Method for improving uniformity of VCP copper plating on insoluble anode based on 6sigma tool

Abstract

The application discloses a method for improving the uniformity of insoluble anode VCP copper plating based on 6sigma tools, relates to a PCB production process, and aims at the problem of great copper plating difference in the prior art. The uniformity of copper plating of an insoluble anode vertical continuous electroplating line is taken as a research object. Through process analysis and early data regression analysis, key influencing factors are found. According to the characteristics of the line, a DOE factor experiment of the key factors is designed. In model analysis, a stepwise method is ingeniously used, a first-order term of each factor is reserved, and a part of second-order or multiple terms with low contribution degrees is reasonably deleted, so that a more effective model can be obtained, and more accurate prediction variables can be achieved. After the final optimization parameters are input into the equipment, test boards and production boards are followed, and actual copper plating is carried out, the difference can meet the expected target.

Description

TECHNICAL FIELD

[0001] The application relates to a method for improving the uniformity of VCP copper plating of insoluble anodes based on 6sigma tools. BACKGROUND

[0002] The PCB insoluble anode VCP process is a process procedure before fine line production, and the copper thickness uniformity directly affects the yield of the fine line process. For a line width / line spacing of 75 mu m and below, the copper thickness range exceeding 8 mu m will bring the risk of local incomplete etching and line thinning coexisting. For a copper plating thickness of 20 mu m, the copper thickness range is expected to be less than 5 mu m, but in the prior art, the ratio of the range greater than 5 mu m is as high as 46.39%, and at least 10% of the range is greater than 8 mu m.

[0003] The range is shown in the formula: Y = max-min, wherein Y is the copper thickness range, max is the maximum value of the copper thickness, and min is the minimum value of the copper thickness. Figure 1 The copper thickness data is measured by using a CMI 700 copper thickness measuring instrument, and nine test points distributed in a nine-square grid are taken on a production board, one being a center point and the other eight being d=5 mm away from the adjacent board edge. The maximum value minus the minimum value of the copper thickness data is the copper thickness range. SUMMARY

[0004] The application aims to provide a method for improving the uniformity of VCP copper plating of insoluble anodes based on 6sigma tools to solve the problems existing in the prior art.

[0005] The method for improving the uniformity of VCP copper plating of insoluble anodes based on 6sigma tools comprises the following steps:

[0006] S1. A fitting regression model is set based on 6sigma tools for the range Y of the copper plating board thickness;

[0007] S2. The factor Xn affecting the range Y is taken as a continuous prediction variable of the fitting regression model, wherein n is the serial number of the factor X;

[0008] S3. The target value of the range Y is set, and the factor XN with the largest contribution rate is found out;

[0009] S4. Dimensional analysis is performed on the factor XN to obtain a plurality of factors Cm of DOE and the extreme value level of each factor Cm, wherein m is the serial number of the factor C;

[0010] S5. The optimal results of each factor Cm of DOE are analyzed and tested;

[0011] S6. The corresponding parameters of the production line are adjusted by using the optimal results, and then copper plating treatment is performed.

[0012] The 6sigma tool is a Minitab software program.

[0013] The factor Xn affecting the polar difference Y includes: plate thickness, current density, copper sulfate concentration, sulfuric acid concentration, chloride ion concentration, float height parameter, and production plate long side size.

[0014] The factor XN is the float height parameter.

[0015] The factor Cm is the serial number corresponding to the copper cylinder, the lowest level in the polar level is the lowest value of the production plate bottom extending into the float bottom; and the highest level in the polar level is the zero value when the production plate bottom is level with the upper end of the float.

[0016] The corresponding high level and low level are adjusted gradually with the increase of the serial number m of the factor C.

[0017] The corresponding high level adjustment of the factor C with the serial number m is: (highest level-9+m); and the corresponding low level adjustment of the factor C with the serial number m is: (lowest level+m-1).

[0018] The method for improving the uniformity of the insoluble anode VCP copper plating based on the 6sigma tool has the advantages that the factor most affecting the copper thickness polar difference is found out through the 6sigma tool, the copper thickness results brought by the changes of each dimension in the factor are analyzed step by step through the DOE, and finally the parameter results capable of optimally adjusting the copper plating uniformity are obtained. The copper thickness polar difference can be maintained below the target value 5μm, and the defective rate caused by the copper thickness polar difference is basically reduced to zero. BRIEF DESCRIPTION OF DRAWINGS

[0019] Figure 1 It is a schematic diagram of the copper thickness polar difference measurement.

[0020] Figure 2 It is a phase relationship diagram of the copper cylinder factor Cm and the polar difference Y in the embodiment of the application.

[0021] Figure 3 It is a comparison diagram of the double-sided results of the existing technology copper plating and the method in the application. DETAILED DESCRIPTION

[0022] The operation steps and principle analysis of the method in the application are specifically as follows.

