compensation strategy. By dynamically time-warping and matching the currently generated attenuation slope evolution trend curve with the primitive library, the distorted segments that deviate from the normal primitive evolution trajectory in the current attenuation trend can be identified. These distorted segments are time-varying perturbation characteristics, reflecting the presence of drift factors in the system.

[0036] For the identified time-varying perturbation features, the previously generated spatial frequency-relaxation feature correlation density distribution is traced back along the time axis to obtain the trend of change of the data segment corresponding to the perturbation feature in the relaxation feature density distribution in terms of time and spatial frequency, thus obtaining the relaxation feature density drift trend. Based on this, time-varying parameter decoupling is performed, the core of which is to distinguish the relatively stable basic part of the polarization correlation factor from the dynamic drift part caused by the slow evolution of the plating solution state. For example, the polarization correlation factor can be considered as a baseline component determined by the plating solution basic formulation and steady-state process conditions, superimposed with a relaxation correction component that slowly changes due to factors such as additive consumption and temperature gradient migration. Based on the direction and amplitude guidance provided by the relaxation feature density drift trend, it is projected onto the current value of the polarization correlation factor, and the dynamic change increment corresponding to the aforementioned slowly changing factors is separated, i.e., the time-varying relaxation correction term affected by the slow change of the plating solution polarization state is extracted. Subsequently, using the change in attenuation slope between adjacent monitoring time windows in the attenuation slope evolution trend curve as the adjustment step size benchmark, the time-varying relaxation correction term is updated successively by amplitude recursion. The updated correction term is then fed back into the generation stage of the spatial frequency-relaxation feature correlation density distribution, so that it affects the solution result of the polarization correlation factor in the next control cycle. This achieves coordinated time-delay adaptive adjustment of the impedance spectrum characteristics and spatial frequency excitation response coupling scale, enabling the system to autonomously offset the influence of process state drift during long-term continuous operation and maintain the long-term stability of compensation accuracy.

[0037] In this embodiment, it also includes: From the two-dimensional phase distribution of the spatial frequency domain deviation spectrum, the phase gradient is calculated pixel by pixel and integrated along the closed loop. Pixels with non-zero integer integral values ​​are located as phase-wound singularities, and the spatial coordinates of the phase-wound singularities and their corresponding winding topological charges are extracted. Based on the spatial coordinates of the phase entanglement singularity, a local phase anomaly subset is extracted from the current characteristic frequency component to be suppressed, centered on the spatial coordinates, and the curl direction of the phase gradient within the local phase anomaly subset is calculated to obtain the spiral phase gradient direction characterizing the spiral twisting direction. Retrieve the known frequency phase vector of each anode block, match the direction of the frequency phase vector with the direction of the spiral phase gradient, calculate the direction cosine between the two, screen out anode blocks whose absolute value of the direction cosine exceeds the preset correlation threshold, and construct a spiral compensation anode group that resonates with the singular topology. Based on the absolute value of the wound topological charge and the frequency phase vector magnitude of each anode block in the spiral compensation anode group, the additional spiral phase modulation depth required for each anode block is determined, generating a non-uniform phase delay distribution consistent with the singular topological charge sign, i.e., a spiral phase mask. The spiral phase mask is superimposed on the dynamic compensation current waveform command of the currently generated low-frequency contour compensation group and high-frequency detail compensation group by multiplying point by point, so that the superimposed compensation current wavefront carries orbital angular momentum, and the spiral compensation current wavefront command is obtained. The spiral compensation current wavefront command is distributed to the spiral compensation anode group, so that each anode block excites the plating solution with a current waveform with spiral phase modulation, generating a spiral deposition compensation field on the cathode plate surface that is opposite to the spiral thickness distortion and cancels out the topological charge in equal amounts, thereby eliminating the vortex thickness deviation of the coating caused by the edge effect.

[0038] It should be noted that in PCB electroplating practice, the edge area of ​​the cathode plate or the densely packed area of ​​via rings often exhibits a special vortex-like thickness distortion due to the concentrated distribution of electric field lines. This distortion is not a random fluctuation in spatial phase distribution, but rather presents a spiral phase accumulation characteristic that continuously rotates around a certain center point. Conventional amplitude compensation strategies cannot identify and cancel this phase topology, and even if the compensation current intensity is sufficient, a spiral deviation will still remain in this area. This embodiment addresses the above problem by introducing a phase winding decoupling and spiral compensation mechanism.

