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

Abstract

The application provides a wafer thinning equipment control method and device, electronic equipment and storage medium, and relates to the technical field of wafer thinning. The method comprises the following steps: real-time monitoring of the spreading and thinning behavior of the spin-on liquid film and generating a residual layer distribution map; obtaining local thickness information by analyzing the fluorescence signal received from the area indicated by the residual layer distribution map according to the area; after scanning the wafer surface with a laser beam, identifying and extracting the micro-vibration signal by analyzing the reflected light signal; dividing the wafer surface into multiple grid areas; evaluating the risk of crack generation in each grid area and generating a wafer surface risk thermodynamic map for controlling the grinding head and the grinding liquid. The method aims to solve a series of problems caused by micro-residual layers or surface energy differences on the wafer surface during the wafer thinning process, realize early and accurate perception of micro-abnormalities on the wafer surface, and guide the control system to make regional and adaptive adjustments to the grinding parameters.

Description

Technical Field

[0001] This invention relates to the field of wafer thinning technology, and more specifically, to a wafer thinning equipment control method, apparatus, electronic device, and storage medium. Background Technology

[0002] Ultrathin wafer manufacturing is a core process in the semiconductor industry, with stringent requirements for processing precision and final product yield. Wafer thinning typically employs mechanical grinding to gradually reduce the wafer thickness to tens of micrometers. However, this high-precision process easily introduces microscopic cracks inside or on the surface of the wafer. These cracks can lead to performance degradation or failure during subsequent chip manufacturing and packaging. Therefore, achieving real-time, accurate monitoring of potential crack risks during the thinning process and adaptively adjusting control parameters to effectively prevent crack formation is a key challenge for improving ultrathin wafer yield.

[0003] Current advanced methods utilize quantum dot thickness sensors and Doppler vibration meters for in-situ monitoring and compensation, and predict crack initiation probabilities through models to adjust control parameters in real time. However, these systems have limitations when faced with specific non-ideal conditions.

[0004] For example, before certain batches of wafers enter the thinning process, residues from the previous cleaning process or fluctuations in drying parameters can cause the formation of extremely thin, non-uniformly distributed organic or inorganic residue layers on the wafer surface that are difficult to detect with the naked eye. These residue layers adhere to the wafer surface at the nanometer scale, and their chemical properties or surface energy differ from those of a standard silicon surface.

[0005] When quantum dot materials are spin-coated onto these wafer surfaces, the residual layer affects the wettability, spreadability, and adhesion of the quantum dot materials, leading to uneven thickness or density of the quantum dot film in micro-regions, or localized agglomeration. During subsequent polishing, the polishing fluid, localized heat, and mechanical stress make the weakly adhered or agglomerated quantum dots more prone to localized detachment, degradation, or fluorescence quenching.

[0006] At this point, the fluorescence signal received by the quantum dot thickness sensor is no longer pure thickness information. In areas where the quantum dot film is damaged or non-uniform, the fluorescence intensity decreases abnormally or the spectral peak shifts. The system's built-in spectrometer, based on an ideal, uniform, and stable quantum dot film design and calibration, misinterprets these signal anomalies caused by impaired film integrity as wafer thickness fluctuations or background noise, and performs conventional filtering. As a result, the instantaneous thickness data fed back by the system shows slight "noise" in local areas, but the overall trend is a smooth decline, failing to trigger an anomaly alarm and masking the true situation where damage to the quantum dot film leads to a decrease in local thickness measurement accuracy.

[0007] Meanwhile, localized inhomogeneities or damage to the quantum dot film lead to changes in the frictional characteristics and local contact stress distribution between the polishing head and the wafer surface, generating localized micro-vibrations with extremely low energy but extremely high frequency. These micro-vibrations indicate localized stress concentration and potential surface damage. However, laser Doppler vibration meters primarily monitor the macroscopic vibration modes of the wafer, and their sampling frequency and signal processing algorithms focus on capturing low-frequency, medium-to-high amplitude vibrations. These high-frequency, low-energy localized micro-vibration signals are overwhelmed by macroscopic vibrations and cannot be identified in the vibration meter's conventional readings; the system reports that the overall wafer vibration is stable.

[0008] Under information misinterpretation, the prediction model within the central processing unit receives steadily decreasing thickness data and a stable overall vibration state. The model fails to identify masked local anomalies, predicts a low probability of crack formation, and suggests the controller maintain efficient grinding parameters. The controller executes the instructions to maintain high-efficiency grinding. However, in areas with damaged quantum dot films, inaccurate local thickness measurements, and high-frequency micro-vibration regions, the wafer surface experiences unexpected local stress, leading to microscopic surface damage or subsurface crack initiation. These latent damages do not immediately cause wafer breakage upon completion of thinning, but in subsequent processes, due to additional mechanical or thermal stress, the microcracks rapidly propagate, ultimately causing chip breakage or failure, resulting in yield loss. Summary of the Invention

[0009] The purpose of this invention is to provide a control method, device, electronic device and storage medium for wafer thinning equipment, which aims to solve the problems of quantum dot film damage, thickness measurement deviation and difficulty in identifying local micro-vibration signals caused by micro-residual layers or surface energy differences on the wafer surface during the wafer thinning process. It enables early and accurate perception of micro-anomalies on the wafer surface and guides the control system to make regional and adaptive adjustments to the grinding parameters.

