Optimization method for double-sided water washing control of plated parts based on computational fluid dynamics space model

Abstract

The present invention relates to an optimization method for double-sided water washing control of plated parts based on a computational fluid dynamics space model, and belongs to the technical field of double-sided semiconductor electronic electroplating. The present invention continuously constructs and improves the CFD space model by combining the computational fluid dynamics model with experimental data feedback, thereby quickly establishing the optimal cleaning production conditions for various semiconductor devices with different structures. After continuous iteration and upgrading of the CFD space model to improve the accuracy of the CFD space model, it is helpful to improve the various parameters of the double-sided cleaning device for double-sided water washing semiconductor devices, thereby obtaining the optimal double-sided water washing control method for plated parts, which can not only achieve the uniformity of double-sided cleaning of semiconductor devices, but also achieve thorough and uniform cleaning of semiconductor devices, meeting the basic requirements of no white spots, no dirt and no color difference on their surfaces.

Description

Technical Field

[0001] The invention relates to an optimization method for double-sided water washing control of plated parts based on a computational fluid dynamics space model, and belongs to the technical field of double-sided semiconductor electronic electroplating. Background Art

[0002] Electronic electroplating forms a metal coating on the plated part through the principle of cathode electrolytic reduction. It is a special manufacturing technology for nano-precision manufacturing of semiconductor devices. Its application areas include semiconductor wafer electroplating, printed circuit board electroplating, lead frame electroplating, connector electroplating, microwave devices and other electronic component manufacturing. It is the only key technology that can realize nano-level electronic logic interconnection and micro-nano structure manufacturing and processing. Electronic electroplating has become a basic, universal and irreplaceable special manufacturing technology in the production of high-end electronic products such as wafer chip manufacturing, three-dimensional integration and device packaging, micro-nano device manufacturing, micro-electromechanical systems, sensors, components, etc. in the semiconductor integrated industry chain.

[0003] In order to make the semiconductor electronic components have reliable and excellent performance, the metal coating formed on the surface of the plated parts needs to have high uniformity, excellent density and strong corrosion resistance in various harsh environments, and can be used for efficient electronic plating equipment and methods for plated parts of various complex shapes. Among them, the double-sided local electronic plating method is an important basis for electronic plating of semiconductor devices.

[0004] With the rapid development of science and technology, the components of semiconductor integrated circuits required by the new energy industry are constantly being updated, and the technical requirements for high-end manufacturing of electronic electroplating are becoming higher and higher. In particular, with the rapid rise of new energy vehicles in recent years, a variety of new performance electronic components are emerging in an endless stream, and the use of conventional electroplating equipment and electroplating methods can no longer meet the performance requirements of new energy electronic components. For example, more and more semiconductor components are limited to local electroplating treatment for the area that transmits signals, while areas with heat dissipation and other functions do not require any electroplating treatment; in the present invention, in order to ensure the heat dissipation function of semiconductor devices, the area of ​​the heat sink is usually quite different from the local electroplating area of ​​the transmission signal, which is several times or dozens of times higher.

[0005] As shown in Table 1, the existing electroplating process technology is used for double-sided local electroplating. There are white spots, color difference and dirt on the copper surface of the non-plated area of ​​the semiconductor device; and there are fatal defects such as nickel plating copper leakage, gold plating dirt and poor plating adhesion in the local plating area. These defects directly lead to a significant reduction in the performance of semiconductor device products, especially the lack of plating adhesion will make the semiconductor device become waste.

[0006] Table 1

[0007]

[0008] Therefore, when more and more semiconductor integrated circuit products only need to be locally electroplated in the area where signals are transmitted, the use of the electroplating process production line of the prior art for the semiconductor integrated circuit products that only need to be electroplated in the area where signals are transmitted will cause the copper material surface area in the non-plated area to: 1) undergo a substitution reaction with other metals in the electroplating solution, resulting in product defects; 2) undergo an oxidation reaction with oxidizing substances in the ambient atmosphere, resulting in the occurrence of performance defects in semiconductor electronic components. The specific defects are: Fig.10a and Fig.10b The copper surface has white spots, Fig.11a and Fig.11b These deficiencies result in the existing electroplating process production lines being unable to meet the development needs of electronic electroplating for an increasing number of semiconductor integrated circuit products that only require electroplating in areas that transmit signals. Summary of the invention

[0009] The present invention focuses on solving the white spot and color difference problems on the copper surface of the non-plated area of ​​semiconductor devices and the metal replacement and copper oxidation problems on the copper surface of semiconductor devices in the existing electronic electroplating cleaning technology by improving the function of the cleaning process tank and adding filtering measures; it can be applied to the nano-precision manufacturing and processing of semiconductor integrated circuit electronic products, and the formed double-sided local electronic electroplated gold film has excellent uniform distribution, morphology, purity and density, and obtains high corrosion resistance and high-precision conductivity; at the same time, the optimization method of the present invention can ensure the design functions such as heat dissipation of the metal material in the non-plated area, so as to provide the semiconductor integrated circuit industry with a variety of high-end precision semiconductor integrated circuit electronic products with high-quality electronic signal transmission performance.

[0010] The present invention provides an optimization method for double-sided water washing control of plated parts based on a computational fluid dynamics space model, which is applied to the double-sided water washing process of semiconductor devices in an electronic electroplating process, and uses a double-sided cleaning device to wash the double-sided surfaces of the semiconductor devices. By using the optimization method to calculate and adjust various cleaning conditions, a double-sided cleaning device with the best double-sided cleaning effect and the best quality of finished semiconductor devices is constructed;

[0011] The optimization method comprises the following steps:

[0012] S1: constructing a database based on cleaning conditions and constructing an initial computational fluid dynamics spatial model linked to the database;

[0013] S2: extracting two different cleaning conditions from the database, performing fluid experiment operations on the semiconductor device for cleaning, and importing the obtained cleaning experiment data into the initial computational fluid dynamics space model;

[0014] S3: According to the imported cleaning experiment data, the initial computational fluid dynamics space model is improved into an initial CFD space model;

[0015] S4: optimizing the simulation data in the initial CFD space model, feeding back and applying it to the fluid experiment operation of cleaning, and importing the cleaning experiment data obtained subsequently into the initial CFD space model, so as to obtain an upgraded CFD space model;

[0016] S5: loop S4 to obtain an iteratively upgraded CFD space model to improve the accuracy of the CFD space model;

[0017] S6: Select a suitable generation of upgraded CFD space model in S5 as the applied CFD space model, use the applied CFD space model to calculate the optimal cleaning condition group, and adjust and control the double-sided cleaning device according to the optimal cleaning condition group.

