A magnetic field design method for a magnetron sputtering system

By defining the magnet layout parameters and boundary curve parameters of the magnetron sputtering system, and combining the constraints and objective functions of the optimization design, multi-parameter optimization iterative calculations are performed, which solves the problems of long development cycle and high cost of traditional magnetron sputtering systems, and achieves efficient and accurate magnetic field optimization.

CN123116782APending Publication Date: 2026-11-17UNIV OF SCI & TECH BEIJING
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611175526.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-08-04
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

The magnetic field design of traditional magnetron sputtering systems relies on empirical formulas and static simulations, resulting in long development cycles and high costs, making it difficult to achieve efficient and accurate magnetic field optimization.

Method used

A magnetic field design method for magnetron sputtering systems is adopted. By defining the first design parameters of the magnet and the parameterized model of the boundary curve, and combining the constraints and objective functions of the optimization design, multi-parameter optimization iterative calculations are performed to achieve automated optimization of the magnet layout.

Benefits of technology

It significantly shortens the magnetic field design cycle, reduces development costs, ensures the physical rationality and process feasibility of the magnet layout, and realizes efficient and precise magnetic field design for magnetron sputtering systems.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123116782A_ABST
    Figure CN123116782A_ABST
Patent Text Reader

Abstract

This invention relates to the fields of thin film technology and surface engineering technology, and particularly to a magnetic field design method for a magnetron sputtering system. The method includes: defining first design parameters for candidate magnets on a magnetron substrate; establishing a parametric model of the boundary curve, and determining second design parameters for the boundary curve based on the parametric model; defining constraints and an objective function for optimization design; performing multi-parameter optimization iterative calculations on the objective function using the constraints as the feasible region until a preset stage termination condition is reached; and evaluating the performance of the optimization results obtained in each optimization stage: if the evaluation passes, proceeding to the next optimization stage or outputting the current optimization result; if the evaluation fails, returning to continue the optimization iteration. This invention models the magnet layout and boundary curve as unified optimization variables, iteratively searching for the optimal configuration within the constrained feasible region, eliminating the need for repeated trial production, significantly shortening the cycle time, reducing costs, and ensuring the rationality of the layout and the feasibility of the process.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the fields of thin film technology and surface engineering technology, and in particular to a magnetic field design method for a magnetron sputtering system. Background Technology

[0002] As a core physical vapor deposition technique, magnetron sputtering's deposition efficiency, film quality, and process stability largely depend on the distribution characteristics of the magnetic field on the sputtering target surface. The magnetic field not only confines the plasma and enhances ionization, but also directly determines the target's etching morphology, sputtering yield, and plasma uniformity, thus fundamentally impacting the film deposition rate, uniformity, and composition. Therefore, magnetic field design is a crucial aspect of magnetron sputtering system development and performance optimization.

[0003] Traditional magnetic field design methods mainly rely on empirical formulas and static magnetic field simulations (such as finite element analysis). The desired magnetic field shape is roughly formed by adjusting the arrangement and size of permanent magnets or electromagnetic coils. This involves an iterative "design-prototype-test-correction" process, resulting in long development cycles and high costs. With increasingly demanding requirements for thin film performance, a more scientific, efficient, and precise magnetic field design method is urgently needed to enable the proactive design and optimization of the core magnetic field of magnetron sputtering systems, breaking through current technological bottlenecks. Summary of the Invention

[0004] To address the technical problems of existing technologies, such as reliance on trial and error, long development cycles, and high costs, this invention provides a magnetic field design method for a magnetron sputtering system. The technical solution is as follows: A magnetic field design method for a magnetron sputtering system is provided. The method includes: defining first design parameters for candidate magnets on a magnetron substrate, the design parameters including discrete variables and position coordinates characterizing the presence and polarity orientation of each candidate magnet; establishing a parameterized model of a boundary curve, and determining second design parameters for the boundary curve based on the parameterized model, the boundary curve dividing the substrate surface into an inner region and an outer region, with magnets located in different regions having different polarity orientations; defining constraints and an objective function for optimization design; performing multi-parameter optimization iterative calculations on the objective function using the constraints as the feasible region until a preset stage termination condition is reached, the optimization iterative calculation including at least one optimization stage, in each optimization stage using at least a portion of the first and second design parameters as optimization variables; and evaluating the performance of the optimization results obtained in each optimization stage: if the evaluation passes, proceed to the next optimization stage or output the current optimization result; if the evaluation fails, return to the current optimization stage to continue optimization iteration.

