A method and system for controlling the concentration of a circuit board etching solution

Abstract

The present application relates to the technical field of integrated circuit manufacturing, in particular to a circuit board etching liquid concentration control method and system, comprising the following steps: real-time monitoring of the circuit board etching liquid, obtaining conductivity readings and temperature readings, comparing with the preset conductivity range and temperature range of the circuit board etching process, checking and screening the data, and establishing an effective etching parameter set. The present application realizes data checking and screening by real-time monitoring of the conductivity and temperature values of the circuit board etching liquid, establishes an effective etching parameter set, improves monitoring accuracy and reduces data abnormality interference, and ensures the reliability of etching liquid concentration estimation; the concentration is real-time calculated and corrected by using the conductivity and etchant concentration reference relationship curve, and the concentration is corrected and probability analyzed in combination with the prior aging characteristics of the etching liquid and the drift probability of the sensor, thereby enhancing the accuracy and stability of the concentration estimation.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of integrated circuit manufacturing, and in particular to a circuit board etching liquid concentration control method and system. BACKGROUND

[0002] The technical field of integrated circuit manufacturing is one of the core technical fields in the electronic industry, which covers the process of manufacturing micron or even nanometer precision structures on semiconductor wafers, substrates and related substrates using microelectronic processing technology, including key steps such as photolithography, thin film deposition, etching, doping diffusion, chemical mechanical polishing (CMP), cleaning and packaging testing, to manufacture integrated circuit chips or electronic components with specific electrical functions.

[0003] In the prior art, the etching liquid concentration is managed mainly by periodic sampling during the integrated circuit manufacturing process, and there is a lack of real-time measurement of parameters such as conductivity and temperature, making it difficult to identify and correct concentration fluctuations caused by sensor drift or etching liquid aging in a timely and effective manner, and thus it is difficult to accurately assess the real-time reactivity of the etching liquid in actual working conditions. Therefore, improvements are needed. SUMMARY

[0004] The purpose of the present application is to solve the shortcomings in the prior art and to provide a circuit board etching liquid concentration control method and system.

[0005] To achieve the above-mentioned purpose, the present application adopts the following technical solution: a circuit board etching liquid concentration control method, comprising the following steps:

[0006] Real-time monitoring of the circuit board etching liquid is performed to obtain conductivity readings and temperature readings, and the data is verified and filtered against the preset conductivity range and temperature range of the circuit board etching process to establish an effective etching parameter set;

[0007] Based on the effective etching parameter set, the conductivity value is called and combined with the conductivity and etchant concentration reference relationship curve to calculate the etchant concentration, generate an etchant concentration estimation result, and based on the etchant concentration estimation result, combine the prior aging characteristics of the circuit board etching liquid and the inherent drift probability of the sensor to establish an etchant concentration probability distribution;

[0008] Based on the etchant concentration probability distribution and the temperature value in the effective etching parameter set, the reactivity of the current solution is evaluated to obtain the current etching liquid activity, and based on the current etching liquid activity, the degree of solution performance decay is determined to obtain a comprehensive etching liquid performance probability point set;

[0009] Based on the comprehensive performance probability point set of the etching liquid, the performance probability point is compared with a minimum acceptable performance threshold and an optimal performance target interval of the circuit board etching, a deviation state of the performance probability point from the threshold or the interval is identified, an exceeding threshold risk point identification result is obtained, a circuit board manufacturing process risk level is matched based on the exceeding threshold risk point identification result, and an etching process deviation early warning code is generated.

[0010] Preferably, the acquisition step of the effective etching parameter set is:

[0011] The conductivity readings and temperature readings of the circuit board etching liquid are collected, the conductivity readings and temperature readings are point-by-point synchronized and paired with time tags, and a synchronized and paired group of conductivity readings and temperature readings is generated;

[0012] Based on the synchronized and paired group of conductivity readings and temperature readings, the preset conductivity range and temperature range of the circuit board etching process are compared, and the interval of each group of conductivity readings and temperature readings in the synchronized and paired group is checked item by item, the synchronized and paired group that falls into the preset conductivity range and temperature range at the same time is screened, and a screened and paired group that meets the preset process condition is obtained;

[0013] Based on the screened and paired group that meets the preset process condition, all conductivity readings and temperature readings in the screened and paired group are extracted and packaged as structured records to obtain an effective etching parameter set.

[0014] Preferably, the acquisition step of the etchant concentration estimation result is:

[0015] Based on the effective etching parameter set, the conductivity value field in each record is extracted, the corresponding numerical value is read and linear interpolation is performed to fill in the missing point, a continuous and complete conductivity value sequence is generated, and each conductivity value is bound to the corresponding time stamp and temperature field to generate a conductivity reading record set containing the conductivity value, time stamp and temperature field;

[0016] According to the conductivity reading record set, a calibrated conductivity-etchant concentration reference relationship curve is called, each conductivity value is taken as an input value to query the table to obtain a concentration output node, if the conductivity value does not directly hit the reference relationship node, a secondary interpolation is performed to generate a corresponding concentration value between adjacent nodes, and the obtained concentration value is combined with the corresponding time stamp and temperature field to write into a concentration fitting result record set to generate an etchant concentration fitting result set;

[0017] Based on the set of etchant concentration fitting results, outlier detection processing is performed on the concentration values in all records, and a 3 standard deviation screening mechanism is used to exclude concentration outliers deviating from the mean value. At the same time, it is verified whether the temperature field within the timestamp corresponding to the fitting concentration value falls within the effective temperature fluctuation range, only the records that meet the conditions are retained, and the concentration values that pass the verification are output in chronological order as the final concentration estimation sequence, and the etchant concentration estimation result is generated.

[0018] Preferably, the etchant concentration probability distribution acquisition step is:

[0019] Based on the etchant concentration estimation result, the aging days and relative humidity value of each record are extracted, the maximum drift amount in the last 30 days is obtained by calling the sensor number matching, and the drift intensity value is recorded. The stirring speed is divided by the reference speed to obtain the stirring ratio value, the aging days are divided by the maximum aging days to obtain the normalized aging factor, and the normalized aging factor, relative humidity value, drift intensity value and stirring ratio value are integrated to generate a set of correction parameters.

[0020] Based on the set of correction parameters, the etchant concentration correction value is calculated.

[0021] Based on the etchant concentration correction value, all etchant concentration correction values are segmented by 0.1 mol / L as a fixed interval starting step, the total number of etchant concentration correction values in each segmented interval is counted in sequence, and the proportion of each segmented item number to the total item number is calculated as the concentration probability value of the interval. Each concentration segmentation interval and the corresponding concentration probability value are paired and output as an ordered concentration-probability combination sequence to generate the etchant concentration probability distribution.

[0022] Preferably, the current etching liquid activity degree acquisition step is:

[0023] Based on the etchant concentration probability distribution, all etchant concentration intervals and corresponding concentration probability values are extracted, each etchant concentration interval is used as an index parameter, and the temperature value corresponding to the same timestamp in the effective etching parameter set is matched to construct a three-tuple set composed of etchant concentration interval, concentration probability value and corresponding temperature value, and a matching record set is generated.

[0024] Based on the matching record set, each set of concentration value and temperature value combination is read, and the standard etching rate value matching or interval fitting is found in the standard copper foil etching rate reference table. The standard etching rate value found is combined with the corresponding concentration probability value in the original three-tuple to form a four-tuple record, and a concentration-temperature-probability-rate matching combination set is obtained.

[0025] Based on the concentration-temperature-probability-rate matching combination set, the probability average level of the rate value in all combinations is calculated, the weighted average rate value is extracted as the etching reaction ability performance value of the current solution under the present concentration and temperature conditions, and the current etching liquid activity degree is generated.

[0026] Preferably, the acquisition step of the etching liquid comprehensive performance probability point set is:

[0027] Based on the current etching liquid activity degree, the standard interval value corresponding to the current etching liquid activity degree is extracted, and the number of processed circuit board batches in the current batch statistical record is read synchronously, the activity degree interval and the batch number are constructed into a one-to-one corresponding group, and the activity degree-batch number comparison set is established under the condition of consistent time labels, and the etching activity batch comparison set is generated;

[0028] Based on the etching activity batch comparison set, the etching liquid performance attenuation ratio of each group is calculated;

[0029] In sequence, it is judged whether each etching liquid performance attenuation ratio exceeds the set attenuation identification threshold value, the corresponding etching liquid activity degree is extracted in the record greater than the attenuation identification threshold value, the appearance probability point of the etching liquid activity degree in the time dimension is constructed, and the etching liquid comprehensive performance probability point set is generated.

[0030] Preferably, the acquisition step of the threshold value risk point identification result is:

[0031] Based on the etching liquid comprehensive performance probability point set, each etching liquid performance probability point is merged with the two end points of the circuit board etching minimum acceptable performance threshold value and the optimal performance target interval, each probability point, the lower limit and the upper limit of the minimum threshold value and the optimal target interval are bound as a four-element relationship, and the performance reference comparison set is generated;

[0032] Based on the performance reference comparison set, the deviation index of each performance probability point is calculated;

[0033] Based on the deviation index, each deviation index is compared with the deviation index in size, the performance probability point number whose deviation index is greater than the deviation judgment threshold value is marked, and the time label and the performance value are output as a record result, and the threshold value risk point identification result is generated.