[0039] During implementation, based on the obtained spatial frequency domain deviation spectrum, the gradient of the two-dimensional phase distribution of the deviation spectrum is calculated pixel by pixel, and the phase gradient is integrated along a closed loop. When the loop integral value is a non-zero integer, it indicates that there is phase entanglement around the pixel, and this point is marked as a phase-entangled singularity. The entanglement direction and number of loops are recorded as the spatial coordinates and entanglement topological charge of the singularity, respectively. Taking the coordinates of each located phase-entangled singularity as the center, an analysis window is defined in the current characteristic frequency component to be suppressed. For example, a range of several pixels in the neighborhood of the singularity is selected, and the phase data within this window is truncated as a local phase anomaly subset. The curl of the phase gradient is calculated for this subset, and the direction of the curl vector is the spiral phase gradient direction, which describes the spatial orientation of the phase spiral distortion in this local region.

[0040] Based on this, the frequency-phase vector of each anode block obtained from previous calculations is retrieved, and the direction of each frequency-phase vector is correlated with the direction of the spiral phase gradient. The degree of coupling between the two vector directions is determined by calculating the cosine of the angle between them. Anode blocks whose absolute cosine value exceeds a preset correlation threshold are selected to form a spiral compensation anode group. For determining the spiral phase modulation depth, the absolute value of the topological charge of the phase-wound singularity is used as the basic modulation magnitude. The larger the value, the larger the required total phase modulation amplitude. This total modulation amplitude is then weighted and allocated according to the proportion of the magnitude of the frequency-phase vector of each anode block in the spiral compensation anode group. Anode blocks with larger magnitudes receive a larger modulation depth, thus obtaining the additional spiral phase modulation depth required for each anode block. Based on the modulation depth, a spiral phase mask with a non-uniform spatial arrangement and a rotation direction consistent with the sign of the topological charge of the singularity is constructed, using the singularity position as the origin of phase distribution generation. The spiral phase mask is superimposed point-by-point onto the dynamically compensated current waveform commands already generated by the low-frequency contour compensation group and the high-frequency detail compensation group. The superimposed compensation current wavefront carries orbital angular momentum, forming a spiral compensation current wavefront command. After this command is distributed to the spiral compensation anode group output, the current field excited by the anode in the plating solution also carries a spiral phase structure matching the singularity. This generates a spiral deposition compensation field in the corresponding area of ​​the cathode plate, which is opposite to the spiral distortion in direction and cancels out the topological charge. This directionally eliminates the spiral thickness deviation of the coating caused by the edge effect. As a result, when dealing with complex edge distortions containing phase singularities, the compensation system can overcome the limitation of traditional amplitude compensation, which can only suppress intensity deviations. It can achieve precise matching and cancellation of deviation morphology at the phase topology level, improving the precision of coating uniformity control and the adaptability of process scenarios under non-uniform electric field distribution.

[0041] In this embodiment, it also includes: From the multi-frequency domain impedance complex spectrum, for each sensing area on the cathode plate, identify the frequency corresponding to the peak value of the imaginary part of the impedance spectrum in the low-frequency diffusion control section, and calibrate the frequency as the finite-length diffusion inflection frequency point of the corresponding sensing area. Simultaneously, the main flow direction vector of the plating solution at each sensing area on the cathode plate is acquired in real time. The position of the finite-length diffusion inflection frequency point of each sensing area is registered with its corresponding main flow direction vector to construct a diffusion inflection frequency-flow direction spatial correlation mapping set. For each sensing node in the diffusion transition frequency-flow direction spatial correlation mapping set, calculate the directional derivative of its finite-length diffusion transition frequency along the direction of the main flow direction vector and the directional derivative along the direction orthogonal to the main flow direction vector. The ratio of the two is defined as the diffusion relaxation anisotropy ratio of the sensing node. Sensing nodes whose diffusion relaxation anisotropy ratio exceeds a preset homogenization threshold are extracted and identified as convection-sensitive nodes. For each convection-sensitive node, the angle between the contour normal of its finite-length diffusion inflection frequency and the main flow direction vector is calculated to generate the anisotropic diffusion relaxation azimuth angle of the convection-sensitive node. The current density diffusion envelope field of each anode block in the low-frequency profile compensation group is retrieved to determine the set of convection sensitive nodes covered by each anode block. The anisotropic diffusion relaxation azimuth angles of all convection sensitive nodes in the set are then averaged by spatial vector to obtain the diffusion relaxation dominant azimuth angle of each anode block. Using the diffusion relaxation dominance azimuth angle as the control angle for the asymmetric envelope, the envelope slope of the preset symmetrical low-frequency profile envelope wave is increased by a preset slope offset on the upstream side along the main flow direction, and the envelope slope is decreased by a preset slope offset on the downstream side along the main flow direction, while keeping the total peak-to-peak amplitude of the envelope wave unchanged, thereby generating the asymmetric envelope slope of the low-frequency profile compensation waveform of the anode block; The asymmetric envelope slope of each anode block is coupled with the low-frequency profile compensation amplitude table to generate an asymmetric low-frequency profile compensation waveform with steep upstream compensation and gentle downstream compensation. This waveform drives the anode blocks in the low-frequency profile compensation group to output differentiated dynamic compensation currents.