[0005] The beneficial effects of the technical solution provided by the embodiments of the present invention include at least the following: by uniformly incorporating magnet layout parameters (position, polarity, presence or absence) and boundary curve parameters into the optimization design variables, defining the feasible region with constraints, and aiming at minimizing coating non-uniformity, a complete automatic magnetic field optimization design framework is constructed; this method adopts multi-parameter optimization iterative calculation, which can actively search for the optimal magnet configuration scheme without repeated trial production and testing, significantly shortening the magnetic field design cycle and reducing development costs; at the same time, by introducing boundary curves to uniformly determine the magnet polarity partition, the physical rationality and process feasibility of the magnet layout are guaranteed, overcoming the defect of lack of quantitative correlation between magnetic field morphology and coating performance in traditional empirical design, and realizing efficient and accurate design of the magnetic field of the magnetron sputtering system. Attached Figure Description

[0006] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0007] Figure 1 This is a general flowchart of a magnetic field design method for a magnetron sputtering system provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the first magnet layout and boundary curve provided in the embodiment of the present invention; Figure 3 This is a flowchart of a single optimization stage of a magnetic field design method for a magnetron sputtering system provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of the second magnet layout and boundary curve provided in an embodiment of the present invention; Figure 5 This is a flowchart of two optimization stages of a magnetic field design method for a magnetron sputtering system provided in an embodiment of the present invention; Figure 6 This is a flowchart of a single optimization stage combined with genetic algorithm and full-process simulation of a magnetic field design method for a magnetron sputtering system provided in an embodiment of the present invention. Detailed Implementation

[0008] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0009] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0010] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0011] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0012] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0013] Please see Figure 1 This invention provides a magnetic field design method for a magnetron sputtering system. The processing flow of this method may include the following steps S101 to S105.

[0014] Step S101: Define the first design parameters for the candidate magnets on the magnetron base plate. These parameters include discrete variables and position coordinates characterizing the presence and polarity orientation of each candidate magnet. In this step, each candidate magnet position on the magnetron base plate corresponds to an independent design variable. The discrete variables control whether a magnet is placed at that position and its polarity orientation, while the position coordinates determine the specific location of the magnet on the base plate. By treating the presence, polarity, and position of the magnets as adjustable design parameters, sufficient degrees of freedom are provided for subsequent optimization algorithms. This allows the control of the magnetic field shape to no longer rely on empirical arrangement but can be systematically searched and optimized within a mathematical framework.

[0015] In one embodiment, such as Figure 2 As shown, in the first design parameters, discrete variables The value is determined as follows: Where 0 indicates that there is no magnet at that location, 1 indicates that there is a magnet and the N pole faces the target, and 2 indicates that there is a magnet and the S pole faces the target; the position coordinates are the center coordinates of the candidate magnet on the base plate. ; Indicates the first One candidate magnet, , This is the preset maximum number of magnets.

[0016] In this embodiment, the position coordinates of each candidate magnet For continuous variables, they can take values ​​freely within the range of the base plate; for discrete variables... This determines the state of the magnet at that location. This design approach is suitable for optimization algorithms searching for the optimal placement of magnets in continuous space, and is particularly well-suited for use in conjunction with global optimization algorithms such as genetic algorithms. Since the magnet's position can move freely in continuous space, this scheme can more precisely match the target magnetic field distribution, theoretically achieving higher coating uniformity. However, correspondingly, its search space is larger, requiring a higher convergence speed from the optimization algorithm.

[0017] In another embodiment, such as Figure 4 As shown, the base plate of the magnetron is divided into There are 1 grid, and the center coordinates of each grid are 1. Each grid corresponds to a candidate magnet; in the first design parameters, discrete variables The value is determined as follows: Where 0 indicates that there is no magnet at that location, 1 indicates that there is a magnet and the N pole faces the target, and 2 indicates that there is a magnet and the S pole faces the target; the position coordinates are the center coordinates of the candidate magnet on the base plate. ; Indicates the first One candidate magnet, .

[0018] In this embodiment, the magnetron base plate is pre-divided into several regular Cartesian grids, with the center of each grid fixed as the location of a candidate magnet. Unlike the previous embodiment, the position coordinates of the candidate magnets are pre-fixed at the grid center and are not included as optimization variables in the first stage of optimization. This approach significantly reduces the dimensionality of the design variables—only the discrete variables of each grid need to be considered. The optimization significantly reduces the search space and accelerates the convergence speed, making it particularly suitable for initial design phases or scenarios with limited computational resources. This implementation typically consists of two stages: the first stage determines the presence and polarity of magnets on a discrete grid to obtain an initial layout scheme; the second stage fine-tunes the positions of the already placed magnets within the grid to further improve uniformity. This two-stage strategy balances global search efficiency with local fine-tuning capabilities.

[0019] For any magnet arrangement, it can be denoted as: ,in, for or , This is the first design parameter. The first design parameter determines the distribution of candidate positions of the magnets on the base plate. When the candidate positions are continuous coordinates, the magnets can be arbitrarily arranged within the base plate area. In this case, the optimization algorithm needs to search for the position, polarity, and presence / absence of the magnets simultaneously. The search space is large, but the flexibility is the highest. Usually, a single-stage optimization can complete the search of all variables. When the candidate position is the center of a discrete grid, the magnet positions are constrained to the grid points. In this case, a two-stage optimization strategy can be adopted: the first stage determines the presence / absence and polarity of the magnets in the discrete space, and the second stage performs local fine-tuning on the positions of the already arranged magnets. Different first design parameter settings correspond to different numbers of optimization stages and search strategies, but all are unified under the overall optimization framework of this invention.