[0034] Preferably, the acquisition step of the etching process deviation early warning code is:

[0035] Based on the threshold value risk point identification result, the time label corresponding to each identified performance probability point, the concentration value, the temperature value and the deviation index value are extracted, and are sorted according to the occurrence order and the source record, the etching condition combination corresponding to the concentration value and the temperature value is read in sequence, and the etching condition deviation feature set is generated;

[0036] According to the etching operation deviation feature set, the concentration value and the temperature value of each record are combined and mapped to a circuit board manufacturing process risk level table, a matching interval or adjacent level is found and marked with a risk level code, all records are bound with the matching risk level code, and a process risk level matching result is generated;

[0037] Based on the process risk level matching result, corresponding early warning information templates are set according to different risk levels, textual early warning contents with time labels and risk level identifiers are constructed, and are uniformly coded into output identifiers of a standard structure, and etching process deviation early warning codes are generated.

[0038] The application provides an etching solution concentration control system, comprising:

[0039] The parameter acquisition module: real-time monitoring of the circuit board etching solution, obtaining conductivity readings and temperature readings, comparing with the preset conductivity range and temperature range of the circuit board etching process, verifying and filtering data, and establishing an effective etching parameter set;

[0040] The concentration estimation module: based on the effective etching parameter set, calling the conductivity value and combining the conductivity and etchant concentration reference relationship curve, calculating the etchant concentration, generating an etchant concentration estimation result, based on the etchant concentration estimation result, combining the prior aging characteristics of the circuit board etching solution and the inherent drift probability of the sensor, establishing an etchant concentration probability distribution;

[0041] The activity analysis module: based on the etchant concentration probability distribution and the temperature value in the effective etching parameter set, evaluating the reaction activity of the current solution, obtaining the current etching solution activity, based on the current etching solution activity, judging the solution performance decay degree, and obtaining a comprehensive etching solution performance probability point set;

[0042] The deviation early warning module: based on the etching solution comprehensive performance probability point set, comparing the performance probability point with the circuit board etching minimum acceptable performance threshold and the optimal performance target interval, identifying the deviation state of the performance probability point and the threshold or interval, obtaining the threshold risk point identification result, based on the threshold risk point identification result, matching the circuit board manufacturing process risk level, and generating the etching process deviation early warning code.

[0043] Compared with the prior art, the application has the advantages and positive effects that:

[0044] The application realizes data checking and screening, improves monitoring precision and reduces data abnormality interference, ensures reliability of etching liquid concentration estimation by monitoring line board etching liquid conductivity and temperature value in real time, establishing effective etching parameter set; the concentration is calculated and corrected in real time by using conductivity and etchant concentration reference relation curve, and the concentration estimation accuracy and stability are enhanced by combining etching liquid prior aging characteristics and sensor drift probability for correction and probability analysis; the real-time reaction activity of etching liquid is evaluated by etchant concentration probability and temperature value, the real-time state change in etching process is captured, and the etching liquid performance fluctuation is beneficial to foresee and master in advance according to the real-time activity degree; the performance state is quantified and the potential risk point is quickly identified based on the comparison of performance attenuation probability point, the lowest performance threshold and the optimal target interval, and the targeted etching process deviation early warning code is dynamically generated after matching the manufacturing process risk level, active intervention and adjustment suggestions are provided in time, the risk probability of line board etching process is reduced, and the process stability and product consistency in line board manufacturing process are improved. BRIEF DESCRIPTION OF DRAWINGS

[0045] Figure 1 The figure is a schematic diagram of the steps of the application. DETAILED DESCRIPTION

[0046] In order to make the purpose, technical scheme and advantages of the application more clear, the application is further described in detail below in combination with the drawings and examples. It should be understood that the specific examples described herein are only used to explain the application and do not limit the application.

[0047] Please refer to Figure 1 The application provides a technical scheme, a line board etching liquid concentration control method, comprising the following steps:

[0048] The line board etching liquid is monitored in real time to obtain conductivity readings and temperature readings, and the data is checked and screened by comparing with the preset conductivity range and temperature range of the line board etching process to establish an effective etching parameter set;

[0049] Based on the effective etching parameter set, the conductivity value is called and combined with the conductivity and etchant concentration reference relation curve to calculate the etchant concentration, generate the etchant concentration estimation result, and establish the etchant concentration probability distribution based on the etchant concentration estimation result, combining the prior aging characteristics of the line board etching liquid and the inherent drift probability of the sensor;

[0050] Based on the etchant concentration probability distribution and the temperature value in the effective etching parameter set, the reaction activity of the current solution is evaluated, the current etching liquid activity is obtained, and the solution performance attenuation degree is judged based on the current etching liquid activity to obtain the comprehensive performance probability point set of the etching liquid;

[0051] The performance probability points are compared with a minimum acceptable performance threshold and an optimal performance target interval of the circuit board etching, a deviation state of the performance probability points and the threshold or the interval is identified, an exceeding threshold risk point identification result is obtained, a circuit board manufacturing process risk level is matched based on the exceeding threshold risk point identification result, and an etching process deviation early warning code is generated.

[0052] The acquisition step of the effective etching parameter set is:

[0053] The conductivity readings and temperature readings of the circuit board etching solution are collected, and the conductivity readings and temperature readings are point-by-point synchronized and paired by matching the time tags, and a synchronized and paired group of conductivity readings and temperature readings is generated;

[0054] Based on the synchronized and paired group of conductivity readings and temperature readings, the preset conductivity range and temperature range of the circuit board etching process are checked, and each group of conductivity readings and temperature readings in the synchronized and paired group is checked by interval, and the synchronized and paired group that falls into the preset conductivity range and temperature range at the same time is screened, and a screened and paired group that meets the preset process condition is obtained;

[0055] Based on the screened and paired group that meets the preset process condition, all conductivity readings and temperature readings in the screened and paired group are extracted and packaged as structured records to obtain an effective etching parameter set.

[0056] Specifically, the conductivity readings and temperature readings of the circuit board etching solution are collected. Specifically, through the online conductivity sensor and temperature sensor deployed in the etching tank, the conductivity analog signal and temperature analog signal of the etching solution are continuously obtained at a preset sampling frequency, for example, 10 times per second. These analog signals are then converted to digital quantities by the analog-to-digital converter built-in or external to the respective sensors, i.e. conductivity raw readings and temperature raw readings. At the same time, a standard clock built-in or synchronized through a network time protocol is used to assign a high-precision time tag to each raw reading, which records the exact time of data collection, for example, 2023-10-26T10:30:05.123. Then, in order to associate the separated conductivity reading stream and temperature reading stream, all collected conductivity readings are traversed, and the temperature reading with the closest time tag is found. A time synchronization tolerance window is set, which is empirically set according to the response speed of the sensor and the data transmission delay, for example, set to 50 milliseconds. The specific calculation method is as follows: if the time tag of the conductivity reading is , and the time tag of the temperature reading is , when the condition is met, where is the preset time synchronization tolerance window (e.g. 50 milliseconds), it is considered that with synchronized, the conductivity readings and temperature readings that meet this condition are paired, if a conductivity reading of a time tag can match multiple temperature readings within a tolerance window, the one with the smallest time difference is selected for pairing, and vice versa, for single side data points that fail to find corresponding readings within this tolerance window, they are marked as unpaired and temporarily shelved or processed according to a preset strategy (such as subsequent interpolation or discard), through this point-by-point timestamp comparison and pairing mechanism, conductivity readings and temperature readings with similar collection times are combined into one data pair after another, eventually forming a collection containing multiple data units composed of conductivity values, corresponding temperature values and shared time tags, generating a synchronized paired group of conductivity readings and temperature readings.

[0057] Based on the synchronized paired group of conductivity readings and temperature readings, the next step is to perform validity check on these preliminary paired data, the core basis of which is the specific requirements of the circuit board etching process on conductivity and temperature, which are embodied in the preset conductivity range and the preset temperature range, the setting of these two ranges is crucial, for example, the preset conductivity range is defined according to historical production data statistical analysis and combined with the etching process guide manual of a specific circuit board model, for example, for the current board being processed, by analyzing the data of the past 200 successful etching batches, after eliminating obvious outliers, the mean conductivity is calculated as , the standard deviation is , then the preset conductivity range may be set as the mean plus or minus twice the standard deviation, i.e. to ; while the preset temperature range is mainly determined according to the chemical specifications provided by the etching chemical supplier and a large number of process verification experimental results, for example, the optimal working temperature range of a commonly used acidic copper chloride etching solution is verified to be to , exceeding this range may result in too fast or too slow etching rate, or even damage the board, therefore the preset temperature range is set as this value, the checking process is to traverse each group of data in the synchronized paired group of conductivity readings and temperature readings, for each synchronized paired group, the conductivity reading and the temperature reading are taken out at the same time, the conductivity reading value is compared with the preset conductivity range (e.g. to ), and the temperature reading value is compared with the preset temperature range (e.g. to ), only when both the conductivity reading and the temperature reading in a synchronized paired group are strictly within their respective preset valid ranges, for example, a group of readings is (conductivity , temperature ), because and Therefore, the group of data is considered valid, and the synchronization pairing group is screened out. If any value exceeds the corresponding preset range, such as (conductivity , temperature ) or (conductivity , temperature ), the synchronization pairing group will be determined as not meeting the process condition and will be rejected. All synchronization pairing groups that pass the double interval check are collected to obtain the screened pairing group that meets the preset process condition.

