Ultrasonic bonding parameter dynamic regulation and control method based on finite element simulation

Abstract

The invention discloses an ultrasonic bonding parameter dynamic regulation and control method based on finite element simulation, and the method comprises the steps: building a multi-physics coupling bonding process simulation model through finite element simulation analysis software in a finite element simulation pre-training stage, considering the key physical phenomena, such as ultrasonic vibration propagation, thermal coupling, etc. Parameterized simulation is carried out by combining orthogonal experimental design and a response surface method, and a quantitative mapping relation and a simulation database of process parameters-quality indexes are constructed. The real-time dynamic regulation and control system constructs a multi-closed-loop cooperative control system of real-time monitoring, simulation prediction and dynamic adjustment, collects data in real time, compares the data with a predicted value of a database, and automatically generates a parameter adjustment scheme when a target is deviated. The method can reduce resource consumption, improve the consistency, stability and reliability of bonding quality, reduce the reject ratio and improve the production efficiency, and can be popularized to other similar connection processes.

Description

Technical Field

[0001] This invention relates to the field of microelectronic packaging technology, and in particular to a method for dynamic control of ultrasonic bonding parameters based on finite element simulation. Background Technology

[0002] Bonding, as a core technology for achieving reliable inter-metal connections in fields such as microelectronic packaging and power device manufacturing, directly impacts product performance and reliability due to its bonding strength. In traditional processes, bonding parameters are often fixed values ​​or statically adjusted based on experience. This makes it difficult to adapt to the dynamic changes in material deformation and interface states over time during bonding, easily leading to problems such as incomplete soldering, insufficient bonding strength, or over-bonding causing device damage. To improve bonding quality stability, parameters need to be dynamically adjusted according to the actual state at different stages of bonding. However, the correlation between parameters and bonding effectiveness is difficult to accurately obtain experimentally, and the cost of trial and error is high. Summary of the Invention

[0003] This invention provides a dynamic control method for ultrasonic bonding parameters based on finite element simulation, which can reduce resource consumption, improve bonding quality consistency, stability and reliability, reduce defect rate, improve production efficiency, and can be extended to other similar bonding processes.

[0004] This invention provides a method for dynamic control of ultrasonic bonding parameters based on finite element simulation, comprising the following steps:

[0005] A simulation model of the multi-physics coupling bonding process was established using finite element simulation analysis software. Key physical phenomena such as ultrasonic vibration propagation and thermo-mechanical coupling were considered. The simulation model of the multi-physics coupling bonding process was parametrically simulated by combining orthogonal experimental design and response surface methodology. A quantitative mapping relationship between process parameters and quality indicators and a simulation database were constructed.

[0006] A multi-closed-loop collaborative control system is established to collect bonding parameter data in real time during the bonding process. The real-time collected bonding parameter data is compared and analyzed with the quality index prediction data in the simulation database. When the current bonding process state deviates from the expected target, bonding parameter adjustment data is generated.

[0007] The bonding process is controlled using the bonding parameter adjustment data until the bonding process achieves the expected goal.

[0008] Optionally, in one embodiment of the present invention, the method further includes:

[0009] During real-time bonding monitoring and bonding parameter data adjustment, the multi-closed-loop collaborative control system is optimized in real time using machine learning algorithms.

[0010] Optionally, in one embodiment of the present invention, the multiphysics coupling bonding process simulation model is used to characterize the vibration characteristics of the ultrasonic transducer, and considers the propagation law of ultrasonic vibration in multilayer materials, the complex thermo-mechanical coupling effect at the contact interface, the nonlinear plastic deformation behavior of the material, and the key physical phenomena of the metallurgical bonding process of the micro-interface.

[0011] Optionally, in one embodiment of the present invention, a simulation model of the multiphysics coupling bonding process is established, including:

[0012] Construct three-dimensional geometric entities of bonding tools, chips, intermediate layers, and substrates;

[0013] Set the material constitutive parameters, plastic hardening parameters, and thermophysical parameters, and configure the contact condition friction coefficient, boundary constraints, and dynamic load parameters;

[0014] The optimal discretization scheme was determined through systematic mesh sensitivity analysis. Local mesh refinement was adopted in the contact area to verify the model convergence and optimize the time step. A finite element simulation model was established to simulate material deformation, temperature distribution and stress state during the actual bonding process.