[0020] Step S102: Establish a parametric model of the boundary curve, and determine the second design parameters of the boundary curve based on the parametric model (which can be denoted as...). The boundary curve divides the base plate surface into inner and outer regions, with magnets located in different regions exhibiting different polarity orientations. The boundary curve is a closed curve that divides the base plate surface into inner and outer regions; its core function is to determine the polarity zoning rules for the magnets—magnets located inside the curve have a uniform polarity orientation, while magnets located outside the curve have a uniform polarity orientation. This is achieved by establishing a parametric model of the boundary curve and incorporating the curve shape parameters (i.e., the second design parameters). By incorporating these into the optimization variables, this invention allows the polarity partitioning of magnets to be automatically determined by an optimization algorithm based on the objective function, rather than relying on manual delineation. The introduction of the boundary curve transforms the polarity selection problem for a large number of individual magnets into an optimization problem for a small number of curve shape parameters, significantly reducing the complexity of the design variables.

[0021] Optionally, in one embodiment of the present invention, the parameterized model of the boundary curve L is represented by a Fourier series in polar coordinates: , , In this system, the origin of the polar coordinate system is the center of rotation of the magnetron. The average radius of the dividing curve, Let represent the highest order of the Fourier series and be a positive integer. The order number is 1 to 1. The integer, the second design parameter includes When using Fourier series to represent the boundary curve, the origin of the polar coordinate system is set at the center of rotation of the magnetron, and the shape of the curve is determined by the average radius. and harmonic coefficients of each order , Determined jointly. Order Control the level of detail in the curve. The larger the value, the more complex the shape the curve can describe. Generally... Using values ​​from 2 to 4 is sufficient for most engineering needs. The Fourier series representation has the advantages of a small number of parameters, smooth curves, and natural closure, making it suitable for situations where the boundary curves are relatively regular in shape and change gradually, resulting in high optimization efficiency.

[0022] Optionally, in one embodiment of the present invention, the parameterized model of the boundary curve is represented by a parameterized closed B-spline curve: , , in, The two-dimensional coordinates of the control points The total number of control points, with the coordinates of the first and last control points being the same to ensure the curve is closed; For node vectors, Let the order be the spline order. The B-spline basis functions are used; the second design parameter includes the coordinates of each independent control point. The B-spline curve representation is achieved through... The shape of the boundary curve is determined by the coordinates of several control points. The curve is attracted by the control points but does not necessarily pass through them. The coordinates of the first and last control points are the same to ensure that the curve is closed. (Spline order) The smoothness of the control curve is usually set to a certain value. (Cuboid B-spline), number of control points Typically, 6 to 12 are chosen. B-spline representation has local support characteristics—moving a control point only affects the shape of a local area of ​​the curve, rather than deforming the whole. This makes the local adjustment of the curve shape more flexible and precise during the optimization process, and is suitable for situations where the boundary curve shape is relatively complex or irregular.

[0023] Optionally, in one embodiment of the present invention, the parameterized model of the boundary curve L can also be determined experimentally. Specifically, by actually measuring experimental data such as the target etching morphology, sputtering yield distribution, or coating uniformity under different magnet layouts, and combining the magnetic field measurement results, the shape of the boundary curve that can reasonably divide the magnets into zones can be derived. For example, several magnets can be arranged on the base plate and the magnetic field distribution on the target surface can be measured. The ideal polarity boundary position can be determined based on the measured magnetic field contour lines or plasma confinement region. Then, these positions can be fitted into a closed curve, and the parameters of the curve can be extracted as the second design parameters. The advantage of the experimental determination method is that it does not rely on a pre-set mathematical model and can truly reflect the optimal boundary form under actual physical conditions. It is especially suitable for scenarios where theoretical models are difficult to accurately describe complex physical effects (such as plasma behavior, edge effects, etc.).

[0024] Step S103: Define the constraints and objective function for the optimization design. The constraints define the feasible region of the optimization search, ensuring that the optimized magnet layout scheme is physically manufacturable and technologically feasible. The objective function provides a quantitative indicator for evaluating the merits of different schemes. This invention defines the constraints and objective function separately, making the mathematical expression of the optimization problem clear and standardized—the optimizer searches for the solution that optimizes the objective function within the range defined by the constraints. The constraints cover the geometric rationality of the magnet layout (no overlap, no boundary violations) and the physical feasibility of magnetized discharge (the horizontal component of the magnetic field is within an effective range), laying a complete mathematical foundation for subsequent optimization calculations.

[0025] Optionally, the constraints include the following five.