[0018] Furthermore, in said S1, the process of constructing the initial computational fluid dynamics space model linked to said database includes using COMSOL Multiphysics finite element simulation software, and adopting a fluid secondary distribution module and a deformation geometry module to simulate the double-sided water washing process of semiconductor devices.

[0019] Furthermore, the double-sided cleaning device used in the double-sided water washing process includes multi-stage water tank units connected in sequence. When the first-stage water tank unit is equipped with a quantitative pure water replenishment pipeline, the water accumulated in the first-stage water tank unit will overflow to the subsequent water tank unit step by step and form a step-by-step overflow structure; each stage of the water tank unit includes an upper pump, an upper diverter plate, an upper water outlet baffle, a lower water outlet baffle, a lower diverter plate, and a lower pump arranged in sequence from top to bottom, and a cavity is arranged between the upper water outlet baffle and the lower water outlet baffle, and the cavity is a double-sided cleaning area; the upper pump and the lower pump are both used to regulate the flow rate of the electroplating solution.

[0020] Preferably, the double-sided cleaning device includes a water tank unit with more than three levels and less than seven levels.

[0021] Furthermore, the S1 also includes the following steps:

[0022] S11: setting a number of standard semiconductor devices, and constructing a corresponding data storage space in the database according to each standard semiconductor device; each standard semiconductor device is associated with at least one cleaning item;

[0023] S12: setting corresponding standard cleaning data for each cleaning item, wherein the standard cleaning data includes a cleaning solution and a cleaning method;

[0024] S13: Importing the size information and contour information of the standard plating area of ​​the corresponding standard semiconductor device into each data storage space; constructing a standard semiconductor device model and a standard plating area model of the corresponding standard semiconductor device according to the size information and contour information;

[0025] S14: Repeat S13, continuously update different size information and contour information to build a variety of standard semiconductor device models and standard plating area models, and iteratively upgrade the data storage space.

[0026] Furthermore, the cleaning conditions include: the double-sided area of ​​the standard semiconductor device, the cleaning time, the angle formed by the outlet plane of the electronic electroplating mold and the flow direction of the electroplating solution, the temperature of the cleaning solution, the specific gravity of the cleaning solution, the pumping frequency, the outlet diameter of the cleaning solution partition, the vertical distance between the horizontal plane of the standard semiconductor device and the solution outlet, and the diameter of the connecting pipe.

[0027] Furthermore, the initial CFD space model, the upgraded CFD space model, and the applied CFD space model all use an infinitesimal fluid element method to construct a set of equations for calculation.

[0028] The infinitesimal fluid element method specifically comprises the following steps:

[0029] In CFD space, along x, y, z The unit vectors of the axes are set to i, j and k , the velocity vector field of the infinitesimal fluid element in CFD space is expressed as: V = ui + vj + wk ;

[0030] in, u,v and w The speed is x , y , z The component of the direction is expressed as:

[0031] u=u ( x , y , z , t )

[0032] v = v ( x , y , z , t )

[0033] w = w ( x , y , z , t )

[0034] Considering a general unsteady flow, u,v and w Both space and time t function; in addition, the scalar density field is set to:

[0035] ρ = ρ ( x , y , z , t );

[0036] Located in t At time 1, the infinitesimal fluid element is located at 1, where at this time, the density of the infinitesimal fluid element is

[0037] ρ 1= ρ 1( x 1, y 1, z 1, t 1);

[0038] Located in t At time 2, the infinitesimal fluid element is located at 2, where at this time, the density of the infinitesimal fluid element is

[0039] ρ 2= ρ 2 ( x 2, y 2, z 2, t 2);

[0040] For functions ρ = ρ ( x , y , z , t ), which can be rearranged into equation A:

[0041]

[0042] (Equation A)

[0043] Then, the following calculations are performed:

[0044]

[0045]

[0046]

[0047] when t 2→ t 1, taking the limit of equation A yields equation B:

[0048]

[0049] (Equation B)

[0050] The expression of the material derivative in the CFD space coordinate system is obtained through equation B, which is equation C:

[0051]

[0052] (Equation C)

[0053] Vector operators in CFD space coordinate system V The definition of is as equation D:

[0054]

[0055] (Equation D)

[0056] Based on equation D, equation C is changed to equation E:

[0057]

[0058] (Equation E)

[0059] Equation E represents the vector symbolic definition of the material derivative, where D / D t It is the material derivative, and its physical meaning is the rate of change of the physical quantity on the moving fluid element with time; ∂ / ∂t It is a local derivative, and its physical meaning is the rate of change of a physical quantity at a fixed position over time; V · V It is the migration derivative. Its physical meaning is the rate of change over time caused by the movement of a fluid element from one position to another in the flow field due to different flow characteristics at different spatial positions.

[0060] Consider an infinitesimal area d S Moving at local speed, due to d S In a time interval Δ t The volume of the control body changes by Δ γ , Δ γ Equal to d S As the base ( V Δ t ) n is the volume of a slender cylinder with n is perpendicular to d S The unit vector of the surface, then Δ γ Expressed as equation F:

[0061] Δγ = {( V Δ t )n} S = ( V Δ t )·d S

[0062] (Equation F)

[0063] The vector d S Defined as d S = n d S , in a time interval Δ t During the process, the total volume change of the entire control volume is equal to the sum of equation F over the entire control surface. S →0, the sum becomes an area integral:

[0064]

[0065] Divide the area integral by Δ t The physical meaning of the result is the rate of change of the control volume over time, which is given by D γ / D t Expressed as equation G:

[0066]

[0067] (Equation G)

[0068] Applying the divergence theorem in vector calculations to the right side of equation G yields equation H:

[0069]

[0070] (Equation H)

[0071] Assume that δγ is small enough that V · V It is basically the same value in the entire δγ region, so when δγ takes the limit 0, the integral of equation H is written as ( V · V )δγ, thus obtaining equation I:

[0072]

[0073] (Equation I)

[0074] Then, equation I is rearranged into equation J:

[0075]

[0076] (Equation J)

[0077] In equation J, the left side is the divergence and the right side is its physical meaning, that is, V ·V The physical meaning of is the relative rate of change of the volume of the control body moving with the fluid element over time;

[0078] According to the above equations A to J, a group of equations is established to serve as the calculation basis for the initial CFD space model, the upgraded CFD space model, and the applied CFD space model.