[0042] The impedance response of the plating solution system in the low-frequency diffusion control section is mainly controlled by the finite-length diffusion process of ions in the diffusion layer on the electrode surface. The impedance spectrum exhibits the characteristic morphology of the typical capacitive arc tail or diffusion tail, which is related to the diffusion mass transfer characteristics.

[0043] It should be noted that in actual electroplating production, the plating solution near the cathode plate is not static, but rather flows along the plate surface. This flow causes a difference in the thickness of the diffusion layer on the upstream and downstream sides; that is, the diffusion layer is thinner and ion mass transfer is faster on the upstream side, while the diffusion layer is thicker and ion mass transfer is slower on the downstream side. If a symmetrical low-frequency profile compensation waveform is used, it is difficult to adapt to this directional mass transfer non-uniformity, resulting in undercompensation or overcompensation of the low-frequency profile compensation in the upstream and downstream directions. In view of this, this embodiment first identifies the frequency corresponding to the peak value of the imaginary part of the impedance in the low-frequency diffusion control section based on the multi-frequency domain impedance complex spectrum of each sensing region. This frequency is calibrated as the finite-length diffusion inflection frequency point, reflecting the characteristic frequency of the diffusion layer transitioning from semi-infinite diffusion to finite-thickness diffusion at that location. At the same time, the flow direction vector of the plating solution at each sensing region is acquired in real time, and the finite-length diffusion inflection frequency point of each point is registered with its flow direction vector to construct a spatial correlation mapping set of diffusion inflection frequency-flow direction. For each sensing node, the directional derivative of the diffusion transition frequency along the flow direction and its orthogonal direction is calculated. The ratio of these two derivatives is used as the diffusion relaxation anisotropy ratio. The greater the deviation of this ratio from 1, the more significant the directional difference in ion mass transfer in the diffusion layer of that node due to convection. Nodes with diffusion relaxation anisotropy ratios exceeding a preset homogenization threshold are extracted as convection-sensitive nodes. The angle between the normal of the finite-length diffusion transition frequency contour line and the flow direction vector is calculated for each node to obtain the anisotropic diffusion relaxation azimuth angle. For each anode block in the low-frequency profile compensation group, the set of convection-sensitive nodes covered by its current density diffusion envelope is taken as the object. The anisotropic diffusion relaxation azimuth angles of each node in the set are weighted by the inverse of spatial distance and then vector averaged to obtain the diffusion relaxation dominant azimuth angle of that anode block. Using this advantageous azimuth angle as the control angle for the asymmetric envelope, when generating the low-frequency profile compensation waveform, the envelope slope of the preset symmetrical envelope wave along the flow direction is increased by a preset offset amount on the upstream side and decreased by the same offset amount on the downstream side, keeping the total peak-to-peak amplitude constant, thus obtaining the asymmetric envelope slope of the anode block. Finally, the asymmetric envelope slope of each anode block is coupled with the low-frequency profile compensation amplitude table through waveform shaping to modulate an asymmetric low-frequency profile compensation waveform with steep compensation on the upstream side and gentle compensation on the downstream side. This drives the anode blocks in the low-frequency profile compensation group to output differentiated dynamic compensation currents, so that the deposition compensation amount along the upstream and downstream of the flow direction matches the difference in diffusion layer thickness, eliminating the low-frequency profile residual deviation caused by convective mass transfer asymmetry.