[0026] (1) The centers of all activated magnets are located within the base plate. For a circular base plate, this constraint is expressed as follows: .in The coordinates of the center of the magnet are: Let be the radius of the circular base plate.

[0027] (2) The center-to-center distance between any two active magnets shall not be less than the preset minimum safety distance. This constraint mainly limits the active magnets from overlapping, and can be expressed by the formula: , It is the sum of the diameter of the cylindrical magnet and the safety clearance (that is, the preset minimum safety distance). .

[0028] (3) The boundary curve lies entirely within the base plate area. Taking a circular base plate as an example, this constraint condition is expressed as follows: ,in Let be any point on the boundary curve L.

[0029] (4) The shortest distance from the center of each activated magnet to the boundary curve is not less than the preset minimum distance threshold, that is, the activated magnet cannot cross the boundary curve L. This can be expressed by the formula: . That is, the preset minimum distance threshold, the value of which can be taken as the magnet radius.

[0030] (5) Horizontal component of magnetic induction intensity on the sputtered surface of the target material It lies between the preset minimum and maximum horizontal component values. This can be expressed by the formula: ,in To maintain the minimum level of magnetic induction intensity required for magnetron discharge (i.e., the preset minimum level component value). To prevent excessive sputtering or uneven horizontal magnetic induction intensity, an upper limit value (i.e., the preset maximum horizontal component value) is set. and The value can be determined experimentally.

[0031] The five constraints mentioned above cover all the hard requirements that the magnet layout scheme must meet. The first four constraints are geometric constraints, ensuring that the design is physically manufacturable—magnets do not exceed the base plate boundary, sufficient gaps are left between them to avoid collisions, the boundary curve itself does not cross the boundary, and no single magnet crosses the polarity boundary line, causing polarity conflict. These geometric constraints can be quickly checked without magnetic field simulation calculations, with extremely low computational cost, and are therefore prioritized in the evaluation process to quickly eliminate obviously unqualified schemes. The fifth constraint is a physical constraint, requiring that the horizontal component of the magnetic induction intensity at each location on the sputtered surface of the target material be within the specified range. to Within the effective range - below Unable to maintain stable magnetron discharge, higher than This can lead to localized over-sputtering or uneven etching of the target material. This constraint can only be verified after calculating the magnetic field distribution. By setting constraint checks in layers, computational resources are saved to the maximum extent while ensuring the feasibility of the solution.

[0032] Optionally, the above objective function can be defined in two ways.

[0033] The first approach: The objective function is to minimize the coating non-uniformity. , in, The non-uniformity of the coating is determined by the distribution of the magnetic flux density B on the sputtered surface of the target material. The magnetic flux density B is determined by the first and second design parameters. This objective function directly uses the final coating performance index—coating non-uniformity—as its objective. This approach serves as the optimization objective. Its advantages lie in its clear optimization direction and physical meaning. The optimizer directly searches for the final performance metric, eliminating the need for additional work on weight coefficient tuning. It is suitable for optimization problems with well-defined constraints and relatively regular feasible regions.

[0034] The second type: The objective function is a composite function that maximizes a penalty term. , in, This represents the coating non-uniformity determined by the distribution of magnetic induction intensity B on the sputtered surface of the target material, where magnetic induction intensity B is determined by the first design parameter and the second design parameter. For each constraint violation penalty item, Let be the weighting coefficients for each penalty term, and each weighting coefficient be a sufficiently large positive number so that any non-zero penalty term can suppress the fluctuations in coating non-uniformity. This objective function addresses the coating non-uniformity... In addition to the above, various penalties for violating constraints have been added. A linear weighted sum. The weight coefficients of each penalty term. Set to a sufficiently large positive number to ensure that any non-zero penalty value will outweigh the penalty. The objective function exhibits normal fluctuations—meaning that if any constraint of a solution is violated, its objective function value becomes extremely poor, leading to its natural elimination during the optimization process. This approach softens the constraints and integrates them into the objective function, allowing the optimization algorithm to search freely throughout the entire variable space (including the constraint violation region) without requiring separate constraint handling outside the algorithm. Its advantage lies in its good compatibility with optimization algorithms, making it particularly suitable for optimization scenarios with narrow feasible regions, complex constraint boundaries, or those where constraints are difficult to handle directly, such as those using genetic algorithms.

[0035] Optionally, penalty items Includes at least one of the following.

[0036] (1) Boundary violation penalty , representing the sum of the squares of the distances the activated magnet extends beyond the base plate, can be expressed by the formula: .

[0037] (2) Overlapping penalty items , which represents the sum of squares of violations where the distance between the centers of any two active magnets is less than the preset minimum safety distance, can be expressed by the formula: .

[0038] (3) Penalty for violation of boundary curve This can be expressed as the sum of the squares of the distances the boundary curve extends beyond the base plate, which can be represented by the formula: , In the formula, This represents the total number of points on the boundary curve L.