[0079] The physical meaning of the left side of equation A is the average rate of change of the density of the infinitesimal fluid element moving from position 1 to position 2 over time. In the limit operation, t 2Close to t 1, we can get:

[0080]

[0081] Among them, D ρ / D t Represents the rate of change of density over time when the semiconductor device passes a certain position. ρ / D t Defined as the material derivative; when D ρ / D t When the infinitesimal fluid element moves in space, the observer locks his eyes on the infinitesimal fluid element and moves with the infinitesimal fluid element, and obtains the rate of change of the density of the infinitesimal fluid element passing through position 1 over time (D ρ / D t )1; and for ( ∂ρ / ∂t )1, the observer fixes his eyes at a static position 1, and obtains the density change caused by instantaneous pulsation in the flow field, D ρ / D t and ∂ρ / ∂t They are physically and numerically different quantities.

[0082] The optimization method for double-sided water washing control of plated parts based on the computational fluid dynamics space model can be applied in nano-precision manufacturing of semiconductor integrated circuit electronic products.

[0083] Beneficial effects of the present invention:

[0084] The present invention continuously constructs and improves the CFD space model by combining the computational space fluid dynamics model with experimental data feedback, thereby quickly establishing the optimal cleaning production conditions for semiconductor devices of various structures. The continuous iteration and upgrading of the CFD space model improves the accuracy of the CFD space model, which helps to improve the various parameters of the double-sided cleaning device used for double-sided water washing semiconductor devices, thereby obtaining the optimal double-sided water washing control method for plated parts, which can not only achieve uniformity in double-sided cleaning of semiconductor devices, but also achieve thorough and uniform cleaning of semiconductor devices to meet the basic requirements of no white spots, no dirt and no color difference on their surfaces. BRIEF DESCRIPTION OF THE DRAWINGS

[0085] Figure 1 It is a schematic structural diagram of a double-sided cleaning device in one embodiment of the present invention.

[0086] Figure 2 It is a schematic diagram of the structure of each water tank unit in one embodiment of the present invention.

[0087] Figure 3 This is a schematic diagram of a water tank unit in a CFD space model in one embodiment of the present invention.

[0088] Figure 4 Schematic diagram of the physical derivative of a fluid element moving with the fluid.

[0089] Figure 5 The core technical concept logic diagram for constructing the CFD space model of the present invention.

[0090] Figure 6-Figure 9 This is a scatter plot trend diagram of the flow distribution of the cleaning solution that effectively contacts the surface of the semiconductor device in some embodiments of the present invention.

[0091] Fig.10a and Fig.10b The white spot pattern images appearing in two comparative examples of the present invention are shown in FIG.

[0092] Fig.11a and Fig.11b These are the color difference samples that appear in two comparative examples of the present invention.

[0093] Fig.12a and Figure 12b This is a sample diagram of color difference occurring in Examples 1 and 2 of the present invention.

[0094] Fig.13a , Fig.13b , Fig.13c , Fig.13d This is a sample diagram of color difference that occurs in Examples 3-6 of the present invention.

[0095] In the figure, 13, double-sided cleaning device; 131, primary water tank unit; 132, secondary water tank unit; 133, tertiary water tank unit; 20, semiconductor device; 210, upper pump; 220, upper diverter plate; 230, upper water outlet baffle; 330, lower water outlet baffle; 320, lower diverter plate; 310, lower pump.

[0096] Y Vertical distance from the upper water outlet baffle to the semiconductor device

[0097] Z The vertical distance from the lower water outlet baffle to the semiconductor device

[0098] d Upper water outlet diameter

[0099] ΦLower water outlet diameter

[0100] δThe angle between the upper outlet and the upper outlet baffle plane

[0101] αThe angle between the lower outlet and the lower outlet baffle plane DETAILED DESCRIPTION

[0102] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0103] In the present invention, unless otherwise clearly specified and limited, the terms "connected", "connected", and "fixed" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, it can be the internal connection of two elements or the interaction relationship between two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0104] The present invention provides an optimization method for double-sided water washing control of plated parts based on a computational fluid dynamics space model, which is applied to the double-sided water washing process of semiconductor devices in an electronic electroplating process, and uses a double-sided cleaning device to wash the semiconductor devices on both sides. By using the optimization method, various cleaning conditions are calculated and adjusted to construct a double-sided cleaning device 13 with the best double-sided cleaning effect and the best quality of semiconductor device finished products;

[0105] like Figure 1-Figure 2 As shown, the double-sided cleaning device 13 used in the double-sided water washing process includes multiple water tank units connected in sequence. When the first-stage water tank unit is equipped with a quantitative pure water replenishment pipeline, the water accumulated in the first-stage water tank unit will overflow to the subsequent water tank unit step by step and form a step-by-step overflow structure; each stage of the water tank unit includes an upper pump 210, an upper diverter plate 220, an upper water outlet baffle 230, a lower water outlet baffle 330, a lower diverter plate 320, and a lower pump 310 arranged from top to bottom, and a cavity is arranged between the upper water outlet baffle 230 and the lower water outlet baffle 330, and the cavity is a double-sided cleaning area; the upper pump and the lower pump are both used to regulate the flow rate of the electroplating solution.

[0106] Preferably, the double-sided cleaning device 13 may include a water tank unit having more than three levels and less than seven levels. Figure 2In the double-sided cleaning device, there are three water tank units, namely, a primary water tank unit 131, a secondary water tank unit 132, and a tertiary water tank unit 133.