[0010] In a first aspect, the present invention provides a method for controlling a wafer thinning device, comprising the following steps:

[0011] S1. During the spin coating of quantum dot materials 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 the comparison results;

[0012] S2. Based on the comparison results, determine whether there is a microscopic residual layer or surface energy difference on the wafer surface, and generate a residual layer distribution map on the wafer surface; the residual layer distribution map on the wafer surface is used to indicate the areas where there is a microscopic residual layer or surface energy difference;

[0013] S3. Based on the region indicated by the residual layer distribution map on the wafer surface, analyze the fluorescence signal received by the quantum dot thickness sensor from the region to obtain local wafer thickness information;

[0014] S4. After the control array laser unit emits a laser beam modulated at a specific frequency to scan the wafer surface, it receives the reflected light signal generated by the scan;

[0015] S5. By analyzing the reflected light signal, identify and extract the micro-vibration signal related to local stress concentration;

[0016] S6. By performing a gridding process on the wafer surface, the wafer surface is divided into multiple grid regions;

[0017] S7. Based on the residual layer distribution map on the wafer surface, the local wafer thickness information, and the micro-vibration signal, assess the risk of crack generation in each of the grid regions, and generate a risk heat map of the wafer surface based on the risk assessment results;

[0018] S8. Based on the risk thermal map of the wafer surface, control the operating parameters of the polishing head and the flow rate of the polishing slurry.

[0019] The core technical concept of the wafer thinning equipment control method provided by this invention lies in introducing real-time optical interference monitoring of spin-coating liquid film changes. By using optical interference technology to monitor the spreading and thinning behavior of the spin-coating liquid film in real time before or during the spin-coating of quantum dot materials, it indirectly and accurately identifies the microscopic residual layers and their distribution on the wafer surface that are difficult to detect with the naked eye. This strategy of early detection at the source of the problem effectively avoids subsequent local damage to the quantum dot film and thickness measurement deviations caused by residual layers, fundamentally solving the problem of information contamination at its starting point. It also provides the most original and accurate surface state information for subsequent identification of local micro-vibration signals and real-time risk zoning assessment of the wafer surface, ultimately guiding the control system to make refined regional grinding parameter adjustments.

[0020] In a second aspect, the present invention provides a wafer thinning equipment control device, comprising:

[0021] The monitoring and comparison module is used to monitor the spreading and thinning behavior of the spin-coated liquid film in real time during the spin-coating of quantum dot materials 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 the comparison results.

[0022] The first generation module is used to determine whether there is a microscopic residual layer or surface energy difference on the wafer surface based on the comparison result, and to generate a residual layer distribution map on the wafer surface; the residual layer distribution map on the wafer surface is used to indicate the areas where there is a microscopic residual layer or surface energy difference.

[0023] The thickness analysis module is used to analyze the fluorescence signal received by the quantum dot thickness sensor from the region indicated by the residual layer distribution map on the wafer surface, and obtain local wafer thickness information.

[0024] The receiving module is used to control the array laser unit to emit a laser beam modulated at a specific frequency to scan the wafer surface and then receive the reflected light signal generated by the scan.

[0025] The identification module is used to identify and extract micro-vibration signals related to local stress concentration by analyzing the reflected light signal;

[0026] The partitioning module is used to divide the wafer surface into multiple grid regions by performing a gridding process on the wafer surface;

[0027] The second generation module is used to assess the risk of crack generation in each grid region based on the residual layer distribution map on the wafer surface, the local wafer thickness information, and the micro-vibration signal, and generate a risk heat map of the wafer surface based on the risk assessment results.

[0028] The control module is used to control the operating parameters of the polishing head and the flow rate of the polishing slurry based on the risk thermal map of the wafer surface.

[0029] The wafer thinning equipment control device provided by this invention can effectively identify and quantify anomalies that are difficult to detect by traditional methods, such as microscopic residual layers on the wafer surface, damage to quantum dot films, and local stress concentration. This overcomes the predicament of information misjudgment in the prior art, realizes early warning and adaptive control of potential crack risks, and significantly improves the yield and reliability of ultrathin wafers.

[0030] Thirdly, the present invention provides an electronic device including a processor and a memory, the memory storing computer-readable instructions, which, when executed by the processor, perform the steps of the wafer thinning device control method provided in the first aspect above.

[0031] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, performs the steps of the wafer thinning device control method provided in the first aspect above.

[0032] As can be seen from the above, the wafer thinning equipment control method provided by this invention achieves early and accurate assessment of the risk of surface cracks by accurately sensing surface micro-anomalies in the early stages of the wafer thinning process and combining multi-source local information for correction and enhancement. This enables the control system to perform fine-grained regional grinding parameter adjustments, thereby fundamentally avoiding the generation of latent micro-cracks and significantly improving the yield and reliability of ultra-thin wafers.

[0033] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing embodiments of the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings. Attached Figure Description

[0034] Figure 1 This is a flowchart of a wafer thinning equipment control method provided in an embodiment of the present invention.

[0035] Figure 2 This is a schematic diagram of a wafer thinning equipment control device provided in an embodiment of the present invention.

[0036] Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention.

[0037] Label Explanation:

[0038] 100. Monitoring and 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 equipment; 1301. Processor; 1302. Memory; 1303. Communication bus. Detailed Implementation

[0039] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0040] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this invention, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0041] Reference Appendix Figure 1 This invention provides a method for controlling a wafer thinning device, comprising the following steps:

[0042] S1. During the spin coating of quantum dot materials on the wafer surface, the spreading and thinning behavior of the spin coating liquid film is monitored in real time using an optical interferometry measurement system. The monitored spreading and thinning behavior of the spin coating liquid film is compared with a preset standard liquid film behavior model to obtain the comparison results.

[0043] S2. Based on the comparison results, determine whether there is a microscopic residual layer or surface energy difference on the wafer surface, and generate a residual layer distribution map on the wafer surface; the residual layer distribution map on the wafer surface is used to indicate the areas where there is a microscopic residual layer or surface energy difference;

[0044] S3. Based on the region indicated by the residual layer distribution map on the wafer surface, analyze the fluorescence signal received by the quantum dot thickness sensor from that region, determine whether the integrity of the quantum dot film is damaged by the spectral characteristics of the fluorescence signal, and calibrate the original thickness measurement data based on the result of the quantum dot film damage determination to obtain local wafer thickness information;

[0045] S4. Control the array-type laser unit integrated below the grinding head to emit a laser beam modulated at a specific frequency to scan the wafer surface, and then receive the reflected light signal generated by the scan;

[0046] S5. By analyzing the phase noise spectral density or higher harmonic components of the reflected light signal, identify and extract the micro-vibration signals related to local stress concentration;

[0047] S6. By performing a gridding process on the wafer surface, the wafer surface is divided into multiple grid regions;

[0048] S7. Based on the residual layer distribution map on the wafer surface, local wafer thickness information, and micro-vibration signals, assess the risk of crack generation in each grid area, and generate a risk thermal map of the wafer surface based on the risk assessment results.

