Wafer thinning equipment control method and device, electronic equipment and storage medium

Abstract

The invention provides a wafer thinning equipment control method and device, electronic equipment and a storage medium, and relates to the technical field of wafer thinning. The method comprises the following steps: monitoring spreading and thinning behaviors of a spin-coating liquid film in real time and generating a residual layer distribution diagram; according to an area indicated by the residual layer distribution map, obtaining local thickness information by analyzing a fluorescence signal received from the area; after a laser beam is used for scanning the surface of the wafer, micro-vibration signals are recognized and extracted by analyzing reflected light signals; dividing the surface of the wafer into a plurality of grid areas; and the risk of generating cracks in each grid area is evaluated, and a wafer surface risk thermodynamic diagram used for controlling the grinding head and the grinding fluid is generated. The method aims at solving a series of problems caused by a microcosmic residual layer on the surface of the wafer or surface energy difference in the wafer thinning process, early and accurate sensing of microcosmic abnormity on the surface of the wafer is achieved, and a control system is guided to conduct regional and self-adaptive grinding parameter adjustment.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of wafer thinning, in particular to a wafer thinning equipment control method and device, electronic equipment and storage medium. BACKGROUND

[0002] Ultra-thin wafer manufacturing is a core link in the semiconductor industry, and has strict requirements on processing precision and final product yield. Wafer thinning processing usually adopts mechanical grinding method to gradually reduce the wafer thickness to tens of microns. However, this high-precision process is easy to introduce micro-cracks in the wafer interior or surface, which may cause device performance degradation or failure in the subsequent chip manufacturing and packaging process. Therefore, it is a key challenge to improve the yield of ultra-thin wafers to realize real-time and accurate monitoring of potential crack risk during the thinning process, and to adaptively adjust the control parameters accordingly to effectively prevent cracks from occurring.

[0003] The current advanced method uses quantum dot thickness sensors and Doppler vibrometers for in-situ monitoring and compensation, and predicts the crack generation probability through a model to adjust the control parameters in real time. However, these systems have limitations when facing certain non-ideal conditions.

[0004] For example, before some batches of wafers enter the thinning process, the residual layer of organic or inorganic residues formed on the wafer surface is difficult to detect with the naked eye, is extremely thin, and is unevenly distributed due to fluctuations in the drying parameters or residual cleaning process in the previous process.

[0005] When the system spin-coats quantum dot material on the surface of these wafers, the residual layer affects the wettability, spreading and adhesion of the quantum dot material, resulting in uneven thickness or density of the quantum dot film in the micro region, or local agglomeration. In the subsequent grinding, the action of the grinding fluid, local heat and mechanical stress make the quantum dots with weak adhesion or agglomeration more likely to locally fall off, degrade or quench fluorescence.

[0006] At this time, the fluorescence signal received by the quantum dot thickness sensor is no longer pure thickness information. In the damaged or uneven area of the quantum dot film, the fluorescence intensity abnormally decreases or the spectral peak shifts. The built-in spectrum analyzer of the system is designed and calibrated based on ideal, uniform and stable quantum dot films, and will misjudge these signal abnormalities caused by damage to the film integrity as wafer thickness fluctuations or background noise, and perform regular filtering processing. As a result, the instantaneous thickness data fed back by the system shows slight "noise" in the local area, but the overall trend is smoothly decreasing, and no abnormal alarm is triggered, which masks the true situation that the local thickness measurement accuracy is reduced due to damage to the quantum dot film.

[0007] Meanwhile, the local non-uniformity or damage of the quantum dot film causes the friction characteristics and local contact stress distribution of the polishing head and the wafer surface to change, resulting in local micro-vibrations with extremely low energy but extremely high frequency. These micro-vibrations indicate local stress concentration and potential surface damage. However, the laser Doppler vibrometer mainly monitors the macro-vibration mode of the wafer, and its sampling frequency and signal processing algorithm focus on capturing low-frequency, medium-high amplitude vibrations. This high-frequency, low-energy local micro-vibration signal is overwhelmed by macro-vibration and cannot be identified on the conventional readings of the vibrometer, which reports that the overall vibration of the wafer is stable.

[0008] Under the misjudgment of information, the prediction model in the central processing unit receives the smoothly decreasing thickness data and the stable overall vibration state. The model cannot identify the hidden local abnormalities, predicts a low probability of crack generation, and suggests that the controller maintain the high-efficiency polishing parameters. The controller executes the instructions and maintains the high-efficiency polishing. However, in the area where the quantum dot film is damaged, the local thickness measurement is inaccurate, and there are high-frequency micro-vibrations, the wafer surface is subjected to local stress beyond expectation, resulting in microscopic surface damage or sub-surface crack initiation. These hidden damages do not immediately cause wafer rupture when the thinning is completed, but in subsequent processes, due to additional mechanical stress or thermal stress, the micro-cracks rapidly expand, eventually leading to chip fragmentation or failure, causing yield loss. SUMMARY

[0009] The purpose of the present application is to provide a wafer thinning equipment control method, device, electronic equipment and storage medium, aiming to solve the problems of quantum dot film damage, thickness measurement deviation and difficulty in identifying local micro-vibration signals caused by microscopic residual layers or surface energy differences on the wafer surface during wafer thinning, and to realize early and accurate perception of microscopic abnormalities on the wafer surface and guide the control system to make regional and adaptive adjustments to the polishing parameters.

[0010] In a first aspect, the present application provides a wafer thinning equipment control method, comprising the following steps: S1. During the process of spin-coating quantum dot material on the wafer surface, the spreading and thinning behavior of the spin-coating liquid film is monitored in real time, and the monitored spreading and thinning behavior of the spin-coating liquid film is compared with a preset standard liquid film behavior model to obtain a comparison result; S2. Determine whether there is a microscopic residual layer or surface energy difference on the wafer surface according to the comparison result, and generate a wafer surface residual layer distribution map; the wafer surface residual layer distribution map is used to indicate the area where there is a microscopic residual layer or surface energy difference; S3. According to the area indicated by the wafer surface residual layer distribution map, analyze the fluorescence signal received by the quantum dot thickness sensor from the area to obtain local wafer thickness information; S4. After controlling the array laser unit to emit a laser beam of a specific frequency modulation to scan the wafer surface, receive the reflected light signal generated by the scanning; S5. Identify and extract the micro-vibration signal related to local stress concentration by analyzing the reflected light signal; S6. Divide the wafer surface into a plurality of grid areas by performing a grid processing on the wafer surface; S7. Evaluate the risk of crack generation of each grid area according to the wafer surface residual layer distribution map, the local wafer thickness information and the micro-vibration signal, and generate a wafer surface risk heat map according to the risk evaluation result; S8. Control the operation parameters of the grinding head and the flow of the grinding liquid according to the wafer surface risk heat map.

