A multi-spectral fusion-based electroplating additive concentration online detection method and system

Abstract

The application belongs to the field of physical property measurement, and discloses a kind of electroplating additive concentration online detection method and system based on multispectral fusion, comprising: guiding flow electroplating liquid through detection cell, control transducer array generates sub-cavitation megahertz standing wave field, utilize acoustic radiation force to drive free gas bubble to gather to wave crest area, make node area form the light path of no scattering interference;Real-time monitoring impedance phase, when reaching resonance extreme point, produce synchronous trigger signal, drive light source to emit ultraviolet wave band light beam, visible wave band light beam and near infrared wave band light beam to light path;Extract the intensity correction absorption brightness parameter of main salt reference wavelength, obtain response sequence and output additive concentration, the application reconstructs the phase distribution of measurement environment, blocks random scattering to the erosion of spectral information, ensures the fidelity of measurement data corresponding to physical fact, solves the problem that trace component signal is hidden in complex matrix.

Description

Technical Field

[0001] This invention relates to an online detection method and system for electroplating additive concentration based on multispectral fusion, belonging to the field of physical property measurement technology. Background Technology

[0002] The concentration ratio of organic additives such as accelerators, inhibitors, and leveling agents directly affects the surface filling quality of electroplated layers. Currently, the industry generally uses cyclic voltammetry to determine the additive concentration, and infers the effective component content by recording the redox current on the electrode surface. This offline measurement method involves multiple chemical titration and electrode pretreatment processes, and the analysis time usually exceeds 30 minutes, with obvious feedback lag, making it difficult to adapt to the high-frequency dynamic changes of continuous electroplating production lines. Although multispectral detection technology has the potential for online sampling, capturing the spectral characteristics of trace components in the complex electroplating fluid environment faces physical constraints. The electroplating solution contains a high concentration of metal salt matrix, which generates high-intensity background absorption covering the entire wavelength range. At the same time, the dynamic circulation of the fluid and the cathodic hydrogen evolution reaction continuously generate a large number of randomly distributed micron-sized bubbles. When the detection beam penetrates the fluid, the free bubbles induce strong random Mie scattering, causing the weak characteristic absorption signal of the target organic component to be submerged in the chaotic physical background noise.

[0003] Research on improving detection accuracy has extended from improving sampling hardware structure to control algorithm compensation and prediction. For example, Chinese invention patent application CN121204790A discloses an intelligent detection and dynamic replenishment control system for electroplating bath composition. It filters and degassing through a sample pretreatment unit and uses a fusion model of PID and machine learning to predict composition trends. However, under high-intensity electroplating conditions, the cathode hydrogen evolution reaction is transient and widespread on the surface. External pretreatment devices are unable to eliminate micron-sized bubbles generated in situ inside the detection pool. Existing technologies usually attempt to obtain characteristic signals by adjusting the optical sampling path or increasing the emission power of the light source, but this easily causes the detector to enter the saturation region and cannot eliminate random coherent interference caused by bubble scattering from a physical level. Some solutions use software algorithms to perform linear or nonlinear compensation on noisy data, but because the bubble distribution and particle size change abruptly with the flow field state in real time, the mathematical model has physical distortion when facing such non-stationary signals, making it impossible for the extracted spectral features to objectively map the true chemical concentration of the material.

[0004] Therefore, the technical problem to be solved by this invention is how to physically reshape the surface phase distribution of the measurement area and filter out random coherent scattering noise in real time while maintaining the dynamic circulation of the electroplating fluid, so as to achieve precise extraction of the spectral characteristics of trace organic components in a high background matrix. Summary of the Invention

[0005] To address the problems in the background art, the technical solution of the present invention is as follows: A method for online detection of electroplating additive concentration based on multispectral fusion includes the following steps: Step 101: The electroplating solution to be tested is guided through the detection flow cell, and a high-frequency excitation electrical signal is applied to the piezoelectric transducer array set on the side wall of the detection flow cell to generate a subcavitated standing wave sound field in the fluid medium inside the detection flow cell. Step 102: Using the spatial gradient acoustic radiation force field generated by the subcavitation standing wave sound field, the free bubbles generated by the hydrogen evolution reaction in the fluid medium are driven to gather in the acoustic pressure antinode region of the subcavitation standing wave sound field, so that the acoustic pressure node region of the subcavitation standing wave sound field forms a light transmission path without bubble scattering phase. Step 103: Real-time acquisition of the impedance phase angle of the piezoelectric transducer array; when the impedance phase angle reaches the preset resonant extreme point, a synchronous trigger level signal is generated. Step 104: In response to the synchronous trigger level signal, drive the multispectral light source to emit detection beams in the ultraviolet, visible and near-infrared bands into the light transmission path, and control the optical acquisition unit to collect the transmitted light intensity data after penetrating the light transmission path. Step 105: Extract the reference transmitted light intensity of the corresponding absorption points of the main salt in the electroplating solution from the transmitted light intensity data, and use the reference transmitted light intensity to compensate for the absorbance parameter of the corresponding band to obtain the absorbance response sequence. Step 106: Substitute the absorbance response sequence into the preset concentration inversion mapping logic, and output the concentration values ​​of different organic additives in the electroplating solution by calculating the linear superposition relationship between the characteristic wavelength absorbance and the absorbance response sequence.

[0006] Preferably, step 104 further includes: monitoring the resonant frequency fluctuation value of the piezoelectric transducer array, adjusting the delay time relative to the synchronous trigger level signal inversely proportional to the resonant frequency fluctuation value, activating the multispectral light source after the delay time is reached, and controlling the pulse width of the detection beam to be less than 10ms.

[0007] Preferably, in step 104, by controlling the collimation aperture of the multispectral light source, the spot diameter of the detection beams in the ultraviolet, visible, and near-infrared bands at the center of the light transmission path is less than 25% of the wavelength of the subcavitating standing wave sound field.

[0008] Preferably, when generating the subcavitary standing wave acoustic field in step 101, the power density of the high-frequency excitation electrical signal is controlled within the range of 0.1 W / cm² to 0.5 W / cm², so that the acoustic pressure amplitude inside the detection flow cell is lower than the cavitation pressure threshold of the electroplating solution.

