Method for optimizing electroplating and etching process parameters

Abstract

The invention relates to a method for optimizing electroplating and etching process parameters, which comprises the following steps of: establishing an interface migration behavior relationship by analyzing the influence of metal ions, doped impurities and surface roughness factors on a reaction interface migration rate and a boundary layer flow behavior in an electroplating and etching process; interface migration behaviors are combined with parameter sensitivity analysis, the influence of process parameters on the interface behaviors is dynamically evaluated, and priority optimization sorting is carried out; dynamic interval zooming is carried out on the sensitive parameters by using the fitted sensitivity function, and a local adjustment range is generated according to material types and graph complexity; a preorder process memory mapping relation is introduced, and morphology evolution information recorded in the electroplating process is transmitted to an etching parameter optimization module; local fine adjustment is carried out on the etching rate, angle control and a disturbance mechanism, and the previous electroplating state is responded; collecting a fluctuation interface feedback signal; establishing a closed-loop optimization control loop of micro-disturbance, response evaluation and parameter correction; and the process consistency and the structure uniformity are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to electroplating and etching process, in particular to the optimization method of electroplating and etching process parameters. BACKGROUND

[0002] At present, the traditional method for optimizing electroplating and etching process parameters, such as the parameter adjustment method of electroplating tool and electroplating method of Chinese patent CN114752986A, has certain improvement effect in the manufacturing process, but still has many technical limitations and systematic disadvantages, which is difficult to adapt to the continuous improvement demand of process response speed, prediction ability and system intelligent level in the current multi-structure, multi-material, high-density integrated micro-nano manufacturing scene. The core path of the existing technology is to adjust the parameters of the electroplating tool based on the thickness distribution result of the metal plating layer after electroplating, and the improvement logic belongs to the posterior feedback adjustment, that is, after completing the electroplating, the thickness of the plating layer on the circuit board is measured and fitted to generate a distribution map, and then the electroplating tool (such as anode plate layout, current density setting, etc.) is corrected and optimized. This strategy can realize local compensation based on the detection result, but it has several fundamental defects. First, the adjustment behavior depends on the measurement after processing, which is a typical post-process correction mechanism, meaning that the system cannot dynamically intervene in the non-uniform behavior during the electroplating process, but only can execute the compensation strategy in the next batch or cycle processing, resulting in response lag and lack of real-time closed-loop capability. Second, the parameter optimization of this method relies on the single physical quantity of plating layer thickness as the evaluation basis, which cannot reflect the comprehensive influence of multi-factor coupling on the interface behavior. In the actual microstructure or complex irregular area electroplating process, such as edge accumulation, local stress concentration, interface morphology mutation, etc., the ability to capture process abnormalities through thickness change is extremely limited, which may cause the phenomenon of controlling the macro but not the local.

[0003] Thirdly, the system adjustment range depends on the structural parameter settings of the electroplating tool, such as electrode shape, shield opening, position correction, etc., and the adjustment of these tool parameters itself has a long cycle, slow mechanical response, strong execution rigidity, and is extremely unsuitable for the current control requirements of micro-adjustment scale, minute-level control frequency in high-speed beat manufacturing process, and lacks soft adjustment flexibility; fourthly, this scheme cannot realize cross-process information transmission, and problems occurring in the electroplating process cannot be automatically transmitted to the subsequent etching stage, resulting in the entire processing flow still being divided into independent stages, and typical problems such as etching overshoot, edge erosion, and out-of-control aspect ratio cannot be predicted and fed forward from the abnormal behavior of the electroplating stage. In addition, from the perspective of data modeling, the comparative document does not propose a technical path for physical modeling or response identification of the electroplating process, its data structure is a two-dimensional static graph of thickness-spatial distribution, lacks time dimension behavior trend analysis, cannot support trend prediction and working condition evolution identification, and also lacks sensitivity sorting and target function optimization capability, which means that in the case of multiple target conflicts, such as simultaneously controlling thickness uniformity and stress release, this method lacks algorithm support and collaborative judgment mechanism; at the same time, it does not involve the etching process at all, and cannot meet the important needs of parameter linkage and upstream and downstream state complementation between electroplating and etching in the current manufacturing process, and cannot realize cross-stage intelligent control from electroplating parameter changes to etching parameter adjustment. SUMMARY

[0004] The purpose of the present application is to provide an electroplating and etching process parameter optimization method, thereby solving some of the problems and deficiencies pointed out in the background art.

[0005] The technical problems solved by the present application are solved by the following technical solutions: an electroplating and etching process parameter optimization method, comprising: establishing an interface migration behavior relationship by analyzing the influence of metal ions, doping impurities, and surface roughness factors on the reaction interface migration rate and boundary layer flow behavior during electroplating and etching; coupling the interface current density gradient, concentration polarization, and electrolyte shear rate physical quantities in a dynamic prediction framework to perceive and predict the position and dynamic changes of the actual reaction front; Combining the interface migration behavior relationship with parameter sensitivity analysis, dynamically evaluating the influence of process parameters including current density, temperature, pH value, and liquid flow rate on the interface behavior and performing priority optimization sorting; using the fitted sensitivity function to implement dynamic interval scaling of sensitive parameters, and generating a local adjustment range according to the material type and pattern complexity; The pre-process memory mapping relationship is introduced to transfer the morphology evolution information recorded in the electroplating process to the etching parameter optimization module; the etching rate, angle control, and disturbance mechanism are locally fine-tuned to respond to the pre-process electroplating state; the surface tension change, potential disturbance, and local reflectivity fluctuation interface feedback signals are collected; and a closed-loop optimization control loop of micro-disturbance, response evaluation, and parameter correction is established.