[0039] (4) Penalty for magnet straddling curve This represents the sum of squares of violations where the shortest distance from the center of the activated magnet to the boundary curve is less than a preset minimum distance threshold. It can be expressed by the formula: .

[0040] (5) Magnetic field violation penalty item , which represents the sum of squares of violations where the horizontal component of the magnetic induction intensity at each monitoring point on the sputtered surface of the target is lower than the preset minimum horizontal component value or higher than the preset maximum horizontal component value, can be expressed by the formula: , In the formula, in the formula, This represents the total number of magnetic field monitoring points set on the sputtered surface of the target material. For the first Location coordinates of each monitoring point Indicates at the monitoring point Calculated value of the horizontal component of magnetic induction intensity.

[0041] The five penalty terms mentioned above correspond to the five hard constraints defined in step S103. When the corresponding constraint is violated, the penalty term takes a positive value, and the greater the degree of violation, the higher the penalty value; when the constraint is satisfied, the penalty term takes a value of zero. By quantifying violations such as geometric boundary violations, magnet overlap, boundary curve violations, magnet straddling boundary curves, and magnetic field horizontal component exceeding limits into continuous penalty functions, the optimization algorithm can perceive the distance of the current solution from the feasible region during the search process, thus being guided to move towards the feasible region. The mathematical expressions of each penalty are all designed as the sum of squares of the violation amount, ensuring the continuous differentiability (or at least the smoothness) of the penalty function, which is beneficial to the stable convergence of the optimization algorithm.

[0042] Step S104: Using the constraints as the feasible region, perform multi-parameter optimization iterative calculations on the objective function until a preset stage termination condition is reached. The optimization iterative calculation includes at least one optimization stage. In each optimization stage, at least a portion of the first and second design parameters are used as optimization variables. This step is the core execution link of the optimization design of this invention. Within the feasible region defined by the constraints defined in step S103, the optimizer systematically searches all design variables (including the presence or absence of magnets, polarity, position coordinates, and boundary curve shape parameters) defined in steps S101 and S102 to find the magnet configuration scheme that optimizes the objective function. The optimization iterative calculation includes at least one optimization stage. Each stage uses at least a portion of the design variables as the optimization object until a preset stage termination condition is reached (such as an upper limit on the number of iterations, no significant improvement in uniformity, etc.). The multi-stage strategy allows the optimization process to be carried out step by step—for example, first quickly locating the optimal solution region in discrete space, and then finely searching for local optima in continuous space, taking into account both global exploration capability and local optimization accuracy.

[0043] Optionally, multi-parameter optimization iterative calculations are performed using a genetic algorithm.

[0044] Optionally, different numbers of optimization stages can be used for designs with different initial design parameters. For example, for... Figure 2 The illustrated parameter design involves only one optimization stage in the iterative optimization calculation, using both the first and second design parameters as optimization variables. The complete flowchart is shown below. Figure 3 As shown. For Figure 4The illustrated parameter design and optimization iterative calculation involve two optimization stages. The first optimization stage uses the coordinates of each grid center as the fixed position of the candidate magnet, and the discrete variables of each candidate magnet... The second design parameter is used as the optimization variable for optimization; in the second optimization stage, the discrete variables of each candidate magnet are maintained. The optimization process remains unchanged, using the center coordinates of each candidate magnet as the optimization variable. The complete flowchart is as follows: Figure 5 As shown.

[0045] Single-stage optimization is suitable when the candidate magnet position is a continuous variable (e.g., Figure 2 As shown in the figure, in this case, all design parameters (including those of each magnet) are... ,coordinate and boundary curve parameters Both phases participate in the optimization iteration simultaneously within the same stage, and the algorithm directly searches for the global optimum within the complete design space. Two-stage optimization is suitable when the candidate magnet positions are pre-fixed at the grid center (e.g., ...). Figure 4 The scenario shown is designed to reduce the search difficulty through a step-by-step strategy: In the first stage, the magnet position is fixed at the center of the grid, and the optimizer only needs to search the discrete variables at each grid point. and boundary curve parameters Because the dimensionality of variables is significantly reduced and the variable types are relatively simple, the algorithm can quickly converge to a better initial layout scheme; the second stage starts from the layout scheme output by the first stage, maintaining the position of each magnet. (i.e., the presence or absence of magnets and their polarity) remains unchanged. Only the position coordinates of each magnet are continuously fine-tuned within a small range within its respective grid, while the boundary curve parameters are precisely adjusted, ultimately obtaining a high-precision optimal configuration. The two-stage strategy utilizes the fast convergence characteristics of discrete optimization while retaining the accuracy advantages of continuous optimization.

[0046] Step S105: Perform performance evaluation on the optimization results obtained in each optimization stage: if the evaluation passes, proceed to the next optimization stage or output the current optimization result; if the evaluation fails, return to the current optimization stage to continue optimization iteration. In this step, if it is a single-stage optimization, the output result is the current optimization result, such as... Figure 3 and Figure 6 As shown; for multi-stage optimization, the output of the last stage is the current optimization result, and the outputs of other stages are input into the next stage for further optimization, such as... Figure 5 As shown.