[0011] The core technical concept of the wafer thinning equipment control method provided by the application is to introduce real-time optical interference monitoring of the spin-coating liquid film. It indirectly and accurately identifies the microscopic residual layer and its distribution on the wafer surface that cannot be detected by the naked eye by monitoring the spreading and thinning behavior of the spin-coating liquid film in real time using optical interference technology before or during the spin-coating of quantum dot materials. This early perception strategy at the source of the problem can effectively avoid the local damage of the quantum dot film and the thickness measurement deviation caused by the residual layer, fundamentally solve the problem of information pollution at the source, and provide the most original and accurate surface state information for the subsequent identification of local micro-vibration signals and real-time update risk zoning evaluation of the wafer surface, and finally guide the control system to make fine regional grinding parameter adjustment.

[0012] In a second aspect, the application provides a wafer thinning equipment control device, comprising: A monitoring and comparison module is configured to monitor the spreading and thinning behavior of the spin-coating liquid film in real time during the spin-coating of quantum dot materials on the wafer surface, compare the monitored spreading and thinning behavior of the spin-coating liquid film with a preset standard liquid film behavior model, and obtain a comparison result; A first generation module is configured to determine whether there is a microscopic residual layer or surface energy difference on the wafer surface according to the comparison result, and generate a wafer surface residual layer distribution map; the wafer surface residual layer distribution map is used to indicate the area where there is a microscopic residual layer or surface energy difference; A thickness analysis module is configured to analyze the fluorescence signal received by the quantum dot thickness sensor from the area indicated by the wafer surface residual layer distribution map, and obtain local wafer thickness information; A receiving module is configured to receive the reflected light signal generated by the scanning after controlling the array laser unit to emit a laser beam of a specific frequency modulation to scan the wafer surface; An identification module is configured to identify and extract the micro-vibration signal related to local stress concentration by analyzing the reflected light signal; The dividing module is configured to divide the wafer surface into a plurality of grid regions by performing a grid processing on the wafer surface. The second generating module is configured to evaluate a risk of crack generation in each of the grid regions according to the wafer surface residual layer distribution map, the local wafer thickness information and the micro-vibration signal, and generate a wafer surface risk heat map according to a risk evaluation result. The control module is configured to control an operation parameter of the grinding head and a flow of the grinding liquid according to the wafer surface risk heat map.

[0013] The wafer thinning equipment control device provided by the application can effectively identify and quantify the micro residual layer on the wafer surface, the damage of the quantum dot film and the local stress concentration and other abnormalities that are difficult to detect by traditional methods, thereby overcoming the information misjudgment dilemma in the prior art, realizing early warning and self-adaptive control of potential crack risks and significantly improving the yield and reliability of the ultra-thin wafer.

[0014] In a third aspect, the application provides an electronic device, comprising a processor and a memory, wherein the memory stores computer readable instructions, and when the computer readable instructions are executed by the processor, the steps in the wafer thinning equipment control method provided in the first aspect are executed.

[0015] In a fourth aspect, the application provides a computer readable storage medium, which stores a computer program, and when the computer program is executed by a processor, the steps in the wafer thinning equipment control method provided in the first aspect are executed.

[0016] As can be seen from the above, the wafer thinning equipment control method provided by the application can accurately perceive the surface micro-abnormalities in the early stage of the wafer thinning process, and correct and enhance the multi-source local information, thereby realizing early and accurate evaluation of the crack risk of the wafer surface. This enables the control system to make fine regional grinding parameter adjustment, thereby fundamentally avoiding the generation of hidden micro-cracks and significantly improving the yield and reliability of the ultra-thin wafer.

[0017] Other features and advantages of the application will be set forth in the following description, and in part will become apparent to those skilled in the art from the description, or can be learned by practice of the application as described in the written description and from the drawings. The purposes and other advantages of the application will be realized and attained by the structures particularly pointed out in the written description and the drawings. BRIEF DESCRIPTION OF DRAWINGS

[0018] Figure 1 A flowchart of the wafer thinning equipment control method provided by the embodiment of the application.

[0019] Figure 2A structural schematic diagram of a wafer thinning equipment control device provided by an embodiment of the present application.

[0020] Figure 3 A structural schematic diagram of an electronic device provided by an embodiment of the present application.

[0021] Label explanation: 100, monitoring comparison module; 200, first generation module; 300, thickness analysis module; 400, receiving module; 500, identification module; 600, division module; 700, second generation module; 800, control module; 13, electronic device; 1301, processor; 1302, memory; 1303, communication bus. DETAILED DESCRIPTION

[0022] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments of the present application. The components of the embodiments of the present application described and shown in the accompanying drawings can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the accompanying drawings is not intended to limit the scope of the claimed present application, but only represents selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of the present application.

[0023] It should be noted that: similar labels and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings. Meanwhile, in the description of the present application, the terms "first", "second", etc. are only used to distinguish the description, and cannot be understood as indicating or implying relative importance.

[0024] With reference to the accompanying drawings, Figure 1 the present application provides a wafer thinning equipment control method, comprising the following steps: S1. In the process of spin-coating quantum dot material on the wafer surface, the spreading and thinning behavior of the spin-coating liquid film is monitored in real time by an optical interference measurement system, and the monitored spreading and thinning behavior of the spin-coating liquid film is compared with a preset standard liquid film behavior model to obtain a comparison result; S2. According to the comparison result, it is judged whether there is a micro residual layer or a surface energy difference on the wafer surface, and a wafer surface residual layer distribution map is generated; the wafer surface residual layer distribution map is used to indicate the area where there is a micro residual layer or a surface energy difference; S3. According to the area indicated by the wafer surface residual layer distribution map, analyze the fluorescence signal received by the quantum dot thickness sensor from the area, judge whether the quantum dot film integrity is damaged through the spectral characteristics of the fluorescence signal, and according to the quantum dot film damage judgment result, calibrate the original thickness measurement data to obtain local wafer thickness information; S4. After controlling the array laser unit integrated below the grinding head to emit a laser beam of a specific frequency modulation to scan the wafer surface, receiving the reflected light signal generated by the scanning; S5. By analyzing the phase noise spectral density or high-order harmonic component of the reflected light signal, identifying and extracting the micro-vibration signal related to local stress concentration; S6. By grid processing the wafer surface, the wafer surface is divided into multiple grid areas; S7. According to the wafer surface residual layer distribution map, local wafer thickness information and micro-vibration signal, evaluate the risk of crack in each grid area, and according to the risk evaluation result, generate a wafer surface risk heat map; S8. According to the wafer surface risk heat map, control the operating parameters of the grinding head and the flow of the grinding liquid; Specifically, it includes regionally adjusting the downforce, rotational speed of the grinding head and local flow of the grinding liquid.

[0025] The present application can effectively identify and quantify the abnormality such as micro residual layer on wafer surface, quantum dot film damage and local stress concentration which is difficult to be detected by traditional methods, so as to overcome the dilemma of information misjudgment in the prior art, realize early warning and self-adaptive control of potential crack risk, and significantly improve the yield and reliability of ultra-thin wafers.