[0006] Further, the method for sensing and predicting the position and dynamic changes of the actual reaction front includes: A dynamic response relationship framework is established for analyzing the position evolution behavior of the process reaction interface under the coupling effect of multiple physical factors in the electroplating or etching process; information flow of the interface potential distribution, solute concentration offset, and fluid disturbance state is obtained and input into the dynamic response relationship framework; Based on the output results, process adjustment instructions oriented to interface control are generated, and parameter adaptive adjustment is performed in the actual processing process to control the uniformity of metal deposition or material removal behavior.

[0007] Further, the process adjustment instructions are prioritized according to the influence weight of physical variables in the interface response relationship framework on the process output results; a cross-stage parameter memory mechanism is established to transfer the residual stress distribution and plating layer non-uniformity information generated in the electroplating stage to the etching stage and dynamically adjust the etching rate and angle control strategy.

[0008] Further, the interface response relationship framework includes identification of the reaction interface curvature change for real-time monitoring of the side etching trend in the etching process and deviation correction through local flow field disturbance; a multi-objective optimization mechanism based on a target function is constructed, which includes multiple objectives of deposition uniformity, etching perpendicularity, reaction interface stability, and overall low stress; Specifically: an interface response identification framework is constructed to identify the local curvature fluctuation formed in the electrochemical process in the reaction area in real time and determine whether there is a risk of side etching; a local fluid disturbance regulation mechanism is introduced to apply directional micro-perturbation flow according to the interface morphology trend to realize dynamic deviation correction; a composite target optimization model is built to guide the optimal configuration of process parameters in the etching and electroplating stages with multiple quality control targets as constraint conditions; and the following composite control function is introduced to achieve the above objectives:

[0009] Wherein: is a composite response evaluation function, reflecting the position of the interface at time t; is the local reaction interface curvature rate of change, capturing the deformation trend of the boundary morphology, is the key factor for predicting the probability of side etching occurrence; is the target function partial derivative index set of the corresponding region, including: deposition uniformity tolerance function , perpendicularity control function , interface stability function , stress index function ; , used to measure the sensitivity of the region to the target function; is the region disturbance response gain factor, used to adjust the intensity of local feedback; is the change rate of the interface control signal (such as tension, potential) per unit time, indicating the difficulty of dynamic adjustment of the response system at the point; is the comprehensive control signal function, reflecting the real-time state of the process reaction interface at position , time , its value is composed of surface tension, potential perturbation and optical reflectivity according to the set weight; By multiplying the curvature fluctuation by the target function partial derivative aggregation term, and coupling with the modulation term of the local feedback rate, the response function is formed, which is calculated in real time. The value of all key regions is calculated periodically: When the value exceeds the set threshold, it indicates that there is a risk of morphology mutation in the region; the control system will fine-tune the parameters of the region through the perturbation mechanism, including fine-tuning the flow direction, changing the polarity, adjusting the local temperature of the electrolyte; at the same time, according to the weight, automatically balance the contradictions between multiple targets, and seek the optimal comprehensive performance.

[0010] Further, the dynamic response relationship framework performs real-time sampling of parameter sensitivity through perturbation heuristic method, and dynamically induces the response trend of different variable inputs.

[0011] Further, the response pre-sequence electroplating state method comprises: In the electroplating stage, the surface to be processed is subjected to staged data acquisition, and the spatiotemporal change information of the material morphology occurring during the information deposition process is recorded; the spatiotemporal change information is encoded in the form of spatial structure data and stored as a parameter input set called by the processing link; in the etching stage, the pre-process affected region is identified according to the data set, and the local processing strategy is parameterized adjusted in combination with the current graphic structure.

[0012] Further, the morphology change information collected in the electroplating stage includes abnormal behaviors of surface accumulation height difference, thickness gradient distribution, stress concentration area, and edge structure discontinuity; the morphology information is used to construct a distribution map of processing sensitive areas in the etching link, to guide setting of etching liquid flow path, flow rate intensity, or local disturbance mode for different areas; spatial marking is implemented on the surface state of the pre-process, and the marked information is called in real time during etching to dynamically adjust etching time, angle, or direction.

[0013] Further, a response rule is established between the morphology information and the etching control strategy through a mapping model, the response rule is matched based on different defect types corresponding to different processing parameter sets; a plurality of collected information sequences of the pre-process are superimposed and analyzed, and trend features evolving over time are extracted to predict potential processing errors and correct etching behaviors in advance.

[0014] Further, the method for establishing a closed-loop optimization control loop comprises: Change information of the interface state including liquid phase tension behavior, electrochemical reaction activity index, and instantaneous response characteristics to external light or energy irradiation under microscale is collected; the change information is analyzed to identify abnormalities of interface response characteristics; local control instructions are generated based on the analysis results to adjust target parameters in the process cycle.

[0015] Further, a disturbance with controllable amplitude is applied to the control variable in the process cycle, and the sensitivity level of the current working condition is judged based on the interface physical response trend; when the interface state change information has unstable behavior, the range adjustment is preferentially implemented on the electrolyte supply rate, temperature distribution, or local charge density.

[0016] The beneficial effects of the present application are: by introducing a dynamic response model based on interface behavior, real-time monitoring and adjustment are performed on various key parameters (such as current density, solution concentration, local stress, etc.) in the electroplating and etching process, the deposition thickness distribution and etching profile are effectively controlled, the overall structure edge flatness, line width consistency, and perpendicularity are significantly improved, and the application is suitable for high-density circuit boards, microstructure devices, and other scenes. A parameter mapping mechanism is established for transferring morphology evolution information of the pre-process (electroplating) to the post-process (etching), so that the system can identify abnormalities in the early deposition process such as thickness fluctuation and stress concentration and respond in advance in the etching link, realize true cross-stage adaptive optimization, and avoid accumulation of pre-process deviations causing post-process defects.