[0047] Performance evaluation serves as a quality check and feedback mechanism for the output results of each optimization stage. For single-stage optimization, performance evaluation is performed once after the entire optimization iteration. If the evaluation passes, the final design result is directly output; if the evaluation fails, the objective function value is calculated and returned to the optimizer to continue iteration. For multi-stage optimization, evaluation is performed after each stage—if the current stage evaluation passes, the next optimization stage is entered for further refinement; if the evaluation fails, the current stage is returned for re-optimization. For evaluations at intermediate stages, the current optimization result should be used as the initial input for the next stage, not the final output; only after the final stage evaluation passes is the magnet configuration result, serving as the final design scheme, output.

[0048] Optionally, the performance evaluation of the optimization results obtained in each optimization stage is performed, including: (1) determining the shape of the boundary curve based on the second design parameters obtained in the current optimization stage; (2) determining whether each candidate magnet is located inside or outside the boundary curve based on the position coordinates of each candidate magnet obtained in the current optimization stage, and assigning corresponding polar orientations to determine the current magnet layout scheme; (3) performing a geometric constraint check based on the shape of the boundary curve and the current magnet layout scheme. If the geometric constraint check fails, the evaluation is deemed unsuccessful; otherwise, proceed to the next step; (4) calculating the magnetic field distribution on the sputtering surface of the target material based on the current magnet layout scheme, and performing a magnetic induction intensity horizontal component constraint check. If the magnetic induction intensity horizontal component constraint check fails, the evaluation is deemed unsuccessful; otherwise, proceed to the next step; (5) calculating the coating uniformity based on the magnetic field distribution. If the coating uniformity meets the preset target, the evaluation is deemed successful; otherwise, the evaluation is deemed unsuccessful, and the objective function value is calculated.

[0049] In this embodiment, although the first design parameters already include discrete variables characterizing the polarity of the magnet, And during the optimization iteration process The value of is constantly adjusted, but the optimization algorithm improves during the search process. The adjustment is flexible—meaning the algorithm may assign different polarity values ​​to the same position in different iterations, resulting in a lack of geometric consistency in the search results. The role of the boundary curve is precisely to provide a unified geometric criterion for polarity assignment: after optimization, based on the final determined shape of the boundary curve, it is re-determined whether each magnet position is inside or outside the curve, and polarity is uniformly assigned accordingly. This approach ensures that the spatial distribution of polarity has a clear geometric regularity (one orientation inside, opposite orientation outside), avoiding the problem of chaotic polarity distribution that may occur during optimization, and ensuring the physical rationality and technological feasibility of the magnet layout scheme. In other words, During the optimization process, it is a temporary variable that assists in the search, while the boundary curve is the basis for the final polarity allocation decision.

[0050] In this embodiment, the geometric constraint check corresponds to the first four constraints, and the magnetic flux density horizontal component constraint check corresponds to the fifth constraint. The geometric constraint checks (the first four constraints) only involve coordinate and distance calculations, requiring no physical field simulation, resulting in extremely low computational cost. Therefore, they are placed at the very beginning of the evaluation process to quickly eliminate obviously unreasonable solutions such as overlapping magnets or exceeding boundaries, avoiding wasting computational resources on invalid solutions. The magnetic flux density horizontal component constraint check (the fifth constraint) can only be performed after calculating the magnetic field distribution, but its computational cost is still far lower than a complete coating uniformity simulation (the latter requires coupling multiple physical processes such as plasma, sputtering, and deposition). Therefore, the magnetic field constraint check is placed before the uniformity calculation as a second screening step; only solutions whose magnetic field horizontal component meets the discharge requirements are worth further calculation of coating uniformity. This layered and progressive evaluation strategy minimizes computational overhead while ensuring the quality of the final solution.

[0051] Optionally, the magnetic field distribution on the sputtered target surface can be calculated using analytical methods, numerical methods, or semi-analytical methods based on the superposition principle; the coating uniformity can be calculated using analytical models that predict film thickness uniformity or numerical methods based on a full-process simulation model of magnetron sputtering. Magnetic field calculations can be performed using analytical methods for rapid estimation, or numerical methods such as the finite element method can be used to obtain high-precision results, or a semi-analytical method based on the superposition principle of magnetic dipoles can be used to achieve a balance between accuracy and speed. The calculation of coating uniformity also offers multiple options—analytical models quickly predict film thickness distribution based on empirical or semi-empirical formulas, suitable for initial screening and rapid iteration; the full-process simulation model sequentially calculates the magnetic field distribution, electric field distribution, plasma behavior, target sputtering process, and the deposition process of sputtered atoms on the substrate, fully coupling the interactions of various physical fields at the physical mechanism level, achieving the highest accuracy, suitable for final scheme verification and high-precision design requirements. Different calculation methods can be flexibly selected according to the optimization stage and accuracy requirements, balancing computational efficiency and accuracy.