[0023] Data collection: through the analysis of the insoluble anode VCP electroplating process, a data table is established, wherein the data includes the factors Xn: X1-plate thickness, X2-current density, X3-copper plating time, X4-copper cylinder temperature, X5-copper sulfate concentration, X6-sulfuric acid concentration, X7-chloride ion concentration, X8-float height parameter, and X9-production plate long side size. The copper thickness polar difference is Y. In this embodiment, the single-sided (A side) copper plating is taken as an example, and the principles of the double-sided (A side and B side) copper plating are the same.

[0024] Data Statistical Analysis: Using the data in the table, a regression model was set up using Minitab software. All changing influencing factors were included in the continuous predictor variables, while fixed-value factors were not considered. The response result was set as the target range Y. Among them, X3 - copper plating time and X4 - copper bath temperature were fixed-value factors. The analysis results are shown in Table 1 (Analysis of Variance). According to the model analysis results, "R-sq" is 86.91%, "R-sq adjusted" is 85.43%, and "R-sq predicted" is 82.96%, all of which are greater than 80%, indicating that the model meets the assumptions of the analysis.

[0025]

[0026] Table 1. Analysis of Variance

[0027] DOE Experiment: By fitting the regression model, the factor with the largest contribution rate (distribution) was identified. In this example, it is the X8-floor height parameter, showing that factor X8 accounts for 50.06% of the range Y. A dimensional analysis of the production line structure using factor X8 was performed. The entire line has eight copper cylinders, and the float height parameter of each cylinder can be controlled independently, with a controllable range of "-50" to "0". Eight factors Cm are defined, where m ≤ 8; two extreme levels are defined, with the lowest level being "-50" and the highest level being "0". The lowest level "-50" indicates that the bottom of the production plate extends 50mm into the bottom of the float, and the highest level "0" indicates that the bottom of the production plate is level with the top of the float. To ensure efficient analysis of each copper cylinder as an independent factor during data processing, the high and low levels of each copper cylinder are fine-tuned. Starting from copper cylinder 2, the low level is increased by "1", and the high level is increased by "1". The DOE factor experiment can study the range of copper plating on one or both sides. In this embodiment, a single-sided experiment is used as an example, and the specific design is shown in Table 2 of the factor experiment.

[0028] C1 C2 C3 C4 C5 C6 C7 C8 Sequence Center point Copper cylinder 1 Copper cylinder 2 Copper cylinder 3 Copper cylinder 4 Copper cylinder 5 Copper cylinder 6 Copper cylinder 7 Copper cylinder 8 Difference Y 1 - -50 -49 -48 -47 -46 -45 -44 -43 4.78 2 - -8 -49 -48 -47 -46 -3 -2 -1 8.81 3 - -50 -7 -48 -47 -4 -45 -2 -1 6.93 4 - -8 -7 -48 -47 -4 -3 -44 -43 12.93 5 - -50 -49 -6 -47 -4 -3 -2 -43 10.21 6 - -8 -49 -6 -47 -4 -45 -44 -1 9.24 7 - -50 -7 -6 -47 -46 -3 -44 -1 5.32 8 - -8 -7 -6 -47 -46 -45 -2 -43 6.60 9 - -50 -49 -48 -5 -4 -3 -44 -1 9.58 10 - -8 -49 -48 -5 -4 -45 -2 -43 10.54 11 - -50 -7 -48 -5 -46 -3 -2 -43 8.05 12 - -8 -7 -48 -5 -46 -45 -44 -1 6.31 13 - -50 -49 -6 -5 -46 -45 -2 -1 7.19 14 - -8 -49 -6 -5 -46 -3 -44 -43 10.39 15 - -50 -7 -6 -5 -4 -45 -44 -43 10.92 16 - -8 -7 -6 -5 -4 -3 -2 -1 19.37 17 Yes -29 -28 -27 -26 -25 -24 -23 -22 7.60 18 Yes -29 -28 -27 -26 -25 -24 -23 -22 6.72 19 Yes -29 -28 -27 -26 -25 -24 -23 -22 6.90

[0029] Table 2 DOE Factor Cm

[0030] Analysis process: Main effects and factor interaction effects. Based on the relationship between the A-surface range and the copper cylinder factor Cm, we can... Figure 2 The results show that copper cylinder 8, i.e., factor C8, exhibits a negative correlation, while other copper cylinders show a strong positive correlation, especially copper cylinder 5, whose strong positive correlation with the range Y is more pronounced than that of other factors. Similarly, this can be used to analyze the relationship between the B-surface range and the copper cylinder factor Cm.