[0017] By introducing a composite objective function, multiple targets such as deposition uniformity, etching verticality, interface stability, and stress minimization are included in the overall control system. The system can dynamically balance the weight relationship between each target, achieve overall balanced optimization of quality control, and meet the process difference requirements of multiple structures and multiple materials. Using the perturbation trial method and closed-loop feedback mechanism, the system can quickly identify parameter sensitivity and give early warning during the process running period. Through fine adjustment of local fluid disturbance, electric field polarity switching, electrolyte replenishment strategy and other measures, the system can actively intervene in the unstable trend of the interface, avoid overreaction or loss of control, and reduce the dependence on human intervention. BRIEF DESCRIPTION OF DRAWINGS

[0018] Figure 1 The flow chart of the electroplating and etching process parameter optimization of the present application.

[0019] Figure 2 The functional relationship diagram of the process optimization based on interface response and sensitivity analysis of the present application.

[0020] Figure 3 The regulation diagram of the present application fusing the preceding information and interface closed-loop feedback.

[0021] Figure 4 The dynamic response and multi-target intelligent regulation flow chart of the high-density PCB electroplating-etching process of Example 1 of the present application.

[0022] Figure 5 The PCB microstructure processing flow chart of Example 2 of the present application across stage state memory and defect feedforward inhibition.

[0023] Figure 6 The PCB process optimization flow chart of Example 3 of the present application microscale multi-modal interface monitoring and priority closed-loop regulation. DETAILED DESCRIPTION

[0024] The specific embodiments of the present application will be described in detail below with reference to the accompanying drawings.

[0025] The specific embodiments of the present application will be described in detail below with reference to the accompanying drawings. Figure 1The application is an electroplating and etching process parameter optimization method, which establishes a multi-physics field coupling model around the evolution behavior of the reaction interface in the metal deposition and material removal process to improve process consistency, stability and precision control ability. Key process variables in the electroplating and etching process are collected through experimental means and process modeling, including metal ion concentration distribution, impurity doping ratio change and substrate surface roughness and other characterization parameters, and then the dominant influence of the above factors on the reaction interface migration rate is analyzed, the relationship between the flow disturbance state and mass transfer efficiency in the boundary layer is identified, and a set of mapping relationship functions describing the dynamic behavior of the interface advancing is constructed. Further, the interface behavior function is embedded into a dynamic predictive control framework, which takes the local current density gradient at the interface as the input of the electric field driving force, the polarization zone response behavior formed by the concentration gradient as the quantitative index of the mass transfer limitation, and the shear rate distribution of the electrolyte in the microscale region as the independent variable, which are considered and coupled at the same time, forming a dynamic model for predicting the advancing trend of the reaction interface in the space-time domain. The model can realize the continuous prediction of the migration path, speed change and morphology evolution mode of the reaction interface at different processing stages, and the prediction results are used for real-time regulation of key process parameters such as current waveform, liquid disturbance direction, electrode polarity switching frequency, etching liquid spraying angle, etc., so as to realize the feedforward control and precision correction of the interface behavior.

[0026] Combining the drawings Figure 2, propose an interface response behavior and parameter sensitivity coupling analysis based on the construction of response-regulation integrated mechanism to realize real-time optimization control of interface stability and processing precision under multi-variable environment. Based on the interface migration behavior model established in the early stage, the characteristic data of interface advancing rate, morphology change and stable boundary under different process conditions are extracted, and the correlation analysis between the data and the variation data of core parameters in the process is carried out; by introducing the multivariable sensitivity analysis mechanism, the influence intensity of multiple control parameters including current density, temperature, pH value and liquid flow rate on interface behavior is dynamically calculated, and based on the influence degree, the parameter response priority matrix is constructed to guide the allocation path of optimization resources; the sensitivity analysis process is based on historical sample regression modeling or real-time multi-parameter micro-disturbance trial method to obtain, by fitting the nonlinear response relationship between parameter change and interface state change, a set of sensitivity function family, on this basis, the dynamic interval scaling control is implemented for the process parameters identified as high sensitivity, that is: under the premise of not affecting the stable operation of the system, the adjustment step is narrowed and the response frequency is improved, so as to enhance the regulation precision; at the same time, combined with the material types (such as metal types, doping system, alloy ratio) and pattern complexity (such as line width change, structure distribution density, edge transition morphology) in the current processing task, the system generates the parameter adjustment range and allowed disturbance amplitude of each local area according to the matching rules, and realizes the spatial differentiation regulation strategy with strong pertinence.

[0027] Combined with the attached Figure 3, a pre-sequence process information transmission and interface level feedback closed loop control are proposed, through the construction of cross-stage data mapping mechanism and multi-signal perception driven intelligent regulation path, the reaction consistency and system adaptability in microstructure processing are significantly improved. The workpiece surface is collected in stages in the electroplating stage, including deposition thickness change trend, local stress distribution, edge accumulation or hole formation and other dynamic behaviors, and these process evolution information is encoded through spatial structure mapping method, stored as a standardized data structure and input to the etching process optimization module; the etching module receives and analyzes the previous electroplating state data, and correspondingly identifies the easy side etching area, accumulation abnormal area or structure transition sensitive area, and adjusts the local etching rate based on this, implements fine correction of the spray angle or ion bombardment direction, or enhances the reaction uniformity of a specific area through fluid disturbance, to realize the accurate matching response of the processing behavior and the electroplating residual state; in order to further strengthen the real-time and accuracy of the response mechanism, a multi-modal interface sensing method is introduced in the etching process, through the embedded micro-sensing array to synchronously collect surface tension fluctuation, electrochemical potential disturbance and regional reflectivity change and other signal characteristics, such signals have very high reaction sensitivity and time domain resolution, and can be used as direct feedback indicators of interface stability and reaction activity changes; on the basis of feedback, a closed loop optimization loop of micro-disturbance-response evaluation-parameter correction is established, that is, the system periodically applies a small amplitude disturbance to the key parameters (such as flow rate, current, liquid temperature, etc.), and evaluates the response amplitude and direction of the interface signal caused by the disturbance, so as to judge whether the current working condition is in a stable window or close to a critical deviation state, if the evaluation result exceeds the set threshold, the control system triggers the regional fine tuning logic, and through the micro-amplitude correction of the operation parameters, the abnormal trend is converged and the ideal processing trajectory is restored, this process does not need to interrupt the process flow, and can realize local closed loop fast correction in the continuous running state of the process.