[0049] S8. Based on the risk thermal map of the wafer surface, control the operating parameters of the polishing head and the flow rate of the polishing slurry; specifically, this includes regional adjustments to the downward pressure, rotation speed, and local flow rate of the polishing head and the polishing slurry.

[0050] This application introduces a multi-dimensional, high-precision real-time monitoring and intelligent analysis mechanism, which can effectively identify and quantify anomalies that are difficult to detect by traditional methods, such as microscopic residual layers on the wafer surface, damage to quantum dot films, and local stress concentration. This overcomes the dilemma of information misjudgment in the prior art, realizes early warning and adaptive control of potential crack risks, and significantly improves the yield and reliability of ultrathin wafers.

[0051] This embodiment provides a control method for wafer thinning equipment, aiming to solve the problem of difficulty in effectively identifying and controlling the risk of cracks caused by microscopic residual layers, quantum dot film damage, and local stress concentration during the ultrathin wafer thinning process. This method integrates multi-source sensor data and performs intelligent analysis to achieve comprehensive perception of the wafer surface state and precise control of the thinning process.

[0052] The "optical interferometry system" is a precision measurement system that utilizes the principle of light wave interference. Its function is to monitor the thickness changes and spreading dynamics of spin-coated liquid films in real time and non-contactly. "Quantum dot material" is a nano-semiconductor material with special optical properties, which is spin-coated onto the wafer surface in this application as a thickness sensing medium. Its fluorescence characteristics are closely related to the film thickness. The "standard liquid film behavior model" refers to a theoretical or empirical model of liquid film spreading and thinning based on ideal spin-coating conditions and material properties, used to compare with actual monitoring data to detect anomalies. "Microscopic residual layer" refers to an extremely thin layer of organic or inorganic material invisible to the naked eye on the wafer surface, 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 different physical or chemical properties, which also affects material wetting and adhesion. The "quantum dot thickness sensor" is a device that uses the fluorescence characteristics of quantum dot materials to measure film thickness, inferring thickness information by analyzing the spectral characteristics of the fluorescence signal. An "array-type laser unit" is a system comprising multiple independently controllable laser emitters and receivers, capable of high-resolution scanning and signal acquisition of the wafer surface. "Phase noise spectral density" and "higher harmonic components" are characteristic parameters of the laser reflection signal, used to characterize the signal's stability, nonlinearity, and weak vibration information. "Micro-vibration signals" refer to extremely low-energy, high-frequency minute vibrations caused by localized stress concentration or abnormal contact. A "wafer surface risk heatmap" is a visualization tool that visually represents the crack risk level of different regions on the wafer surface through color depth or numerical value.

[0053] The core of the wafer thinning equipment control method in this embodiment lies in achieving accurate identification and adaptive control of potential crack risks during the wafer thinning process through multi-dimensional data acquisition and intelligent analysis.

[0054] In step S1, during the spin-coating of quantum dot material onto the wafer surface, the spreading and thinning behavior of the spin-coated liquid film is monitored in real time using an optical interferometry system. The monitored spreading and thinning behavior is compared with a preset standard liquid film behavior model to obtain the comparison results. Specifically, the optical interferometry system can employ a white light interferometer or a laser interferometer. For example, a white light interferometer emits broadband white light to illuminate the surface of the spin-coated liquid film and receives the reflected light. Since 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. Simultaneously, the system can also capture the dynamic changes at the edge of the liquid film using a high-speed camera, thereby monitoring its spreading behavior. These real-time monitored parameters, such as liquid film thickness, spreading speed, and uniformity, are compared with a preset standard liquid film behavior model. This model can be established based on extensive experimental data and fluid dynamics theory, describing the expected spreading and thinning trajectory of the liquid film under ideal conditions. The comparison results can be expressed as deviation values, abnormal area markers, etc.

[0055] In step S2, the system determines whether a microscopic residual layer or surface energy difference exists on the wafer surface based on the comparison results, and generates a residual layer distribution map on the wafer surface. This residual layer distribution map indicates areas where microscopic residual layers or surface energy differences exist. For example, if the comparison results show that the liquid film spreads significantly slower than the model prediction in certain areas, or that the liquid film thickness exhibits abnormal fluctuations or unevenness in local areas, it can be inferred that microscopic residual layers or surface energy differences may exist in these areas. These differences alter the wettability of the wafer surface, thereby affecting the spread of the liquid film. The system can generate a residual layer distribution map on the wafer surface based on the location and extent of these abnormal areas. This map can be presented as a two-dimensional image, where different colors or grayscale values ​​represent the severity of the residual layer or surface energy difference.

[0056] In step S3, based on the region indicated by the residual layer distribution map on the wafer surface, the fluorescence signal received by the quantum dot thickness sensor from that region is analyzed. The spectral characteristics of the fluorescence signal are used to determine whether the integrity of the quantum dot film is damaged. Based on the damage assessment result, the original thickness measurement data is calibrated to obtain local wafer thickness information. For example, in the region indicated by the residual layer distribution map, the quantum dot thickness sensor will receive a fluorescence signal. By analyzing the spectral characteristics of these fluorescence signals, such as fluorescence intensity, peak position, and full width at half maximum (FWHM), it can be determined whether the quantum dot film is damaged due to residual layer or surface energy differences, such as local detachment, degradation, or aggregation. If damage is determined, it indicates that the original thickness measurement data may be biased. In this case, the original data needs to be calibrated according to the degree of damage. The calibration method may include establishing a correction model that correlates the degree of quantum dot film damage with the thickness measurement error, thereby 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 a wafer thinning equipment control method, when the quantum dot film is damaged, the thickness analysis module will receive a fluorescence signal. The correction model can be pre-trained using experimental data to establish the relationship between the degree of damage to the quantum dot film (e.g., fluorescence intensity decay rate, peak shift) and the thickness measurement error. Assuming the fluorescence intensity decay rate is X, the degree of thermally 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, for example, f(X,Y) = aX + bY + cXY + d, where a, b, c, and d are model coefficients. Once 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.