[0058] Based on the screened pairing group that meets the preset process condition, further processing of the verified data is performed. First, from each data pair in the screened pairing group that meets the preset process condition, the valid conductivity reading and the valid temperature reading paired with it are extracted, while the original or synchronized shared time label is retained. This extraction process ensures that the data points used for subsequent analysis are collected within the process allowed parameter window. Then, in order to facilitate subsequent calculation, storage and model input, the extracted scattered valid data points (i.e. time label, conductivity reading, temperature reading triplets) need to be uniformly structured and packaged. Specifically, a standardized data record structure can be created for each valid data point, for example, a record format containing the following fields is defined: field one is "collection timestamp", data type is high-precision time format (such as ISO8601 string or Unix timestamp), used to record the precise time of data collection; field two is "valid conductivity value", data type is floating point number, unit is mS / cm, storing the conductivity measurement result that passes the check; field three is "valid temperature value", data type is floating point number, unit is Celsius (℃), storing the temperature measurement result that passes the check. The system will traverse all data pairs in the screened pairing group, and generate corresponding records according to the above defined structured record format, for example, if the screened pairing group has a data pair {time: "2023-10-26T10:31:00.500", conductivity: 21.2, temperature: 50.5}, it will be converted into a structured record containing these three specific values. All generated structured records are organized to form an ordered (usually sorted by timestamp) collection, which is the final required data compilation containing all relevant measurement information under valid process parameter conditions, obtaining the valid etching parameter set.

[0059] The etchant concentration estimation result obtaining step is:

[0060] Based on the effective etching parameter set, the conductivity value field in each record is extracted, the corresponding numerical value is read, and linear interpolation is performed to fill in the missing point positions to generate a continuous and complete conductivity value sequence. Each conductivity value is bound to the corresponding time stamp and temperature field to generate a conductivity reading record set containing conductivity value, time stamp and temperature field;

[0061] According to the conductivity reading record set, the calibrated conductivity and etchant concentration reference relationship curve is called, each conductivity value is taken as the input value to query the table to obtain the concentration output node, if the conductivity value does not directly hit the reference relationship node, secondary interpolation is performed to generate the corresponding concentration value between adjacent nodes, and the obtained concentration value is combined with the corresponding time stamp and temperature field to write into the concentration fitting result record set to generate an etchant concentration fitting result set;

[0062] Based on the etchant concentration fitting result set, outlier detection processing is performed on the concentration values in all records, a 3 times standard deviation screening mechanism is used to exclude concentration outliers deviating from the mean value, and it is verified whether the temperature field corresponding to the fitting concentration value falls within the effective temperature fluctuation range within the time stamp. Only the record entries meeting the conditions are retained, and the concentration values passing the verification are output in time sequence as the final concentration estimation sequence to generate the etchant concentration estimation result.

[0063] Specifically, based on the effective etching parameter set, the system first traverses each structured record in the parameter set, extracts the value of the 'effective conductivity value' field from each record, and the associated 'acquisition time stamp' and 'effective temperature value' fields. Because of the instantaneous failure of the sensor, loss of signal transmission or incomplete synchronization tolerance window in the previous processing during actual acquisition, some conductivity readings are missing at some time points. In order to obtain a continuous conductivity data stream, the system will perform linear interpolation to fill in the missing point positions. The specific operation is as follows: first, sort all effective conductivity values according to the time stamp, then detect whether there is an expected conductivity data point missing between consecutive time stamps. The expectation here is based on the normal sampling frequency, for example, 10 points per second corresponds to one point per 100 milliseconds. If there is one or more data points missing between two effective data points (t1, G1) and (t2, G2) according to the time stamps t1, t2 and the preset sampling interval, interpolation is performed on these missing points. There is a prerequisite for interpolation, that is, the time interval between two known data points cannot exceed a "maximum interpolation interval threshold". The threshold is set according to the stability of the process and the acceptable data quality loss, for example, set to 500 milliseconds. If the interval is less than or equal to this threshold, for the missing conductivity values located in (t1, t2) (where t1 ), the missing conductivity values are interpolated to generate a continuous conductivity value sequence. The conductivity value is bound to the corresponding time stamp and temperature field to generate a conductivity reading record set containing conductivity value, time stamp and temperature field. According to the conductivity reading record set, the calibrated conductivity and etchant concentration reference relationship curve is called, each conductivity value is taken as the input value to query the table to obtain the concentration output node, if the conductivity value does not directly hit the reference relationship node, secondary interpolation is performed to generate the corresponding concentration value between adjacent nodes, and the obtained concentration value is combined with the corresponding time stamp and temperature field to write into the concentration fitting result record set to generate an etchant concentration fitting result set; Based on the etchant concentration fitting result set, outlier detection processing is performed on the concentration values in all records, a 3 times standard deviation screening mechanism is used to exclude concentration outliers deviating from the mean value, and it is verified whether the temperature field corresponding to the fitting concentration value falls within the effective temperature fluctuation range within the time stamp. Only the record entries meeting the conditions are retained, and the concentration values passing the verification are output in time sequence as the final concentration estimation sequence to generate the etchant concentration estimation result.​​​ which is calculated as In this way, all the missing points within the allowed range are filled in one by one, thus forming a continuous and complete conductivity value sequence without any break in the time axis. For the original existing valid conductivity values, their corresponding time stamp and temperature value are directly retained. For the newly generated conductivity values through linear interpolation, their time stamp is the time stamp of the interpolation point itself The corresponding temperature value is also interpolated in a similar way, using and the valid temperature values at the time points If the temperature change is relatively flat, each conductivity value (whether original or interpolated) is recombined with its corresponding time stamp and temperature field to form a new record, generating a set of conductivity reading records containing conductivity values, time stamps, and temperature fields.

[0064] According to the conductivity reading record set, the system then converts the processed conductivity data into etchant concentration information. This conversion process relies on a core reference tool, namely the "calibrated conductivity-etchant concentration reference curve". This reference curve is pre-established through systematic experiments in a controlled laboratory environment. The specific establishment process includes preparing a series of etchant standard samples with known different accurate concentrations (for example, measured by weighing and titration, with a concentration range covering all possible cases that may be encountered in the actual process, such as from to , with an interval of ), then measuring the conductivity values of these standard samples using a high-precision conductivity meter at a constant temperature consistent with the actual process conditions (for example, the central temperature required by the process specification), thus obtaining a set of "conductivity-concentration" paired data points, such as ( , ), ( , ), ( , ) etc. These data points are then fitted by a numerical method (e.g. least square method to fit a polynomial curve, or piecewise linear connection) to form a reference relation curve. During processing, each record in the conductivity reading record set is traversed, and the conductivity value in the record is taken as input to look up the corresponding etchant concentration on the reference relation curve. If the input conductivity value happens to be exactly the same as the conductivity value of a calibration node on the curve, the concentration value of the node is directly taken as output. However, more commonly, the input conductivity value is between two calibration nodes. In this case, if the input conductivity value does not hit a calibration node on the reference relation curve, a quadratic interpolation is performed to generate the corresponding concentration value. This is done by linearly interpolating between the two calibration nodes (denoted as node 1: conductivity , concentration ; node 2: conductivity , concentration ) on the reference relation curve on both sides of the input conductivity value, and the formula is: wherein is the input conductivity value, and is the calculated corresponding concentration value. Each such calculated concentration value is merged with the time stamp and temperature fields in the original record to create a new record entry, and these new entries are sequentially written into a set to generate an etchant concentration fitting result set.

[0065] Based on the etchant concentration fitting result set, further cleaning and verification is performed on the preliminary concentration data. First, an outlier detection process is performed to identify and exclude unreliable concentration estimates caused by accidental interference or transient sensor abnormalities. A 3-sigma screening mechanism is used, which operates as follows. All concentration values in a sliding window (e.g. the last 100 concentration data points collected or calculated) or a fixed batch are selected, and the arithmetic mean ( ) and standard deviation ( ) of these concentration values are calculated. Then each concentration record in the window or batch is traversed, and if the concentration value of a record satisfies the condition , the concentration value of the record is judged to be abnormal and is removed from the data set. For example, if the average concentration is and the standard deviation is , any concentration value less than or greater than All concentration values ​​will be considered outliers. After outlier screening, the system will perform a correlation check, which verifies whether the temperature field of these fitted concentration values ​​at the corresponding timestamp is within a reasonable, preset "effective temperature fluctuation range." This range is usually narrower than the initial process window and is derived from statistical analysis of temperature data from highly stable and excellent etching stages in historical production. For example, if the target temperature during normal stable etching is... Therefore, the effective temperature fluctuation range may be set to arrive The basis for setting this range is to ensure the accuracy of concentration estimation, because drastic temperature fluctuations may indicate process abnormalities or affect the existing relationship between conductivity and concentration. Only those records whose concentration values ​​pass the three-times-standard-deviation test and whose corresponding temperature values ​​fall within this preset effective temperature fluctuation range are ultimately retained. All concentration values ​​that have passed the above double verification are arranged strictly according to their original timestamp order to form a time series. This series is the final concentration estimation data used for subsequent analysis, generating the etchant concentration estimation result.