[0107] The optimization method comprises the following steps:

[0108] S1: Construct a database based on cleaning conditions, wherein the cleaning conditions include: double-sided area of ​​standard semiconductor devices, cleaning time, angle between the outlet plane of the electronic electroplating mold and the flow direction of the electroplating solution, cleaning solution temperature, cleaning solution specific gravity, pump frequency, cleaning solution baffle outlet diameter, vertical distance between the horizontal plane of the standard semiconductor device and the solution outlet, and connecting pipe diameter; use COMSOL Multiphysics finite element simulation software, fluid secondary distribution module and deformation geometry module to simulate the double-sided water washing process of semiconductor devices, thereby constructing an initial computational fluid dynamics space model linked to the database;

[0109] S2: extracting two different cleaning conditions from the database, performing fluid experiment operations on the semiconductor device for cleaning, and importing the obtained cleaning experiment data into the initial computational fluid dynamics space model;

[0110] S3: According to the imported cleaning experiment data, the initial computational fluid dynamics space model is improved into an initial CFD space model;

[0111] S4: optimizing the simulation data in the initial CFD space model, feeding back and applying it to the fluid experiment operation of cleaning, and importing the cleaning experiment data obtained subsequently into the initial CFD space model, so as to obtain an upgraded CFD space model;

[0112] S5: loop S4 to obtain an iteratively upgraded CFD space model to improve the accuracy of the CFD space model;

[0113] S6: Select a suitable generation of upgraded CFD space model in S5 as the applied CFD space model, use the applied CFD space model to calculate the optimal cleaning condition group, and adjust and control the double-sided cleaning device according to the optimal cleaning condition group.

[0114] Furthermore, the S1 also includes the following steps:

[0115] S11: setting a number of standard semiconductor devices, and constructing a corresponding data storage space in the database according to each standard semiconductor device; each standard semiconductor device is associated with at least one cleaning item;

[0116] S12: setting corresponding standard cleaning data for each cleaning item, wherein the standard cleaning data includes a cleaning solution and a cleaning method;

[0117] S13: Importing the size information and contour information of the standard plating area of ​​the corresponding standard semiconductor device into each data storage space; constructing a standard semiconductor device model and a standard plating area model of the corresponding standard semiconductor device according to the size information and contour information;

[0118] S14: Repeat S13, continuously update different size information and contour information to build a variety of standard semiconductor device models and standard plating area models, and iteratively upgrade the data storage space.

[0119] Furthermore, the initial CFD space model, the upgraded CFD space model, and the applied CFD space model all use an infinitesimal fluid element method to construct a set of equations for calculation.

[0120] The infinitesimal fluid element method specifically comprises the following steps:

[0121] In CFD space, along x, y, z The unit vectors of the axes are set to i, j and k , the velocity vector field of the infinitesimal fluid element in CFD space is expressed as: V = ui + vj + wk ;

[0122] in, u,v and w The speed is x , y , z The component of the direction is expressed as:

[0123] u=u ( x , y , z , t )

[0124] v = v ( x , y , z , t )

[0125] w = w ( x , y , z , t )

[0126] Considering a general unsteady flow, u,v and w Both space and time t function; in addition, the scalar density field is set to:

[0127] ρ = ρ( x , y , z , t );

[0128] Located in t At time 1, the infinitesimal fluid element is located at 1, where at this time, the density of the infinitesimal fluid element is

[0129] ρ 1= ρ 1( x 1, y 1, z 1, t 1);

[0130] Located in t At time 2, the infinitesimal fluid element is located at 2, where at this time, the density of the infinitesimal fluid element is

[0131] ρ 2= ρ 2 ( x 2, y 2, z 2, t 2);

[0132] For functions ρ = ρ ( x , y , z , t ), a Taylor series expansion can be performed at position 1:

[0133]

[0134] Divide both sides by t 2— t 1. Ignore high-order small quantities and rearrange it into equation A:

[0135]

[0136] (Equation A)

[0137] The physical meaning of the left side of equation A is the average rate of change of the density of the infinitesimal fluid element moving from position 1 to position 2 over time. In the limit operation, t 2Close to t 1, we can get:

[0138]

[0139] Among them, D ρ / D t Represents the rate of change of density over time when the semiconductor device passes a certain position. ρ / D t Defined as the material derivative; when D ρ / D t When the infinitesimal fluid element moves in space, the observer locks his eyes on the infinitesimal fluid element and moves with the infinitesimal fluid element, and obtains the rate of change of the density of the infinitesimal fluid element passing through position 1 over time (D ρ / D t )1; and for ( ∂ρ / ∂t )1, the observer fixes his eyes at a static position 1, and obtains the density change caused by instantaneous pulsation in the flow field, D ρ / D t and ∂ρ / ∂t They are physically and numerically different quantities.

[0140] Then, the following calculations are performed:

[0141]

[0142]

[0143]

[0144] when t 2→ t 1, taking the limit of equation A yields equation B:

[0145]

[0146] (Equation B)

[0147] The expression of the material derivative in the CFD space coordinate system is obtained through equation B, which is equation C:

[0148]

[0149] (Equation C)

[0150] Vector operators in CFD space coordinate system V The definition of is as equation D:

[0151]

[0152] (Equation D)

[0153] Based on equation D, equation C is changed to equation E:

[0154]

[0155] (Equation E)

[0156] Equation E represents the vector symbolic definition of the material derivative, where D / Dt It is the material derivative, and its physical meaning is the rate of change of the physical quantity on the moving fluid element with time; ∂ / ∂t It is a local derivative, and its physical meaning is the rate of change of a physical quantity at a fixed position over time; V · V It is the migration derivative. Its physical meaning is the rate of change over time caused by the movement of a fluid element from one position to another in the flow field due to different flow characteristics at different spatial positions.