[0026] The embodiment provides a wafer thinning equipment control method, which aims to solve the problem that crack risk caused by micro residual layer, quantum dot film damage and local stress concentration is difficult to be effectively identified and controlled in the process of thinning ultra-thin wafers. The method realizes comprehensive perception of wafer surface state and fine regulation and control of thinning process by integrating multi-source sensing data and intelligent analysis.

[0027] Among them, "optical interferometry system" is a system that uses the principle of light wave interference for precise measurement, which functions to monitor the thickness change and spreading dynamics of the spin-on liquid film in real time and non-contact. "Quantum dot material" is a kind of nanometer semiconductor material with special optical properties, which is spin-coated on the wafer surface as a thickness sensing medium in this application, and its fluorescence characteristics are closely related to the film thickness. "Standard liquid film behavior model" refers to the theoretical or empirical model of liquid film spreading and thinning based on ideal spin coating conditions and material properties, which is used for comparison with actual monitoring data to find abnormalities. "Microscopic residual layer" refers to the extremely thin layer of organic or inorganic substances on the wafer surface that cannot be seen by the naked eye, which may affect subsequent processes. "Surface energy difference" refers to the difference in surface free energy exhibited by different regions of the wafer surface due to differences in physical or chemical properties, which also affects the wetting and adhesion of materials. "Quantum dot thickness sensor" is a device that uses the fluorescence characteristics of quantum dot materials to measure film thickness, and analyzes the spectral characteristics of the fluorescence signal to infer thickness information. "Array laser unit" is a system that contains multiple independent controllable laser transmitters and receivers, which can perform high-resolution scanning and signal acquisition on the wafer surface. "Phase noise spectral density" and "high-order harmonic component" are characteristic parameters of laser reflection signals, used to represent signal stability, nonlinearity and weak vibration information. "Micro-vibration signal" refers to extremely low energy and high frequency micro-vibrations caused by local stress concentration or abnormal contact. "Wafer surface risk thermodynamic map" is a visualization tool that visually represents the crack risk level of different regions of the wafer surface through color depth or numerical size.

[0028] The wafer thinning equipment control method of the embodiment is characterized by multi-dimensional data acquisition and intelligent analysis, which realizes accurate identification and adaptive control of potential crack risks in the wafer thinning process.

[0029] In step S1, during the process of spin-coating quantum dot material on the wafer surface, the spreading and thinning behavior of the spin-coating liquid film is monitored in real time by an optical interferometry system, and the monitored spreading and thinning behavior of the spin-coating liquid film is compared with a preset standard liquid film behavior model to obtain a comparison result. Specifically, the optical interferometry system can use a white light interferometer or a laser interferometer. For example, the white light interferometer irradiates broadband white light onto the surface of the spin-coating liquid film and receives the reflected light. Because the reflected light from the upper and lower surfaces of the liquid film interferes, the thickness of the liquid film can be calculated in real time by analyzing the spectral information of the interference pattern. At the same time, the system can also capture the dynamic changes of the liquid film edge by a high-speed camera, thereby monitoring its spreading behavior. These real-time monitored liquid film thickness, spreading speed, uniformity and other parameters are compared with the preset standard liquid film behavior model. The model can be established based on a large amount of experimental data and fluid mechanics theory, and describes the spreading and thinning trajectory of the liquid film under ideal conditions. The comparison result can be in the form of deviation value, abnormal area marking, etc.

[0030] In step S2, whether there is a micro residual layer or surface energy difference on the wafer surface is determined according to the comparison result, and a wafer surface residual layer distribution map is generated. The wafer surface residual layer distribution map is used to indicate the area where there is a micro residual layer or surface energy difference. For example, if the comparison result shows that the spreading speed of the liquid film in some areas is significantly slower than the model prediction, or the liquid film thickness in local areas appears abnormal fluctuation or unevenness, it can be inferred that these areas may have a micro residual layer or surface energy difference. These differences will change the wettability of the wafer surface, thereby affecting the spreading of the liquid film. The system can generate a wafer surface residual layer distribution map according to the position and degree of these abnormal areas, which can be presented in the form of a two-dimensional image, where different colors or gray values represent the severity of the residual layer or surface energy difference.

[0031] In step S3, the quantum dot thickness sensor analyzes the fluorescence signals received from the area indicated by the wafer surface residual layer distribution map. By analyzing the spectral characteristics of the fluorescence signals, such as fluorescence intensity, peak position, and half-width, the system can determine whether the quantum dot film has been damaged by the residual layer or surface energy differences. If the film is damaged, the original thickness measurement data may be inaccurate. In this case, the original data needs to be calibrated according to the degree of damage. The calibration method can include establishing a correction model that relates the degree of damage to the thickness measurement error, and then correcting the original measurement data to obtain more accurate local wafer thickness information. The correction model can be a polynomial regression model. For example, in the wafer thinning device control method, when the quantum dot film is damaged, the thickness analysis module receives fluorescence signals. The correction model can be trained using experimental data to establish a relationship between the degree of damage to the quantum dot film (such as fluorescence intensity decay rate, peak shift) and the thickness measurement error. Assuming that the fluorescence intensity decay rate is X, the degree of thermal-induced degradation is Y, and the original thickness measurement data is Z. The correction model can be expressed as: calibrated thickness = Z + f(X, Y), where f(X, Y) is a correction function, such as f(X, Y) = aX + bY + cXY + d, where a, b, c, and d are model coefficients. When the system determines that the film is damaged and quantifies X and Y, the correction model calculates the corresponding correction value and applies it to the original thickness measurement data Z, thereby obtaining the calibrated local wafer thickness information.

[0032] In step S4, after controlling the array laser unit integrated below the polishing head to emit a laser beam modulated at a specific frequency to scan the wafer surface, the reflected light signals generated by the scanning are received. For example, the array laser unit can be composed of multiple independent semiconductor lasers, each emitting a laser beam modulated at a specific frequency. These laser beams scan the wafer surface at a predetermined scanning path and speed. When the laser beams hit the wafer surface, they produce reflected light. The photodetectors in the array laser unit receive these reflected light signals and convert them into electrical signals for subsequent processing.

[0033] In step S5, the micro-vibration signals related to local stress concentration are identified and extracted by analyzing the phase noise spectral density or higher harmonic components of the reflected light signals. For example, the received reflected light signals contain modulation information of the micro-vibration on the wafer surface. By performing spectral analysis on these signals, their phase noise spectral density or higher harmonic components can be calculated. The micro-vibration caused by local stress concentration usually appears as high-frequency, low-energy signals, which produce specific peaks or patterns in the phase noise spectral density or higher harmonic components. Through signal processing algorithms such as wavelet transform, Fourier transform combined with band-pass filtering, etc., these micro-vibration signals related to local stress concentration can be identified and extracted from the complex reflected light signals.

[0034] In step S6, the wafer surface is divided into multiple grid regions through grid processing. For example, the entire wafer surface can be divided into a two-dimensional grid, each grid region has a predetermined size, such as 1mm*1mm. This grid processing helps to manage and analyze the parameters of the wafer surface locally, providing a spatial reference for subsequent risk assessment.