[0057] In step S4, the array-type laser unit integrated below the grinding head emits a laser beam modulated at a specific frequency to scan the wafer surface, and then receives the reflected light signals generated by the scan. For example, the array-type laser unit can consist of multiple independent semiconductor lasers, each emitting a laser beam modulated at a specific frequency. These laser beams scan the wafer surface at a preset scanning path and speed. When the laser beam irradiates the wafer surface, reflected light is generated. The photodetector in the array-type laser unit receives these reflected light signals and converts them into electrical signals for subsequent processing.

[0058] In step S5, 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 signal. For example, the received reflected light signal may contain modulation information of micro-vibrations on the wafer surface. By performing spectral analysis on these signals, their phase noise spectral density or higher harmonic components can be calculated. Micro-vibrations caused by local stress concentration typically manifest 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 bandpass filtering, these micro-vibration signals related to local stress concentration can be identified and extracted from complex reflected light signals.

[0059] In step S6, the wafer surface is divided into multiple grid regions by performing a gridding process. For example, the entire wafer surface can be divided into a two-dimensional grid, with each grid region having a preset size, such as 1mm*1mm. This gridding process facilitates the localized management and analysis of various parameters on the wafer surface, providing spatial reference for subsequent risk assessment.

[0060] In step S7, the risk of crack formation in each grid region is assessed based on the residual layer distribution map on the wafer surface, local wafer thickness information, and micro-vibration signals. A risk heatmap of the wafer surface is generated based on the risk assessment results. For example, for each grid region, the severity of the residual layer in its corresponding residual layer distribution map, the thickness uniformity or outliers in the local wafer thickness information, and the stress concentration in the micro-vibration signals can be comprehensively considered. A multi-parameter risk assessment model can be established, which weights these input parameters and outputs a crack risk score. Specifically, the parametric risk assessment model can be a mathematical model based on weighted summation. This model receives three quantitative indices 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 model's calculation formula is: Crack Risk Score = (Residual Layer Influence Index × 0.3) + (Thickness Inhomogeneity Index × 0.4) + (Stress Concentration Intensity Index × 0.3). Each index ranges from 0 to 10, where 0 represents no risk and 10 represents high risk. For example, if a grid region has a residual layer influence index of 7, a thickness inhomogeneity index of 6, and a stress concentration intensity index of 8, then the crack risk score for that region is (7 × 0.3) + (6 × 0.4) + (8 × 0.3) = 2.1 + 2.4 + 2.4 = 6.9. This score is then used to generate a wafer surface risk heatmap. A higher risk score indicates a greater risk of crack formation in that region. Based on the risk scores of all grid regions, a wafer surface risk heatmap can be generated, where different colors or brightness levels represent different risk levels, visually displaying high-risk areas.

[0061] In step S8, the polishing head and polishing slurry are controlled according to the risk thermal map of the wafer surface. Specifically, this includes regional adjustments to the downward pressure and rotation speed of the polishing head, as well as the local flow rate of the polishing slurry. For example, for high-risk areas shown in the risk thermal map, the control system can instruct the polishing head to reduce the downward pressure and rotation speed in that area, or increase the local flow rate of the polishing slurry in that area through a microfluidic device, thereby reducing polishing intensity, reducing stress concentration, and effectively preventing crack formation. For low-risk areas, a higher polishing efficiency can be maintained. This regional adaptive adjustment enables fine-grained control of the thinning process.

[0062] The wafer thinning equipment control method in this embodiment introduces a multi-dimensional, high-precision real-time monitoring and intelligent analysis mechanism, which can effectively identify and quantify anomalies that are difficult to detect by traditional methods, such as microscopic residual layers on the wafer surface, quantum dot film damage, and local stress concentration. This overcomes the dilemma of information misjudgment in the prior art, realizes early warning and adaptive control of potential crack risks, and significantly improves the yield and reliability of ultrathin wafers.

[0063] Specifically, this application achieves a breakthrough over existing technologies through the following core innovations:

[0064] First, during the spin-coating of quantum dot materials, this application utilizes an optical interferometry system to monitor the spreading and thinning behavior of the spin-coated liquid film in real time and compares it with a pre-defined standard liquid film behavior model. This step enables the early detection of microscopic residual layers or surface energy differences on the wafer surface, differences that are difficult to detect using traditional methods but have a crucial impact on the uniformity and adhesion of subsequent quantum dot films. Compared to existing technologies that rely solely on quantum dot thickness sensors for thickness measurement, this application intervenes in the surface state assessment at an earlier stage of the process, thus laying the foundation for subsequent accurate measurements and risk assessments.

[0065] Secondly, this application analyzes the fluorescence signal received by the quantum dot thickness sensor based on the region indicated by the residual layer distribution map on the wafer surface, and determines whether the integrity of the quantum dot film is damaged by the spectral characteristics of the fluorescence signal, thereby calibrating the original thickness measurement data. This innovation directly solves the problem of thickness measurement data distortion caused by quantum dot film damage in existing technologies. Traditional methods often misjudge signal anomalies caused by quantum dot film damage as thickness fluctuations or background noise, leading to a decrease in local thickness measurement accuracy. This application, through in-depth analysis of spectral characteristics, can distinguish between film damage and actual thickness changes, thereby obtaining more accurate local wafer thickness information and avoiding the dilemma of "information misjudgment".