[0157] When the control volume (semiconductor device) moves with the fluid, the control volume is always composed of the same fluid mass points that move with the flow, so its mass is fixed and does not change with time. However, its volume γ and control surface S It changes with time because it has different densities when it moves to different areas. ρ Therefore, the mass of the control body of this motion is constant, its volume may increase or decrease, and its shape may change due to different flow characteristics. Figure 4 As shown. Consider an infinitesimal area d S Moving at local speed, due to d S In a time interval Δ t The volume of the control body changes by Δ γ , Δ γ Equal to d S As the base ( V Δ t ) n is the volume of a slender cylinder with n is perpendicular to d S The unit vector of the surface, then Δ γ Expressed as equation F:

[0158] Δγ = {( V Δ t ) n} S = ( V Δ t )·d S

[0159] (Equation F)

[0160] The vector d S Defined as d S = n d S , in a time interval Δ t During the process, the total volume change of the entire control volume is equal to the sum of equation F over the entire control surface. S→0, the sum becomes an area integral:

[0161]

[0162] Divide the area integral by Δ t The physical meaning of the result is the rate of change of the control volume over time, which is given by D γ / D t Expressed as equation G:

[0163]

[0164] (Equation G)

[0165] Applying the divergence theorem in vector calculations to the right side of equation G yields equation H:

[0166]

[0167] (Equation H)

[0168] Assume that δγ is small enough that V · V It is basically the same value in the entire δγ region, so when δγ takes the limit 0, the integral of equation H is written as ( V · V )δγ, thus obtaining equation I:

[0169]

[0170] (Equation I)

[0171] Then, equation I is rearranged into equation J:

[0172]

[0173] (Equation J)

[0174] In equation J, the left side is the divergence and the right side is its physical meaning, that is, V · V The physical meaning of is the relative rate of change of the volume of the control body moving with the fluid element over time;

[0175] According to the above equations A to J, a group of equations is established to serve as the calculation basis for the initial CFD space model, the upgraded CFD space model, and the applied CFD space model.

[0176] The description of the calculation equation group of the above-mentioned CFD space model is only one of the embodiments. The data derived from the initial CFD space model, upgraded CFD space model, and applied CFD space model in the present invention are a fusion of three-dimensional simulation and experiment based on theory and the computational fluid dynamics space model constructed with the semiconductor device of the present invention as the control body. Therefore, the present invention has the function of optimizing the obtained simulation data and feeding it back to the fluid cleaning experiment; more experimental data are repeatedly imported into the corresponding cleaning condition database to obtain a continuously iterative upgraded CFD space model; in this cyclic process, the more experimental data, the higher the accuracy of the obtained applied CFD space model.

[0177] In the following embodiments, we use the above method to optimize the various parameters and conditions of the double-sided cleaning device, and apply the optimized double-sided cleaning device to the electronic electroplating process, which helps to obtain the best finished product quality.

[0178] like Figure 3 As shown, we define:

[0179] Y Vertical distance from the upper water outlet baffle to the semiconductor device

[0180] Z The vertical distance from the lower water outlet baffle to the semiconductor device

[0181] d Upper water outlet diameter

[0182] ΦLower water outlet diameter

[0183] δThe angle between the upper outlet and the upper outlet baffle plane

[0184] αThe angle between the lower outlet and the lower outlet baffle plane

[0185] First, the angle δ between the upper water outlet and the plane of the upper water outlet baffle, and the angle α between the lower water outlet and the plane of the lower water outlet baffle, can be selected in the range of preferably 30°≤δ or α≤90° based on the simulation quantitative experimental results derived from the applied CFD spatial model constructed by the above method.

[0186] The specific parameters and conditions of Example 1 and Example 2 are shown in Table 2:

[0187] Table 2

[0188]

[0189] After using the double-sided cleaning device under the above conditions to wash the semiconductor device on both sides, it was observed under a 40x microscope and found that the copper surface was Fig.12aThe "color difference" phenomenon shown in the figure occurs; the cleaning effect obtained under the condition of 30° outlet angle in Example 1 is far from meeting the requirements of product appearance price, and the condition of 90° in Example 2, such as Figure 12b As shown, although the color difference is not very obvious, the color difference can still be distinguished by careful observation.

[0190] Therefore, it is necessary to further optimize the water outlet angle of the partition in order to achieve the cleaning of the semiconductor device and make its surface uniform, smooth, free of bubbles, white spots, dirt, and scratches.

[0191] The experimental data of Example 1 and Example 2 are deeply analyzed through numerical calculation of the computational fluid dynamics model to obtain simulated quantitative experimental results that fuse the simulation data with the experimental data.

[0192] By applying the CFD spatial model constructed by the present invention, the flow of the cleaning solution of the double-sided cleaning device is deeply calculated and analyzed under the cleaning conditions set in Example 1 and Example 2, and the simulated quantitative experimental results are obtained, that is, the distribution trend of the flow of the cleaning solution that effectively contacts the surface of the semiconductor device is as follows: Figure 5 shown.

[0193] from Figure 6 It can be seen that the scatter plot of the fluid outlet velocity distribution using the CFD spatial model is a parabola with the outlet pointing upward; the closer the point position is to the central axis (vertical line), the greater the outlet flow rate of the corresponding point, that is, the greater the flow rate at the center of the circle. The closer the point position is to the outlet edge, the smaller the outlet flow rate of the corresponding point, and the closer to the surrounding edges, the closer the corresponding flow rate is to zero.

[0194] Secondly, for 0≤δ and α≤90°, two sets of actual test data, Example 1 and Example 2, are introduced into the CFD space model. Through the deep analysis of the numerical calculation of the CFD space model, the simulation quantitative experimental results of the mutual fusion of simulation data and experimental data are derived, that is, the scatter point trend of the flow distribution of the cleaning solution that effectively contacts the surface of the semiconductor device is as follows: Figure 7 shown.

[0195] from Figure 7It can be seen that the experimental data of fluid flow distribution at δ=30° and α=90° in Example 1 and Example 2 have good matching with the simulation data of deep analysis using CFD spatial model. Since the cleaning result of the semiconductor device in Example 1 is not good, its burr surface has obvious color difference, while no color difference is detected on the bright surface. Therefore, the condition of 30° water washing outlet angle fails to meet the product requirements of thorough cleaning; the result of Example 2 shows that there is no obvious color difference on the burr surface, and only a faint color difference on the surface can be distinguished, while no color difference is detected on the bright surface. Therefore, the condition of 90° water washing outlet angle is close to meeting the product requirements.