[0035] In step S7, according to the wafer surface residual layer distribution map, local wafer thickness information and micro-vibration signal, the risk of crack generation in each grid area is evaluated, and a wafer surface risk heat map is generated according to the risk evaluation result. For example, for each grid area, the severity of the residual layer in the corresponding residual layer distribution map, the thickness uniformity or abnormal value in the local wafer thickness information, and the stress concentration degree in the micro-vibration signal can be considered comprehensively. A multi-parameter risk evaluation model can be established, which calculates the input parameters by weighting and outputs a crack risk score; the parameter risk evaluation model can be a mathematical model based on weighted summation. The model receives three quantitative indicators as input: residual layer influence index, thickness non-uniformity index and stress concentration intensity index. Each index is assigned a weight, for example, the weight of the residual layer influence index is 0.3, the weight of the thickness non-uniformity index is 0.4, and the weight of the stress concentration intensity index is 0.3. The calculation formula of the model is: crack risk score=(residual layer influence index x 0.3)+(thickness non-uniformity index x 0.4)+(stress concentration intensity index x 0.3). The value range of each index is 0 to 10, 0 represents no risk, and 10 represents high risk. For example, if the residual layer influence index of a certain grid area is 7, the thickness non-uniformity index is 6, and the stress concentration intensity index is 8, then the crack risk score of the area is (7 x 0.3)+(6 x 0.4)+(8 x 0.3)=2.1+2.4+2.4=6.9. The score is then used to generate a wafer surface risk heat map. The higher the risk score, the greater the risk of crack generation in the area. According to the risk scores of all grid areas, a wafer surface risk heat map can be generated, in which different colors or brightnesses represent different risk levels, and high-risk areas are displayed intuitively.

[0036] In step S8, according to the wafer surface risk heat map, the grinding head and the grinding liquid are controlled; specifically including regional adjustment of the downforce, rotational speed of the grinding head and the local flow of the grinding liquid. For example, for the high-risk areas displayed in the risk heat map, the control system can instruct the grinding head to reduce the downforce in the area, reduce the rotational speed, or increase the local flow of the grinding liquid in the area through the microfluidic device to reduce the grinding intensity and reduce stress concentration, thereby effectively preventing the generation of cracks. For low-risk areas, higher grinding efficiency can be maintained. This regional adaptive adjustment can achieve fine control of the thinning process.

[0037] The wafer thinning equipment control method of the embodiment can effectively identify and quantify abnormal conditions such as micro residual layer on the wafer surface, quantum dot film damage, and local stress concentration, which are difficult to detect by traditional methods, by introducing a multi-dimensional, high-precision real-time monitoring and intelligent analysis mechanism, thereby overcoming the information misjudgment dilemma in the prior art, realizing early warning and adaptive control of potential crack risks, and significantly improving the yield and reliability of ultra-thin wafers.

[0038] Specifically, the present application realizes a breakthrough in the prior art through the following core innovations: Firstly, during the spin coating of quantum dot materials, the present application uses an optical interferometry system to monitor the spreading and thinning behavior of the spin coating liquid film in real time and compares it with a pre-set standard liquid film behavior model. This step can early detect micro residual layers or surface energy differences on the wafer surface, which are difficult to detect by traditional methods but have a key impact on the uniformity and adhesion of the subsequent quantum dot film. Compared to the prior art which only relies on quantum dot thickness sensors for thickness measurement, the present application intervenes in the evaluation of the surface state at an earlier process stage, thereby laying a foundation for subsequent accurate measurement and risk assessment.

[0039] Secondly, the present application analyzes the fluorescence signal received by the quantum dot thickness sensor according to the area indicated by the wafer surface residual layer distribution map, and judges whether the quantum dot film integrity is damaged by the spectral characteristics of the fluorescence signal, and calibrates the original thickness measurement data accordingly. This innovation directly solves the problem of distorted thickness measurement data caused by damaged quantum dot films in the prior art. Traditional methods often misjudge the signal anomaly caused by damaged quantum dot films as thickness fluctuations or background noise, leading to a decrease in local thickness measurement accuracy. By in-depth analysis of the spectral characteristics, the present application can distinguish between film damage and actual thickness changes, thereby obtaining more accurate local wafer thickness information and avoiding the dilemma of "information misjudgment".

[0040] Furthermore, the present application controls the array laser unit integrated below the grinding head to emit laser beams of a specific frequency modulation to scan the wafer surface, and identifies and extracts micro-vibration signals related to local stress concentration by analyzing the phase noise spectrum density or high-order harmonic components of the reflected light signals. This technical means breaks through the limitation of traditional laser Doppler vibrometers in capturing high-frequency, low-energy local micro-vibration signals. Through fine analysis of the reflected light signals, the present application can identify weak signals that indicate local stress concentration and potential surface damage, providing key early warning information for crack risk assessment and making up for the shortcomings of the prior art in local stress monitoring.

[0041] Finally, the application synthesizes the wafer surface residual layer distribution map, local wafer thickness information and micro-vibration signal to evaluate the risk of crack generation in each grid area, and generates a wafer surface risk heat map, and then adjusts the down pressure, rotation speed of the grinding head and local flow of the grinding liquid according to the heat map. The risk assessment mechanism of multi-source information fusion combined with regional adaptive control enables the thinning process to be fine-tuned according to the actual local conditions of the wafer surface. Compared with the method of overall adjustment based on macroscopic data in the prior art, the method can more effectively prevent the generation of microscopic cracks, and significantly improves the yield and reliability of ultra-thin wafers.

[0042] In summary, by introducing a multi-dimensional and high-precision real-time monitoring and intelligent analysis mechanism in the wafer thinning equipment control method, the application can effectively identify and quantify the abnormality such as wafer surface microscopic residual layer, quantum dot film damage and local stress concentration which is difficult to detect by traditional methods, thereby overcoming the information misjudgment dilemma in the prior art, realizing early warning and adaptive control of potential crack risk, and significantly improving the yield and reliability of ultra-thin wafers.

[0043] In some embodiments, according to the quantum dot film damage judgment result, the step of calibrating the original thickness measurement data to obtain the local wafer thickness information includes: S31. Obtain local instantaneous temperature information of the grinding interface; S32. Obtain multi-dimensional quantum dot fluorescence characteristics; S33. According to the local instantaneous temperature information, the multi-dimensional quantum dot fluorescence characteristics and the quantum dot film damage judgment result, decouple and quantify the influence degree of residual layer damage and thermal-induced degradation on the state of the quantum dot film; S34. According to the quantified residual layer damage influence degree and thermal-induced degradation influence degree, adjust the correction parameter of the original thickness measurement data, and calibrate the original thickness measurement data using the correction parameter to obtain the local wafer thickness information.

[0044] Specifically, in step S31, the local instantaneous temperature information of the grinding interface can be obtained by integrating micro temperature sensors near the grinding head or the wafer surface. These sensors can be non-contact infrared sensors, or contact thermocouples or resistance temperature detectors, which are intended to accurately capture the temperature fluctuations in the local area during the grinding process.