[0031] The analysis process adopts the step-by-step selection method of terms, and the terms in each model include: copper cylinder 1, copper cylinder 2, copper cylinder 3, copper cylinder 4, copper cylinder 5, copper cylinder 6, copper cylinder 7, and copper cylinder 8. The model summary analysis of Y (A-face difference) is as follows: "R-sq" is 99.54%, "R-sq adjusted" is 97.92%, and "R-sq predicted" is 84.01%. The model summary analysis of Y (B-face difference) is as follows: "R-sq" is 99.78%, "R-sq adjusted" is 99.00%, and "R-sq predicted" is 99.42%. The total contribution rate of the model is high, and the predictability is greater than 80%. It is indicated that the DOE experiment model can accurately predict the change of "Y" through the change of "X".

[0032] According to the DOE factor analysis test, the model summary result shows that the factor table can well fit the target value. The target value of copper plating difference is set to be less than 5 μm, the overall desirability D obtained by parameter optimization is 0.9068, and the fitting values are 0.0175 and 0.8741, respectively. The optimal solution of fitting is as follows: copper cylinder 1 is -27.9394, copper cylinder 2 is -49, copper cylinder 3 is -48, copper cylinder 4 is -47, copper cylinder 5 is -46, copper cylinder 6 is -45, copper cylinder 7 is -44, and copper cylinder 8 is -1.

[0033] Improvement verification: according to the optimal parameter solution of the DOE response optimization, the corresponding parameters are respectively entered according to different cylinders, the copper plating thickness is 20 μm, and the actual production result of the actual production plate after the optimization of parameters is compared with the time sequence chart before the optimization. The result is shown in the following table. Figure 3 The copper plating difference data of 10 batches randomly selected from 50 batches of data collected before the improvement are compared with the copper plating difference data of 20 batches randomly selected from 100 batches of data collected after the improvement. It is found that the copper plating difference of the actual production plate after the improvement of parameters can be less than the expected target of 5 μm. Moreover, it is found that the method described in the application can not only be used for single-side copper plating, but also be used for double-side copper plating optimization.

[0034] After the comparison and confirmation of the process capability after the optimization, the overall Z value of the process capability before the optimization is only 0.09, and the overall defective rate is 46.39%. After the optimization, the overall Z value reaches 1.65, and the overall defective rate is 4.93%. The defective rate is reduced by 41.46%, and the improvement effect is remarkable.

[0035] In the application, the 6sigma tool is used, and the copper plating uniformity of the insoluble anode vertical continuous electroplating line is taken as the research object. Through process analysis and early data regression analysis, the key influencing factors are found. According to the characteristics of the line, the key factor DOE factor experiment is designed. In the model analysis, the step-by-step method of terms is skillfully used, the first-order term of each factor is reserved, and part of the second-order or multiple terms with low contribution degree is reasonably deleted, so that a more effective model can be obtained, and the variable can be more accurately predicted. After the final optimization parameter is entered into the equipment, the test plate and the actual copper plating of the production plate can meet the expected target.

[0036] As can be apparent to those skilled in the art in the light of the foregoing description, all the changes and modifications of the technical solutions and concepts described above can be made, and all these changes and modifications shall fall within the scope of protection of the claims of the present application.

Claims

1. A method for improving the uniformity of VCP copper plating on insoluble anodes based on 6 sigma tools, characterized in that, Includes the following steps: S1. For the range Y of the copper plating thickness, a fitting regression model is set based on the 6sigma tool; S2. The factor Xn that affects the range Y is used as a continuous predictor variable of the fitted regression model, where n is the index of factor X; S3. Set the target value for the range Y and find the factor XN with the largest contribution rate; S4. Perform dimensional analysis on factor XN to obtain several factors Cm of DOE and the extreme value level of each factor Cm, where m is the index of factor C; S5. Analyze and test to obtain the optimal results for each factor Cm of DOE; S6. Adjust the corresponding parameters of the production line using the optimal results and then perform copper plating. The factors Xn that affect the range Y include: plate thickness, current density, copper sulfate concentration, sulfuric acid concentration, chloride ion concentration, float height parameter, and the long side dimension of the production plate. The factor XN is the floating platform height parameter; The factor Cm is the serial number corresponding to the copper cylinder. The lowest level in the extreme value level is the lowest value at which the bottom of the production plate extends into the bottom of the floating frame; the highest level in the extreme value level is the zero value when the bottom of the production plate is level with the top of the floating frame. The corresponding high and low levels are gradually adjusted as the index m of factor C increases; Factor C with index m is adjusted to a high level as: (highest level -9 + m); Factor C with index m is adjusted to a low level as: (lowest level + m - 1).

2. The method for improving the uniformity of insoluble anode VCP copper plating based on a 6 Sigma tool according to claim 1, characterized in that, The 6sigma-based tool is the Minitab software program.

Images