[0196] In summary, Figure 7 It can be seen that the simulated fluid flow distribution in the cleaning water outlet angle range of 50° to 80°, compared with Example 1 and Example 2, has a flow rate that effectively contacts the surface of the semiconductor device, which is not only much higher than the cleaning effect of the setting conditions with obvious color difference in Example 1, but also has an advantage over the cleaning effect of the setting conditions with only slight color difference in Example 2.

[0197] Further changing δ and α, while keeping other conditions unchanged, as Example 3-6, the details are shown in Table 3:

[0198] Table 3

[0199]

[0200] Among them, the burr surface of the product obtained in Example 3 has color difference, such as Fig.13a As shown; the burr surface of the product obtained in Example 4 has a slight color difference, such as Fig.13b As shown; the copper surface of the product obtained in Example 5 has no color difference, as Fig.13c As shown; the burr surface of the product obtained in Example 6 has a slight color difference, such as Fig.13d shown.

[0201] According to the results, Examples 3 and 6 each have 1 sample with burr surface color difference, and the qualified rate is 90%, while the qualified rates of Examples 4 and 5 are both 100%. Among them, Examples 4 and 6 each have 1 sample with slight color difference on the burr surface, and 10 samples of Examples 4 and 5 are all color-free and no other appearance defects are detected. Therefore, the most preferred water outlet angle is 70°, that is, the condition of Example 5, and the cleaning conditions of Example 5 can be applied to the double-sided cleaning device.

[0202] When the outlet angle is selected as 70°, we further explore the upper and lower outlet diameters d and Φ, and import the data of the lower outlet diameter Φ7mm of the 6 heat sinks in the lower baffle gradually reduced to 1mm into the CFD space model, while the lower outlet diameter Φ6mm remains unchanged. Through the deep analysis of the numerical calculation of the CFD space model, the simulation quantitative experimental results of the simulation data are derived, that is, the scatter point trend of the flow distribution of the aqueous solution that effectively contacts the surface of the semiconductor device is as follows Figure 7 As shown, under the condition that the flow rates of the upper and lower pumps 210 and 310 are both 70 L / min, when the lower outlet diameter Φ of the 6 heat sink areas of the lower partition of the burr surface of the semiconductor device varies in the range of 0.0~7.0 mm, the flow distribution trend curve presented by the applied CFD spatial model of the double-sided cleaning device has a preferred range of 4.0~5.5 mm.

[0203] Different Φs are set as Examples 7-10 for comparison, and the other conditions are the same as those of Example 5. The specific parameters are shown in Table 4:

[0204] Table 4

[0205]

[0206] It can be seen from Table 4 that the upper water outlet diameter d is the same as that in Example 5 and does not need to be replaced. It is only necessary to replace the lower water outlet diameter Φ of the 6 heat sinks in the lower baffle plate with a value of 4.0-5.5 mm.

[0207] According to the observation results, Examples 7 and 10 each had one sample with burr surface color difference, and the qualified rate was 90%, while Examples 8 and 9 had no color difference and other appearance defects, and the qualified rate was 100%. Therefore, the most preferred conditions for the lower water outlet diameter Φ of the 6 fins area of ​​the lower water outlet baffle are 4.5 and 5.0 mm, that is, Example 8 is the most excellent cleaning condition.

[0208] Furthermore, when the water outlet angle is selected to be 70° and the lower water outlet diameter Φ is set to 4.5 and 5.0 mm, and the vertical distance Z between the lower baffle cleaning water outlet and the semiconductor device is changed in the range of 0.0 to 60 mm, the flow distribution trend curve presented by the CFD spatial model of the double-sided cleaning device is as follows: Figure 8 As shown, the preferred range is 25~45mm.

[0209] Different Zs were set for comparison as Examples 11-15, and the other conditions were the same as those of Example 8. The specific parameters are shown in Table 5.

[0210] Table 5

[0211]

[0212] After observation, Examples 11 and 15 each had one sample with burr surface color difference, and the qualified rate was 90%, while Examples 12 to 14 had no color difference or other appearance defects, and the qualified rate was 100%. Therefore, the most preferred condition for the vertical distance Z from the lower water outlet baffle to the semiconductor device is 30 to 40 mm, that is, Example 13 is the most excellent semiconductor device double-sided cleaning condition.

[0213] Further, when the water outlet angle is selected to be 70°, the lower water outlet diameter Φ is set to 4.5 and 5.0 mm, and the vertical distance Z from the lower water outlet baffle to the semiconductor device is selected to be 30-40 mm, when the frequency of the lower pump 310 is changed in the range of 0.0-60 Hz, the flow distribution trend presented by the applied CFD space model of the double-sided cleaning device is as follows: Fig. 9 As shown, the preferred range is 25-45 Hz, and different Zs are set as Examples 16-20 for comparison. The other conditions are the same as those of Example 13. The specific parameters are shown in Table 6:

[0214] Table 6

[0215]

[0216] After observation, there is one sample with burr surface color difference in each of Examples 16 and 20, and the qualified rate is 90%, while no color difference and other appearance defects were found in Examples 17 to 19, and the qualified rate is 100%. Therefore, the most preferred condition for the frequency of the lower pump 310 is 30 to 40 Hz, that is, Examples 17 to 19 are the best preferred cleaning conditions, and the most suitable cleaning condition is 35 Hz in Example 18.

[0217] In summary, we can conclude from Examples 1-20 that the optimal conditions for the double-sided cleaning device are shown in Table 7, corresponding to Example 18.

[0218] Table 7

[0219]

[0220] Comparative Examples 1 to 5 (Experiments for comparing and verifying the electronic electroplating production process flow of the prior art with the present invention)

[0221] The electronic electroplating production processes of Comparative Examples 1-5 do not include a double-sided cleaning device and optimization of cleaning conditions.

[0222] In addition, the prior art cleaning water tank is only equipped with one pump, and cannot independently control the upper and lower cleaning water supply. The outlet angle α is only designed to be fixed at 90°, which limits the possibility of obtaining the best cleaning conditions through the optimization of the outlet angle α. The specific parameters of the water tank are shown in Table 8.