[0066] The steps for obtaining the etchant concentration probability distribution are as follows:

[0067] Based on the etchant concentration estimation results, the aging days and relative humidity values ​​of each record are extracted. The sensor number is called to match and obtain the maximum drift amount in the last 30 days and record it as the drift intensity value. The stirring speed is divided by the reference speed to obtain the stirring ratio value. The aging days are divided by the maximum aging days to obtain the normalized aging factor. The normalized aging factor, relative humidity value, drift intensity value and stirring ratio are integrated to generate a set of correction parameters.

[0068] Based on the set of correction parameters, the etchant concentration correction value is calculated using the following formula:

[0069] ;

[0070] in, For the first The etchant concentration correction value (in mol / L) for each record. For the first Estimated etchant concentration (in mol / L) for each record. For the first The record normalizes the aging factor (dimensionless, aging days divided by the maximum aging days of the system). For the first Record the relative humidity values ​​(dimensionless, original percentage divided by 100). For the first Record drift intensity values ​​(unit: mol / L). For the first The stirring ratio (dimensionless, stirring rate divided by reference rate) is recorded for each The buffer constant (dimensionless) is set empirically;

[0071] Based on the etchant concentration correction value, all etchant concentration correction values are segmented by numerical value with 0.1 mol / L as a fixed interval starting step, and the total number of entries containing etchant concentration correction values in each segmented interval is counted in turn. The proportion value of each segmented entry number to the total entry number is calculated as the concentration probability value of the interval. Each concentration segmented interval is paired with the corresponding concentration probability value and output as an ordered concentration-probability combination sequence to generate the etchant concentration probability distribution.

[0072] Specifically, based on the etchant concentration estimation result, which is a series of etchant concentration estimation values with time stamps, the system further collects and calculates the auxiliary parameters required for correction for each record. First, for the extraction of aging days, the system queries the start date or last replacement date of the current etchant batch, for example, the record is May 1, 2025 08:00. Then subtract the start date from the time stamp of the current record (for example, May 8, 2025 10:30), to get the actual aging days, which is 7 days and 2.5 hours here. Take the integer or accurate record as 7.1 days. Second, the relative humidity value is obtained by deploying an environmental monitoring sensor near the etching equipment on the production line to collect data in real time. The sensor reports data every minute, for example. The system matches the effective humidity reading at the same time point or the nearest time point according to the time stamp of the current etchant concentration estimation record, for example, the relative humidity is 65%. Third, the maximum drift in the last 30 days is obtained by calling the sensor number to match. It refers to the unique number of the conductivity sensor from which the current concentration estimation value is derived. In the sensor maintenance and calibration history database, find the maximum absolute deviation value that the sensor has appeared when compared with the standard concentration solution in the past 30 working days. This value is recorded as the drift intensity value. For example, for the sensor with serial number SN007, the maximum drift recorded is , which is the drift intensity value, then, the actual running speed of the stirrer in the current etching tank, for example, 95 revolutions per minute, is read through the device interface, and is divided by the preset standard reference speed of the process, for example, 100 revolutions per minute, to obtain the stirring ratio value, which is 0.95 in this example, then, the aging days obtained in the previous step (7.1 days) is divided by the preset maximum aging days of the system, for example, 14 days, which is the upper limit of the period for replacing the etchant determined according to the chemical stability of the etchant and historical experience, to obtain the normalized aging factor, which is 7.1 divided by 14, approximately equal to 0.507 in this example, finally, the normalized aging factor (0.507), the relative humidity value (65 percent, which needs to be converted to 0.65 for use), the drift intensity value (0.05), and the stirring ratio value (0.95) are integrated, and a complete set of correction factors is provided for each etchant concentration estimation result, and a correction parameter set is generated. )and the stirring ratio value (0.95) are integrated, and a complete set of correction factors is provided for each etchant concentration estimation result, and a correction parameter set is generated.

[0073] Formula: The formula has the advantages that the formula comprehensively considers the etchant aging, the influence of external environmental humidity on the sensor, the inherent drift of the sensor itself, and the solution stirring state and other factors that affect the accuracy of concentration measurement in actual production, and corrects the original concentration estimation value through a structured mathematical model, the molecular part converges the combined influence of aging, humidity and drift by square sum and square root, and the denominator part introduces the stirring effect and empirical constant for adjustment, so that the correction is more in line with the actual working condition, thereby improving the accuracy and reliability of the etchant concentration monitoring, and providing more accurate data support for subsequent process control and decision-making;

[0074] is the etchant concentration estimation value of the kth record. The obtaining step of the etchant concentration estimation value of the kth record is as follows: the parameter is directly derived from the output sequence of the previous step "generate etchant concentration estimation result", is the concentration estimation obtained based on the conductivity and temperature readings, through the reference curve conversion and the preliminary outlier processing, represents the etchant concentration without further correction of environmental factors and aging effect, for example, for the kth record currently processed, the corresponding concentration value read from the "etchant concentration estimation result" is Therefore .

[0075] is the etchant concentration estimation value of the kth record. The obtaining step of the record normalization aging factor is: the parameter is calculated by first obtaining the actual "aging days" of the current etching solution batch since the start or the last replacement, and then obtaining a preset "system maximum aging days". The actual "aging days" are calculated from the etching solution replacement time (for example, May 1, 2025, 08:00) recorded in the production management system to the time (for example, May 10, 2025, 12:00) of the first record, which is 9 days and 4 hours, i.e. The "system maximum aging days" is a fixed value set according to the chemical properties of the etching solution, the supplier's recommendation, and the critical usage time length of the etching solution in long-term production practice when the etching efficiency significantly decreases or the quality does not meet the standard, for example, 15 days. This value ensures that the etching solution is replaced before it fails. The calculation formula of the normalization aging factor is: .

[0076] The obtaining step of the relative humidity value of the first record is: the parameter is obtained by reading the measurement value of the environmental relative humidity sensor installed in the operating area of the etching equipment at the corresponding time point of the first record. The relative humidity sensor usually sends readings to the data acquisition system regularly (for example, every minute). The system matches the closest humidity record according to the timestamp, and the obtained original relative humidity is a percentage value, for example, at the time point of the first record, the environmental relative humidity sensor measures the humidity as 58%. To convert it to a dimensionless value for formula calculation, it needs to be divided by 100, .

[0077] The obtaining step of the drift strength value of the first record is: the parameter represents the maximum measurement deviation of the sensor (usually a conductivity sensor) used to measure the concentration of the current etching solution in the near future. Its obtaining method is: first identify the sensor number that generates the first record original data, then query the historical calibration and maintenance record database of the sensor, extract the maximum absolute deviation value recorded by the sensor in the past 30 natural days when compared with the standard etching solution with known concentration, and this deviation value is the drift strength value. For example, the sensor number SCD-003 has the highest calibration deviation record of in the past 30 days, then .

[0078] The obtaining step of the drift strength value of the first The steps for obtaining the stirring ratio (dimensionless) of each record are as follows: This parameter reflects the comparison between the actual and ideal state of solution stirring during the etching process. This is achieved by obtaining the stirring ratio of the first record. The actual operating speed of the stirrer in the etching tank at the corresponding time point is recorded and divided by a preset "reference stirring rate". The "reference stirring rate" is the optimal stirring rate that ensures etching uniformity and efficiency, determined according to process specifications or experimental optimization, for example, set to 120 revolutions per minute (RPM). If the current time is... At the recorded time, the actual speed sensor reading of the stirrer was 110 revolutions per minute. Therefore, the formula for calculating the stirring ratio is: .

[0079] The steps for obtaining the empirically set buffer constant are as follows: This parameter is a constant obtained through experiments and long-term application experience. Its function is to adjust the sensitivity of the entire correction term, preventing excessive correction amplitude or oscillations due to parameter fluctuations, and ensuring the stability of the correction model. Its setting is usually done during the system debugging phase by comparing the degree of agreement between the concentration values ​​before and after correction and the offline test values, and by observing the stability of the correction system during continuous operation. For example, the initial setting is 0.1; if large fluctuations in the corrected concentration values ​​are found, the setting is appropriately increased. Adjust the value (e.g., to 0.15 or 0.2) to enhance the buffering effect; conversely, if the correction response is sluggish, the value can be appropriately reduced. In this case, the value was determined through long-term operational data analysis and process verification. The system is the most stable and accurate.