[0044] like Figure 2As shown, the second aspect of the present invention discloses a dynamic compensation control system for the uniformity of copper plating in PCB, including a memory and a processor. The memory stores a dynamic compensation control method program for the uniformity of copper plating in PCB. When the dynamic compensation control method program for the uniformity of copper plating in PCB is executed by the processor, the steps of the method described in any one of the present invention are implemented.

[0045] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for dynamic compensation and control of the uniformity of copper plating layer in PCB, characterized in that, include: S1. Real-time acquisition of coating thickness distribution on cathode plate surface, conversion of coating thickness distribution into spatial frequency domain deviation spectrum through two-dimensional spatial frequency domain decomposition, and extraction of the characteristic frequency component to be suppressed that constitutes the dominant deviation of uniformity from the spatial frequency domain deviation spectrum. S2. Based on the spatial layout of each anode block, construct the topological mapping relationship between each anode block and different spatial frequency response ranges, and dynamically identify the anode blocks as low-frequency contour compensation group and high-frequency detail compensation group according to the extracted characteristic frequency components to be suppressed. S3. Simultaneously apply multi-frequency micro-perturbation electrochemical impedance spectroscopy measurement during the electroplating process, extract the real-time impedance spectrum characteristic parameters of the plating solution, and calculate the polarization correlation factor related to the spatial frequency based on the impedance spectrum characteristic parameters. S4. Using the polarization correlation factor, the current-deposition transfer gain of the anode block to be suppressed characteristic frequency components is corrected, and combined with the spatial frequency domain deviation spectrum inversion, dynamic compensation current waveform commands are generated for the low-frequency contour compensation group and the high-frequency detail compensation group respectively. S5. Continuously monitor the attenuation slope of the characteristic frequency amplitude in the spatial frequency domain deviation spectrum, and adjust the time-varying parameter in the polarization correlation factor according to the feedback of the attenuation slope to complete the online collaborative self-correction of impedance spectrum characteristics and spatial frequency topology mapping.

2. The method for dynamic compensation and control of copper layer uniformity in PCB electroplating according to claim 1, characterized in that, Specifically, S1 is: Real-time acquisition of instantaneous measurements of coating thickness at discrete sampling points on the cathode plate surface to construct the original thickness distribution matrix in a two-dimensional spatial domain; The original thickness distribution matrix is ​​subjected to two-dimensional high-frequency texture envelope filtering processing, namely, morphological opening operation is performed along the transfer direction of the plated part and the transverse direction of the plate surface to extract the inherent non-defect fine texture envelope surface of the plate surface, and the original thickness distribution matrix is ​​subtracted from the texture envelope surface to obtain the substrate thickness deviation distribution field after removing the substrate texture interference. A two-dimensional spatial Fourier transform is performed on the substrate thickness deviation distribution field to map the thickness deviation information in the spatial location domain to the spatial frequency domain, generating a global spatial frequency domain deviation spectrum characterized by the amplitude and phase of different spatial frequency components. The standard spatial frequency amplitude envelope sequence stored in the historical database is retrieved, and statistical generalization processing is performed on each standard spatial frequency amplitude envelope sequence to fit and generate a historical stationary generalization boundary benchmark that characterizes the allowable range of normal sedimentary fluctuations. The real-time amplitude of each spatial frequency component in the global spatial frequency domain deviation spectrum is compared with the historical stationary generalization boundary benchmark frequency by frequency. Spatial frequency components whose real-time amplitude continuously exceeds the historical stationary generalization boundary benchmark are identified and marked as abnormal frequency domain deviation feature clusters. The frequency component with the largest amplitude proportion is selected from the abnormal frequency domain deviation feature clusters as the characteristic frequency component to be suppressed that constitutes the dominant deviation of uniformity.

3. The method for dynamic compensation and control of copper layer uniformity in PCB electroplating according to claim 1, characterized in that, Specifically, S2 is: The spatial geometric center coordinates and boundary contour dimensions of each independent controllable anode block in the electroplating tank are obtained. Based on the boundary contour dimensions of each anode block, electric field edge diffusion envelope modeling is adopted to generate the current density diffusion envelope field of each anode block on the cathode plate. The current density diffusion envelope field characterizes the spatial coverage of the current excitation of each position on the cathode plate by the anode block. Two-dimensional spatial Fourier transforms are performed on the current density diffusion envelope field of each anode block to extract the frequency response bandwidth and main response band of each anode block in the spatial frequency domain, and a spatial frequency excitation response coupling scale between each anode block and different spatial frequency response ranges is established. Based on the frequency values ​​of the extracted characteristic frequency components to be suppressed, frequency matching search is performed in the spatial frequency excitation response coupling scale to screen out the anode blocks that generate response coupling with the characteristic frequency components to be suppressed, forming a set of candidate compensation anode blocks. Calculate the low-frequency-high-frequency dominant deviation vector of the slow-changing deviation characteristic frequency and the fluctuating deviation characteristic frequency in the characteristic frequency component to be suppressed for each anode block in the candidate compensation anode block set; Based on the response dominance trend characterized by the low-frequency to high-frequency response dominance deviation vector, the anode blocks are dynamically identified as a low-frequency profile compensation group and a high-frequency detail compensation group.