[0009] Preferably, the absorbance response sequence includes characteristic absorbance response values ​​corresponding to the accelerator, inhibitor, and leveling agent in the electroplating solution, respectively; wherein, the ultraviolet band corresponds to the absorption peak of the conjugated double bond electronic transition of the accelerator, and the near-infrared band corresponds to the absorption peak of the ether bond overtone vibration of the inhibitor.

[0010] Preferably, step 101 further includes an adaptive cleaning operation: during non-detection cycles, the excitation phase difference of the piezoelectric transducer array is changed to switch the subcavitated standing wave acoustic field to an asymmetric traveling wave field, and the directional acoustic flow generated by the asymmetric traveling wave field is used in conjunction with the fluid shear force to peel off the deposits on the surface of the light-transmitting window of the detection flow cell.

[0011] Preferably, step 103 further includes: performing frequency sweep sampling on the piezoelectric transducer array, extracting the peak frequency of the admittance curve to determine the preset resonant extreme point, and setting the phase value corresponding to the preset resonant extreme point as the trigger threshold for generating a synchronous trigger level signal.

[0012] Preferably, step 102 further includes: controlling the excitation frequency of the high-frequency excitation electrical signal to be in the range of 1.5MHz to 3.0MHz, and using a pulse group excitation mode with a duty cycle of less than 20% to drive the piezoelectric transducer array.

[0013] Preferably, step 105 further includes: identifying a fixed characteristic wavelength in the transmitted light intensity data where the rate of change of absorbance with respect to the concentration of the electroplating solution approaches 0, and determining the fixed characteristic wavelength as the absorption point of the main salt, etc.

[0014] An online detection system for electroplating additive concentration based on multispectral fusion is disclosed. This system implements an online detection method for electroplating additive concentration based on multispectral fusion, comprising: a detection flow module, a sound field modulation module, an excitation light source module, a signal detection module, and an analysis and control module. The sound field modulation module is located on the side of the fluid channel of the detection flow module. It is used to generate a subcavitated standing wave sound field inside the detection flow module and use the spatial gradient acoustic radiation force field to drive the free bubbles generated by the hydrogen evolution reaction in the fluid medium to gather towards the acoustic pressure antinode region, so as to form a light transmission path without bubble scattering phase in the acoustic pressure node region. The excitation light source module and the signal detection module are located on opposite sides of the light transmission path, respectively. The analysis and control module is connected to the sound field modulation module, the excitation light source module, and the signal detection module, respectively. It is used to acquire the impedance phase angle of the sound field modulation module in real time and generate a synchronous trigger level signal when the impedance phase angle reaches the preset resonant extreme point. The excitation light source module is used to respond to the synchronous trigger level signal and emit detection beams in the ultraviolet, visible, and near-infrared bands into the light transmission path. The analysis and control module is also used to extract the reference transmitted light intensity of the corresponding absorption points of the main salt in the electroplating solution from the transmitted light intensity data, use the reference transmitted light intensity to compensate the absorbance parameters of the corresponding band to obtain the absorbance response sequence, and substitute the absorbance response sequence into the preset concentration inversion mapping logic to output the concentration values ​​of different organic additives in the electroplating solution. The signal detection module is used to collect transmitted light intensity data after passing through the light transmission path.

[0015] Compared with the prior art, the beneficial effects of the present invention are: 1. In online detection of electroplating additive concentration, to address the Mie scattering interference from bubbles generated during the dynamic circulation of the electroplating solution and electrolysis, a subcavitated mega-frequency standing wave field is excited at the orthogonal section of the measurement optical path. Utilizing the difference in compressibility coefficients between the fluid medium and bubbles in the acoustic radiation force field, the free bubbles are pulled by the force field pointing towards the antinodes of the acoustic pressure wave, causing them to migrate in a directional manner. This constructs a bubble-free physical light-transmitting window in situ in the acoustic pressure node region. Combined with real-time monitoring of the phase angle of the piezoelectric ceramic impedance, a multispectral light source is driven to perform nanosecond-level synchronous excitation within the resonant window established by the acoustic field. This allows the detection beam to penetrate the physical light-transmitting window, obtaining a pure light signal with physically filtered-out scattering noise at the front end of the measurement environment, eliminating baseline random drift caused by fluid turbulence and hydrogen evolution reaction.

[0016] 2. By extracting the transmission intensity response of the multi-band detection beam at the absorption wavelength position of the main salt in the electroplating solution, and using it as a dynamic baseline compensation reference in the online detection process, the absorbance parameters of each additive characteristic detection band are physically differentially calculated with the reference intensity. This directly offsets the broadband background absorption attenuation introduced by electroplating solution aging, main salt concentration fluctuations, and system thermal effects during the calculation process. The final extracted multispectral differential feature vector is exclusively associated with the intrinsic electronic transitions of organic additive molecules, solving the problem of trace component characteristic signals being submerged by strong background absorption in complex chemical matrices.

[0017] 3. By adopting a burst pulse driving strategy in the subcavitation frequency band, combined with the dynamic impedance feedback of the acoustic transducer array, the acoustic energy density inside the fluid medium is maintained below the preset subcavitation threshold. This eliminates the high temperature and high pressure generated by local transient cavitation and the resulting breakage effect on the long-chain molecules of organic additives. Thus, while intervening in the physical arrangement of bubble phases, the molecular structure integrity of the electroplating solution's chemical components is maintained. This physical intervention mechanism has natural compatibility with the electroplating process environment, avoiding the introduction of chemical stability risks into precision electronic manufacturing production lines. Attached Figure Description

[0018] Figure 1 This is a flowchart of the concentration detection method based on subcavitation acoustic field modulation and multispectral fusion of the present invention; Figure 2 This is a diagram showing the module relationships of the detection system for acoustic-optical closed-loop control and detection analysis of the present invention.