[0015] Optionally, in one embodiment of the present invention, the multiphysics coupling bonding process simulation model is parametrically simulated by combining orthogonal experimental design and response surface methodology, constructing a quantitative mapping relationship between process parameters and quality indicators and a simulation database, including:

[0016] Orthogonal experimental design and response surface methodology were employed to conduct full-factor simulation experiments on key process parameters such as ultrasonic amplitude, bonding pressure, and bonding time. The effects of different material combinations and surface roughness were also considered, and a simulation database covering a wide range of process windows was established. The simulation database contains predicted data of quality indicators such as bonding strength, interfacial bonding rate, and residual stress under various parameter combinations, as well as the spatiotemporal evolution laws of temperature field and stress field, establishing a quantitative mapping relationship between process parameters and quality indicators.

[0017] Optionally, in one embodiment of the present invention, real-time acquisition of bonding parameter data during the bonding process includes:

[0018] Real-time acquisition of vibration signals, pressure data, and stress information during the bonding process.

[0019] The finite element method for dynamic control of ultrasonic bonding parameters, as described in this invention, offers significant technical advantages compared to traditional fixed-parameter bonding processes. Firstly, it systematically studies the influence of process parameters through finite element simulation, avoiding the resource consumption of traditional trial-and-error methods. Secondly, the dynamic control mechanism can respond promptly to changes in process conditions, significantly improving the consistency of bonding quality. In practical applications, this invention can substantially enhance the stability and reliability of ultrasonic bonding. Particularly in the field of high-precision microelectronic packaging, this method can effectively reduce defect rates and improve production efficiency. Furthermore, this technical approach can be extended to other similar bonding processes.

[0020] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0021] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein:

[0022] Figure 1 A flowchart illustrating a method for dynamic control of ultrasonic bonding parameters based on finite element simulation according to an embodiment of the present invention;

[0023] Figure 2 This is a flowchart illustrating the simulation model of the multiphysics coupling bonding process in an embodiment of the present invention.

[0024] Figure 3 This is a flowchart illustrating the bonding parameter-state feature mapping relationship in an embodiment of the present invention.

[0025] Figure 4 This is a flowchart illustrating the dynamic control during the bonding process in an embodiment of the present invention. Detailed Implementation

[0026] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0027] Figure 1 This is a flowchart of a method for dynamic control of ultrasonic bonding parameters based on finite element simulation, according to an embodiment of the present invention.

[0028] like Figure 1 As shown, the method for dynamic control of ultrasonic bonding parameters based on finite element simulation includes the following steps:

[0029] In step S101, a simulation model of the multiphysics coupling bonding process is established using finite element simulation analysis software. Key physical phenomena such as ultrasonic vibration propagation and thermo-coupling are considered. The simulation model of the multiphysics coupling bonding process is parametrically simulated by combining orthogonal experimental design and response surface methodology, and a quantitative mapping relationship between process parameters and quality indicators and a simulation database are constructed.

[0030] In step S102, a multi-closed-loop collaborative control system is established to collect bonding parameter data in real time during the bonding process. The real-time collected bonding parameter data is compared and analyzed with the quality index prediction data in the simulation database. When the current bonding process state deviates from the expected target, bonding parameter adjustment data is generated.

[0031] In step S103, the bonding process is controlled using bonding parameter adjustment data until the bonding process reaches the expected goal.

[0032] In embodiments of the present invention, the method further includes: optimizing the multi-closed-loop collaborative control system in real time using machine learning algorithms during real-time bonding monitoring and bonding parameter data adjustment.

[0033] Specifically, the finite element simulation-based dynamic control method for ultrasonic bonding parameters of this invention mainly includes three key steps: finite element simulation pre-training, a real-time dynamic control system, and an adaptive optimization mechanism. Finite element simulation technology can construct a mechanical model of the bonding process to simulate stress distribution, plastic deformation, and other characteristics at different stages, providing a theoretical basis and optimization direction for dynamic parameter control. In the finite element simulation pre-training process, a high-precision multi-physics coupled bonding process simulation model needs to be established in the finite element simulation analysis software (ANSYS). This model not only needs to accurately characterize the vibration characteristics of the ultrasonic transducer but also needs to fully consider the propagation law of ultrasonic vibration in multilayer materials, the complex thermo-mechanical coupling effects at the contact interface (including frictional heat generation, heat conduction, and thermal expansion), the nonlinear plastic deformation behavior of materials, and the key physical phenomena such as the metallurgical bonding process of the micro-interface. Through systematic parametric simulation research, orthogonal experimental design and response surface methodology are used to conduct full-factor simulation experiments on key process parameters such as ultrasonic amplitude, bonding pressure, and bonding time. Simultaneously, the influence of different material combinations and surface roughness is considered, establishing a simulation database covering a wide range of process windows. This database not only contains predicted data on quality indicators such as bonding strength, interfacial bonding rate, and residual stress under various parameter combinations, but also records the spatiotemporal evolution laws of temperature field, stress field, etc. It establishes a quantitative mapping relationship between process parameters and quality indicators through machine learning algorithms, providing a reliable theoretical basis and prediction model for subsequent real-time control.