[0045] In step S32, the multi-dimensional quantum dot fluorescence characteristics can be understood as including spectral information of multiple dimensions such as fluorescence lifetime, fluorescence peak position drift, fluorescence half-peak width change, in addition to fluorescence intensity. These characteristics can be collected by a high-sensitivity spectrometer or a time-resolved fluorescence detection system, and the purpose is to provide richer and more detailed quantum dot film state information to more accurately reflect the damage situation.

[0046] In actual application, in step S33, the influence degree of residual layer damage and heat-induced degradation on the quantum dot film state is decoupled and quantified, which is specifically realized by establishing a multi-factor coupling model. For example, machine learning algorithms can be used, and local instantaneous temperature information, multi-dimensional quantum dot fluorescence characteristics, and quantum dot film damage judgment results are taken as inputs. By training a pre-set damage model, the fluorescence signal attenuation caused by residual layer and the fluorescence quenching or spectral change caused by heat induction are distinguished and quantified. The purpose is to accurately identify and separate the influence of different damage mechanisms on the performance of the quantum dot film; Specifically, the multi-factor coupling model can be realized by machine learning algorithms such as support vector machine (SVM), neural network or decision tree integrated model. First, a training data set containing multiple damage conditions is needed, wherein each sample includes local instantaneous temperature information, multi-dimensional quantum dot fluorescence characteristics (such as fluorescence intensity, peak position, half-width, lifetime, etc.) and the true value of the corresponding residual layer damage degree and heat-induced degradation degree. For example, quantum dot film samples with only residual layer damage, only heat-induced degradation and both can be manufactured by controlling experimental conditions, and the damage degree is measured using high-precision analysis equipment. Then, use these data to train the machine learning model, so that it can learn and identify the mapping relationship between different input feature combinations and different damage types and degrees. After training is completed, when new real-time data is input into the model, the model can output the quantified values of the influence degree of residual layer damage and heat-induced degradation on the quantum dot film state.

[0047] Further, in step S34, according to the quantified residual layer damage influence degree and heat-induced degradation influence degree, the correction parameters of the original thickness measurement data are adjusted, which is specifically realized by consulting a pre-set correction parameter database or by real-time calculation of a correction function. For example, when the influence degree of residual layer damage is high, the correction of the deviation caused by signal attenuation in the original thickness measurement data can be increased; when the influence degree of heat-induced degradation is high, the fluorescence efficiency correction parameter related to temperature can be adjusted. The purpose is to ensure that the calibration process can finely compensate for different types of damage, thereby improving the accuracy of thickness measurement.

[0048] The scheme of the present application introduces local transient temperature information of the grinding interface and multi-dimensional quantum dot fluorescence characteristics, and further decouples and quantifies the influence of residual layer damage and thermal-induced degradation on the state of the quantum dot film, thereby solving the problem that the traditional method fails to fully consider the coupling of multiple complex damage mechanisms when calibrating the original thickness measurement data. It is precisely because the specific influence of different damage sources (such as residual layer and thermal effect) on the state of the quantum dot film can be accurately distinguished and quantified that the subsequent correction parameter adjustment is more targeted. In this way, calibration deviation caused by a single damage judgment result can be avoided, ensuring that the local wafer thickness information is more accurate and reliable in a complex grinding environment.

[0049] Through the above technical scheme, the measurement accuracy and reliability of the local wafer thickness information can be significantly improved. Because the influence of residual layer damage and thermal-induced degradation on the state of the quantum dot film can be finely decoupled and quantified, the calibration process of the original thickness measurement data is more targeted and accurate. This high-precision thickness information is crucial for subsequent wafer surface crack risk assessment, which can effectively reduce the risk of misjudgment caused by inaccurate thickness measurement, thereby providing more precise and safe control basis for wafer thinning process, and ultimately improving the yield and product quality of wafer thinning.

[0050] In some embodiments, the step of identifying and extracting the micro-vibration signal related to local stress concentration by analyzing the phase noise spectral density or high-order harmonic component of the reflected light signal includes: S51. Synchronously collecting the phase noise spectral density or high-order harmonic component of the reflected light signal at different positions on the wafer surface using multiple receiving points of the array laser unit to obtain multi-point reflected signals; S52. Performing time-space synchronous analysis on the collected multi-point reflected signals to obtain time-space synchronous analysis results; S53. Determining the origin position of the micro-vibration signal and identifying the spatial propagation mode related to local stress concentration according to the time-space synchronous analysis results; S54. According to the origin position and spatial propagation mode of the micro-vibration signal, by distinguishing the micro-vibration caused by wafer surface damage from background noise, the micro-vibration signal related to local stress concentration is identified and extracted.

[0051] Specifically, in step S51, multiple receiving points of the arrayed laser unit are configured to simultaneously receive reflected light signals from different regions of the wafer surface. These receiving points can be arranged in a matrix or a ring to cover a large area of the wafer. Synchronous acquisition refers to recording the phase noise spectral density or higher harmonic components of the reflected light signals received by all receiving points simultaneously within the same time window. The phase noise spectral density can characterize the frequency stability of the signal, while the higher harmonic components can reflect nonlinear effects, both of which can be sensitive indicators of micro-vibration signals. Through multi-point synchronous acquisition, micro-vibration information at different positions on the wafer surface can be obtained, forming a multi-point reflected signal data set.

[0052] Further, in step S52, the collected multi-point reflected signals are subjected to spatio-temporal synchronous analysis. Spatio-temporal synchronous analysis refers to correlating and comparing signals collected by different receiving points at different time points. This can use signal processing techniques such as correlation analysis, Fourier transform, wavelet analysis, etc. to reveal the propagation characteristics of the signals in space and time. For example, by calculating the phase difference and amplitude attenuation between signals from different receiving points, the propagation direction and speed of the micro-vibration wave can be inferred.

[0053] Thus, in step S53, according to the results of spatio-temporal synchronous analysis, the origin position of the micro-vibration signal can be determined, and the spatial propagation pattern related to local stress concentration can be identified. The origin position of the micro-vibration signal usually corresponds to the region of stress concentration or potential damage. The spatial propagation pattern refers to the way the micro-vibration wave spreads on the wafer surface, such as point source diffusion, line source diffusion, or more complex patterns. These patterns can be compared with known stress distribution models to improve the accuracy of identification. For example, a finite element analysis (FEA) model can be established to simulate the micro-vibration response and stress distribution of the wafer under different local stress concentration conditions. The FEA model can include various typical stress concentration patterns, such as edge cracks, internal defects, surface scratches, etc., and predict the propagation characteristics of the micro-vibration wave under these patterns. When the system identifies that the actual micro-vibration signal originates from the edge of the wafer and exhibits a specific spatial propagation pattern that spreads along the edge to both sides, this pattern can be matched with the pre-set edge crack stress distribution model in the FEA model. If the matching degree is high, it can be confirmed that the micro-vibration signal is indeed related to the edge stress concentration.