[0019] The objectives, features, and advantages of this invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0020] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0021] An online detection method for electroplating additive concentration based on multispectral fusion includes the following steps: Step 101: The electroplating solution to be tested is guided through the detection flow cell, and a high-frequency excitation electrical signal is applied to the piezoelectric transducer array set on the side wall of the detection flow cell to generate a subcavitated standing wave sound field in the fluid medium inside the detection flow cell. Step 102: Using the spatial gradient acoustic radiation force field generated by the subcavitation standing wave sound field, the free bubbles generated by the hydrogen evolution reaction in the fluid medium are driven to gather in the acoustic pressure antinode region of the subcavitation standing wave sound field, so that the acoustic pressure node region of the subcavitation standing wave sound field forms a light transmission path without bubble scattering phase. Step 103: Real-time acquisition of the impedance phase angle of the piezoelectric transducer array; when the impedance phase angle reaches the preset resonant extreme point, a synchronous trigger level signal is generated. Step 104: In response to the synchronous trigger level signal, drive the multispectral light source to emit detection beams in the ultraviolet, visible and near-infrared bands into the light transmission path, and control the optical acquisition unit to collect the transmitted light intensity data after penetrating the light transmission path. Step 105: Extract the reference transmitted light intensity of the corresponding absorption points of the main salt in the electroplating solution from the transmitted light intensity data, and use the reference transmitted light intensity to compensate for the absorbance parameters of the corresponding band to obtain the absorbance response sequence. Step 106: Substitute the absorbance response sequence into the preset concentration inversion mapping logic, and output the concentration values ​​of different organic additives in the electroplating solution by calculating the linear superposition relationship between the characteristic wavelength absorbance and the absorbance response sequence.

[0022] Preferably, step 104 further includes: monitoring the resonant frequency fluctuation value of the piezoelectric transducer array, adjusting the delay time relative to the synchronous trigger level signal inversely proportional to the resonant frequency fluctuation value, activating the multispectral light source after the delay time is reached, and controlling the pulse width of the detection beam to be less than 10ms.

[0023] Preferably, in step 104, by controlling the collimation aperture of the multispectral light source, the spot diameter of the detection beams in the ultraviolet, visible, and near-infrared bands at the center of the light transmission path is less than 25% of the wavelength of the subcavitating standing wave sound field.

[0024] Preferably, when generating the subcavitary standing wave acoustic field in step 101, the power density of the high-frequency excitation electrical signal is controlled within the range of 0.1 W / cm² to 0.5 W / cm², so that the acoustic pressure amplitude inside the detection flow cell is lower than the cavitation pressure threshold of the electroplating solution.

[0025] Preferably, the absorbance response sequence includes characteristic absorbance response values ​​corresponding to the accelerator, inhibitor, and leveling agent in the electroplating solution, respectively; wherein, the ultraviolet band corresponds to the absorption peak of the conjugated double bond electronic transition of the accelerator, and the near-infrared band corresponds to the absorption peak of the ether bond overtone vibration of the inhibitor.

[0026] Preferably, step 101 further includes an adaptive cleaning operation: during non-detection cycles, the excitation phase difference of the piezoelectric transducer array is changed to switch the subcavitated standing wave acoustic field to an asymmetric traveling wave field, and the directional acoustic flow generated by the asymmetric traveling wave field is used in conjunction with the fluid shear force to peel off the deposits on the surface of the light-transmitting window of the detection flow cell.

[0027] Preferably, step 103 further includes: performing frequency sweep sampling on the piezoelectric transducer array, extracting the peak frequency of the admittance curve to determine the preset resonant extreme point, and setting the phase value corresponding to the preset resonant extreme point as the trigger threshold for generating a synchronous trigger level signal.

[0028] Preferably, step 102 further includes: controlling the excitation frequency of the high-frequency excitation electrical signal to be in the range of 1.5MHz to 3.0MHz, and using a pulse group excitation mode with a duty cycle of less than 20% to drive the piezoelectric transducer array.

[0029] Preferably, step 105 further includes: identifying a fixed characteristic wavelength in the transmitted light intensity data where the rate of change of absorbance with respect to the concentration of the electroplating solution approaches 0, and determining the fixed characteristic wavelength as the absorption point of the main salt, etc.

[0030] An online detection system for electroplating additive concentration based on multispectral fusion includes a detection flow module, a sound field modulation module, an excitation light source module, a signal detection module, and an analysis and control module. The sound field modulation module is located on the side of the fluid channel of the detection flow module. It is used to generate a subcavitated standing wave sound field inside the detection flow module and use the spatial gradient acoustic radiation force field to drive the free bubbles generated by the hydrogen evolution reaction in the fluid medium to gather towards the acoustic pressure antinode region, so as to form a light transmission path without bubble scattering phase in the acoustic pressure node region. The excitation light source module and the signal detection module are located on opposite sides of the light transmission path, respectively. The analysis and control module is connected to the sound field modulation module, the excitation light source module, and the signal detection module, respectively. It is used to acquire the impedance phase angle of the sound field modulation module in real time and generate a synchronous trigger level signal when the impedance phase angle reaches the preset resonant extreme point. The excitation light source module is used to respond to the synchronous trigger level signal and emit detection beams in the ultraviolet, visible, and near-infrared bands into the light transmission path. The analysis and control module is also used to extract the reference transmitted light intensity of the corresponding absorption points of the main salt in the electroplating solution from the transmitted light intensity data, use the reference transmitted light intensity to compensate the absorbance parameters of the corresponding band to obtain the absorbance response sequence, and substitute the absorbance response sequence into the preset concentration inversion mapping logic to output the concentration values ​​of different organic additives in the electroplating solution. The signal detection module is used to collect transmitted light intensity data after passing through the light transmission path.

[0031] Example 1: In a high-density printed circuit board high-frequency dynamic copper electroplating production line operating continuously, cathodic electrolysis reactions continuously occur inside the electroplating solution matrix, precipitating micron-sized free bubbles. At the same time, the fluid contains a transition metal main salt with a concentration greater than 50 g / L. When a conventional detection beam penetrates the fluid medium with high turbidity and high concentration of matrix, the characteristic absorption signal of trace organic additives suffers from the dual physical erosion caused by Mie scattering induced by free bubbles and the broadband background absorption of the main salt. This results in the optical features characterizing the intrinsic electronic transitions of organic additive molecules being submerged in physical background noise. There is a physical contradiction between real-time online continuous measurement and the inherent high background absorption and high scattering physical microenvironment of the electroplating solution. Traditional technical approaches substitute the collected multispectral data into chemometrics or blind source separation models in order to extract signals through mathematical noise reduction. However, the aforementioned models that rely on data smoothing will produce boundary failures and baseline shifts when faced with non-stationary physical phenomena such as bubble distribution and flow field morphology transient distortion with production load, causing physical distortion of the extracted spectral features.