[0028] Example 1: Combined with the attached Figure 4In this embodiment, a method for sensing and predicting the position and dynamic changes of the actual reaction front is adopted, and through the construction of a dynamic response relationship framework, real-time analysis and regulation of the reaction interface under multiple physical coupling effects are realized. In the manufacturing of a certain high-density interconnection structure PCB, the goal in this manufacturing scenario is to form a uniform thickness, no edge over-deposition, and unobstructed inner hole copper plating layer on the surface of a 10-layer multi-via circuit board, followed by fine pattern etching to obtain a 50-micron line width conductor structure. The entire process requires high edge consistency, low stress distribution, and minimization of etching residues. During the plating process, the dynamic response relationship framework is used to model the reaction interface. This framework forms a complete interface behavior mapping by collecting information streams such as potential distribution at every 5-millimeter grid point in the work area, copper ion concentration offset in the electrolyte, and shear flow state generated by local stirring. The system discovers through the sensing module that in the board edge area, the potential rise speed is slightly faster, and the local concentration offset is significantly higher than the central area (maximum offset reaches 0.7 mol / L), and the shear disturbance is weak. It is preliminarily judged that there is a trend of local over-thick plating due to current density concentration. In response to this trend, the system reduces the current density in the edge area by 12% at the 20th minute, increases the local circulation flow rate of the electrolyte by 15%, and introduces pulse current modulation, adjusting the pulse frequency to 20 Hz to improve the uniformity of charge distribution. The monitoring data after processing shows that the edge thickness growth rate decreases by about 18%, and the potential fluctuation tends to be stable, effectively controlling the risk of over-deposition. After plating is completed, the etching process stage is entered, and the system transfers the interface response data collected in the previous stage to the etching control module. In the graphical area, a compensation strategy of reducing the etching speed by 10% is set for the area where the previous plating thickness is slightly higher, and the spray angle is changed by 6 degrees from the standard direction to slow down the side etching tendency. At the same time, a local fluid disturbance mechanism is started in the hole position to avoid hole mouth residues. During the etching process, the system continues to track the reflectivity change as a reference index for the advancement of the interface front. When the system finds that the reflectivity change curve of a certain area tends to plateau within 5 seconds, it means that etching is slowing down, and the control system increases the etching liquid pressure in that area by 3%. Finally, the conductor structure edge is steep and the morphology is clear, with a line width error of ±1.5 microns, and the hole residue rate is less than 0.2%.

[0029] In this manufacturing procedure, the system performs priority evaluation on the reaction interface parameter influence after the electroplating is completed, and according to the historical data in the dynamic response relationship framework, the collected variables, including the current density distribution, the copper ion concentration gradient in the electrolyte, the potential fluctuation trend and the flow field shear rate intensity, are weighted and sorted. By comparing the variable with the historical response data of the plating layer uniformity, interface roughness and stress residual rate, the system analysis shows that the influence coefficient of current density on the final plating layer thickness difference is 0.52, the influence weight of concentration polarization is 0.31, and the correlation between shear rate and stress distribution is the lowest, only 0.12. Accordingly, the system identifies the current density change as the first-level core variable, and the corresponding control logic is preferentially deployed in the next-stage etching strategy; at the same time, the concentration offset and the boundary layer disturbance are set as the secondary reference parameters for dynamically fine-tuning the etching liquid component compensation strategy. Before entering the etching stage, the system calls the cross-stage memory module established in the electroplating process to read the residual stress distribution map and the plating layer thickness contour map fitted by the visual system on each PCB. Taking the 32nd PCB as an example, due to the local potential being too high in the early stage of electroplating, a micro-arch-shaped stress band is formed in the 7th to 10th line area, with the maximum in-plane residual stress value being 48 MPa (higher than the upper limit of process control 40 MPa), and the edge plating layer thickness is 6 microns higher than the center on average. After the system identifies this structure, it specially adjusts the etching strategy in this area: on the one hand, the spraying angle is adjusted from the standard vertical direction to outward deflection of 7 degrees to alleviate the lateral expansion trend of the stress band area and prevent the etching recess from enlarging; on the other hand, the etching rate is reduced to 85% of the standard value to give the edge area more reaction time to gradually balance the stress gradient. In the actual execution process, the system continues to collect the surface tension and reflectivity fluctuation, and detects that the reflectivity fluctuation of the stress band area tends to be stable during the 11th to 12th minute, judging that the morphology response has completed the transition, so the etching rate is restored to the normal value. The final test results show that the line width of the 32nd plate in this area remains at 48.7±1.1 microns, the side etching depth is controlled within 2.3 microns, and the residual stress is reduced to 22 MPa after etching, indicating that the local adjustment strategy driven by the cross-stage memory has achieved effective response and structure repair to the previous morphology abnormality.