[0054] Finally, in step S54, according to the origin position and spatial propagation pattern of the micro-vibration signal, the micro-vibration signal related to local stress concentration is identified and extracted by distinguishing the micro-vibration caused by wafer surface damage from background noise. Background noise may come from equipment vibration, environmental interference or other non-stress concentration caused micro-disturbance. By analyzing the characteristics of the signal (such as frequency, amplitude, duration) and its spatial propagation behavior, the real stress concentration caused by micro-vibration signal can be effectively separated from the noise. For example, the micro-vibration caused by damage usually has a specific frequency range and propagation path, while the background noise may show a more random or uniform distribution.

[0055] The scheme of the present application effectively solves the problems of insufficient precision and noise interference in identifying micro-vibration signals related to local stress concentration by introducing multi-point synchronous acquisition, space-time synchronous analysis and identification of origin position and spatial propagation pattern. Specifically, the use of multiple receiving points of the array laser unit to synchronously acquire reflected light signals can provide more comprehensive micro-vibration information of the wafer surface, avoiding the loss of local information that may be caused by single-point measurement. By performing space-time synchronous analysis on these multi-point signals, the propagation path and origin of the micro-vibration wave can be accurately tracked, so that the real stress concentration area can be distinguished from the non-stress concentration area. Further, the origin position and spatial propagation pattern of the micro-vibration signal are determined, so that the system can effectively distinguish the micro-vibration caused by wafer surface damage from random background noise according to these unique physical characteristics, thereby ensuring that the extracted micro-vibration signal more accurately reflects the local stress concentration state inside the wafer.

[0056] Through the above technical scheme, the accuracy and reliability of identifying and extracting micro-vibration signals related to local stress concentration can be significantly improved. The combination of multi-point synchronous acquisition and space-time synchronous analysis enables the origin position and propagation pattern of the micro-vibration signal to be accurately determined, so that background noise and non-stress concentration caused interference can be more effectively excluded. As a result, the obtained micro-vibration signal can more truly reflect the local stress concentration inside the wafer, providing more accurate data support for subsequent crack risk assessment, and thus improving the safety and yield of the wafer thinning process.

[0057] In some embodiments, in step S7, according to the wafer surface residual layer distribution map, local wafer thickness information and micro-vibration signal, the step of evaluating the risk of each grid area to produce cracks includes: S71. Obtain the quantitative indicators corresponding to the wafer surface residual layer distribution map, local wafer thickness information and micro-vibration signal; S72. Assign weights to each quantitative indicator and perform comprehensive calculation to obtain a crack risk score as the risk assessment result of the corresponding grid area to produce cracks.

[0058] The quantification of the wafer surface residual layer distribution map, the local wafer thickness information, and the micro-vibration signal corresponds to the original data obtained from different sources, such as the residual layer thickness or density in the wafer surface residual layer distribution map, the thickness deviation or gradient in the local wafer thickness information, and the amplitude or frequency characteristics in the micro-vibration signal, which are converted into a unified quantitative form that can be numerically compared and calculated by a specific method. Specifically, the wafer surface residual layer distribution map can be quantified as a residual layer coverage or average thickness value; the local wafer thickness information can be quantified as a local thickness non-uniformity index; and the micro-vibration signal can be quantified as a stress concentration intensity index. The purpose is to standardize heterogeneous data for subsequent comprehensive analysis. These quantitative indicators are numerical representations of the original information to facilitate subsequent risk assessment algorithms.

[0059] Further, assigning weights to each of the quantitative indicators and performing comprehensive calculation to obtain a crack risk score as the risk assessment result of the corresponding grid area is to assign a weight coefficient to each quantitative indicator according to the influence degree of different indicators on crack formation, and to perform comprehensive calculation on these weighted indicator values by weighted summation or other multi-factor decision model, so as to obtain a single numerical value representing the crack risk of the grid area, i.e. the crack risk score. For example, the different contributions of residual layer, thickness non-uniformity and stress concentration to crack formation can be determined according to historical data or expert experience, and the corresponding weights can be set accordingly. The purpose is to provide a comprehensive and quantitative risk assessment result to guide the subsequent adjustment of grinding parameters.

[0060] The scheme of the present application converts the three types of key data, i.e. wafer surface residual layer distribution map, local wafer thickness information, and micro-vibration signal, into quantifiable indicators, and performs weighted comprehensive calculation on these indicators, thereby realizing systematic evaluation of the crack risk of each grid area. Specifically, step S71 ensures that data of different sources and different forms can be uniformly processed and compared, laying a data foundation for subsequent risk assessment. It is precisely because these key factors are quantified that they can be included in a unified evaluation framework. On this basis, step S72 reflects the relative importance of different indicators in inducing cracks by assigning weights to them, for example, the residual layer in some areas may contribute more to cracks than the local thickness non-uniformity, and the weight adjustment can more accurately reflect this difference. Through comprehensive calculation, the one-sidedness of single indicator evaluation is avoided, and the crack risk score can more comprehensively and accurately reflect the real risk situation of each area on the wafer surface.

[0061] By the technical solution, the wafer thinning crack risk assessment method is more refined and accurate. Specifically, through comprehensive analysis of multi-dimensional quantitative indicators, false positives or false negatives caused by single factor evaluation can be effectively avoided, and the accuracy and reliability of crack risk prediction are significantly improved. Therefore, the crack risk score obtained can more truly reflect the potential crack risk of each grid area on the wafer surface, providing a more scientific and accurate basis for subsequent regional adjustment of the grinding head and the grinding liquid, thereby effectively reducing the probability of cracks in the wafer thinning process and improving the yield and production efficiency of the wafer.

[0062] Please refer to Figure 2 , Figure 2 is a wafer thinning device control device in some embodiments of the application, which is integrated in the form of a computer program in a rear-end control device, comprising: The monitoring and comparison module 100 is used to monitor the spreading and thinning behavior of the spin-coated liquid film in real time during the process of spin-coating quantum dot material on the wafer surface, and compare the monitored spreading and thinning behavior of the spin-coated liquid film with a preset standard liquid film behavior model to obtain a comparison result. The first generation module 200 is used to determine whether there is a micro residual layer or surface energy difference on the wafer surface according to the comparison result, and generate a wafer surface residual layer distribution map; the wafer surface residual layer distribution map is used to indicate the area where there is a micro residual layer or surface energy difference. The thickness analysis module 300 is used to analyze the fluorescence signal received by the quantum dot thickness sensor from the area indicated by the wafer surface residual layer distribution map to obtain local wafer thickness information. The receiving module 400 is used to receive the reflected light signal generated by scanning after controlling the array laser unit to emit a laser beam of a specific frequency modulation to scan the wafer surface. The identification module 500 is used to identify and extract the micro-vibration signal related to local stress concentration by analyzing the reflected light signal. The division module 600 is used to divide the wafer surface into a plurality of grid areas by grid processing the wafer surface. The second generation module 700 is used to evaluate the risk of crack generation in each grid area according to the wafer surface residual layer distribution map, the local wafer thickness information and the micro-vibration signal, and generate a wafer surface risk heat map according to the risk evaluation result. The control module 800 is used to control the operating parameters of the grinding head and the flow of the grinding liquid according to the wafer surface risk heat map.