[0032] To address the distortion caused by algorithm-dependent fitting, the system abandons mathematical noise reduction calculations on the received random scattered noise. Instead, it constructs a detection flow cell containing a multispectral light source, an optical acquisition unit, and a piezoelectric transducer array. A high-frequency excitation signal is applied to the piezoelectric transducer array mounted on the sidewall of the detection flow cell, exciting a subcavitary standing wave acoustic field on the orthogonal cross section of the fluid medium. Utilizing the difference in compressibility between the fluid medium and free bubbles in the standing wave field, the free bubbles generated by the hydrogen evolution reaction are driven to gather towards the acoustic pressure antinode region through a spatial gradient acoustic radiation force field. This physical action occurs in the acoustic pressure nodal region. A bubble-free scattering phase light transmission path is constructed in situ. Based on this, the impedance phase angle of the piezoelectric transducer array is acquired in real time. Combined with the phase value corresponding to the peak frequency extracted by frequency sweep sampling of the admittance curve, a fixed trigger threshold is set. When the impedance phase angle reaches the trigger threshold and a resonant extreme point is generated, a nanosecond-level synchronous trigger level signal is output to drive a multispectral light source to emit ultraviolet, visible, and near-infrared light beams with a spot diameter less than 25% of the standing wave sound field wavelength into the light transmission path. The specific spot aperture constraint allows the energy of the detection beam to penetrate the low-density medium. In the acoustic pressure nodal region, the optical acquisition unit synchronously acquires transmitted light intensity data penetrating the physical light transmission path. The fluid spatial physical phase arrangement characteristics constructed by this subcavitary standing wave acoustic field provide a low-scattering physical channel prerequisite for multispectral beams. Based on the spatiotemporal synchronous excitation characteristics of the resonant phase threshold, the transmitted light intensity is converted into a high-fidelity optical signal relying on this channel prerequisite. These two technical features constitute a coupled causal chain of spatial phase shaping and precise temporal sampling, stripping random scattering factors from the optical path system within a single physical device architecture. The piezoelectric transducer array is connected to the fluid channel inside the detection flow cell. The acoustic impedance matching layer is made of polytetrafluoroethylene (PTFE) to isolate the piezoelectric crystal from the chemical damage caused by the highly corrosive electroplating solution matrix. The physical thickness of the acoustic impedance matching layer is set to one-quarter of the propagation wavelength of the sound wave inside the PTFE material corresponding to the center frequency of the high-frequency excitation electrical signal. By utilizing specific physical dimensions, the reflection of the interface sound wave caused by the difference in acoustic impedance between the fluid medium and the end face of the piezoelectric crystal is eliminated. This allows the mechanical vibration potential energy excited by the high-frequency excitation electrical signal to penetrate the acoustic impedance matching layer and be conducted into the fluid medium to form a subcavitary standing wave sound field, maintaining the long-term resonant output of the piezoelectric transducer array under fluid load conditions.

[0033] After acquiring transmitted light intensity data in the region penetrating the acoustic pressure wave node, the system identifies a fixed characteristic wavelength in the transmitted light intensity data where the rate of change of absorbance with respect to the electroplating solution concentration approaches 0 as the isoabsorption point of the main salt. The system extracts the reference transmitted light intensity corresponding to this isoabsorption point to establish a dynamic baseline attenuation benchmark. The absorbance parameters acquired in the ultraviolet, visible, and near-infrared bands are subtracted from the quantized value corresponding to this dynamic baseline attenuation benchmark to offset the broadband background absorption attenuation introduced by the main salt concentration drift. This process yields a multispectral differential feature vector characterizing the intrinsic transition absorption of organic additives, and outputs the measured concentration value of a specific organic additive in the electroplating solution. This process relies on the physical filtering of random Mie scattering by standing wave fields. It eliminates systematic errors introduced by the slow drift effect of the fluid medium system by using the baseline of the main salt isoabsorption point established by optical laws to counteract the path. This allows the analysis and control module to output quantitative parameters that map the physical facts of the organic additive concentration in a high-turbidity fluid test environment. The system reconstructs the spatial phase distribution order of the medium by acoustic radiation force field and anchors the reference optical path transmittance in real time by isoabsorption points. This changes the material boundary of optical intervention, eliminating the mathematical noise reduction problem of high scattering under the path of micro-environment purification. It transforms the noise reduction at the data level into a system engineering mechanism for constructing the underlying physical phase order.