[0066] Furthermore, this application controls an array-type laser unit integrated below the grinding head to emit a laser beam modulated at a specific frequency to scan the wafer surface. By analyzing the phase noise spectral density or higher harmonic components of the reflected light signal, it identifies and extracts micro-vibration signals related to local stress concentration. This technique overcomes the limitation of traditional laser Doppler vibrators in capturing high-frequency, low-energy local micro-vibration signals. Through refined analysis of the reflected light signal, this application can identify weak signals that predict local stress concentration and potential surface damage, providing crucial early warning information for crack risk assessment and compensating for the shortcomings of existing technologies in local stress monitoring.

[0067] Finally, this application integrates the residual layer distribution map on the wafer surface, local wafer thickness information, and micro-vibration signals to assess the risk of crack formation in each grid region and generate a risk heat map of the wafer surface. Based on this heat map, the downward pressure, rotation speed, and local flow rate of the polishing head are adjusted regionally. This multi-source information fusion risk assessment mechanism, combined with regional adaptive control, allows for precise control of the thinning process based on the actual local conditions of the wafer surface. Compared to existing methods that rely on macroscopic data for overall adjustment, this approach more effectively prevents the formation of micro-cracks, significantly improving the yield and reliability of ultra-thin wafers.

[0068] In summary, this application introduces a multi-dimensional, high-precision real-time monitoring and intelligent analysis mechanism into the wafer thinning equipment control method. This mechanism can effectively identify and quantify anomalies that are difficult to detect by traditional methods, such as microscopic residual layers on the wafer surface, quantum dot film damage, and local stress concentration. This overcomes the predicament of information misjudgment in the prior art, realizes early warning and adaptive control of potential crack risks, and significantly improves the yield and reliability of ultrathin wafers.

[0069] In some embodiments, the step of calibrating the original thickness measurement data based on the quantum dot film damage assessment result to obtain local wafer thickness information includes:

[0070] S31. Obtain local instantaneous temperature information of the grinding interface;

[0071] S32. Obtain multi-dimensional quantum dot fluorescence characteristics;

[0072] S33. Based on local instantaneous temperature information, multi-dimensional quantum dot fluorescence characteristics, and quantum dot film damage assessment results, decouple and quantify the degree of influence of residual layer damage and thermally induced degradation on the state of quantum dot films;

[0073] S34. Based on the quantified impact of residual layer damage and thermally induced degradation, adjust the correction parameters of the original thickness measurement data, and use the correction parameters to calibrate the original thickness measurement data to obtain local wafer thickness information.

[0074] Specifically, in step S31, the instantaneous local temperature information of the grinding interface can be acquired in real time by integrating miniature temperature sensors near the grinding head or wafer surface. These sensors can be non-contact infrared sensors or contact thermocouples or resistance temperature detectors, with the aim of accurately capturing temperature fluctuations in local areas during the grinding process.

[0075] In step S32, the multidimensional quantum dot fluorescence characteristics can be understood as including spectral information in multiple dimensions, such as fluorescence lifetime, fluorescence peak position shift, and fluorescence half-width variation, in addition to fluorescence intensity. These characteristics can be acquired using a high-sensitivity spectrometer or a time-resolved fluorescence detection system, with the aim of providing richer and more detailed information on the state of the quantum dot film to more accurately reflect its damage status.

[0076] In practical applications, step S33, decoupling and quantifying the impact of residual layer damage and thermally induced degradation on the state of quantum dot films, is specifically achieved by establishing a multi-factor coupling model. For example, machine learning algorithms can be used, taking local instantaneous temperature information, multi-dimensional quantum dot fluorescence characteristics, and quantum dot film damage assessment results as inputs. By training a pre-defined damage model, the attenuation of fluorescence signals caused by residual layers and fluorescence quenching or spectral changes caused by thermal induction can be distinguished and quantified. The aim is to accurately identify and separate the effects of different damage mechanisms on the performance of quantum dot films.

[0077] Specifically, multi-factor coupling models can be implemented using machine learning algorithms, such as Support Vector Machines (SVM), neural networks, or decision tree ensemble models. First, a training dataset containing various damage scenarios needs to be constructed. Each sample includes local instantaneous temperature information, multi-dimensional quantum dot fluorescence characteristics (such as fluorescence intensity, peak position, full width at half maximum, lifetime, etc.), and the corresponding true values ​​of residual layer damage and thermally induced degradation. For example, by controlling experimental conditions, quantum dot film samples with only residual layer damage, only thermally induced degradation, and both can be fabricated, and their damage levels can be measured using high-precision analysis equipment. Then, this data is used to train the machine learning model, enabling it to learn and recognize the mapping relationship between different combinations of input features and different damage types and degrees. After training, when new real-time data is input into the model, it can output quantitative values ​​of the influence of residual layer damage and thermally induced degradation on the state of the quantum dot film.

[0078] Further, in step S34, the correction parameters of the original thickness measurement data are adjusted based on the quantified impact of residual layer damage and thermally induced degradation. This is achieved either by consulting a preset correction parameter database or by calculating a correction function in real time. For example, when the impact of residual layer damage is high, the correction for deviations caused by signal attenuation in the original thickness measurement data can be increased; when the impact of thermally induced degradation is high, the temperature-related fluorescence efficiency correction parameter can be adjusted. The purpose is to ensure that the calibration process can provide fine-grained compensation for different types of damage, thereby improving the accuracy of thickness measurement.

[0079] This application's solution addresses the problem of traditional methods failing to adequately consider the coupling effects of multiple complex damage mechanisms when calibrating original thickness measurement data. This is achieved by introducing local instantaneous temperature information at the grinding interface and multi-dimensional quantum dot fluorescence characteristics, and further decoupling and quantifying the impact of residual layer damage and thermally induced degradation on the quantum dot film state. The ability to accurately distinguish and quantify the specific effects of different damage sources (such as residual layers and thermal effects) on the quantum dot film state makes subsequent correction parameter adjustments more targeted. This approach avoids calibration deviations caused by single damage assessment results, ensuring more accurate and reliable acquisition of local wafer thickness information under complex grinding environments.