[0223] Table 8

[0224]

[0225] It can be seen from Table 8 that in the cleaning conditions of Comparative Example 1, the lower cleaning water outlet angle α is vertical 90°, the lower cleaning water outlet diameter Φ is the commonly used 6 mm, the cleaning time is 10 seconds, the same as in the embodiment, and the vertical distance Z between the semiconductor device and the lower cleaning water outlet is 35 mm, which has the best effect. Since there is only one pump, its frequency is 60 Hz.

[0226] The detection method of Comparative Example 1 is the same as that of Example 18. After cleaning 10 semiconductor devices, the results are obtained by the detection method of Example 1: there are 2 samples with color difference, 6 samples with white spots, and only 2 samples with qualified appearance. The details are shown in Table 9.

[0227] Table 9

[0228]

[0229] It can be seen from Table 9 that the cleaning defect rate of semiconductor devices in Comparative Example 1 is as high as 80%, and the qualified rate of finished products is only 20%. Therefore, in order to improve the qualified rate of double-sided cleaning of semiconductor devices, the method of increasing the cleaning time is tried. The specific cleaning time is shown in Table 10.

[0230] Table 10

[0231]

[0232] As shown in Table 10, the cleaning time of Comparative Examples 2 to 5 is increased by 1.5 to 3 times compared with Comparative Example 1. The cleaning method is the same as that of Comparative Example 1. After each cleaning process, 10 semiconductor devices are processed, and the same appearance inspection method is used. The results are shown in Table 11.

[0233] Table 11

[0234]

[0235] It can be seen from Table 11 that the cleaning pass rates of Comparative Examples 1 to 5 are only 20% to 50%. Even if the cleaning time is increased by 3 times, the pass rate can only be increased to 50%. This proves that the cleaning tank structure and configuration of the prior art are imperfect when processing semiconductor double-sided local series products.

[0236] In summary, the cleaning process method of the prior art has defects in the cleaning effect because the cleaning tank device is not the best optimized solution. Therefore, there is a considerable gap between the low cleaning pass rate of 20%~50% in Examples 1~5 and the pass rate of 80%~100% in the preferred range of cleaning based on computational fluid dynamics of the present invention. This gap will lead to a serious decline in the quality of the final electroplated products of electronic electroplating, because color differences and white spots are generated not only on the copper surface of the non-plated area, but also on the surface of the local plated area. Therefore, if the copper surface has produced the white spots and color differences before electronic electroplating, it will lead to the inability to electroplate on the missing copper surface after electronic electroplating, thus resulting in copper leakage and loss. On the other hand, after the base layer of nickel is plated and thoroughly cleaned, it is set in the gold-plating mold, and the gold-plating solution remaining on the surface of the mold will not only contact the copper surface of the non-plated area, but also contact the nickel-plated surface of the plated area. Therefore, this uneven contact will not only lead to the loss of the copper surface of the non-plated area, but also the loss of the nickel-plated surface; the performance of the electronic electroplated products on both sides of the semiconductor device will be lost, resulting in unqualified products.

[0237] In contrast, the present invention constructs and applies a CFD space model to quickly establish the optimal cleaning production conditions for semiconductor devices of various structures.

[0238] Although the present invention has been disclosed as above in the form of a preferred embodiment, it is not intended to limit the present invention. Anyone familiar with this technology can make various changes and modifications without departing from the spirit and scope of the present invention. Therefore, the scope of protection of the present invention should be based on the definition of the claims.

Claims

1. An optimization method for double-sided water washing control of plated parts based on a computational fluid dynamics space model, which is applied to the double-sided water washing process of semiconductor devices in electronic electroplating technology, characterized in that: Using a double-sided cleaning device to wash the semiconductor device on both sides, and using the optimization method to calculate and adjust various cleaning conditions to construct a double-sided cleaning device with the best double-sided cleaning effect and the best quality of semiconductor device finished products; The optimization method comprises the following steps: S1: constructing a database based on cleaning conditions and constructing an initial computational fluid dynamics spatial model linked to the database; S2: extracting two different cleaning conditions from the database, performing fluid experiment operations on the semiconductor device for cleaning, and importing the obtained cleaning experiment data into the initial computational fluid dynamics space model; S3: According to the imported cleaning experiment data, the initial computational fluid dynamics space model is improved into an initial CFD space model; S4: optimizing the simulation data in the initial CFD space model, feeding back and applying it to the fluid experiment operation of cleaning, and importing the cleaning experiment data obtained subsequently into the initial CFD space model, so as to obtain an upgraded CFD space model; S5: loop S4 to obtain an iteratively upgraded CFD space model to improve the accuracy of the CFD space model; S6: Select a suitable generation of upgraded CFD space model in S5 as the applied CFD space model, use the applied CFD space model to calculate the optimal cleaning condition group, and adjust and control the double-sided cleaning device according to the optimal cleaning condition group.

2. According to the optimization method of the double-sided water washing control of plated parts based on the computational fluid dynamics space model in claim 1, it is characterized in that: In S1, the process of constructing the initial computational fluid dynamics space model linked to the database includes using COMSOL Multiphysics finite element simulation software, and adopting the fluid secondary distribution module and the deformation geometry module to simulate the double-sided water washing process of the semiconductor device.

3. The optimization method for double-sided water washing control of plated parts based on computational fluid dynamics space model according to claim 2 is characterized in that: The double-sided cleaning device used in the double-sided water washing process includes multiple water tank units connected in sequence. When the first-stage water tank unit is equipped with a quantitative pure water replenishment pipeline, the water in the first-stage water tank unit will accumulate and overflow to the subsequent water tank unit step by step to form a step-by-step overflow structure; each stage of the water tank unit includes an upper pump, an upper diverter plate, an upper water outlet baffle, a lower water outlet baffle, a lower diverter plate, and a lower pump arranged in sequence from top to bottom, and a cavity is arranged between the upper water outlet baffle and the lower water outlet baffle, and the cavity is a double-sided cleaning area; the upper pump and the lower pump are both used to regulate the flow of the electroplating solution.