[0034] Example 2: In the high-load dynamic electroplating solution of printed circuit boards, the high-speed hydrogen evolution reaction at the cathode generates micron-sized free bubble clusters, causing random fluctuations in the transmittance of the fluid medium. Conventional absorbance measurements face the challenge of signal distortion. The experimental data in this example comes from a dynamic fluid spectroscopy testing platform with an optical resolution of 0.1 nm and a piezoelectric excitation frequency coverage of 1.0 MHz to 5.0 MHz. This platform highly replicates the real industrial fluid shear environment through a mechanical pump-controlled flow path. Technical considerations for setting the high-frequency excitation electrical signal power density of the piezoelectric transducer array are also discussed. The goal is to achieve a balance between the spatial driving force of free bubbles and the avoidance of secondary cavitation in the fluid medium. When the dynamic viscosity of the monitored fluid increases and the base flow velocity is high, the power density tends to be at the upper limit of the limited range to overcome the hysteresis effect of fluid drag on bubble arrangement. Based on this judgment model and the current base fluid viscosity at a main salt concentration of 50 g / L, the power density of the high-frequency excitation electrical signal of the sample group of this invention is set to 0.3 W / cm², and the duty cycle of the pulse group excitation mode is set to 45.0%. To verify the anti-interference limit of the scheme in an industrial environment, the fluid pumping module holds... Gaussian white noise with a signal-to-noise ratio of 20dB is continuously injected, and a 50Hz power frequency vibration harmonic disturbance is superimposed in the acoustic field micro-environment. It is important to clarify that the 45.0% duty cycle set for this experimental environment refers to the proportion of energy release time of the overall waveform envelope signal that includes a relatively long rest period in the control logic. At the actual drive output of the system, the surface high-frequency carrier signal actually applied to the piezoelectric element within this overall envelope period is intercepted by the secondary forced blocking logic of the gated switch array to generate an extremely short acoustic burst of energy, which is then transmitted through the aforementioned underlying layer. The microsecond-level time slicing mechanism clamps and maintains the effective duty cycle of the surface layer, where the fluid medium directly bears substantial sound pressure work, within a preset red line range of less than 20%. Through a dual spatiotemporal decoupling mechanism—maintaining the continuity of bubble expulsion through overall arrangement and blocking thermodynamic accumulation through surface cutoff—the system eliminates the physical possibility of secondary acoustic cavitation nuclei at the bottom layer. In the fluid dynamics control process of the flow cell, the axial flow velocity setting of the fluid medium corresponds to the acoustic migration characteristics of free bubbles. Specific setting procedures include using a Doppler velocimeter to measure the axial flow velocity of the fluid medium flowing through the piezoelectric transducer array's operating region. Based on the diameter of the detection beam spot and the axial velocity of the fluid. Ratio calculation of bubble residence time The average acoustic migration time of free bubbles driven to the acoustic pressure antinode by the spatial gradient acoustic radiation force field is extracted by combining the application of high-frequency excitation electrical signals. Adjusting the fluid output of the fluid pumping unit to control the bubble residence time Greater than the average acoustic migration time Therefore, the specific operating range of the influent flow rate is determined to be 0.5 L / min to 2.0 L / min, forming an engineering boundary condition that maintains the bubble-free phase state of the acoustic pressure wave node. In the above-mentioned fluid dynamics control procedure, the system can independently complete the extraction of the migration parameter in closed loop without the need for any external optical or high-speed photography methods. During the construction of the acoustic pressure field, the analysis and control module continuously samples the dynamic electroacoustic impedance spectrum during the operation of the piezoelectric transducer array at high frequency. It utilizes the evolution of the equivalent physical density of the local mixed medium caused by the directional aggregation of a large number of free bubbles towards the acoustic pressure wave antinode to capture the damping slip characteristics of the resonant impedance trajectory. The system precisely extracts the steady-state arrival time difference from the start of the applied high-frequency excitation electrical signal until the impedance trajectory attenuation rate approaches zero, and explicitly links the time taken for the acoustic load to reach equilibrium state as the aforementioned average acoustic migration time. .

[0035] The experiment constructed a multi-dimensional control system including a comparative sample group, an out-of-range control group, and the sample group of the present invention to test the intermediate physical response data of the light transmission path. The comparative sample group was set to a static flow cell measurement mode without applying a subcavitary standing wave acoustic field. The out-of-range control group had its high-frequency excitation electrical signal power density set to 0.8 W / cm², exceeding the preferred range. The sample group of the present invention was loaded with a high-frequency excitation electrical signal of 0.3 W / cm². The measurement data showed that, under the dual perturbation of bubbles and power frequency vibration harmonics, the transmittance of the central optical path of the comparative sample group exhibited disordered oscillations between 42.5% and 86.7%. Under the same fluid disturbance conditions, the free bubbles in the sample group of the present invention were repelled by the spatial gradient acoustic radiation force field. The accumulation of acoustic pressure antinodes causes the transmittance of the light transmission path region of the acoustic pressure node to increase and remain stable within the range of 98.2% to 98.9%. The test results of the out-of-range control group show that when the power density exceeds the threshold, the transmittance of the light transmission path does not maintain a linear relationship, but decreases nonlinearly to 54.3%. The physical driving force of this nonlinear inflection point phenomenon is that the excessively high sound intensity directly induces a secondary cavitation phase transition in the fluid medium, and the emerging micro cavitation bubbles reconstruct the strong scattering physical barrier. The intermediate state data link of this group confirms that the parameter setting of 0.3 W / cm² is the preferred working window to avoid cavitation traps and overcome fluid drag. The sample group of this invention relies on the acoustic pressure node structure to strip away the physical noise substrate formed by random Mie scattering.

[0036] Based on the established scatter-free physical window, the experiment introduced a problem intensity gradient control system to address the slow drift interference of the main salt concentration. Three physical states were set as the main salt concentration fluctuation benchmarks: 40 g / L for a low gradient, 50 g / L for a medium gradient, and 60 g / L for a high gradient. The measurement deviation rate of the organic additive concentration retrieved by multispectral differential eigenvector inversion was tested. For the partially missing control group lacking dynamic baseline hedging logic, the relative standard deviations of the measured organic additive concentrations reached 5.4%, 12.8%, and 21.6% at the low, medium, and high main salt concentration gradients, respectively, showing a distortion trend that worsens with increasing matrix absorption. After acquiring multispectral transmitted light intensity data, the sample group of this invention… The reference transmitted light intensity is extracted from the characteristic wavelength where the rate of change of absorbance with respect to the concentration of the electroplating solution approaches 0. The absorbance parameters obtained in the ultraviolet and near-infrared bands are subtracted from the quantitative value of the reference transmitted light intensity. The output results show that under the three concentration perturbation gradients of low, medium and high, the relative standard deviation of the organic additive concentration measured by the sample group of this invention converges to 1.1%, 1.3% and 1.5%, respectively. The converged deviation response data confirms that the baseline offsetting path at the isoabsorption point smooths out the broadband background absorption attenuation. The spatiotemporal synchronous excitation and differential compensation mechanism maintain the optical resolution stability of the measurement system in the evolving physical flow field, transforming the measurement of trace physical properties under physical background into deterministic quantitative output.