[0052] It should be noted that the step divisions of the various methods described above are only for clarity. In practice, they can be combined into one step or some steps can be broken down into multiple steps. As long as they contain the same logical relationship, they are all within the scope of protection of this patent. Adding insignificant modifications or introducing insignificant designs to the algorithm or process, without changing the core design of the algorithm and process, are also within the scope of protection of this patent. For example... Figure 1 In the overall flowchart shown, step S103 is: defining the constraints and objective function for the optimization design. And... Figure 3 , Figure 5 as well as Figure 6The specific flowcharts illustrate the constraints and objective functions. Although the flowchart representations differ, the overall concept behind these solutions remains the same.

[0053] In summary, the magnetic field design method for magnetron sputtering systems provided by this invention achieves proactive and automated search for magnetic field design by unifying magnet layout parameters and boundary curve parameters into a mathematical optimization framework, completely eliminating the reliance on empirical formulas and repeated trial production in traditional methods. This method has the following significant advantages: First, it significantly shortens the development cycle of magnetron sputtering systems, transforming the traditional serial iterative "design-trial production-testing-correction" model into a computationally guided parallel optimization model, reducing R&D costs; Second, by parametrically modeling the boundary curve, it transforms magnet polarity allocation into a geometric partitioning problem, ensuring the physical rationality and technological feasibility of the optimization results; Third, by adopting a multi-stage optimization strategy, it can perform step-by-step optimization according to different types of design variables (discrete / continuous) and search accuracy, balancing global exploration capability with local convergence accuracy; Fourth, through a hierarchical and progressive performance evaluation mechanism (geometric constraints), it achieves optimal results. (Magnetic field constraint → coating uniformity) While ensuring that the final design meets all constraints and performance targets, it saves computational resources to the greatest extent. Fifth, it provides a complete computational link from magnetic field calculation to coating uniformity prediction. In particular, the coupled magnetron sputtering full-process simulation model establishes a predictable and quantifiable relationship between magnetic field configuration and final coating performance from the physical mechanism level, providing a reliable design tool for the development of high-performance magnetron sputtering systems.

[0054] The above embodiments can be implemented, in whole or in part, by software, hardware (such as circuits), firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via a wired connection. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0055] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. A and B can be singular or plural. Additionally, the character " / " in this article generally indicates an "or" relationship between the preceding and following related objects, but it can also represent an "and / or" relationship. Please refer to the context for a more accurate understanding.

[0056] In this invention, "at least one" means one or more, and "more than one" means two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of a single item or a plurality of items. For example, at least one of a, b, or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be a single item or multiple items.

[0057] It should be understood that, in various embodiments of the present invention, the order of the above-mentioned process numbers does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0058] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0059] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the devices, apparatuses, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0060] In the several embodiments provided by this invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0061] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0062] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0063] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0064] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method of designing a magnetic field of a magnetron sputtering system, characterized in that, The method includes: Define first design parameters for candidate magnets on the magnetron base plate, the design parameters including discrete variables and position coordinates used to characterize the presence and polarity orientation of each candidate magnet; A parametric model of the boundary curve is established, and a second design parameter of the boundary curve is determined based on the parametric model. The boundary curve divides the surface of the base plate into an inner region and an outer region, and magnets located in different regions have different polarity orientations. Define the constraints and objective function for the optimization design; Using the constraints as the feasible region, the objective function is subjected to multi-parameter optimization iterative calculation until a preset stage termination condition is reached. The optimization iterative calculation includes at least one optimization stage. In each optimization stage, at least a portion of the first design parameter and the second design parameter are used as optimization variables. The performance of the optimization results obtained in each optimization stage is evaluated: if the evaluation passes, the process proceeds to the next optimization stage or the current optimization result is output; if the evaluation fails, the process returns to the current optimization stage to continue the optimization iteration.

2. The magnetic field design method for the magnetron sputtering system according to claim 1, characterized in that, In the first design parameter, the discrete variable is in the following form: where 0 means no magnet at the position, 1 means a magnet exists and the N pole faces the target material, and 2 means a magnet exists and the S pole faces the target material; the position coordinate is the center coordinate of the candidate magnet on the base plate ; represents the th candidate magnet, , is the preset maximum number of magnets; The optimization iterative calculation includes only one optimization stage, and uses all of the first design parameters and the second design parameters as optimization variables.