[0034] During the construction phase of the real-time dynamic control system, a multi-closed-loop collaborative control system integrating "real-time monitoring, simulation prediction, and dynamic adjustment" needs to be developed. This system collects vibration signals, pressure data, and stress information in real time during the bonding process. This real-time monitoring data is compared and analyzed with predicted values ​​in a database. When the process state deviates from the expected target, the system automatically generates parameter adjustment schemes. For example, when the interface temperature is detected to be lower than the simulation prediction value, the system intelligently adjusts the ultrasonic amplitude and bonding pressure, rather than simply using fixed compensation values. This dynamic control strategy based on simulation data enables more precise process control.

[0035] The adaptive optimization mechanism not only dynamically adjusts based on real-time monitoring data but also continuously optimizes the control model through machine learning algorithms. In the initial stage, the system primarily relies on a parameter optimization model established through simulation pre-training. As actual production data accumulates, on-site operating condition data is gradually incorporated into the training set, and the model parameters are continuously updated using incremental learning. This mechanism enables the system to automatically adapt to changes in actual conditions such as bond head wear and material property fluctuations, with control accuracy gradually improving over time. Simultaneously, the system records abnormal operating condition data and explores optimal adjustment strategies through reinforcement learning, ultimately forming an adaptive control system that balances theoretical predictions and practical experience.

[0036] like Figure 2 As shown, this invention first requires establishing a multiphysics coupling model of the ultrasonic bonding process in finite element simulation software. This model must fully consider the interactions of thermo-mechanical-ultrasonic multiphysics fields. First, it is necessary to accurately construct three-dimensional geometric entities of the bonding tool, chip, interlayer, and substrate. Parameters such as the radius of curvature at the tip of the bonding tool must be consistent with reality, and the dimensions of the chip, interlayer, and substrate must be set according to actual dimensions. Second, material constitutive parameters are set, including basic parameters such as elastic modulus, yield strength, Poisson's ratio, and coefficient of thermal expansion, as well as temperature-related plastic hardening parameters (Johnson-Cook model) and thermophysical parameters (thermal conductivity, specific heat capacity). Next, contact condition friction coefficient, boundary constraints, and dynamic load parameters, including ultrasonic amplitude, frequency, and bonding pressure, are configured. Then, the optimal discretization scheme is determined through system mesh sensitivity analysis, and local mesh refinement is adopted in the contact area. Finally, model convergence verification and time step optimization are performed to obtain a finite element simulation model that can accurately simulate material deformation, temperature distribution, and stress state during the actual bonding process.

[0037] A mapping between bonding process parameters and state characteristics is established through systematic finite element simulation experiments. For example... Figure 3As shown, orthogonal experimental design and Latin hypercube sampling were first employed to design multiple simulation conditions within the parameter space of ultrasonic amplitude, ultrasonic frequency, pressure, and time. Batch processing was performed using ANSYS LS-DYNA, and a Python script was used to automatically extract key feature data such as contact area, interfacial stress, and temperature distribution for each condition, as shown in Table 1. A response surface model was constructed based on the random forest algorithm, transforming four-dimensional process parameters into multi-objective quality indicators such as shear strength, temperature, and strain. This ultimately formed a process database containing a massive amount of data points, providing a foundation for parameter optimization in subsequent real-time control.

[0038] Table 1

[0039] Feature categories Specific parameters Extraction method Morphological analysis Contact area ratio, bump height Image processing Mechanical characteristics Interfacial shear stress, equivalent strain Unit integration point data statistics Thermal characteristics Maximum temperature, temperature gradient Node temperature output Energy characteristics Plastic dissipation energy, frictional work Energy balance calculation

[0040] like Figure 4 As shown, real-time dynamic control of bonding parameters first requires collecting real-time state data during the bonding process to ensure data accuracy and completeness. Next, the collected real-time state data is compared and analyzed with features in the knowledge base to identify potential problems or areas for optimization through in-depth data comparison. Then, algorithms process this data to predict the optimal parameters, providing a scientific basis for subsequent control operations. Control instructions are then generated based on the predicted optimal parameters and used for precise control of the bonding process. Parameter adjustment operations are then executed, strictly following the generated instructions to change the relevant parameters. Finally, the effectiveness of this control needs to be verified to evaluate whether the control has achieved the expected goals. If the control is successful, the entire bonding process is completed smoothly, and the entire process ends. If the control is unsuccessful, the response surface model needs to be optimized, the data in the knowledge base updated, and the above series of processes repeated continuously until the bonding process is successfully controlled.