[0030] Further introduce the curvature recognition mechanism and multi-objective function control strategy of interface response behavior to realize real-time perception and dynamic correction of the etching stage side etching trend, and combine the quality targets of the whole process to build a composite optimization path to ensure the dual requirements of process overall stability and local precision. In the application process, for the 44th PCB sample, the edge area of the electroplating layer presents irregular convex features in the microscopic scanning, and after fitting the curvature change rate, the local curvature change peak value is , which is much higher than the average level of the central area . The system identifies this area as a high-risk zone for potential lateral erosion. At this point, the control system activates the interface response recognition module, based on a pre-defined composite function:

[0031] Real-time assessment of the area, including The objective function includes four weighted index functions, derived from historical data analysis of the sample area: Depositional homogeneity objective function. =0.36, Etching Verticality Control Function =0.41, Interface stability function =0.33, stress response index function =0.27, calculated as follows:

[0032] Simultaneously, the local disturbance response gain factor Set in range The current region is determined to be 1.05 by system diagnosis; interface integrated control signal function. Rate of change per unit time at t=12min The value is 0.024 (dimensionless after normalization). Combined with the curvature value, the current region can be obtained:

[0033] The system's response risk threshold was set to 0.065, therefore this area was determined to be in an over-threshold state, triggering the dynamic fine-tuning procedure. The system immediately shifted the local jet direction of the etching solution 3° away from its original central axis, while simultaneously increasing the liquid temperature in the area to 107% of the original set value to improve the reaction speed and adjust reaction uniformity. Furthermore, the polarity switching frequency was increased from the standard 20Hz to 25Hz to stabilize the electric field distribution. After the entire adjustment lasted 4 minutes, the system detected... The value dropped to 0.058, below the threshold, indicating a noticeable correction effect. After etching, the linewidth in the area measured by a laser profilometer converged to 49.2 μm, and the lateral etching depth was 1.9 μm, a reduction of 47.2% compared to the average lateral etching depth of 3.6 μm in the unoptimized control sample, resulting in a 12% improvement in structural edge smoothness. This process was also simultaneously incorporated into the overall quality control objectives, among which... The weights of each objective function under this item can be adjusted according to actual process requirements. The recommended value range is as follows: , , , The total is maintained between 1.2 and 1.6 to ensure the responsiveness of multi-objective optimization.

[0034] In the process of electroplating and etching optimization control of the 44th PCB sample, the system successfully controlled the local side etching problem through curvature recognition and multi-objective optimization function control. To further improve the adaptive ability and scene generalization level of the control system, a micro-disturbance exploration mechanism was enabled to dynamically sample and respond to multiple control variables, establishing an accurate variable sensitivity model. During operation, the control system selected four core process parameters as exploration variables: current density, local liquid flow rate, electrolyte temperature, and spray angle. In the main processing interval from the 5th to 20th minute, the system applied a minimum amplitude disturbance to these four variables every 180 seconds, with the amplitude controlled within ±3% of the standard set value, ensuring that the main process target was not disturbed while stimulating the system's micro-reaction behavior. For example, at the 8th minute, the system adjusted the edge area current density from the original set value to and observed the resulting surface tension, potential disturbance, and reflectivity fluctuation changes. Statistical measurements showed that the surface tension decreased by 0.017 N / m, the potential changed by 14 mV within 2 seconds, and the reflectivity increased by 0.4%. In the same area, at the 11th minute, the liquid flow rate was disturbed by an increase of 2.5%, but the corresponding interface response changes were very weak, with a tension decrease of only 0.004 N / m and a potential fluctuation of less than 5 mV. The system thus determined that, under the given material and structure parameter combination, the liquid flow rate had a relatively small impact on the interface reaction state, while the current density change had a significant impact on the interface sensitivity. Through this mechanism, the system constructed a local variable influence weight database, mapped the response curve convergence results of different variables to a process behavior model, and iteratively updated the sensitivity ranking table in real time, which was used to rearrange the adjustment priorities in the next round of target function control. Taking the data measured in this test phase as an example, in the under-etching area at x = 14.5 mm, the system identified that the current density was a first-level high-sensitive variable, the spray angle was a second-level high-sensitive variable, and the temperature and flow rate had limited impact. Subsequently, in the etching refinement stage in this area, the system only implemented the main adjustment mechanism for the current density and supplemented the linkage correction for the spray angle, while the remaining variables remained at the original set values, avoiding parameter coupling disorder caused by excessive intervention. This mechanism also adapted a completely different priority ranking in the center area of the 46th PCB than in the previous sample, where the electrolyte temperature had the greatest impact on the verticality control in high-line density areas. The system accordingly prioritized temperature adjustment and suppressed the thermal stress concentration trend. Finally, through comparison tests on five groups of samples, it was found that after adopting the optimization control strategy with micro-disturbance exploration and sensitivity induction, the overall line width consistency improved by 9.8%, the local stress fluctuation decreased by 23%, and the process adjustment frequency decreased by about 31%, significantly improving the process stability and control efficiency.