[0063] In some embodiments, the thickness analysis module 300 performs the following when it is used to analyze the fluorescence signal received by the quantum dot thickness sensor from the area indicated by the wafer surface residual layer distribution map to obtain the local wafer thickness information: According to the spectral characteristics of the fluorescence signal, it is determined whether the quantum dot film integrity is damaged, and a quantum dot film damage judgment result is obtained; According to the quantum dot film damage judgment result, the original thickness measurement data is calibrated to obtain the local wafer thickness information.

[0064] In some embodiments, the thickness analysis module 300 performs the following when it is used to analyze the fluorescence signal received by the quantum dot thickness sensor from the area indicated by the wafer surface residual layer distribution map to obtain the local wafer thickness information: S31. Obtain local transient temperature information of the polishing interface; S32. Obtain multi-dimensional quantum dot fluorescence characteristics; S33. According to the local transient temperature information, the multi-dimensional quantum dot fluorescence characteristics, and the quantum dot film damage judgment result, the influence degree of the residual layer damage and the thermal-induced degradation on the quantum dot film state is decoupled and quantified; S34. According to the quantified residual layer damage influence degree and thermal-induced degradation influence degree, the correction parameter of the original thickness measurement data is adjusted, and the original thickness measurement data is calibrated using the correction parameter to obtain the local wafer thickness information.

[0065] In some embodiments, the identification module 500 performs the following when it is used to identify and extract the micro-vibration signal related to local stress concentration by analyzing the reflected light signal: By analyzing the phase noise spectral density or high-order harmonic component of the reflected light signal, the micro-vibration signal related to local stress concentration is identified and extracted.

[0066] In some embodiments, the identification module 500 performs the following when it is used to identify and extract the micro-vibration signal related to local stress concentration by analyzing the phase noise spectral density or high-order harmonic component of the reflected light signal: S51. Synchronously collect the phase noise spectral density or high-order harmonic component of the reflected light signal at different positions on the wafer surface using multiple receiving points of the array laser unit to obtain multi-point reflected signals; S52. Time and space synchronization analysis is performed on the collected multi-point reflected signals to obtain a time and space synchronization analysis result; S53. According to the time and space synchronization analysis result, the origin position of the micro-vibration signal is determined, and the spatial propagation mode related to the local stress concentration is identified; S54. According to the origin position and spatial propagation mode of the micro-vibration signal, the micro-vibration signal related to the local stress concentration is identified and extracted by distinguishing the micro-vibration caused by wafer surface damage from the background noise.

[0067] In some embodiments, the second generation module 700 performs the following steps when generating a risk thermal map of the wafer surface for controlling the polishing head and the polishing liquid according to the wafer surface residual layer distribution map, the local wafer thickness information and the micro-vibration signal: S71. Obtain the quantization indexes corresponding to the wafer surface residual layer distribution map, the local wafer thickness information and the micro-vibration signal; the quantization indexes are the digital representations of the wafer surface residual layer distribution map, the local wafer thickness information and the micro-vibration signal; S72. Assign weights to each quantization index and perform a comprehensive calculation to obtain a crack risk score as the risk assessment result of the corresponding grid area producing cracks.

[0068] In some embodiments, the control module 800 performs the following steps when controlling the polishing head and the polishing liquid according to the wafer surface risk thermal map: S81. According to the wafer surface risk thermal map, the downforce, the rotation speed of the polishing head and the local flow of the polishing liquid are adjusted regionally.

[0069] Please refer to Figure 3 , Figure 3A structural schematic diagram of an electronic device provided by the embodiment of the present application, the present application provides an electronic device 13, comprising: a processor 1301 and a memory 1302, the processor 1301 and the memory 1302 are interconnected and communicate with each other through a communication bus 1303 and / or other forms of connection mechanism (not marked), the memory 1302 stores computer readable instructions executable by the processor 1301, when the electronic device runs, the processor 1301 executes the computer readable instructions to execute the wafer thinning device control method in any optional implementation manner of the above-mentioned embodiment, to realize the following functions: in the process of spin coating quantum dot material on the wafer surface, the spreading and thinning behavior of the spin coating liquid film is monitored in real time, and the monitored spreading and thinning behavior of the spin coating liquid film is compared with the preset standard liquid film behavior model to obtain a comparison result; according to the comparison result, it is judged whether there is micro residual layer or surface energy difference on the wafer surface, and a wafer surface residual layer distribution map is generated; the wafer surface residual layer distribution map is used to indicate the area where there is micro residual layer or surface energy difference; according to the area indicated by the wafer surface residual layer distribution map, the fluorescence signal received by the quantum dot thickness sensor from the area is analyzed to obtain local wafer thickness information; after the array type laser unit emits a laser beam of a specific frequency modulation to scan the wafer surface, the reflected light signal generated by the scanning is received; by analyzing the reflected light signal, the micro vibration signal related to local stress concentration is recognized and extracted; the wafer surface is grid processed to divide the wafer surface into a plurality of grid areas; according to the wafer surface residual layer distribution map, the local wafer thickness information and the micro vibration signal, the risk of crack of each grid area is evaluated, and according to the risk evaluation result, a wafer surface risk heat map is generated; according to the wafer surface risk heat map, the operating parameters of the grinding head and the flow of the grinding liquid are controlled.

[0070] The embodiment of the present application provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to perform the wafer thinning device control method in any optional implementation manner of the above embodiment, so as to realize the following functions: in the process of spin coating quantum dot material on the wafer surface, the spreading and thinning behavior of the spin coating liquid film is monitored in real time, the monitored spreading and thinning behavior of the spin coating liquid film is compared with a preset standard liquid film behavior model, and a comparison result is obtained; whether there is a micro residual layer or a surface energy difference on the wafer surface is judged according to the comparison result, and a wafer surface residual layer distribution map is generated; the wafer surface residual layer distribution map is used to indicate the area where the micro residual layer or the surface energy difference exists; local wafer thickness information is obtained by analyzing the fluorescence signal received by the quantum dot thickness sensor from the area indicated by the wafer surface residual layer distribution map; after the array type laser unit emits a laser beam of a specific frequency modulation to scan the wafer surface, the reflected light signal generated by the scanning is received; by analyzing the reflected light signal, a micro vibration signal related to local stress concentration is recognized and extracted; the wafer surface is subjected to grid processing, and the wafer surface is divided into a plurality of grid areas; according to the wafer surface residual layer distribution map, the local wafer thickness information and the micro vibration signal, the risk of crack generation of each grid area is evaluated, and according to the risk evaluation result, a wafer surface risk heat map is generated; according to the wafer surface risk heat map, the operating parameters of the grinding head and the flow of the grinding liquid are controlled.