[0041] The proposed method for dynamic control of ultrasonic bonding parameters based on finite element simulation in this invention solves the problem of unstable bonding quality caused by fixed parameters in traditional bonding processes. In the finite element simulation pre-training stage, a multi-physics coupled bonding process simulation model is established using ANSYS, considering key physical phenomena such as ultrasonic vibration propagation and thermo-coupling. Parametric simulation is performed using orthogonal experimental design and response surface methodology, constructing a quantitative mapping relationship between process parameters and quality indicators, as well as a simulation database. The real-time dynamic control system constructs a multi-closed-loop collaborative control system of "real-time monitoring - simulation prediction - dynamic adjustment," collecting data in real time and comparing it with predicted values ​​in the database. When deviations from the target occur, a parameter adjustment scheme is automatically generated. Compared with traditional processes, this method reduces resource consumption, improves bonding quality consistency, stability, and reliability, reduces defect rates, increases production efficiency, and can be extended to other similar bonding processes.

[0042] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0043] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "N" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0044] Any process or method described in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or N executable instructions for implementing custom logic functions or processes, and the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of this application pertain.

Claims

1. A method for dynamic control of ultrasonic bonding parameters based on finite element simulation, characterized in that, Includes the following steps: A simulation model of the multi-physics coupling bonding process was established using finite element simulation analysis software. Key physical phenomena such as ultrasonic vibration propagation and thermo-mechanical coupling were considered. The simulation model of the multi-physics coupling bonding process was parametrically simulated by combining orthogonal experimental design and response surface methodology. A quantitative mapping relationship between process parameters and quality indicators and a simulation database were constructed. A multi-closed-loop collaborative control system is established to collect bonding parameter data in real time during the bonding process. The real-time collected bonding parameter data is compared and analyzed with the quality index prediction data in the simulation database. When the current bonding process state deviates from the expected target, bonding parameter adjustment data is generated. The bonding process is controlled using the bonding parameter adjustment data until the bonding process achieves the expected goal.

2. The method according to claim 1, characterized in that, The method further includes: During real-time bonding monitoring and bonding parameter data adjustment, the multi-closed-loop collaborative control system is optimized in real time using machine learning algorithms.

3. The method according to claim 1, characterized in that, The simulation model of the multiphysics coupling bonding process is used to characterize the vibration characteristics of the ultrasonic transducer, and considers the propagation law of ultrasonic vibration in multilayer materials, the complex thermo-mechanical coupling effect at the contact interface, the nonlinear plastic deformation behavior of the material, and the key physical phenomena of the metallurgical bonding process of the micro-interface.

4. The method according to claim 1, characterized in that, Establish a simulation model for the multiphysics coupling bonding process, including: Construct three-dimensional geometric entities of bonding tools, chips, intermediate layers, and substrates; Set the material constitutive parameters, plastic hardening parameters, and thermophysical parameters, and configure the contact condition friction coefficient, boundary constraints, and dynamic load parameters; The optimal discretization scheme was determined through systematic mesh sensitivity analysis. Local mesh refinement was adopted in the contact area to verify the model convergence and optimize the time step. A finite element simulation model was established to simulate material deformation, temperature distribution and stress state during the actual bonding process.

5. The method according to claim 1, characterized in that, The simulation model of the multiphysics coupled bonding process is parametrically simulated by combining orthogonal experimental design and response surface methodology, and a quantitative mapping relationship between process parameters and quality indicators and a simulation database are constructed, including: Orthogonal experimental design and response surface methodology were employed to conduct full-factor simulation experiments on key process parameters such as ultrasonic amplitude, bonding pressure, and bonding time. The effects of different material combinations and surface roughness were also considered, and a simulation database covering a wide range of process windows was established. The simulation database contains predicted data of quality indicators such as bonding strength, interfacial bonding rate, and residual stress under various parameter combinations, as well as the spatiotemporal evolution laws of temperature field and stress field, establishing a quantitative mapping relationship between process parameters and quality indicators.

6. The method according to claim 1, characterized in that, Real-time acquisition of bonding parameter data during the bonding process, including: Real-time acquisition of vibration signals, pressure data, and stress information during the bonding process.

Images