4. The optimization method for double-sided water washing control of plated parts based on computational fluid dynamics space model according to claim 3 is characterized in that: The double-sided cleaning device includes a water tank unit with more than three levels and less than seven levels.

5. The optimization method for double-sided water washing control of plated parts based on computational fluid dynamics space model according to claim 3 is characterized in that: The S1 also includes the following steps: S11: setting a number of standard semiconductor devices, and constructing a corresponding data storage space in the database according to each standard semiconductor device; each standard semiconductor device is associated with at least one cleaning item; S12: setting corresponding standard cleaning data for each cleaning item, wherein the standard cleaning data includes a cleaning solution and a cleaning method; S13: Importing the size information and contour information of the standard plating area of ​​the corresponding standard semiconductor device into each data storage space; constructing a standard semiconductor device model and a standard plating area model of the corresponding standard semiconductor device according to the size information and contour information; S14: Repeat S13, continuously update different size information and contour information to build a variety of standard semiconductor device models and standard plating area models, and iteratively upgrade the data storage space.

6. The optimization method for double-sided water washing control of plated parts based on computational fluid dynamics space model according to claim 5 is characterized in that: The cleaning conditions include: the double-sided area of ​​the standard semiconductor device, the cleaning time, the angle formed by the outlet plane of the electronic electroplating mold and the flow direction of the electroplating solution, the temperature of the cleaning solution, the specific gravity of the cleaning solution, the pumping frequency, the outlet diameter of the cleaning solution partition, the vertical distance between the horizontal plane of the standard semiconductor device and the solution outlet, and the diameter of the connecting pipe.

7. The optimization method for double-sided water washing control of plated parts based on computational fluid dynamics space model according to claim 1 is characterized in that: The initial CFD space model, the upgraded CFD space model, and the applied CFD space model all use the infinitesimal fluid element method to construct a set of equations for calculation.

8. The optimization method for double-sided water washing control of plated parts based on computational fluid dynamics space model according to claim 7 is characterized in that: When the infinitesimal fluid element method is used to construct the equation group for calculation, the specific steps include: In CFD space, the unit vectors along the x, y, and z axes are set as i, j, and k, respectively. The velocity vector field of the infinitesimal fluid element in CFD space is expressed as: V = ui + vj + wk; Among them, u, v and w are the components of velocity in the x, y and z directions respectively, expressed as: u=u(x,y,z,t) v = v(x, y, z, t) w=w(x,y,z,t) Considering a general unsteady flow, u, v, and w are functions of space and time t; in addition, the scalar density field is set to: ρ = ρ(x, y, z, t); Assume that at time t1 the infinitesimal fluid element is located at 1, at this point in time, the density of the infinitesimal fluid element is ρ1=ρ1(x1, y1, z1, t1); Assume that at time t2 the infinitesimal fluid element is located at position 2, at this point in time, the density of the infinitesimal fluid element is ρ2 = ρ2(x2, y2, z2, t2); For the function ρ = ρ(x, y, z, t), it can be organized into equation A: Then, the following calculations are performed: When t2→t1, equation A takes the limit and obtains equation B: The expression of the material derivative in the CFD space coordinate system is obtained through equation B, which is equation C: The definition of the vector operator V in the CFD space coordinate system is taken as equation D: Based on equation D, equation C is changed to equation E: Equation E represents the vector symbol definition of material derivative, where D / Dt is the material derivative, and its physical meaning is the rate of change of the physical quantity on the moving fluid element with time; is the local derivative, and its physical meaning is the rate of change of the physical quantity at a fixed position over time; V·V is the migration derivative, and its physical meaning is the rate of change over time caused by the movement of the fluid element from one position to another in the flow field due to different flow characteristics at different spatial positions; Consider an infinitesimal area dS moving at a local velocity. Due to the movement of dS in a time interval Δt, the volume of the control volume changes by Δγ, which is equal to the volume of a slender cylinder with dS as the base and (VΔt)·n as the height, where n is the unit vector perpendicular to the surface of dS. Then Δγ is expressed as equation F: Δγ={(VΔt)·n}dS=(VΔt)·dS Define the vector dS as dS = ndS. During a time interval Δt, the total volume change of the entire control volume is equal to the sum of equation F over the entire control surface. At the limit of the infinitesimal area dS → 0, the sum becomes the area integral: The physical meaning of the area integral divided by Δt is the rate of change of the control volume over time, which is expressed as equation G by Dγ / Dt: Applying the divergence theorem in vector calculations to the right side of equation G yields equation H: Assuming that δγ is small enough so that V·V is essentially the same value throughout the δγ region, then when δγ takes the limit 0, the integral of equation H is written as (V·V)δγ, which gives equation I: Then, equation I is rearranged into equation J: In equation J, the left side is the divergence and the right side is its physical meaning, that is, the physical meaning of V·V is the relative rate of change of the volume of the control body moving with the fluid element over time; According to the above equations A to J, a set of equations is established to serve as the calculation basis for the initial CFD space model, the upgraded CFD space model, and the applied CFD space model.

9. The optimization method for double-sided water washing control of plated parts based on computational fluid dynamics space model as claimed in claim 8, characterized in that: The physical meaning of the left side of equation A is the average rate of change of density over time when an infinitesimal fluid element moves from position 1 to position 2. In the limit operation, when t2 is close to t1, we get: Where Dρ / Dt represents the density change rate over time when the semiconductor device passes a certain position, and Dρ / Dt is defined as the material derivative; when Dρ / Dt is the infinitesimal fluid element moving in space, the observer locks his eyes on the infinitesimal fluid element and moves with the infinitesimal fluid element, and obtains the density change rate over time (Dρ / Dt)1 when the infinitesimal fluid element passes position 1; and for The observer fixes his eyes at a stationary position 1 and obtains the density changes caused by instantaneous pulsation in the flow field, Dρ / Dt and They are physically and numerically different quantities.

10. Application of the optimization method for double-sided water washing control of plated parts based on a computational fluid dynamics space model as described in any one of claims 1 to 9 in nano-precision manufacturing of semiconductor integrated circuit electronic products.

Images