3. The magnetic field design method for the magnetron sputtering system according to claim 1, characterized in that, The base plate of the magnetron is divided into There are 1 grid, and the center coordinates of each grid are 1. Each grid corresponds to one candidate magnet; In the first design parameter, the discrete variable is determined in the following way: where 0 means no magnet at the position, 1 means a magnet exists and the N pole faces the target material, and 2 means a magnet exists and the S pole faces the target material; the position coordinate is the center coordinate of the candidate magnet on the base plate ; represents the th candidate magnet, ; The optimization iterative calculation includes two optimization stages. The first optimization stage uses the coordinates of each grid center as the fixed position of the candidate magnet, and uses the discrete variables of each candidate magnet... The second design parameter is used as the optimization variable for optimization; The second optimization stage maintains the discrete variables of each candidate magnet. The optimization is performed with the center coordinates of each candidate magnet remaining unchanged.

4. The magnetic field design method for the magnetron sputtering system according to claim 1, characterized in that, The parameterized model of the boundary curve is represented by a Fourier series in polar coordinates: , , In this system, the origin of the polar coordinate system is the center of rotation of the magnetron. The average radius of the boundary curve is... Let represent the highest order of the Fourier series and be a positive integer. The order number is 1 to 1. The integer, the second design parameter includes .

5. The magnetic field design method for the magnetron sputtering system according to claim 1, characterized in that, The parameterized model of the boundary curve is represented by a parameterized closed B-spline curve: , , in, The two-dimensional coordinates of the control points The total number of control points, with the coordinates of the first and last control points being the same to ensure the curve is closed; For node vectors, Let the order be the spline order. The second design parameter is the B-spline basis function; it includes the coordinates of each independent control point.

6. The magnetic field design method for the magnetron sputtering system according to claim 1, characterized in that, The constraints include: The center of all activating magnets is located within the area of ​​the base plate; The center-to-center distance between any two active magnets shall not be less than the preset minimum safety distance; The boundary curve is entirely located within the area of ​​the base plate; The shortest distance from the center of each activated magnet to the boundary curve is not less than a preset minimum distance threshold; The horizontal component of the magnetic induction intensity on the sputtered surface of the target material is between the preset minimum and maximum horizontal component values.

7. The magnetic field design method for the magnetron sputtering system according to claim 1, characterized in that, The objective function is to maximize a composite function that includes a penalty term: , in, This represents the coating non-uniformity determined by the distribution of magnetic induction intensity B on the sputtered surface of the target material, wherein the magnetic induction intensity B is determined by the first design parameter and the second design parameter; For each constraint violation penalty item, Let be the weighting coefficients corresponding to each penalty term. Each weighting coefficient is a sufficiently large positive number, such that any non-zero penalty term can suppress the fluctuations in coating unevenness.

8. The magnetic field design method for the magnetron sputtering system according to claim 7, characterized in that, The penalty item Includes at least one of the following: The boundary violation penalty is represented as the sum of the squares of the distances the activated magnet extends beyond the base plate. The overlap penalty term is represented as the sum of the squares of the violations when the distance between the centers of any two active magnets is less than the preset minimum safe distance. The penalty for violating the boundary of the dividing curve is expressed as the sum of the squares of the distances from which the dividing curve extends beyond the base plate. The magnet riding the curve penalty term is represented as the sum of squares of violations where the shortest distance from the center of the activated magnet to the boundary curve is less than a preset minimum distance threshold. The magnetic field violation penalty is represented as the sum of squares of the violation amount where the horizontal component of the magnetic induction intensity at each monitoring point on the sputtered surface of the target material is lower than the preset minimum horizontal component value or higher than the preset maximum horizontal component value.

9. The magnetic field design method for the magnetron sputtering system according to claim 1, characterized in that, The performance of the optimization results obtained in each optimization stage is evaluated, including: The shape of the boundary curve is determined based on the second design parameters obtained in the current optimization phase; Based on the position coordinates of each candidate magnet obtained in the current optimization stage, determine whether it is located inside or outside the boundary curve, and assign a corresponding polarity orientation to determine the current magnet layout scheme. Based on the shape of the boundary curve and the current magnet layout scheme, a geometric constraint check is performed. If the geometric constraint check fails, the evaluation is deemed unsuccessful; otherwise, proceed to the next step. Based on the current magnet layout scheme, calculate the magnetic field distribution on the sputtering surface of the target material and perform a constraint check on the horizontal component of the magnetic induction intensity. If the constraint check on the horizontal component of the magnetic induction intensity fails, the evaluation is deemed unsuccessful; otherwise, proceed to the next step. The coating uniformity is calculated based on the magnetic field distribution. If the coating uniformity meets the preset target, the evaluation is deemed to have passed; otherwise, the evaluation is deemed to have failed, and the target function value is calculated.

10. The magnetic field design method for the magnetron sputtering system according to claim 9, characterized in that, The multi-parameter optimization iterative calculation is performed using a genetic algorithm; the magnetic field distribution on the sputtered surface of the target material is calculated using analytical methods, numerical methods, or semi-analytical methods based on the superposition principle; the coating uniformity is calculated using an analytical model for predicting film thickness uniformity or a numerical method based on a full-process simulation model of magnetron sputtering.