[0080] The above technical solution significantly improves the measurement accuracy and reliability of local wafer thickness information. By precisely decoupling and quantifying the impact of residual layer damage and thermally induced degradation on the quantum dot film state, the calibration process for the original thickness measurement data becomes more targeted and accurate. This high-precision thickness information is crucial for subsequent wafer surface crack risk assessment, effectively reducing the risk of misjudgments due to inaccurate thickness measurements. This provides a more refined and safer control basis for the wafer thinning process, ultimately improving the yield and product quality of wafer thinning.

[0081] In some embodiments, the step of identifying and extracting micro-vibration signals related to local stress concentration by analyzing the phase noise spectral density or higher-order harmonic components of the reflected light signal includes:

[0082] S51. The phase noise spectral density or higher harmonic components of the reflected light signal at different positions on the wafer surface are simultaneously acquired by multiple receiving points of an array laser unit to obtain a multi-point reflection signal.

[0083] S52. Perform spatiotemporal synchronization analysis on the collected multi-point reflection signals to obtain the spatiotemporal synchronization analysis results;

[0084] S53. Based on the spatiotemporal synchronization analysis results, determine the origin location of the micro-vibration signal and identify the spatial propagation mode related to local stress concentration;

[0085] S54. Based on the origin location and spatial propagation mode of the micro-vibration signal, by distinguishing between micro-vibrations caused by wafer surface damage and background noise, identify and extract micro-vibration signals related to local stress concentration.

[0086] Specifically, in step S51, multiple receiving points of the array-type 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 ring to cover a large area of ​​the wafer. Synchronous acquisition refers to all receiving points simultaneously recording the phase noise spectral density or higher-order harmonic components of the received reflected light signals within the same time window. Phase noise spectral density can characterize the frequency stability of the signal, while higher-order harmonic components can reflect nonlinear effects; both can serve as sensitive indicators of micro-vibration signals. Through multi-point synchronous acquisition, micro-vibration information at different locations on the wafer surface can be obtained, forming a multi-point reflected signal dataset.

[0087] Further, in step S52, spatiotemporal synchronization analysis is performed on the acquired multi-point reflection signals. Spatiotemporal synchronization analysis refers to correlating and comparing signals acquired by different receiving points at different time points. This can be achieved using signal processing techniques such as correlation analysis, Fourier transform, and wavelet analysis to reveal the propagation characteristics of the signal in space and time. For example, by calculating the phase difference and amplitude attenuation between signals from different receiving points, the propagation direction and velocity of the micro-vibration wave can be inferred.

[0088] Therefore, in step S53, based on the spatiotemporal synchronization analysis results, the origin location of the micro-vibration signal can be determined, and the spatial propagation mode related to local stress concentration can be identified. The origin location of the micro-vibration signal usually corresponds to the area of ​​stress concentration or potential damage. The spatial propagation mode refers to the way the micro-vibration wave diffuses on the wafer surface, such as point source diffusion, line source diffusion, or a more complex mode. These modes 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. This FEA model can include a variety of typical stress concentration modes, such as edge cracks, internal defects, surface scratches, etc., and predict the propagation characteristics of micro-vibration waves under these modes. When the system identifies that the actual micro-vibration signal originates from the wafer edge and exhibits a specific spatial propagation mode that diffuses along the edge to both sides, this mode can be matched with the preset 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 edge stress concentration.

[0089] Finally, in step S54, based on the origin location and spatial propagation mode of the micro-vibration signal, the micro-vibration signal related to local stress concentration is identified and extracted by distinguishing between micro-vibrations caused by wafer surface damage and background noise. Background noise may originate from equipment vibration, environmental interference, or other minor disturbances not caused by stress concentration. By analyzing the characteristics of the signal (such as frequency, amplitude, and duration) and its spatial propagation behavior, the micro-vibration signal caused by true stress concentration can be effectively separated from the noise. For example, micro-vibrations caused by damage typically have a specific frequency range and propagation path, while background noise may exhibit a more random or uniform distribution.

[0090] This application's solution effectively addresses the accuracy limitations and noise interference issues that may exist in traditional methods when identifying micro-vibration signals related to local stress concentrations by introducing multi-point synchronous acquisition, spatiotemporal synchronous analysis, and the identification of origin location and spatial propagation mode. Specifically, by synchronously acquiring reflected light signals from multiple receiving points of an array-type laser unit, more comprehensive micro-vibration information of the wafer surface can be provided, avoiding the local information loss that may result from single-point measurements. Through spatiotemporal synchronous analysis of these multi-point signals, the propagation path and origin of micro-vibration waves can be accurately tracked, thereby distinguishing true stress concentration areas from non-stress concentration areas. Furthermore, by determining the origin location and spatial propagation mode of the micro-vibration signal, the system can effectively distinguish micro-vibrations caused by wafer surface damage from random background noise based on these unique physical characteristics, thus ensuring that the extracted micro-vibration signals more accurately reflect the local stress concentration state inside the wafer.

[0091] The above technical solutions significantly improve the accuracy and reliability of identifying and extracting micro-vibration signals related to local stress concentration. The combination of multi-point synchronous acquisition and spatiotemporal synchronous analysis allows for precise determination of the origin and propagation mode of the micro-vibration signals, effectively eliminating background noise and interference from non-stress concentration sources. Consequently, the obtained micro-vibration signals more accurately reflect the local stress concentration within the wafer, providing more precise data support for subsequent crack risk assessment, and ultimately improving the safety and yield of the wafer thinning process.

[0092] In some embodiments, step S7, which assesses the risk of crack formation in each grid region based on the residual layer distribution map on the wafer surface, local wafer thickness information, and micro-vibration signals, includes:

[0093] S71. Obtain the distribution map of residual layer on the wafer surface, local wafer thickness information, and quantitative indicators corresponding to micro-vibration signals;

[0094] S72. Assign weights to each quantitative indicator and perform comprehensive calculations to obtain a crack risk score as the risk assessment result for cracks to occur in the corresponding grid area.