[0080] Calculation process: For the th The records contain the following parameter settings: ;

[0081] Substitute these values ​​into the formula:

[0082] ;

[0083] Calculate the terms in the numerator: ;

[0084] Square root of the sum of squares of the numerators:

[0085] ;

[0086] Calculate the terms in the denominator: Denominator = ;

[0087] Calculate the correction term: Correction term = ;

[0088] Calculate the final : ;

[0089] The result (Round to three decimal places) is the [number]th [number]. This record shows the etchant concentration value after comprehensive factor correction. This value represents a fine adjustment to the original concentration estimate after considering various influencing factors such as the aging degree of the etchant, current ambient humidity, inherent sensor drift, and stirring efficiency. It is a more refined result than the original... ( The decrease indicates that, under the current parameter settings, the combined effect of these factors led to an overestimation of the original value, and the corrected concentration value is closer to the true active concentration of the etching solution.

[0090] Based on a series of calculated etchant concentration correction values, such as those calculated in the previous paragraph. Given numerous data points, the system will next transform these discrete concentration values ​​into a statistically significant probability distribution. First, a fixed concentration interval step size will be determined, which is set here as [value missing]. The choice of this step size is based on considerations of the etching process accuracy requirements and the fluctuation range of historical data. A moderate step size can ensure resolution while avoiding excessively sparse data. For example, if the concentration correction values ​​are mainly distributed in... arrive Between these points, numerical segmentation will produce results such as , Until This series of continuous concentration ranges is followed by the system iterating through all etchant concentration correction values, assigning each value to its corresponding concentration range, and accumulating the number of correction value entries falling into each range. For example, after statistics, it is found that the number of correction value entries falling into each range is... There are 150 concentration correction values ​​for each interval, while the total number of valid etchant concentration correction value entries in the entire dataset is 1000. Next, the proportion of entries contained in each interval to the total number of entries is calculated; this proportion is defined as the "concentration probability value" for that concentration interval. In the example above, the interval... The concentration probability value is 150 divided by 1000, which is 0.15. The system performs this calculation for all concentration intervals. Finally, each concentration interval (e.g., represented by the midpoint or lower limit) is paired with its corresponding calculated concentration probability value to form an ordered sequence of combinations that reflects the probability of different concentration levels occurring, for example (" ",0.02), (" "0.05), ..., (" "0.15), ..., this sequence is the desired etchant concentration probability distribution that can quantify the current etchant concentration state.

[0091] The current steps for obtaining the activity of the etching solution are as follows:

[0092] Based on the probability distribution of etchant concentration, all etchant concentration intervals and their corresponding concentration probability values ​​are extracted. Each etchant concentration interval is used as an index parameter to match the temperature value corresponding to the same timestamp in the set of valid etching parameters. A set of triplets consisting of etchant concentration intervals, concentration probability values ​​and corresponding temperature values ​​is constructed to generate a matching record set.

[0093] Based on the matching record set, each combination of concentration and temperature values ​​is read one by one, and the standard etching rate value that matches or is fitted in the standard copper foil etching rate reference table is found. The standard etching rate value found is combined with the corresponding concentration probability value in the original triplet to form a quadruple record, thus obtaining the concentration-temperature-probability-rate matching combination set.

[0094] Based on the concentration-temperature-probability-rate matching combination set, the probability average level of rate values ​​in all combinations is calculated, and the weighted average rate value is extracted as the etching reaction capacity performance value of the current solution under the current concentration and temperature conditions, thus generating the current etching solution activity.

[0095] Specifically, based on the etchant concentration probability distribution, which consists of a series of paired etchant concentration ranges and their corresponding concentration probability values, for example, including ( The system first extracts each etchant concentration range and its associated concentration probability value from this probability distribution, including items such as (probability 0.15). Next, for each extracted etchant concentration range, the system needs to match it with a representative process temperature value at that concentration level. This temperature value comes from the "effective etching parameter set" established in the previous steps. This parameter set contains timestamp records synchronized with the original effective conductivity readings and effective temperature readings. The specific matching logic is as follows: For the currently processed etchant concentration range (e.g., [example missing]), The system reviews the sequence of "etcher concentration correction values" that fall within the given concentration range when generating the probability contribution. Each such "etcher concentration correction value" retains its original acquisition timestamp. Using these timestamps, the system searches for and extracts "effective temperature values" that perfectly match these timestamps from the "effective etching parameter set." If multiple "etcher concentration correction values" contribute within a concentration range, multiple corresponding effective temperature values ​​will be obtained. The system then calculates the arithmetic mean of these temperature values ​​and sets this average temperature (e.g., if the corresponding temperatures are...) , , The average temperature is As a representative temperature associated with the current etchant concentration interval, by this operation, the original concentration interval is augmented with the corresponding average temperature value from the concentration probability value, thus constructing a triplet for each concentration interval containing the etchant concentration interval itself, the concentration probability value of the interval, and the calculated corresponding representative average temperature value, for example, 0.15, ), all of which together constitute a new set of matching records.

[0096] Based on the matching record set, where each record is a triplet consisting of an etchant concentration interval, a concentration probability value, and a corresponding representative average temperature value, the system next processes these records one by one, with the goal of querying the etching rate of standard copper foil under specific concentration and temperature conditions. This query operation relies on a pre-established "standard copper foil etching rate reference table," which is established through rigorous laboratory calibration work: select a uniform specification (for example, thickness 1 ounce, purity 99.9%) of copper foil as a standard sample, under the combined conditions of a series of different etchant concentration gradients (for example, from increasing to , every a gradient point) and different operating temperature gradients (for example, from increasing to , every a gradient point), perform timed etching experiments, calculate the average etching rate (for example, in microns / minute) under each specific concentration and temperature combination by measuring the weight loss of the copper foil before and after etching or observing the change in copper layer thickness using a microscope. These experimental data points are arranged, statistically analyzed, and smoothed as necessary to form a two-dimensional lookup table or a computable empirical model as the "standard copper foil etching rate reference table." When processing each triplet in the matching record set, the system extracts the etchant concentration interval (usually using the midpoint value of the interval as the query input, for example, for the interval , use ) and the corresponding representative average temperature value (for example ), the system looks up the corresponding standard etching rate value in the reference table indexed by these two values, if the input values do not exactly match any of the calibration points in the reference table, then the system uses numerical methods such as bilinear interpolation to estimate the standard etching rate under the current conditions based on the neighboring calibration points, for example, if the table lookup and interpolation calculation results in a standard etching rate of 25 micrometers per minute, then the system combines this derived standard etching rate value with the existing etchant concentration interval, representative average temperature value, and concentration probability value from the original triad record to form a quad record containing the etchant concentration interval, representative average temperature value, concentration probability value, and derived standard etching rate value, for example, , , 0.15, 25 micrometers per minute), all of these newly generated quad records are aggregated to form a concentration-temperature-probability-rate matching combination set.

[0097] Based on the concentration-temperature-probability-rate matching combination set, which contains multiple quad records composed of an etchant concentration interval, representative average temperature value, concentration probability value, and a derived standard etching rate value under the given concentration and temperature conditions, the system then proceeds to calculate the overall etching reaction capability of the current etching solution under all possible combinations of concentrations and temperatures. The calculation process involves taking the probability-weighted average of all the standard etching rate values. Operationally, the system iterates through each quad record in the concentration-temperature-probability-rate matching combination set, extracts the standard etching rate value (denoted as ) and the corresponding concentration probability value (denoted as ), multiplies these two values to obtain the weighted contribution of the etching rate under that specific condition to the overall average rate (denoted as ), for example, for a record ( , , 0.15, 25 micrometers per minute), the weighted contribution is 25 micrometers per minute multiplied by 0.15, which equals 3.75 (micrometers per minute)·probability units. The system performs this multiplication operation for all quad records in the set, then sums up all the calculated weighted contribution values. Since the probability values for each concentration interval are derived from a complete probability distribution, theoretically, the sum of all is 1, therefore this cumulative sum itself represents the expected etching rate of the entire etching solution under the current state. This final calculated weighted average rate value, for example, if the cumulative sum of all weighted contribution values is 28.5 micrometers per minute, is extracted and recorded as a single indicator representing the etching reaction capability of the current solution under the current overall concentration distribution and corresponding temperature conditions, generating the current etching solution activity.

[0098] The acquisition step of the comprehensive performance probability point set of the etching solution is as follows:

[0099] Based on the current etching solution activity, the standard interval value corresponding to the current etching solution activity is extracted, and the number of processed circuit board batches in the current batch statistical record is read synchronously. The activity interval and the batch number form a one-to-one corresponding group. Under the condition of consistent time label, the activity-batch number comparison set is established, and the etching activity batch comparison set is generated;

[0100] Based on the etching activity batch comparison set, the etching solution performance decay ratio of each group is calculated, and the calculation formula is as follows:

[0101]

[0102] Among them, is the etching solution performance decay ratio corresponding to the first group of samples, is the initial etching solution activity (unit: mol / s·cm²) of the first group of samples, is the current activity estimate (unit: mol / s·cm²) of the first group of samples under the influence of the number of processed circuit boards;

[0103] Each etching solution performance decay ratio is sequentially judged whether it exceeds the set decay recognition threshold. The etching solution activity corresponding to the record greater than the decay recognition threshold is extracted to construct the appearance probability point of the etching solution activity in the time dimension, and the comprehensive performance probability point set of the etching solution is generated.