4. The method for dynamic compensation and control of PCB electroplated copper layer uniformity according to claim 3, characterized in that, The low-frequency to high-frequency response dominant deviation vector of the slow-changing deviation characteristic frequency and the fluctuating deviation characteristic frequency in the characteristic frequency component to be suppressed is calculated for each anode block in the candidate compensation anode block set. Specifically: From the characteristic frequency components to be suppressed, the slow-change deviation characteristic frequency, which characterizes the overall thickness tilt trend of the coating, and the fluctuation deviation characteristic frequency, which characterizes the local coating undulation, are separated according to the spatial frequency value. The current density diffusion envelope field of each anode block in the candidate compensation anode block set is retrieved. Low-frequency driving induction excitation corresponding to the slow-changing deviation characteristic frequency and high-frequency driving induction excitation corresponding to the fluctuation deviation characteristic frequency are applied to the current density diffusion envelope field of each anode block respectively. The first driving induction coupled electromagnetic field harmonic response spectrum of each anode block under low-frequency driving induction excitation and the second driving induction coupled electromagnetic field harmonic response spectrum under high-frequency driving induction excitation are obtained. The main peak of the anode response amplitude and the anode response phase shift under low-frequency induction excitation are extracted from the harmonic response spectrum of the first induction coupled electromagnetic field, and the secondary peak of the anode response amplitude and the anode response phase shift under high-frequency induction excitation are extracted from the harmonic response spectrum of the second induction coupled electromagnetic field. The amplitude ratio of the main peak and the secondary peak of the anode response amplitude of the same anode block is normalized to generate the driving-induction dominant amplitude ratio of the corresponding anode block. The driving-induction dominant amplitude ratio is then calculated with the anode response phase offset to generate the frequency-phase vector value that characterizes the dominant bias of the anode block's response to the slow-changing deviation characteristic frequency and the fluctuation deviation characteristic frequency. The frequency phase vector values ​​of each candidate compensation anode block are normalized by performing deviation calculation to generate a numerical low-frequency to high-frequency response dominant deviation vector for each anode block, so as to characterize the degree of response bias of the anode block to slow-changing deviation or fluctuating deviation.

5. The method for dynamic compensation and control of copper layer uniformity in PCB electroplating according to claim 1, characterized in that, Specifically, S3 is: During the cathode deposition phase of the electroplating process, a multi-frequency perturbation excitation signal composed of multiple sinusoidal perturbation currents of different frequencies is simultaneously injected into the plating solution through anode blocks, and the potential response signal of each sensing area on the cathode plate is simultaneously acquired. The potential response signal and the multi-frequency perturbation excitation signal are decoupled in the frequency domain to establish a multi-frequency domain impedance complex spectrum characterizing the impedance amplitude and phase relationship of the plating solution system at multiple frequencies. The cathode plate is divided into multiple local micro-elements of electrode surface reaction according to the spatial resolution that matches the spatial frequency value in the characteristic frequency component to be suppressed. The impedance amplitude and phase of each frequency point in the multi-frequency domain impedance complex spectrum are mapped and distributed to each local micro-element of electrode surface reaction according to the spatial distribution of the potential response signal of each sensing area of ​​the cathode plate, thus forming the local relaxation characteristic parameter spectrum of each local micro-element of electrode surface reaction at multiple frequencies. Spatial frequency-relaxation feature correlation analysis is performed on the local relaxation feature parameter spectrum of the local reaction local micro-element on each electrode surface. That is, the local relaxation feature parameter spectrum of the local reaction local micro-element on each electrode surface is arranged in a spatial sequence along the spatial direction of the cathode plate surface, and the fluctuation features of the local relaxation feature parameter as the spatial position changes are extracted to generate the spatial frequency-relaxation feature correlation density distribution. From the spatial frequency-relaxation feature correlation density distribution, query the relaxation feature density value that matches the spatial frequency value in the feature frequency component to be suppressed. Perform bias normalization operation on the relaxation feature density value and the base impedance reference value of the plating solution to generate the polarization correlation factor associated with the spatial frequency.