[0035] Example 2: In combination with the attached Figure 5, based on embodiment 1, a state memory and regulation linkage mechanism across process stages is established, so that the morphology evolution data left over from the electroplating stage can be accurately identified and efficiently called in the etching process, thereby realizing targeted processing strategy adjustment and defect risk pre-feeding inhibition. In the electroplating stage of the 51st PCB sample, the system enabled a staged data acquisition mechanism, periodically acquiring data such as thickness, stress, morphology roughness, and edge accumulation trend of the board area every 10 minutes during the entire 90-minute plating process; for example, in the 30th to 60th minute interval, the system detected that the plating layer thickness growth rate of the 4th to 6th column conductor unit area was slightly higher than that of the central area, with an average growth rate of 0.43 μm / min, about 16% higher than the full-board average, and the local maximum tensile stress value in the area stress analysis reached 62 MPa, more than 15% above the recommended upper limit. Based on these data, the system generates a spatial structure coding matrix, compresses and encodes information such as area location, thickness increment, stress gradient, and roughness coefficient, and uploads it to the central control module, marking it as a high-risk area for edge accumulation as control input for the subsequent etching stage. After entering the etching process, the system calls the aforementioned coded data and matches it with the current graphic structure model in space, confirming that the area corresponds to a dense segment of graphic path corners, and reasoning from the historical etching model database that such areas are prone to nonlinear side etching and path deviation problems during etching. Therefore, the control system preferentially deploys local strategy adjustment in this area, including reducing the standard etching liquid spray angle from 90 degrees to 83 degrees to more stably cover the non-flat area, increasing the etching liquid pressure by 12% to enhance the solution update frequency in the boundary area to prevent bubble retention, and extending the process time for this segment by 9% to give enough reaction time. The system monitors the interface reflectivity and electrochemical disturbance signals in real time throughout the process, and when it finds that the adjusted 5th unit area has a stable reflectivity change trend at 13 minutes, it determines that the edge structure has stabilized, and then the system restores the area to the standard setting to ensure overall processing efficiency. After etching is completed, the laser profiler detection shows that the edge etching structure depth of the 5th unit is controlled at 3.1 μm, and the side etching angle is controlled at 87.4 degrees, while in the 50th sample without the optimization scheme, the corresponding area side etching angle is only 83.2 degrees, and the structure edge appears discontinuous cracks, indicating that the parameterized fine-tuning strategy based on the electroplating state response successfully realizes the prediction and processing of defect risk areas. Through comparative analysis, in the continuous processing of 10 batches of samples, this mechanism reduces the etching error rate by an average of 27% and improves the edge consistency index by 13.4%.

[0036] In the electroplating process, the system uses a high-precision laser confocal measuring instrument to sample the three-dimensional topography data of the key areas on the board surface every 5 minutes, and combines with the stress infrared mapping and reflectivity imaging system to obtain the material state evolution trajectory. In the deposition window of the 45th to 75th minute, the system collects multiple key data: the surface accumulation height difference in the intersection area of x=42.8mm and y=18.4mm is recorded as 7.6μm, which is 28% higher than the adjacent area; the thickness gradient in this area fluctuates more than 4.3μm within 5mm, and the stress monitoring data shows that the maximum concentrated stress reaches 68MPa, which is much higher than the average value of 43MPa in the central area; at the same time, this edge area is also accompanied by a micro-scale crack propagation trend, which is identified as an edge structure discontinuity anomaly by optical imaging. The system encodes the above topography information into a four-dimensional structure marker data package, including position coordinates, anomaly type, amplitude level and influence range, and stores it in the central control module in a spatial hash structure to form a complete processing sensitive area distribution map, which is marked as a priority response area. In the etching stage, the system loads the distribution map and compares it with the current wire pattern grid to confirm a consensus area of 5 blocks. The system uses the abnormal characteristic values of the marked areas as parameter control basis, and at x=42.8mm, because it identifies that there is a high accumulation and high stress superposition behavior in this area, the control system actively increases the initial flow rate of the etching liquid in this area by 22%, and adjusts the angle combination of the three-way spray head to form a 15-degree skewed spiral flushing to improve the local flow shear effect and enhance the boundary solution replacement efficiency. In addition, considering the stress directionality, the system dynamically adjusts the etching time of this area to 110% of the standard time to weaken the residual stress of the microstructure in stages and prevent the edge erosion effect induced by stress release. At the 16th minute, the system updates the surface tension gradient curve according to real-time feedback and retrieves the position data of the high crack sensitive area marked in the previous space, and adjusts the electrolyte direction angle at y=18.4mm from the standard vertical to 12° deflection, and reduces the voltage pulse amplitude by 3%, to achieve a more moderate interface advancing strategy to reduce the secondary defects induced by topography mutation. After etching, the system outputs a comparison report showing that the average edge linewidth of the marked sensitive area is 49.1μm, with a standard deviation controlled within ±0.8μm, and the edge micro-crack propagation rate is 0.4%, which is much lower than the average value of 1.3% in the unmarked area. In addition, in the five priority response areas guided by the space marking and etching path and parameter fine-tuning, the overall side etching angle is improved to 87.2 degrees, and there is no obvious stress fracture defect.

[0037] By establishing a learnable and inductive mapping matching system and an evolution trend recognition algorithm, the system achieves classification responses to various electroplating anomalies and feedforward correction of etching paths. The system labels and classifies key morphological data collected during the preceding electroplating stage, identifying three typical defect types: the first type is the surface buildup excess zone, characterized by a local thickness difference exceeding 4.5 μm; the second type is the gradient stress zone, with an average stress change rate greater than 12 MPa / mm; and the third type is the edge crack initiation zone, with a decrease in surface optical continuity exceeding 2.8%. Based on a response rule mapping model trained using previous case data, the system calls a pre-set parameter matching strategy set to bind processing parameter combinations to these three types of defects: for example, for the buildup excess zone, the matching strategy is to increase the local etching solution exchange frequency, enhance the disturbance flow rate, and appropriately extend the processing time; for the gradient stress zone, the strategy is to reduce the electrolyte temperature, slow down the electric field change amplitude, and implement angled spraying; and for the crack-sensitive zone, the strategy is to set a low-impact etching path, flexible waveform pulses, and high-frequency detection circuit monitoring. For this sample, over 780 sets of process data sequences were collected between the 40th and 90th minutes of electroplating. The system processed these time-series data using a superimposed sliding window analysis method to form a characteristic evolution trajectory map. It identified a continuously increasing thickness fluctuation sequence between x=36.4mm and 42.8mm, with an average growth rate of 0.29μm / min. Based on this, the system predicted that this area would have potential edge erosion or etching-through risks during etching, and marked this area as an evolution warning zone, allowing for early intervention and corrective measures. During the etching process, the system implemented a three-stage response strategy in this area: the first stage reduced the etchant concentration by 2.6% to decrease the initial corrosion rate; the second stage used a 7-degree angle spray to target and weaken the corrosion between the 10th and 14th minutes; and the third stage dynamically monitored reflectivity, restoring normal parameters once the fluctuation stabilized. After etching, the test results showed that the etching depth at the edge of the warning area was stable at 3.0 μm, with no continuity breaks, and the etching uniformity error was ±1.2%. However, in sample No. 48, which did not implement predictive correction, two small-sized bitten edge structures existed in a similar area, affecting conductivity. Further statistics showed that during the trial production of 12 consecutive batches of samples, after adopting the mapping model and evolution prediction method, the system achieved a 93.2% hit rate in handling different anomaly types, a 78.5% early correction rate, and an average 10.7% improvement in etching yield.