[0104] Specifically, based on the current etching solution activity, which is a single value representing the overall reaction ability of the current etching solution, for example, in the previous step, the calculation is , the system first needs to classify this specific activity value into a predefined standard activity interval. These standard intervals, such as the "high activity interval" ( ), "medium activity interval" ( to ), and "low activity interval" ( ), are determined based on long-term production data statistical analysis, combined with etching quality standards and process windows. They reflect the general impact of different activity levels on production results. For example, if the current etching solution activity is ​, then its corresponding standard interval value is "high activity interval", at the same time, the system synchronously reads from the production batch management system the number of circuit board batches that have been processed since the current etching solution batch was started, for example, the system records show that the current etching solution has processed 25 batches of circuit boards, then the system extracts the standard interval value corresponding to the current etching solution activity degree ("high activity interval") and the number of processed circuit board batches (25 batches) to form a corresponding group, in order to track the change of etching solution activity with the increase of processing batches, the system repeats the extraction and matching process every time a new current etching solution activity degree is calculated (usually this is associated with a certain monitoring period or a specific production event, and is time-labeled), and records the formed corresponding group (activity interval, batch number) together with the corresponding time label, and continuously accumulates these records, thereby establishing a data set containing the corresponding relationship between activity interval and the number of batches processed at the time at multiple time points under the condition of consistent time label, and this set is the etching activity batch control set.

[0105] Formula: The advantage of the formula is that the formula provides a method for quantitatively evaluating the degree of etching solution performance decay, which not only considers the absolute difference between the current activity and the initial activity, but also amplifies the case where the activity degree drops more, making the decay effect more sensitive, the term in the denominator normalizes the decay degree based on the initial activity in a certain form, and on the other hand, by adding "1", it ensures that the denominator is always positive when the initial activity exists, avoiding the risk of division by zero, and providing a certain buffer for the case where the value is small, so that the value can stably reflect the relative loss of performance;

[0106] The initial etching solution activity degree of the first group of samples is obtained by: this parameter represents the reference etching ability of a specific batch of etching solution in a new or just replaced state, i.e. when the number of processed circuit board batches is zero, its value is measured and determined by measuring the newly configured etching solution using the same method as calculating the "current etching solution activity degree" (i.e. by monitoring its conductivity, temperature, calculating the concentration distribution, and then calculating the weighted average etching rate based on the standard etching rate table), this initial activity degree is measured and recorded when the etching solution batch is replaced, as the reference for decay comparison throughout the life cycle of the batch, for example, for the newly replaced etching solution in the large tank, the initial etching solution activity degree measured before use (when the number of processed batches is 0) is , then for all sample points belonging to this large tank etching solution , the values are all taken from the .

[0107] the th group of samples, the parameter is the "current etching solution activity" calculated in the previous step, which reflects the actual etching ability of the etching solution at a certain sampling or evaluation time point after processing a certain number of batches of circuit boards. This value is extracted from the "etching activity batch control set", which contains the "current etching solution activity" at a certain time point and the "number of processed batches of circuit boards" at that time. For example, for the th group of samples, which corresponds to an evaluation point after processing 25 batches of circuit boards, the "current etching solution activity" obtained from the "etching activity batch control set" or directly calculated from the previous step is , then . .

[0108] The calculation process is as follows: for the th group of samples, the parameter values are as follows: ;

[0109] Substitute these values into the formula:

[0110] ;

[0111] Calculate the term in the numerator: ;

[0112] Calculate the term in the denominator: denominator = ;

[0113] The result is the etching solution performance decay ratio corresponding to the th group of samples.

[0114] Based on the etching activity batch control set, the system calculates the etching solution performance decay ratio, such as the value calculated in the previous paragraph . The system then evaluates these decay ratios one by one to determine whether they exceed a pre-set "decay recognition threshold". The setting of this "decay recognition threshold" is based on the accumulation of production experience and the specific requirements of product quality. For example, through long-term production data monitoring and analysis, it is found that when the etching solution performance decay ratio reaches , the threshold value is When the calculation results (of comparable magnitude) show a noticeable decrease in the etching uniformity of the circuit board, or a slight upward trend in the scrap rate, the process engineer will consider both economic benefits and quality control when determining this value. The threshold for determining whether the etching solution's effectiveness has significantly decreased is set. During this determination, the system will calculate each... The value is compared with this threshold, if The value is greater than this "attenuation recognition threshold" (e.g., ), then it is considered that The corresponding etching solution state has shown significant performance degradation. For these records identified as having significant performance degradation, the system will extract the "current etching solution activity" at the time of its occurrence (i.e., calculate the value). The time used The values ​​(time stamps, etchant activity at significant decay) and the corresponding timestamps of the record are combined. These combinations (time stamps, etchant activity at significant decay) constitute a series of "points" marked on the time axis. These points together depict the historical trajectory and distribution of the etchant activity when it reaches a significant decay state, thus forming a set of probability points for the comprehensive performance of the etchant.

[0115] The steps for obtaining the results of risk point identification exceeding the threshold are as follows:

[0116] Based on the comprehensive performance probability point set of etching solution, each etching solution performance probability point is merged with the two ends of the minimum acceptable performance threshold and the optimal performance target interval of the circuit board etching. Each probability point, minimum threshold, lower limit and upper limit of the optimal target interval are bound into a quaternary relationship to generate a performance reference comparison set.

[0117] Based on the performance reference comparison set, the deviation index for each performance probability point is calculated using the following formula:

[0118] ;

[0119] in, For the first Deviation index of each efficiency probability point For the first The performance value at each performance probability point. This represents the minimum acceptable performance threshold for circuit board etching. This is the median value of the optimal performance target range;

[0120] Based on the deviation index, each deviation index is compared with the other deviation index, and the efficiency probability point number of the deviation index is marked when it exceeds the deviation judgment threshold. The time label and efficiency value are output as the recording result to generate the risk point identification result that exceeds the threshold.

[0121] Specifically, based on the probability point set of the comprehensive performance of the etching solution, this set consists of a series of paired timestamps and the etching solution performance value marked at that timestamp due to significant performance decay (i.e., "current etching solution activity"), for example, including ("2025-05-10 14:00:00"). The system will then assign a set of reference benchmarks for risk assessment to each performance probability point (i.e., each performance value and its timestamp) in this set. These benchmarks include the upper and lower limits of the "minimum acceptable performance threshold for PCB etching" and the "optimal performance target range". First, the setting of the "minimum acceptable performance threshold for PCB etching" is determined based on strict process baselines and product quality standards. For example, through extensive production practice and failure analysis, it has been confirmed that when the activity of the etching solution is lower than... At this point, the etching rate will be severely insufficient, leading to a sharp increase in the risk of fatal defects such as incomplete circuit patterns and open circuits, and the production cycle will be unacceptably extended. Therefore, The minimum acceptable performance threshold is set as the baseline. Secondly, the setting of the "optimal performance target range" aims to define an ideal etching solution activity range that achieves the best product quality, highest production efficiency, and most economical operating costs. For example, the process department, through statistical regression analysis of historical batch data and combined with small-batch verification experiments, found that when the etching solution activity is maintained at a certain level... to Within this range, the etch uniformity of the circuit board is optimal, the defect rate is lowest, and the reagent consumption rate is within the economic range. Therefore, this range is set as the optimal performance target range, with its lower limit being [missing value]. The upper limit is Subsequently, the system iterates through each performance probability point in the "etching solution comprehensive performance probability point set" (let its performance value be...). ), and the performance value The minimum acceptable performance threshold set above ( ), lower limit of the optimal performance target interval ( ) and the upper limit of the optimal performance target range ( These four values ​​are bound together to form a quaternion relation group, while retaining the relationship with... The associated original timestamp information, all of these forming quaternion groups (with timestamps) are aggregated to generate a performance reference comparison set.

[0122] formula: The advantage of this formula lies in the fact that it provides an effective means of quantitatively assessing the combined deviation of the etching solution's performance at a specific moment from its ideal and minimum acceptable states. This is achieved by calculating the performance values. To the lowest acceptable performance threshold and the median of the optimal performance target interval The average of the sums of absolute distances can simultaneously reflect the degree to which current effectiveness deviates from both the "safety baseline" and the "optimal target." The use of absolute values ​​ensures that the direction of deviation does not affect the measurement of the magnitude of deviation, while the summation and averaging provide a comprehensive deviation indicator. The larger the value, the further the current performance is from the ideal and safe range, and the higher the potential process risk.