6. The method for dynamic compensation and control of copper layer uniformity in PCB electroplating according to claim 1, characterized in that, Specifically, S4 is: Retrieve the generated polarization correlation factor, and perform gain bias calibration on the polarization correlation factor and the basic current-deposition transfer gain coefficient of the anode block in the unpolarized state. That is, use the polarization correlation factor as a gain bias correction term to compensate the bias of the basic current-deposition transfer gain coefficient, and generate the dynamic current-deposition transfer gain coefficient of the anode block after polarization correction. From the spatial frequency domain deviation spectrum, based on the frequency values ​​in the characteristic frequency components to be suppressed, the deviation amplitude and deviation phase at the deviation frequency are extracted, and the deviation amplitude and deviation phase are used to construct a two-dimensional component matrix of the spatial spectrum at the deviation frequency. The deviation amplitude and deviation phase corresponding to the slowly changing deviation characteristic frequency in the two-dimensional component matrix of the spatial spectrum at the deviation frequency are used as the low-frequency contour compensation amplitude table and the low-frequency contour compensation phase reference. Through frequency-guided signal waveform synthesis, a low-frequency contour envelope wave with the low-frequency contour compensation amplitude table as the amplitude envelope and the low-frequency contour compensation phase reference as the phase reference is generated. Meanwhile, the deviation amplitude and deviation phase corresponding to the wave-like deviation characteristic frequency in the two-dimensional component matrix of the spatial spectrum at the deviation frequency are decomposed into frequency bands with the main response frequency band of the corresponding anode block in the spatial frequency excitation response coupling scale as the center frequency band. The wavelet components of each frequency band are then synthesized into high-frequency detail ripple primitives. The low-frequency contour envelope wave and the high-frequency detail ripple primitive are waveform-shaped and superimposed, and the superimposed waveform is amplitude-distributed by the dynamic current-deposition transfer gain coefficient to generate dynamic compensation current waveform commands that act on the low-frequency contour compensation group and the high-frequency detail compensation group respectively.

7. The method for dynamic compensation and control of copper layer uniformity in PCB electroplating according to claim 1, characterized in that, Specifically, S5 is: Within multiple consecutive monitoring time windows after the execution of the dynamic compensation current waveform command, the coating thickness distribution on the cathode plate is successively acquired and converted into a spatial frequency domain deviation spectrum. The characteristic frequency amplitude corresponding to the characteristic frequency component to be suppressed is extracted from the spatial frequency domain deviation spectrum of each conversion, and the evolution trend curve of the characteristic frequency amplitude with the attenuation slope of the monitoring time window is generated. The decay slope evolution trend curve is dynamically time-warped and matched with the multi-scale historical decay slope trend primitive library pre-stored in the database to identify time-varying perturbation features in the decay slope evolution trend curve that deviate from the historical decay slope trend primitive library. Based on the identified time-varying perturbation features, the spatial frequency-relaxation feature correlation density distribution is successively time-delayed tracked to extract the relaxation feature density drift trend associated with the time-varying perturbation features in the relaxation feature density distribution. The relaxation characteristic density drift direction is decoupled from the polarization correlation factor by time-varying parameters, and the time-varying relaxation correction term affected by the gradual change of the polarization state of the plating solution is extracted from the polarization correlation factor. Using the attenuation slope change amplitude of each monitoring time window in the attenuation slope evolution trend curve as the adjustment step size benchmark, the amplitude of the time-varying relaxation correction term is updated by recursion, and the updated time-varying relaxation correction term is fed back to the generation stage of spatial frequency-relaxation feature correlation density distribution to complete the coordinated time-delay adaptive adjustment of impedance spectrum features and spatial frequency excitation response coupling scale.

8. A dynamic compensation control system for the uniformity of copper plating layer in PCB, characterized in that, The device includes a memory and a processor. The memory stores a program for a dynamic compensation control method for the uniformity of copper plating on a PCB. When the program for the dynamic compensation control method for the uniformity of copper plating on a PCB is executed by the processor, the steps of the method as described in any one of claims 1 to 7 are implemented.

Images