[0095] The acquisition of the wafer surface residual layer distribution map, the local wafer thickness information, and the corresponding quantitative indicators of the micro-vibration signal involves converting raw data from different sources—such as residual layer thickness or density in the wafer surface residual layer distribution map, thickness deviation or gradient in the local wafer thickness information, and amplitude or frequency characteristics in the micro-vibration signal—into a unified quantitative form that allows for numerical comparison and calculation. Specifically, the wafer surface residual layer distribution map can be quantified as residual layer coverage or average thickness; 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 raw information to facilitate processing by subsequent risk assessment algorithms.

[0096] Furthermore, weights are assigned to each of the aforementioned quantitative indicators and a comprehensive calculation is performed to obtain a crack risk score, which serves as the risk assessment result for crack formation in the corresponding grid region. This means that based on the degree of influence of different indicators on crack formation, a weight coefficient is assigned to each quantitative indicator, and these weighted indicator values ​​are comprehensively calculated through weighted summation or other multi-factor decision models to obtain a single numerical value representing the crack risk of the grid region, i.e., the crack risk score. For example, based on historical data or expert experience, the different contributions of residual layer, thickness inhomogeneity, and stress concentration to crack formation can be determined, and corresponding weights can be set accordingly. The purpose is to provide a comprehensive and quantitative risk assessment result to guide subsequent grinding parameter adjustments.

[0097] This application's solution transforms three key data types—residual layer distribution maps on the wafer surface, local wafer thickness information, and micro-vibration signals—into quantifiable indicators, and then performs weighted comprehensive calculations on these indicators, thereby achieving a systematic assessment of crack risk for each grid region. Specifically, step S71 ensures that data from different sources and in different formats can be processed and compared uniformly, laying a data foundation for subsequent risk assessment. It is precisely because these key factors are quantified that they can be incorporated into a unified assessment framework. Based on this, step S72 assigns weights to different indicators, reflecting their relative importance in inducing cracks. For example, residual layers in certain regions may contribute more to crack induction than local thickness inhomogeneities; weight adjustments can more accurately reflect this difference. Through comprehensive calculations, the one-sidedness of single-indicator assessments is avoided, enabling crack risk scores to more comprehensively and accurately reflect the true risk status of each region on the wafer surface.

[0098] Through the above technical solution, this application provides a more refined and accurate method for assessing wafer thinning crack risk. Specifically, by comprehensively analyzing multi-dimensional quantitative indicators, misjudgments or omissions caused by assessments based on a single factor can be effectively avoided, significantly improving the accuracy and reliability of crack risk prediction. Therefore, the obtained crack risk score can more realistically reflect the potential crack risk in each grid region of the wafer surface, providing a more scientific and accurate basis for subsequent regional adjustments of the polishing head and polishing slurry. This effectively reduces the probability of cracks forming on the wafer during the thinning process, improving wafer yield and production efficiency.

[0099] Please refer to Figure 2 , Figure 2 This invention provides a wafer thinning equipment control device in some embodiments, which is integrated into a back-end control device in the form of a computer program, comprising:

[0100] 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 materials 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 the comparison results.

[0101] The first generation module 200 is used to determine whether there is a microscopic residual layer or surface energy difference on the wafer surface based on the comparison results, and to generate a residual layer distribution map on the wafer surface; the residual layer distribution map on the wafer surface is used to indicate the areas where there is a microscopic residual layer or surface energy difference.

[0102] The thickness analysis module 300 is used to analyze the fluorescence signal received by the quantum dot thickness sensor from the region indicated by the residual layer distribution map on the wafer surface to obtain local wafer thickness information.

[0103] The receiving module 400 is used to control the array laser unit to emit a laser beam modulated by a specific frequency to scan the wafer surface and then receive the reflected light signal generated by the scan.

[0104] The identification module 500 is used to identify and extract micro-vibration signals related to local stress concentration by analyzing the reflected light signal;

[0105] The partitioning module 600 is used to divide the wafer surface into multiple grid regions by performing gridding processing on the wafer surface;

[0106] The second generation module 700 is used to assess the risk of cracking in each grid area based on the residual layer distribution map on the wafer surface, local wafer thickness information, and micro-vibration signals, and generate a risk heat map of the wafer surface based on the risk assessment results.

[0107] The control module 800 is used to control the operating parameters of the polishing head and the flow rate of the polishing slurry based on the risk thermal map of the wafer surface.

[0108] In some embodiments, the thickness analysis module 300 performs the following steps when analyzing the fluorescence signal received by the quantum dot thickness sensor from the region indicated by the residual layer distribution map on the wafer surface to obtain local wafer thickness information:

[0109] The integrity of the quantum dot film is determined based on the spectral characteristics of the fluorescence signal, and the result of the quantum dot film damage assessment is obtained.

[0110] Based on the assessment of quantum dot film damage, the original thickness measurement data is calibrated to obtain local wafer thickness information.

[0111] In some embodiments, the thickness analysis module 300 performs the following when calibrating the original thickness measurement data based on the quantum dot film damage assessment result to obtain local wafer thickness information:

[0112] S31. Obtain local instantaneous temperature information of the grinding interface;

[0113] S32. Obtain multi-dimensional quantum dot fluorescence characteristics;

[0114] S33. Based on local instantaneous temperature information, multi-dimensional quantum dot fluorescence characteristics, and quantum dot film damage assessment results, decouple and quantify the degree of influence of residual layer damage and thermally induced degradation on the state of quantum dot films;

[0115] S34. Based on the quantified impact of residual layer damage and thermally induced degradation, adjust the correction parameters of the original thickness measurement data, and use the correction parameters to calibrate the original thickness measurement data to obtain local wafer thickness information.