[0071] The computer readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0072] In the embodiments of the present application, it should be understood that the disclosed apparatus and method can be implemented in other manners. The embodiments described above are merely exemplary, for example, the division of the units is only a logical function division, and there can be another division manner in actual implementation; for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the displayed or discussed mutual couplings or direct couplings or communication connections can be indirect couplings or communication connections through some interfaces, devices or units, and can be in electrical, mechanical or other forms.

[0073] In addition, the units described as separate components can or can not be physically separate, and the components displayed as units can or can not be physical units, i.e., can be located in one place, or can be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purposes of the embodiments of the present application.

[0074] In addition, the functional modules in each of the embodiments of the present application can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0075] In this article, the relationship terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations.

[0076] The above description is only some embodiments of the present application, and is not used to limit the protection scope of the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A wafer thinning apparatus control method characterized by comprising: The method comprises the following steps: S1. Real-time monitoring of the spreading and thinning behavior of the spin coating liquid film during the process of spin coating quantum dot material on the wafer surface, and comparing the monitored spreading and thinning behavior of the spin coating liquid film with a preset standard liquid film behavior model to obtain a comparison result; S2. Judging whether there is a micro residual layer or surface energy difference on the wafer surface according to the comparison result, and generating a wafer surface residual layer distribution map; the wafer surface residual layer distribution map is used to indicate the area where there is a micro residual layer or surface energy difference; S3. According to the area indicated by the wafer surface residual layer distribution map, analyzing the fluorescence signal received by the quantum dot thickness sensor from the area to obtain local wafer thickness information; S4. After the array laser unit emits a laser beam of a specific frequency modulation to scan the wafer surface, the reflected light signal generated by the scanning is received; S5. By analyzing the reflected light signal, the micro-vibration signal related to local stress concentration is identified and extracted; S6. By performing grid processing on the wafer surface, the wafer surface is divided into multiple grid areas; S7. According to the wafer surface residual layer distribution map, the local wafer thickness information and the micro-vibration signal, the risk of each grid area to produce cracks is evaluated, and a wafer surface risk heat map is generated according to the risk evaluation result; S8. According to the wafer surface risk heat map, the operating parameters of the grinding head and the flow of the grinding liquid are controlled.

2. The wafer thinning apparatus control method according to claim 1, wherein The specific steps in step S3 include: According to the spectral characteristics of the fluorescence signal, it is judged whether the integrity of the quantum dot film is damaged, and a quantum dot film damage judgment result is obtained; According to the quantum dot film damage judgment result, the original thickness measurement data is calibrated to obtain the local wafer thickness information.

3. The wafer thinning apparatus control method according to claim 2, wherein The step of calibrating the original thickness measurement data according to the quantum dot film damage judgment result to obtain the local wafer thickness information includes: S31. Obtain local instantaneous temperature information of the grinding interface; S32. Obtain multi-dimensional quantum dot fluorescence characteristics; S33. According to the local instantaneous temperature information, the multi-dimensional quantum dot fluorescence characteristics and the quantum dot film damage judgment result, the influence degree of residual layer damage and thermal induced degradation on the state of quantum dot film is decoupled and quantified; S34. According to the quantified residual layer damage influence degree and thermal induced degradation influence degree, the correction parameters of the original thickness measurement data are adjusted, and the original thickness measurement data is calibrated by using the correction parameters to obtain the local wafer thickness information.

4. The wafer thinning apparatus control method according to claim 1, wherein The specific steps in step S5 include: By analyzing the phase noise spectral density or high-order harmonic component of the reflected light signal, the micro-vibration signal related to local stress concentration is identified and extracted.

5. The wafer thinning apparatus control method according to claim 4, wherein By analyzing the phase noise spectral density or high-order harmonic component of the reflected light signal, the micro-vibration signal related to local stress concentration is identified and extracted. The steps include: S51. Synchronously collect the phase noise spectral density or high-order harmonic component of the reflected light signal at different positions on the wafer surface by using multiple receiving points of the array laser unit to obtain multi-point reflected signals; S52. Perform a time-space synchronization analysis on the collected multi-point reflection signals to obtain a time-space synchronization analysis result; S53. Determine the origin position of the micro-vibration signals according to the time-space synchronization analysis result, and identify a spatial propagation mode related to the local stress concentration; S54. Identify and extract the micro-vibration signals related to the local stress concentration by distinguishing the micro-vibration caused by the wafer surface damage from the background noise according to the origin position of the micro-vibration signals and the spatial propagation mode.

6. The wafer thinning apparatus control method according to claim 1, wherein In step S7, the step of evaluating the risk of crack generation in each grid area according to the wafer surface residual layer distribution map, the local wafer thickness information, and the micro-vibration signals includes: S71. Obtain a quantitative index corresponding to the wafer surface residual layer distribution map, the local wafer thickness information, and the micro-vibration signals; the quantitative index is a digital representation of the wafer surface residual layer distribution map, the local wafer thickness information, and the micro-vibration signals; S72. Assign a weight to each quantitative index and perform a comprehensive calculation to obtain a crack risk score as the risk evaluation result of crack generation in the corresponding grid area.

7. The wafer thinning apparatus control method according to claim 1, wherein The specific steps in step S8 include: S81. According to the wafer surface risk thermal map, regionally adjust the downforce, rotation speed of the grinding head, and local flow of the grinding liquid.

8. A wafer thinning apparatus control device characterized by comprising: Comprise: A monitoring comparison module, configured to monitor the spreading and thinning behavior of the spin-coating liquid film in real time during the process of spin-coating quantum dot materials on the wafer surface, and compare the monitored spreading and thinning behavior of the spin-coating liquid film with a preset standard liquid film behavior model to obtain a comparison result; A first generation module, configured to determine whether there is a micro residual layer or surface energy difference on the wafer surface according to the comparison result, and generate a wafer surface residual layer distribution map; the wafer surface residual layer distribution map is used to indicate the area where there is a micro residual layer or surface energy difference; A thickness analysis module, configured to analyze the fluorescence signal received by the quantum dot thickness sensor from the area indicated by the wafer surface residual layer distribution map to obtain local wafer thickness information; A receiving module, configured to receive the reflected light signal generated by scanning the wafer surface after the array laser unit emits a laser beam of a specific frequency modulation; An identification module, configured to identify and extract the micro-vibration signals related to the local stress concentration by analyzing the reflected light signal; A division module, configured to divide the wafer surface into a plurality of grid areas by performing a grid processing on the wafer surface; A second generation module, configured to evaluate the risk of crack generation in each grid area according to the wafer surface residual layer distribution map, the local wafer thickness information, and the micro-vibration signals, and generate a wafer surface risk thermal map according to the risk evaluation result; A control module, configured to control the operating parameters of the grinding head and the flow of the grinding liquid according to the wafer surface risk thermal map.

9. An electronic device, comprising: Comprise a processor and a memory, the memory stores computer readable instructions, when the computer readable instructions are executed by the processor, the steps in the wafer thinning equipment control method of any one of claims 1-7 are run.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program, when executed by a processor, performs the steps of the wafer thinning apparatus control method of any of claims 1-7.

Images