[0038] Example 3: Combined with appendix Figure 6, based on embodiment 2, to solve the potential etching unevenness problem caused by the edge over-deposition and stress gradient accumulation in the previous electroplating stage, the system enables a micro-scale interface monitoring device, six groups of multi-modal detection units are arranged on the working surface, each group integrates tension micro-sensor, potential disturbance probe and optical reflectivity sensor, and collects real-time state change information of the reaction interface at different times and different areas. At the 7th minute of etching, the system monitors the detection node at x = 28.4mm position to detect that the liquid phase tension fluctuation decreases from the original value of 29.7mN / m to 26.1mN / m, the short-period oscillation frequency of potential increases from 8Hz to 12Hz, and the reflectivity response signal decreases by 0.9%. The system internal identification module compares the model database and judges that the interface is in an unstable early warning state, the main risk is that the etching depth deviation increases and the edge stress induces erosion. The control system immediately generates local control instructions based on the abnormal signal, first increases the etching liquid flow rate of this area by 8.5% to increase the boundary shear rate; secondly, adjusts the electrolyte temperature from the standard 42℃ to 39℃ to slow down the excessive reaction rate; thirdly, adjusts the electric field polarity switching frequency from 15Hz to 18Hz to suppress the asymmetry of interface charge accumulation. After that, the system continues to track the state of this area, at the 11th minute, the tension returns to 28.6mN / m, the potential oscillation tends to be stable, and the reflectivity curve rises to the pre-offset level, the system determines that the area is stable and recovered, and the instructions withdraw all temporary adjustments and restore the normal parameter process. After the etching of this area is completed, the profilometer detection result shows that the etching depth error is ±1.1μm, which is much better than the control area of ±2.6μm without closed-loop adjustment. At the same time, after analyzing the data of the whole plate, it is found that the average etching error of the sample plate with closed-loop optimization mechanism is reduced by 32%, the micro-crack density of the structure edge is reduced by about 41%, the system feedback adjustment frequency is 4.2 times per plate on average, and the average delay time of control instructions is only 1.3 seconds.

[0039] In the 52th processing of the batch template, the system enabled the micro-perturbation heuristic mechanism in the electroplating stage, and applied a perturbation amplitude of no more than ±2.5% to the current density, electrolyte flow rate and bath temperature in each 20-minute process section to observe the changes in the interfacial tension, potential perturbation frequency and reflectance signal response. Among them, at about the 35th minute of electroplating, the system increased the liquid flow rate at the node of y=16.2mm by 2.1%, and after 180 seconds of perturbation, the system observed that the tension response at this point decreased by 1.9mN / m, the potential perturbation frequency increased from 9Hz to 12Hz, and the reflectance amplitude reached 1.1%. All three signals deviated significantly from the standard trend band, and after evaluation by the built-in sensitivity judgment model of the control system, this area was marked as a high sensitivity area. When the process was pushed to the 42nd minute, even without applying new perturbations, the system still monitored that the charge distribution in this area was abnormally concentrated, and the potential difference fluctuated by 41mV within 60 seconds, which was judged by the system as a potential unstable trend, so the priority regulation logic was started. The electrolyte supply system was adjusted first: the liquid replenishment frequency of this tank area was increased from every 2 minutes to every 1.5 minutes, making the local concentration gradient more balanced; secondly, the system lowered the electrolyte temperature in this area from the original setting of 42°C to 39.5°C to reduce the reaction rate and suppress abnormal active reactions; in addition, the system also adjusted the polarity switching frequency in the electrode plate configuration from the standard setting of 15Hz to temporarily increase to 19Hz to improve the local charge balance and deposition structure symmetry of this area. After the above intervention treatment, the system detected that the tension returned to 29.2mN / m at the 50th minute of electroplating, the potential perturbation returned to the normal range of 9.3Hz, and the reflectance fluctuation amplitude fell to 0.3%, indicating that the interface response had returned to stability, and the system switched the control state to monitoring mode and withdrew the perturbation adjustment. Finally, the thickness consistency of the electroplated layer in this area was controlled within ±1.7μm, and the average micro-stress was 43MPa, which was significantly lower than the average of 61MPa in the high sensitivity area, indicating that the strategy of micro-perturbation identification + priority control successfully avoided the risk of potential defect generation. Further comparison of the results of 10 continuously processed templates found that the structure defect rate caused by abnormal stress in the process was reduced by 39.4% after using the method of the application, and the flatness consistency index after electroplating was improved by 12.6%, proving that controlled perturbation testing of control variables during the process cycle can actively identify the sensitivity of the system, and potential deviation risks can be mastered in advance, so that the maximum stability benefit can be achieved with the minimum adjustment intervention.