[0037] Example 3: This example combines Figures 1 to 2 A description of an online detection method and system for electroplating additive concentration based on multispectral fusion, such as... Figure 1 As shown, an online detection method for electroplating additive concentration based on multispectral fusion includes guiding the electroplating solution to be tested and directing the fluid medium through a detection flow cell. Subsequently, a subcavitated standing wave acoustic field is generated to apply a high-frequency excitation signal to a piezoelectric transducer array. During this process, a synchronous trigger level signal is generated through real-time feedback, and the impedance phase angle of the transducer array is acquired to reach a preset resonant extreme point. Simultaneously, the generated acoustic field environment forms a bubble-free light transmission path, driving hydrogen evolution bubbles to converge towards the acoustic pressure antinode region to provide a scattering-free phase state. Then, under the action of synchronous triggering, transmitted light intensity data is acquired, and a multi-band detection beam is emitted in response to the synchronous triggering to penetrate the path. Next, the absorbance response sequence is obtained, and the reference transmitted light intensity of absorption points such as the main salt is extracted for compensation. Finally, the concentration value of the organic additive is output and substituted into the concentration inversion mapping logic to calculate the linear superposition relationship. Figure 2As shown, an online detection system for electroplating additive concentration based on multispectral fusion includes an analysis and control module, an acoustic field modulation module, an excitation light source module, a signal detection module, and a detection flow module. The analysis and control module serves as the core of the system, with its output pointing to the acoustic field modulation module and the excitation light source module to transmit control signals. The outputs of the acoustic field modulation module and the excitation light source module converge and point to the detection flow module, thereby causing the output of the detection flow module to point to the signal detection module for data acquisition. The output of the signal detection module ultimately closes the loop and feeds back to the analysis and control module.

[0038] Example 4: In the online continuous monitoring of the dynamic electroplating solution for printed circuit boards, temperature fluctuations and sudden changes in flow velocity within the fluid medium cause a shift in the resonant state of the piezoelectric transducer array. The spectral responses of the accelerator, inhibitor, and leveling agent overlap in the frequency domain. The trigger control and mapping model exhibit timing mismatch and component decoupling failures. The analysis and control module constructs a synchronous trigger correction link to match the sampling timing. The module continuously reads the admittance spectrum of the piezoelectric transducer array and extracts the real-time resonant frequency. It calculates the resonant frequency fluctuation value relative to the unloaded reference resonant frequency. The analysis and control module multiplies this resonant frequency fluctuation value by a value calibrated using the sound velocity-temperature gradient of the fluid medium. The hysteresis compensation coefficient, used to obtain the dynamic correction time, incorporates the fixed spatial optical path parameter from the piezoelectric transducer array to the center of the light transmission path, as well as the inherent sound velocity of the electroplating solution under a specific temperature reference state. Essentially, it is a physical transfer constant that enables dimensional transformation. The resonant frequency fluctuation value obtained by the analysis and control module represents the ratio parameter of the relative change in sound wave propagation rate with temperature. Through direct algebraic conversion between this ratio and the aforementioned constant containing distance-time attributes, the state offset in the frequency domain dimension is completely equivalently converted into the hysteresis increment of the acoustic transit time required for the light beam to penetrate the fluid medium. This ensures that the physical dimensions of the control model in the overall computational chain accurately converge to time. The analysis and control module calculates the sum of the inherent hardware response time and the dynamic correction time, inversely adjusting and determining the delay time relative to the synchronous trigger level signal. After the impedance phase angle reaches the preset resonant extreme point and a synchronous trigger level signal is generated, the analysis and control module drives the multispectral light source to emit detection beams in the ultraviolet, visible, and near-infrared bands into the light transmission path after the delay time. The pulse width of the detection beam is controlled to be less than 10ms. The optical acquisition unit collects the transmitted light intensity data after penetrating the light transmission path, extracts the reference transmitted light intensity corresponding to the absorption points of the main salt in the electroplating solution from the transmitted light intensity data, and uses the reference transmitted light intensity to compensate for the absorbance parameters of the corresponding band to obtain the absorbance. After obtaining the absorbance response sequence, the analysis and control module calls the concentration inversion mapping logic. The concentration inversion mapping logic contains a preset regression coefficient matrix. This regression coefficient matrix is ​​generated by using partial least squares to extract the characteristic wavelength absorbance of a standard electroplating solution sample of known concentration and performing dimensionality reduction training. The analysis and control module calculates the linear superposition relationship between the characteristic wavelength absorbance elements inside the regression coefficient matrix and the absorbance response sequence. The linear superposition relationship calculation process cancels the cross-absorption interference between the absorption peak of conjugated double bond electronic transition and the absorption peak of ether bond overtone vibration. Based on the calculation result of the linear superposition relationship, the analysis and control module outputs the concentration values ​​of accelerator, inhibitor and leveling agent in the electroplating solution.

[0039] Example 5: In the field deployment scenario, the analysis and control module initiates the calibration data acquisition procedure to construct the concentration inversion mapping logic. The fluid pumping unit dispenses multiple sets of standard electroplating solution samples with known absolute concentrations, containing gradient transition metal main salts and organic additive standards. The fluid pumping unit sequentially injects these samples into the detection flow cell. The analysis and control module drives the piezoelectric transducer array and multispectral light source to acquire the absorbance response sequence of each standard electroplating solution sample penetrating the acoustic pressure wave node region. The analysis and control module establishes the absolute concentration of the standard electroplating solution sample as the target dependent variable matrix and the absorbance response sequence as the independent variable matrix. The block calls the partial least squares algorithm to solve for the covariance structure of the independent variable matrix and the target dependent variable matrix. The analysis and control module extracts the principal components with a cumulative variance contribution rate greater than 95% from the covariance structure. Based on the principal components, the analysis and control module calculates the characteristic wavelength absorbance regression coefficient matrix and writes it into the memory address space to establish the basic mapping dataset for concentration feature decoupling. When the analysis and control module executes the concentration inversion mapping logic, it uses the standard normal variable transformation algorithm to preprocess the acquired absorbance response sequence to eliminate system optical path fluctuations and baseline physical tilt caused by residual surface suspended matter. During the multiplicative dimensionality reduction training phase, the top three principal components with the highest cumulative variance contribution rate are extracted from the covariance structure. The first principal component maps to the electronic transition characteristics of conjugated double bonds in the ultraviolet band corresponding to accelerators, the second principal component maps to the overtone vibration characteristics of ether bonds in the near-infrared band corresponding to inhibitors, and the third principal component maps to the residual of main salt concentration fluctuations. The analysis and control module constructs a feature wavelength absorbance regression coefficient matrix based on the extracted three principal components, quantifies and offsets the frequency domain cross-overlap interference of multi-band feature absorption peaks. To prevent semantic breakage between the dimensionality reduction abstract vector from a purely algebraic perspective and the actual chemical surface molecular structure, the system embeds material after the covariance solution process. The physical load mapping verification procedure involves the analysis and control module calculating the array of spectral load weight coefficients corresponding to the eigenvectors of each principal component in the wavelength domain, and calling the built-in standard pure component prior spectral library to perform characteristic band inner product cross-correlation comparison. When a principal component exhibits an extreme load weight distribution in a specific conjugated double bond ultraviolet characteristic band or ether bond near-infrared band, and its peak envelope shape and Spearman correlation coefficient with the quantum transition characteristics of the standard pure product cross the verification threshold, the algorithm formally archives it to the specific component sequence according to the physical mapping decoupling rule, thereby ensuring that the highly abstract statistical orthogonal dimensionality reduction space is rigidly constrained by the real physicochemical properties of the underlying material.