[0123] For the first The steps for obtaining the performance value of each performance probability point are as follows: This parameter comes from the "etching solution comprehensive performance probability point set" generated in the previous step. Each point in this set contains a timestamp and an etching solution performance value that is identified as having significant performance degradation at that time point. That is, the first one Such performance values, for example, are selected from the "etching solution comprehensive performance probability point set". Each point records a performance value of [number]. ,but .

[0124] The steps for obtaining the minimum acceptable performance threshold for circuit board etching are as follows: This parameter is a key process parameter defined in the previous step. Its value is set based on the minimum etching quality requirements of a specific type of circuit board on the production line and the absolute bottom line of the process operation. It is combined with the critical etchant activity at which etching defects (such as incomplete etching, excessive side etching, etc.) occur in historical production data, and determined after process verification. For example, as mentioned above, this value has been set as follows: ,therefore, .

[0125] The steps for obtaining the median value of the optimal performance target range are as follows: This parameter represents the center point of the most ideal etchant performance state expected to be achieved in the process, and its calculation depends on the lower limit of the "optimal performance target range" defined in the previous step of this method. ) and upper limit ( For example, in the aforementioned case, the lower limit of the optimal performance target range is set as follows: The upper limit is set to Then the median value of the optimal performance target interval .

[0126] Calculation process: For the selected first... For each performance probability point, the parameter settings are as follows:

[0127] ;

[0128] Substitute these values ​​into the deviation index In the calculation formula:

[0129] ;

[0130] Calculate the absolute value of each term: Absolute value of the first term: The absolute value of the second term: ;

[0131] Calculate the sum of absolute values: sum = ;

[0132] Calculate deviation index : ;

[0133] The result It is the first The deviation index value of each performance probability point comprehensively measures the current performance value. At the same time, it deviates from the minimum acceptable threshold. and the optimal target median value The average degree, in this example, ( (Below the minimum threshold) ( (and is far below the optimal median value) ( The deviation index calculated Value This value will be used to compare with the "deviation judgment threshold" to determine the risk level of the current performance point. The larger the value, the higher the risk.

[0134] The deviation index of each performance probability point is calculated based on the performance reference comparison set, such as the first one calculated in the previous paragraph. Deviation indicators The system will then evaluate each such deviation indicator. The core operation involves comparing each calculated deviation indicator with a preset "deviation judgment threshold." This threshold is a key parameter used in process management to define risk levels. Its setting considers requirements for production stability, acceptable fluctuations in product quality, and the statistical correlation between deviation indicators and actual process problems in historical data. For example, through retrospective analysis of historical data, the process team might find that under similar unit systems, when the deviation indicator... The value exceeds At this time, the probability of quality problems in subsequent production batches increases significantly, or the frequency of needing manual intervention and adjustment increases significantly. Therefore, it is advisable to... Set to this "deviation judgment threshold", the system will read each performance probability point corresponding to the deviation index when judging , and compare it with the threshold value . If the deviation index of a performance probability point is greater than the "deviation judgment threshold" (for example ), the system will mark this performance probability point as having a higher risk. Subsequently, the system will extract the original information corresponding to all performance probability points marked as high risk, i.e. the time label and the actual performance value at that time (i.e. ), and output these time labels and performance values as a set of records, which together constitute a risk event set indicating that the etching liquid performance has exceeded the acceptable deviation range, generating an out-of-threshold risk point identification result.

[0135] The acquisition steps of the etching process deviation warning code are:

[0136] Based on the out-of-threshold risk point identification result, the time label, concentration value, temperature value and deviation index value corresponding to each identified performance probability point are extracted, and the concentration value and temperature value corresponding to each record are read in order of occurrence and source record, and the etching condition combination is generated, and the etching condition deviation feature set is generated;

[0137] According to the etching condition deviation feature set, the concentration value and temperature value combination of each record is mapped to the printed circuit board manufacturing process risk level table, the matching interval or adjacent level is found and the risk level code is marked, all records and their matching risk level codes are bound, and the process risk level matching result is generated;

[0138] Based on the process risk level matching result, set the corresponding warning information template according to different risk levels, construct the text warning content with time label and risk level identification, and uniformly encode it into the output identification of standard structure, and generate the etching process deviation warning code.

[0139] Specifically, based on the out-of-threshold risk point identification result, the result is a series of performance probability points judged to have higher process risk, each point contains at least a time label and a corresponding performance value , and the deviation index value that causes it to be identified, the system first extracts the time label and deviation index value from these risk point records, and for each risk point corresponding to the performance value , i.e. the "current etchant activity" at that time), the system needs to trace back and extract the representative etchant concentration value and temperature value that best characterizes the etchant's overall state at the time when the performance value was calculated, for example, the most probable concentration interval midpoint value from the "etchant concentration probability distribution" can be selected as the representative concentration, and the corresponding representative average temperature of that concentration interval is matched, the specific values are, for example, the time tag is "2025-05-1015:30:00", the representative concentration value is , the representative temperature value is , the performance value is , and the deviation index value is , the system will collect this complete set of data (time tag, representative concentration value, representative temperature value, performance value , deviation index value ) for each identified risk point, and then sort all these collected risk point data records in the order of their time tags, if the records also contain data source information (such as specific production line number or equipment number), they can also be used as secondary sorting basis, after sorting, the system obtains a time-evolving risk event list containing detailed working condition parameters, where the combination of representative concentration value and representative temperature value in each record constitutes the specific etching condition combination at that time, and the set of these combinations forms the etching working condition deviation feature set.

[0140] According to the etching working condition deviation feature set, where each record describes the etching conditions (representative concentration and temperature) at an identified risk moment and the related performance and deviation evaluation results in detail, the system will then compare these actually occurring working condition deviations with the preset risk criteria, the core basis is a "circuit board manufacturing process risk level table", which is prepared in advance by the process management department based on a large amount of historical production data, product defect analysis reports, process experiment results and experienced engineers' experience, its content defines in detail the possible process risks under different combinations of etchant concentration intervals and temperature intervals, for example, the table stipulates that when the etchant concentration is lower than and the temperature is higher than , excessive side etching is easy to occur, the risk level is "high" (code H03); when the concentration is in the range of to , and the temperature is in the range of to , it may cause slow etching rate, the risk level is "medium" (code M02); and when the concentration is in the range of to , and the temperature is in the range of to When the time belongs to the ideal process window, the risk level is "low" (code L01), and the system traverses each record in the etching process deviation feature set, extracts the representative concentration value and the representative temperature value, and then uses the combination as the query condition to search in the "PCB manufacturing process risk level table". The system will first try to match the exact interval, and if it fails to match, it will determine the most suitable risk level according to the preset adjacent logic (for example, select the smallest risk interval containing the point, or if the point is at the intersection of multiple intervals, take the higher risk level), and mark the code corresponding to the risk level (such as H03, M02, L01) to the record being processed. After the risk level code binding of all records is completed, the process risk level matching result is generated.

[0141] Based on the process risk level matching result, the result is that each record in the "etching process deviation feature set" is supplemented with the corresponding risk level code. The system will then generate a specific warning prompt based on this information. First, the system internally presets standardized "warning information templates" for different risk level codes. These templates are designed based on the possible consequences of each risk level, recommended response measures, and key information to be delivered. For example, for the risk level code "H03" (high risk), the template might be: "Severe warning! Time: {time label}, etching liquid state high risk (H03), current concentration: {concentration value} mol / L, temperature: {temperature value} °C, performance value: {performance value}, deviation index: {deviation index value}. This state is extremely prone to excessive side etching. Please check the etching parameters immediately and suspend production if necessary and adjust the etching liquid!" For medium risk "M02", the template might be: "Attention: Time: {time label}, etching liquid state medium risk (M02), current concentration: {concentration value} mol / L, temperature: {temperature value} °C. Etching rate may be slow. Please pay attention to subsequent batch quality and prepare for inspection." The system will traverse each record in the process risk level matching result, select the corresponding warning information template based on its risk level code, and then fill in the specific data such as the time label, representative concentration value, representative temperature value, performance value ( ), deviation index value ( ) in the template to the corresponding placeholders, thereby constructing a complete, time-stamped, and risk level-identified text warning content. Finally, to facilitate automated processing and interaction between different systems, these generated text warning contents will be further packaged or converted into a standard structured output identifier with uniform coding, such as a JSON object or XML record containing specific fields (such as alert level, timestamp, message body, parameter details, etc.). This standard structured output is the final etching process deviation warning code.

[0142] The above merely describes the preferred embodiments of the present application, and is not intended to limit the present application in other forms. Any skilled person in the art can modify or change the disclosed technical content into equivalent embodiments with equivalent changes, and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments according to the technical essence of the present application, without departing from the technical solution content of the present application, still falls within the protection scope of the present application.