[0040] The above shows and describes the basic principles, main features and advantages of the present application. Those skilled in the art should understand that the present application is not limited to the above-mentioned embodiments, and the above-mentioned embodiments and descriptions in the specification are only to illustrate the principles of the present application. Without departing from the spirit and scope of the present application, various changes and improvements can be made to the present application, and these changes and improvements all fall within the scope of the claimed present application. The scope of protection of the present application is defined by the appended claims and their equivalents.

Claims

1. A method for optimizing plating and etching process parameters, characterized in that The method comprises the following steps: By analyzing the influence of metal ions, doping impurities and surface roughness factors on the reaction interface migration rate and boundary layer flow behavior during electroplating and etching, the interface migration behavior relationship is established; the current density gradient, concentration polarization and electrolyte shear rate physical quantities at the interface are coupled in the dynamic prediction framework to perceive and predict the position and dynamic change of the actual reaction front; The interface migration behavior relationship is combined with parameter sensitivity analysis to dynamically evaluate the influence of process parameters including current density, temperature, pH value and liquid flow rate on the interface behavior and to prioritize and optimize the sorting; The sensitivity function after fitting is used to implement dynamic interval scaling of sensitive parameters, and local adjustment range is generated according to the material type and pattern complexity; The morphology evolution information recorded during electroplating is transferred to the etching parameter optimization module by introducing the pre-process memory mapping relationship; the etching rate, angle control and disturbance mechanism are locally adjusted to respond to the pre-electroplating state; the interface feedback signals of surface tension change, potential disturbance and local reflectivity fluctuation are collected; and a closed-loop optimization control loop of micro-disturbance, response evaluation and parameter correction is established.

2. The method of optimization of electroplating and etching process parameters according to claim 1, wherein The method for perceiving and predicting the position and dynamic change of the actual reaction front comprises: A dynamic response relationship framework is established for analyzing the position evolution behavior of the process reaction interface under the coupling effect of multiple physical factors during electroplating or etching; information flow of interface potential distribution, solute concentration deviation and fluid disturbance state is obtained and input into the dynamic response relationship framework; Process adjustment instructions oriented to interface control are generated based on the output results, and parameter adaptive adjustment is performed in the actual processing process to control the uniformity of metal deposition or material removal behavior.

3. The method of claim 2, wherein the method is performed by a computer system. The process adjustment instructions are prioritized according to the influence weight of physical variables in the interface response relationship framework on the process output results; a cross-stage parameter memory mechanism is established to transfer the residual stress distribution and non-uniformity information of the plated layer generated in the electroplating stage to the etching stage, and dynamically adjust the etching rate and angle control strategy.

4. The method of optimization of electroplating and etching process parameters according to claim 3, wherein The interface response relationship framework comprises identification of the curvature change of the reaction interface for real-time monitoring of the side etching trend during etching and deviation correction by local flow field disturbance; a multi-objective optimization mechanism based on a target function is constructed, and the target function comprises multiple objectives of deposition uniformity, etching perpendicularity, reaction interface stability and overall low stress.

5. The method of claim 4, wherein the method is performed in a computer system. The dynamic response relationship framework performs real-time sampling of parameter sensitivity by micro-disturbance trial method and dynamically induces the response trend of different variable inputs.

6. The method of claim 1, wherein the method is performed in a computer system. The method for responding to the pre-electroplating state comprises: In the electroplating stage, the surface to be processed is subjected to staged data collection, and the spatiotemporal change information of the material morphology during the information deposition process is recorded; the spatiotemporal change information is encoded in the form of spatial structure data and stored as a parameter input set called by the processing link; in the etching stage, the pre-process influence area is identified according to the data set, and the local processing strategy is parameterized and adjusted in combination with the current pattern structure.

7. The method of claim 6, wherein the method further comprises: determining a target thickness of the metal layer; and determining a target thickness of the metal layer after the etching process. The morphology change information collected in the electroplating stage includes abnormal behaviors of surface accumulation height difference, thickness gradient distribution, stress concentration area, and edge structure discontinuity; the morphology information is used to construct a distribution map of processing sensitive areas in the etching stage, to guide the setting of etching liquid flow path, flow rate intensity, or local disturbance mode for different areas; the surface state of the pre-process is spatially labeled, and the labeled information is called in real time during etching to dynamically adjust the etching time, angle, or direction.

8. The method of claim 7, wherein the method further comprises: determining a target thickness of the metal layer; and determining a target thickness of the metal layer after the etching process. A response rule is established between the morphology information and the etching control strategy through a mapping model, the response rule is based on different defect types corresponding to different processing parameter sets for matching; a plurality of collected information sequences of the pre-process are superimposed and analyzed, and the trend characteristics evolving over time are extracted to predict potential processing errors and correct the etching behavior in advance.

9. The method of claim 1, wherein the method is used for optimizing electroplating and etching process parameters. The method for establishing a closed-loop optimization control loop comprises: Collecting the change information of the reaction interface state under microscale, including the liquid phase tension behavior of the material surface, the electrochemical reaction activity index, and the instantaneous response characteristics to external light or energy irradiation; analyzing the change information to identify the abnormality of the interface response characteristics; generating a local control instruction based on the analysis result to adjust the target parameters in the process cycle.

10. The method of claim 9, wherein the method is performed in a computer system. A controllable disturbance is applied to the control variables in the process cycle, and the sensitivity level of the current working condition is judged based on the interface physical response trend; when the interface state change information has unstable behavior, the range adjustment of the electrolyte supply rate, temperature distribution, or local charge density is preferentially implemented.

Images