[0040] During the initial system commissioning, the analysis and control module adjusts the physical temperature of the fluid medium in 5-degree increments within the 20°C to 60°C range. The impedance analysis link sends a wideband sweep excitation signal to the piezoelectric transducer array. The analysis and control module extracts the admittance spectrum data for each temperature step and locks the peak frequency of the admittance curve. It calculates the linear slope between the temperature change and the peak frequency offset, outputting the hysteresis compensation coefficient corresponding to the sound velocity-temperature gradient of the fluid medium. A multispectral light source emits a continuous spectrum into the detection flow cell. The optical acquisition unit records the rate of change of transmitted light intensity under the cross-state of transition metal main salt concentration and temperature. Based on this rate of change sequence, the analysis and control module searches for a fixed characteristic wavelength where the rate of change of absorbance with respect to the electroplating solution concentration approaches zero. In the specific computational implementation architecture, the analysis and control module utilizes a discrete-time dimension-based approach. The central difference numerical approximation algorithm uses the continuously increasing calibrated concentration variable as the denominator reference data for calculation. The system performs first-order difference calculation based on the spectral absorbance matrix array in the continuous wavelength domain, and obtains and caches the first-order partial derivative spectral vector of absorbance shift with concentration. The microprocessor retrieves the cached sequence and performs finite difference iteration again to accurately obtain the second-order partial derivative array output at the spectral level. Finally, the system calls the absolute value minimization boundary optimization function to retrieve the zero-fluctuation anchor wavelength in the full-band dataset where the second-order value is lower than the tolerance lower limit threshold throughout the complete concentration perturbation cycle. The analysis and control module anchors the fixed characteristic wavelength as the absorption point wavelength value of the main salt of the electroplating solution. The analysis and control module synchronously writes the hysteresis compensation coefficient and the absorption point wavelength value of the main salt of the electroplating solution into the configuration register, and sets the acoustic-optical synchronization triggering sequence and dynamic baseline attenuation reference.

[0041] Example 6: In a scenario where continuous circulation of the electroplating fluid leads to the accumulation of deposits on the surface of the light-transmitting window of the detection flow cell and physical attenuation of the optical measurement baseline, the analysis and control module initiates an adaptive cleaning operation. The analysis and control module extracts the reference transmitted light intensity corresponding to the absorption points of the main salt in the electroplating solution from the transmitted light intensity data, and calculates the attenuation ratio of the current reference transmitted light intensity relative to the initial clean state calibration reference transmitted light intensity. When this attenuation ratio exceeds the cleaning trigger threshold of 10%, the analysis and control module disconnects the output link of the synchronous trigger level signal, setting the system to enter a non-detection cycle. During the non-detection cycle, the analysis and control module changes the excitation phase difference of the piezoelectric transducer array, changing the excitation phase difference between array elements from 0°. The phase constant is switched to 90°. At the physical structural level for this acoustic field mode switching, the piezoelectric transducer array, mounted on the sidewall of the detection flow cell, consists of multiple independent oscillator elements connected in series at equal intervals along the fluid flow axis. The spatial physical geometric spacing between the central nodes of adjacent piezoelectric elements is designed to be one-quarter of the ultrasonic wavelength corresponding to the current high-frequency excitation center frequency. This specific spatial quarter-wavelength array arrangement allows the electrical phase difference to generate a wave dynamic interference superposition response with the spatial geometric phase difference between the elements when the analysis and control module actively induces a 90-degree time lag phase in the electronic drive link, thereby breaking the original symmetry. Interference standing wave node balance excites asymmetric traveling wave energy with overall momentum transport capability to the specific fluid discharge side. This excitation phase difference switching action switches the subcavitated standing wave sound field inside the detection flow cell to an asymmetric traveling wave field. This asymmetric traveling wave field generates a directional acoustic flow along the sound wave propagation axis. The momentum gradient carried by this directional acoustic flow is superimposed on the axial fluid shear force generated by the electroplating solution pump circulation. The analysis and control module uses the directional acoustic flow generated by this asymmetric traveling wave field in conjunction with the fluid shear force to peel off the deposits on the surface of the light-transmitting window of the detection flow cell. During the peeling of the deposits, the analysis and control module drives the multispectral light source to emit a detection beam at fixed time intervals and monitors the recovery value of the reference transmitted light intensity. When the reference transmitted light intensity recovery value is greater than 98% of the initial clean state calibrated reference transmitted light intensity, the analysis and control module eliminates the excitation phase difference of the piezoelectric transducer array and restores the asymmetric traveling wave field to a subcavitated standing wave sound field. The analysis and control module restores the output link of the synchronous trigger level signal and enters the spectral detection cycle. This adaptive cleaning operation starts the excitation phase difference switching based on the comparison result of the reference transmitted light intensity attenuation ratio and the 10% cleaning trigger threshold. It uses the directional acoustic flow generated by the asymmetric traveling wave field in conjunction with the fluid shear force to peel off the deposits on the surface of the light-transmitting window of the detection flow cell, and restores the subcavitated standing wave sound field based on the comparison result of the reference transmitted light intensity recovery value and the 98% transmittance recovery threshold.