Claims

1. A method of controlling the concentration of a circuit board etching solution, characterized by, The method comprises the following steps: Real-time monitoring of the circuit board etching solution, obtaining conductivity readings and temperature readings, comparing with the preset conductivity range and temperature range of the circuit board etching process, checking and screening the data, and establishing an effective etching parameter set; Based on the effective etching parameter set, the conductivity value is called and combined with the conductivity and etchant concentration reference relationship curve to calculate the etchant concentration, generate the etchant concentration estimation result, and based on the etchant concentration estimation result, combine the prior aging characteristics of the circuit board etching solution and the inherent drift probability of the sensor to establish the etchant concentration probability distribution; Based on the etchant concentration probability distribution and the temperature value in the effective etching parameter set, the reactivity of the current solution is evaluated, the current etching solution activity is obtained, and based on the current etching solution activity, the solution performance decay degree is judged, and the etching solution comprehensive performance probability point set is obtained; Based on the etching solution comprehensive performance probability point set, the performance probability point is compared with the circuit board etching minimum acceptable performance threshold and the optimal performance target interval, the deviation state of the performance probability point and the threshold or interval is identified, the threshold risk point identification result is obtained, the circuit board manufacturing process risk level is matched based on the threshold risk point identification result, and the etching process deviation early warning code is generated.

2. The method of claim 1, wherein The effective etching parameter set is obtained by: Collecting the conductivity readings and temperature readings of the circuit board etching solution, matching the time tags to synchronously pair the conductivity readings and temperature readings point by point, and generating the conductivity readings and temperature readings synchronous pairing group; Based on the conductivity readings and temperature readings synchronous pairing group, the preset conductivity range and temperature range of the circuit board etching process are compared, and the interval of each conductivity reading and temperature reading in the synchronous pairing group is checked item by item, the synchronous pairing group falling into the preset conductivity range and temperature range is screened, and the screened pairing group meeting the preset process conditions is obtained; Based on the screened pairing group meeting the preset process conditions, all conductivity readings and temperature readings in the screened pairing group are extracted and packaged as structured records to obtain the effective etching parameter set.

3. The method of claim 1, wherein the concentration of the etching solution is controlled by the following equation: ###0001### wherein, C is the concentration of the etching solution, t is the elapsed time, and k is a constant. The etchant concentration estimation result is obtained by: Based on the effective etching parameter set, the conductivity value field in each record is extracted, the corresponding numerical value is read and the missing point is filled by linear interpolation, a continuous and complete conductivity value sequence is generated, and each conductivity value is bound with the corresponding time stamp and temperature field to generate a conductivity reading record set containing conductivity value, time stamp and temperature field; According to the conductivity reading record set, the calibrated conductivity and etchant concentration reference relationship curve is called, each conductivity value is taken as an input value to obtain the concentration output node by table lookup, if the conductivity value does not directly hit the reference relationship node, a secondary interpolation is performed to generate the corresponding concentration value between adjacent nodes, and the obtained concentration value is combined with the corresponding time stamp and temperature field and written into the concentration fitting result record set to generate the etchant concentration fitting result set; Based on the etchant concentration fitting result set, outlier detection processing is performed on the concentration values in all records, a 3 standard deviation screening mechanism is used to exclude concentration outliers deviating from the mean value, and it is checked whether the temperature field in the fitting concentration value corresponding timestamp falls within the effective temperature fluctuation range, only the records meeting the conditions are retained, the concentration values passing the check are output in time sequence as the final concentration estimation sequence, and the etchant concentration estimation result is generated.

4. The method of claim 1, wherein The etchant concentration probability distribution acquisition step is: Based on the etchant concentration estimation result, the aging days and relative humidity value of each record are extracted, the sensor number is matched to obtain the maximum drift amount in the last 30 days and record it as the drift intensity value, the stirring speed is divided by the reference speed to obtain the stirring ratio value, the aging days are divided by the maximum aging days to obtain the normalized aging factor, and the normalized aging factor, relative humidity value, drift intensity value and stirring ratio value are integrated to generate a correction parameter set; Based on the correction parameter set, the etchant concentration correction value is calculated; Based on the etchant concentration correction value, all etchant concentration correction values are segmented by 0.1 mol / L as a fixed interval starting step, the total number of etchant concentration correction values in each segmented interval is counted, and the proportion of each segmented item number to the total item number is calculated as the concentration probability value of the interval. Each concentration segmentation interval and the corresponding concentration probability value are paired and output as an ordered concentration-probability combination sequence to generate the etchant concentration probability distribution.

5. The method of claim 1, wherein the concentration of the etching solution is controlled by the following equation: ###0001### wherein, C is the concentration of the etching solution, t is the elapsed time, and k is a constant. The acquisition step of the current etching liquid activity degree is: Based on the etchant concentration probability distribution, all etchant concentration intervals and corresponding concentration probability values are extracted, each etchant concentration interval is used as an index parameter, and the temperature value corresponding to the same timestamp in the effective etching parameter set is matched to construct a three-tuple set composed of etchant concentration interval, concentration probability value and corresponding temperature value, and a matching record set is generated; Based on the matching record set, each group of concentration value and temperature value combination is read, and the standard etching rate value matching or interval fitting in the standard copper foil etching rate reference table is found, the found standard etching rate value and the corresponding concentration probability value in the original three-tuple are combined into a four-tuple record, and a concentration-temperature-probability-rate matching combination set is obtained; Based on the concentration-temperature-probability-rate matching combination set, the probability average level of the rate value in all combinations is calculated, the weighted average rate value is extracted as the etching reaction ability performance value of the current solution under the current concentration and temperature conditions, and the current etching liquid activity degree is generated.

6. The method of claim 1, wherein The acquisition step of the etching liquid comprehensive performance probability point set is: Based on the current etching liquid activity degree, the standard interval value corresponding to the current etching liquid activity degree is extracted, and the number of processed circuit board batches in the current batch is read synchronously, the activity degree interval and the batch number are constructed into a one-to-one corresponding group, and the activity degree-batch number comparison set is established under the condition of consistent time label to generate the etching activity batch comparison set; Based on the etching activity batch comparison set, the etching liquid performance decay ratio value of each group is calculated; The method comprises the following steps: sequentially judging whether each etching liquid performance attenuation ratio exceeds a set attenuation identification threshold value, extracting corresponding etching liquid activity in records greater than the attenuation identification threshold value, constructing etching liquid activity occurrence probability points in time dimension, and generating an etching liquid comprehensive performance probability point set.

7. The method of claim 1, wherein The step of obtaining the threshold value exceeding risk point identification result is: Based on the etching liquid comprehensive performance probability point set, each etching liquid performance probability point is merged with the two end points of the minimum acceptable performance threshold value and the optimal performance target interval of the circuit board etching, each probability point, the lower limit and the upper limit of the minimum threshold value and the optimal target interval are bound as a four-element relationship, and a performance reference comparison set is generated; Based on the performance reference comparison set, the deviation index of each performance probability point is calculated; Based on the deviation index, each deviation index is compared with the deviation index in size, the number of performance probability points greater than the deviation judgment threshold value is marked, and the time label and the performance value are output as a record result, and a threshold value exceeding risk point identification result is generated.

8. The method of claim 1, wherein The step of obtaining the etching process deviation early warning code is: Based on the threshold value exceeding risk point identification result, the time label, the concentration value, the temperature value and the deviation index value corresponding to each identified performance probability point are extracted, and are sorted according to the occurrence order and the source record, the concentration value and the temperature value corresponding to each record are read, and an etching process deviation feature set is generated; According to the etching process deviation feature set, the concentration value and the temperature value combination of each record is mapped to the circuit board manufacturing process risk level table, the matching interval or the adjacent level is found and marked, the risk level code is bound, and the process risk level matching result is generated; Based on the process risk level matching result, the corresponding early warning information template is set according to different risk levels, the text early warning content with time label and risk level identification is constructed, and the standard structure output identification is uniformly coded, and the etching process deviation early warning code is generated.

9. The etching solution concentration control system of the method for controlling the concentration of a circuit board etching solution according to any one of claims 1 to 8, characterized by, It comprises: A parameter acquisition module: real-time monitoring of the circuit board etching liquid to obtain conductivity readings and temperature readings, comparing with the preset conductivity range and temperature range of the circuit board etching process, verifying and screening the data, and establishing an effective etching parameter set; A concentration estimation module: based on the effective etching parameter set, the conductivity value is called and combined with the conductivity and etchant concentration benchmark relationship curve to calculate the etchant concentration, generate the etchant concentration estimation result, based on the etchant concentration estimation result, combine the prior aging characteristics of the circuit board etching liquid and the inherent drift probability of the sensor to establish the etchant concentration probability distribution; An activity analysis module: based on the etchant concentration probability distribution and the temperature value in the effective etching parameter set, the reaction activity of the current solution is evaluated, the current etching liquid activity is obtained, and the etching liquid comprehensive performance probability point set is obtained based on the current etching liquid activity. The deviation early warning module: based on the etching liquid comprehensive performance probability point set, compare the performance probability point with the circuit board etching minimum acceptable performance threshold and the optimal performance target interval, identify the deviation state of the performance probability point and the threshold or interval, get the threshold exceeding risk point identification result, based on the threshold exceeding risk point identification result, match the circuit board manufacturing process risk level, generate etching process deviation early warning code.

Images