[0116] In some embodiments, the identification module 500 performs the following actions when identifying and extracting micro-vibration signals related to local stress concentration by analyzing the reflected light signal:

[0117] By analyzing the phase noise spectral density or higher harmonic components of the reflected light signal, micro-vibration signals related to local stress concentration can be identified and extracted.

[0118] In some embodiments, the identification module 500 performs this function when identifying and extracting micro-vibration signals associated with local stress concentration by analyzing the phase noise spectral density or higher-order harmonic components of the reflected light signal:

[0119] S51. The phase noise spectral density or higher harmonic components of the reflected light signal at different positions on the wafer surface are simultaneously acquired by multiple receiving points of an array laser unit to obtain a multi-point reflection signal.

[0120] S52. Perform spatiotemporal synchronization analysis on the collected multi-point reflection signals to obtain the spatiotemporal synchronization analysis results;

[0121] S53. Based on the spatiotemporal synchronization analysis results, determine the origin location of the micro-vibration signal and identify the spatial propagation mode related to local stress concentration;

[0122] S54. Based on the origin location and spatial propagation mode of the micro-vibration signal, by distinguishing between micro-vibrations caused by wafer surface damage and background noise, identify and extract micro-vibration signals related to local stress concentration.

[0123] In some embodiments, the second generation module 700 is executed when assessing the risk of crack formation in each grid region based on the residual layer distribution map on the wafer surface, local wafer thickness information, and micro-vibration signals:

[0124] S71. Obtain the distribution map of residual layer on the wafer surface, local wafer thickness information, and the quantization index corresponding to the micro-vibration signal; the quantization index is the digital representation of the distribution map of residual layer on the wafer surface, local wafer thickness information, and micro-vibration signal;

[0125] S72. Assign weights to each quantitative indicator and perform comprehensive calculations to obtain a crack risk score as the risk assessment result for cracks to occur in the corresponding grid area.

[0126] In some embodiments, the control module 800 performs the following when controlling the polishing head and polishing slurry based on the wafer surface risk thermal map:

[0127] S81. Based on the risk thermal map of the wafer surface, the downward pressure, rotation speed, and local flow rate of the polishing head are adjusted regionally.

[0128] Please refer to Figure 3 , Figure 3This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. The present invention provides an electronic device 13, including: a processor 1301 and a memory 1302. The processor 1301 and the memory 1302 are interconnected and communicate with each other via a communication bus 1303 and / or other forms of connection mechanisms (not shown). The memory 1302 stores computer-readable instructions executable by the processor 1301. When the electronic device is running, the processor 1301 executes the computer-readable instructions to execute the wafer thinning device control method in any optional implementation of the above embodiments, to achieve the following functions: during the spin coating of quantum dot material on the wafer surface, the spreading and thinning behavior of the spin-coated liquid film is monitored in real time, and the monitored spreading and thinning behavior of the spin-coated liquid film is compared with a preset standard liquid film behavior model to obtain a comparison result; based on the comparison result, it is determined whether there is a microscopic residual layer or surface layer on the wafer surface. The surface energy difference is analyzed to generate a residual layer distribution map on the wafer surface. This map indicates areas with microscopic residual layers or surface energy differences. Based on the areas indicated by the residual layer distribution map, the fluorescence signal received by the quantum dot thickness sensor from those areas is analyzed to obtain local wafer thickness information. An array-type laser unit emits a laser beam modulated at a specific frequency to scan the wafer surface, and the reflected light signal generated by the scan is received. By analyzing the reflected light signal, micro-vibration signals related to local stress concentration are identified and extracted. The wafer surface is divided into multiple grid regions through gridding. Based on the residual layer distribution map, local wafer thickness information, and micro-vibration signals, the risk of crack formation in each grid region is assessed, and a wafer surface risk heat map is generated based on the risk assessment results. The operating parameters of the polishing head and the flow rate of the polishing slurry are controlled according to the wafer surface risk heat map.

[0129] This invention provides a computer-readable storage medium storing a computer program. When executed by a processor, the computer program performs a wafer thinning device control method in any optional implementation of the above embodiments to achieve the following functions: During the spin-coating of quantum dot material onto a wafer surface, the spreading and thinning behavior of the spin-coated liquid film is monitored in real time, and the monitored spreading and thinning behavior is compared with a preset standard liquid film behavior model to obtain a comparison result; based on the comparison result, it is determined whether there is a microscopic 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 areas where there is a microscopic residual layer or surface energy difference; based on the wafer surface residual layer distribution map... The system analyzes the fluorescence signal received by the quantum dot thickness sensor from the indicated area to obtain local wafer thickness information; it controls the array laser unit to emit a laser beam modulated at a specific frequency to scan the wafer surface and receives the reflected light signal generated by the scan; by analyzing the reflected light signal, it identifies and extracts micro-vibration signals related to local stress concentration; it divides the wafer surface into multiple grid regions by meshing; based on the wafer surface residual layer distribution map, local wafer thickness information, and micro-vibration signals, it assesses the risk of cracking in each grid region and generates a wafer surface risk heat map based on the risk assessment results; and it controls the operating parameters of the polishing head and the flow rate of the polishing slurry based on the wafer surface risk heat map.

[0130] 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 Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0131] In the embodiments provided by this invention, it should be understood that the disclosed apparatus and method can be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. Furthermore, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Additionally, the displayed or discussed mutual couplings, direct couplings, or communication connections may be through some communication interfaces; indirect couplings or communication connections between devices or units may be electrical, mechanical, or other forms.

[0132] Furthermore, the units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0133] Furthermore, the functional modules in the various embodiments of the present invention 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.

[0134] In this document, relational terms such as first and second are used only to distinguish one entity or operation from another entity or operation, without necessarily requiring or implying any such actual relationship or order between these entities or operations.

[0135] The above description is merely an embodiment of the present invention and is not intended to limit the scope of protection of the present invention. For those skilled in the art, the present invention can have various modifications and variations. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

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