[0042] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0043] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for online detection of electroplating additive concentration based on multispectral fusion, characterized in that, Includes the following steps: Step 101: The electroplating solution to be tested is guided through the detection flow cell, and a high-frequency excitation electrical signal is applied to the piezoelectric transducer array set on the side wall of the detection flow cell to generate a subcavitated standing wave sound field in the fluid medium inside the detection flow cell. Step 102: Using the spatial gradient acoustic radiation force field generated by the subcavitation standing wave sound field, the free bubbles generated by the hydrogen evolution reaction in the fluid medium are driven to gather in the acoustic pressure antinode region of the subcavitation standing wave sound field, so that the acoustic pressure node region of the subcavitation standing wave sound field forms a light transmission path without bubble scattering phase. Step 103: Real-time acquisition of the impedance phase angle of the piezoelectric transducer array; when the impedance phase angle reaches the preset resonant extreme point, a synchronous trigger level signal is generated. Step 104: In response to the synchronous trigger level signal, drive the multispectral light source to emit detection beams in the ultraviolet, visible and near-infrared bands into the light transmission path, and control the optical acquisition unit to collect the transmitted light intensity data after penetrating the light transmission path. Step 105: Extract the reference transmitted light intensity of the corresponding absorption points of the main salt in the electroplating solution from the transmitted light intensity data, and use the reference transmitted light intensity to compensate for the absorbance parameters of the corresponding band to obtain the absorbance response sequence. Step 106: Substitute the absorbance response sequence into the preset concentration inversion mapping logic, and output the concentration values ​​of different organic additives in the electroplating solution by calculating the linear superposition relationship between the characteristic wavelength absorbance and the absorbance response sequence.

2. The method for online detection of electroplating additive concentration based on multispectral fusion according to claim 1, characterized in that, Step 104 further includes: monitoring the resonant frequency fluctuation value of the piezoelectric transducer array, adjusting the delay time relative to the synchronous trigger level signal inversely proportional to the resonant frequency fluctuation value, starting the multispectral light source after the delay time is reached, and controlling the pulse width of the detection beam to be less than 10ms.

3. The method for online detection of electroplating additive concentration based on multispectral fusion according to claim 1, characterized in that, In step 104, by controlling the collimation aperture of the multispectral light source, the spot diameter of the detection beams in the ultraviolet, visible, and near-infrared bands at the center of the light transmission path is made less than 25% of the wavelength of the subcavitating standing wave sound field.

4. The method for online detection of electroplating additive concentration based on multispectral fusion according to claim 1, characterized in that, When generating the subcavitary standing wave acoustic field in step 101, the power density of the high-frequency excitation electrical signal is controlled within the range of 0.1 W / cm² to 0.5 W / cm², so that the acoustic pressure amplitude inside the detection flow cell is lower than the cavitation pressure threshold of the electroplating solution.

5. The method for online detection of electroplating additive concentration based on multispectral fusion according to claim 1, characterized in that, The absorbance response sequence includes characteristic absorbance response values ​​corresponding to accelerators, inhibitors, and leveling agents in the electroplating solution, respectively; among them, the ultraviolet band corresponds to the absorption peak of the conjugated double bond electronic transition of the accelerator, and the near-infrared band corresponds to the absorption peak of the ether bond overtone vibration of the inhibitor.

6. The method for online detection of electroplating additive concentration based on multispectral fusion according to claim 1, characterized in that, Step 101 also includes an adaptive cleaning operation: during non-detection cycles, the excitation phase difference of the piezoelectric transducer array is changed to switch the subcavitated standing wave acoustic field to an asymmetric traveling wave field, and the directional acoustic flow generated by the asymmetric traveling wave field is used in conjunction with the fluid shear force to peel off the deposits on the surface of the light-transmitting window of the detection flow cell.

7. The method for online detection of electroplating additive concentration based on multispectral fusion according to claim 1, characterized in that, Step 103 further includes: performing frequency sweep sampling on the piezoelectric transducer array, extracting the peak frequency of the admittance curve to determine the preset resonant extreme point, and setting the phase value corresponding to the preset resonant extreme point as the trigger threshold for generating a synchronous trigger level signal.

8. The method for online detection of electroplating additive concentration based on multispectral fusion according to claim 1, characterized in that, Step 102 further includes: controlling the excitation frequency of the high-frequency excitation electrical signal to be in the range of 1.5MHz to 3.0MHz, and using a pulse group excitation mode with a duty cycle of less than 20% to drive the piezoelectric transducer array.

9. The method for online detection of electroplating additive concentration based on multispectral fusion according to claim 1, characterized in that, Step 105 further includes: identifying a fixed characteristic wavelength in the transmitted light intensity data where the rate of change of absorbance with respect to the concentration of the electroplating solution approaches 0, and determining the fixed characteristic wavelength as the absorption point of the main salt, etc.

10. An online detection system for electroplating additive concentration based on multispectral fusion, used to implement the online detection method for electroplating additive concentration based on multispectral fusion as described in claim 1, characterized in that, It includes a detection and flow module, a sound field modulation module, an excitation light source module, a signal detection module, and an analysis and control module. The sound field modulation module is located on the side of the fluid channel of the detection flow module. It is used to generate a subcavitated standing wave sound field inside the detection flow module and use the spatial gradient acoustic radiation force field to drive the free bubbles generated by the hydrogen evolution reaction in the fluid medium to gather towards the acoustic pressure antinode region, so as to form a light transmission path without bubble scattering phase in the acoustic pressure node region. The excitation light source module and the signal detection module are located on opposite sides of the light transmission path, respectively. The analysis and control module is connected to the sound field modulation module, the excitation light source module, and the signal detection module, respectively. It is used to acquire the impedance phase angle of the sound field modulation module in real time and generate a synchronous trigger level signal when the impedance phase angle reaches the preset resonant extreme point. The excitation light source module is used to respond to the synchronous trigger level signal and emit detection beams in the ultraviolet, visible, and near-infrared bands into the light transmission path. The analysis and control module is also used to extract the reference transmitted light intensity of the corresponding absorption points of the main salt in the electroplating solution from the transmitted light intensity data, use the reference transmitted light intensity to compensate the absorbance parameters of the corresponding band to obtain the absorbance response sequence, and substitute the absorbance response sequence into the preset concentration inversion mapping logic to output the concentration values ​​of different organic additives in the electroplating solution. The signal detection module is used to collect transmitted light intensity data after passing through the light transmission path.

Images