A semiconductor aging test method and system

Abstract

This invention discloses a semiconductor aging test method and system, relating to the fields of artificial intelligence and automated testing. The method includes: collecting semiconductor performance parameters and preprocessing them; building a quantum platform; deploying a nanosensor network to monitor the semiconductor and obtain semiconductor monitoring data; constructing an aging prediction model; determining the aging degree of the semiconductor based on the collected performance parameters; parsing the semiconductor monitoring data to obtain semiconductor state information; and combining the aging degree and state information to form a complete semiconductor test result. Furthermore, a short-time Fourier transform is used to generate a time-frequency graph of the semiconductor monitoring data in the time-frequency domain. This method captures the trend of microstructural changes, which helps in the early detection of potential problems.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence and automated testing, and in particular to a semiconductor aging test method and system. Background Technology

[0002] Aging testing of semiconductor devices is a crucial step in ensuring their long-term reliability, especially in modern electronic devices where the performance and lifespan of semiconductor devices directly impact the overall stability and safety of the equipment. In recent years, with the rapid development of semiconductor technology, the requirements for aging testing methods have also become increasingly stringent.

[0003] Commonly used semiconductor aging testing methods mainly include electrical parameter testing and thermal analysis. Electrical parameter testing can only provide macroscopic information and is difficult to reveal changes in microstructure. Thermal analysis indirectly infers the stress distribution inside the device by monitoring temperature distribution, but this method has low sensitivity for detecting early defects and cannot monitor in real time. In addition, most existing machine learning methods are limited to classical computing platforms and fail to fully utilize the powerful processing capabilities of quantum computing. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a semiconductor aging test method that solves the problems of traditional semiconductor aging test methods lacking microstructure changes and being insufficiently comprehensive.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] In a first aspect, the present invention provides a semiconductor aging test method, comprising,

[0008] Collect semiconductor performance parameters and perform preprocessing;

[0009] A quantum platform was built, and a network of nanosensors was deployed to monitor semiconductors and obtain semiconductor monitoring data.

[0010] An aging prediction model is constructed to determine the degree of semiconductor aging based on the collected performance parameters;

[0011] Analyze semiconductor monitoring data to obtain semiconductor status information;

[0012] By combining the aging degree and condition information of semiconductors, a complete semiconductor test result is formed;

[0013] Based on semiconductor test results, maintenance recommendations and visual icons are provided.

[0014] In a preferred embodiment of the semiconductor aging test method described in this invention, the semiconductor performance parameters are collected and preprocessed, specifically including the following steps.

[0015] Real-time acquisition of semiconductor operating voltage, current, temperature, and humidity as performance parameters;

[0016] The collected performance parameters are cleaned, converted in format, outliers are removed, and missing values ​​are filled.

[0017] The cleaned data is then standardized.

[0018] In a preferred embodiment of the semiconductor aging test method described in this invention, the following steps are included: constructing a quantum platform and deploying a nanosensor network to monitor the semiconductor and obtain semiconductor monitoring data.

[0019] IBM Quantum was chosen as the quantum cloud service platform.

[0020] Configure the hardware interface between the classical computer and the quantum processor, and write an initialization script to set the initial state vector;

[0021] Nanoscale sensor arrays are arranged at key locations in semiconductors to form a densely distributed sensing network;

[0022] Nanoscale sensor networks will monitor the state of semiconductors in real time and obtain monitoring data.

[0023] As a preferred embodiment of the semiconductor aging test method of the present invention, the following steps are included: constructing an aging prediction model and obtaining the degree of semiconductor aging based on the collected performance parameters.

[0024] Define a feature extraction function that captures hidden patterns in the acquired performance parameters through integral transform; its expression is:

[0025]

[0026] Where F(x) represents the result of capturing the latent pattern, x represents the performance parameter vector, α represents the regularization coefficient, t represents time, n represents the number of performance parameters, and x' i Let dt represent the i-th standardized performance parameter, and dt represent the differential sign.

[0027] Mapping the results of hidden modes to quantum states;

[0028] Based on the results of the hidden modes mapped to quantum states, the results of the hidden modes of multiple quantum states are introduced. The inner product similarity of the results of any two quantum state hidden modes in the quantum state space is calculated using the quantum kernel function. The inner product similarity between the results of all quantum state hidden modes is used to form a quantum kernel matrix.

[0029] Define an information filtering function to extract feature similarity information related to aging features from the quantum kernel matrix, and emphasize highly similar feature pairs. Its expression is:

[0030]

[0031] Where G(K) represents the feature similarity information related to aging characteristics extracted from the quantum kernel matrix K, and K represents the quantum kernel matrix. ij The inner product similarity of the results of the i-th and j-th quantum state hidden modes in the quantum state space is represented by β, which represents the temperature parameter K. ii K represents the similarity between the result of the hidden mode of the i-th quantum state and itself. jj The result representing the hidden mode of the j-th quantum state is similar to itself;

[0032] Feature similarity information related to aging characteristics is extracted from the quantum kernel matrix and combined into a feature similarity information vector, which is then input into the fully connected layer. The softmax function is applied to obtain the probability of the semiconductor being in different aging states, and its expression is as follows:

[0033]

[0034] Where, P(z) a Let G(K) represent the probability that a semiconductor is in the a-th aging state based on the feature similarity information vector z. a This represents the feature similarity information related to the a-th aging state extracted from the feature similarity information related to aging characteristics from the quantum kernel matrix K, where L represents the total number of aging states and l represents the index variable;

[0035] The degree of semiconductor aging is obtained based on the probability of the semiconductor being in different aging states.

[0036] In a preferred embodiment of the semiconductor aging test method described in this invention, the process of parsing semiconductor monitoring data to obtain semiconductor state information specifically includes the following steps.

[0037] Use short-time Fourier transform to generate a time-frequency plot of semiconductor monitoring data in the time-frequency domain;

[0038] Extract the maximum peak frequency and average power spectral density from the time-frequency plot and combine them into a microscopic feature vector;

[0039] Based on microscopic feature vectors, a convolutional neural network is introduced;

[0040] By using supervised learning and cross-entropy as the loss function, a convolutional neural network is trained to obtain the state information of the semiconductor.

[0041] The state information of the semiconductor includes changes in the semiconductor's microstructure and the initiation of defects.

[0042] As a preferred embodiment of the semiconductor aging test method of the present invention, the method combines the aging degree and state information of the semiconductor to form a complete semiconductor test result, specifically including the following steps.

[0043] Define a comprehensive evaluation function that combines the aging degree and condition information of the semiconductor to form a comprehensive semiconductor test score. Its expression is:

[0044] Z = λ1 × A(P(z)) a )+λ2×Y;

[0045] Where Z represents the overall semiconductor test score, λ1 represents the weight of the semiconductor aging degree, and A(P(z)) a ) represents the aging degree of the semiconductor, λ2 represents the weight of the semiconductor state information, and Y represents the semiconductor state information;

[0046] Test thresholds are set based on historical semiconductor data, and complete semiconductor test results are obtained based on the comprehensive semiconductor test score falling within the range of the test thresholds.

[0047] In a preferred embodiment of the semiconductor aging test method described in this invention, maintenance suggestions and visual icons are provided based on the semiconductor test results, specifically including the following steps.

[0048] Based on the semiconductor test results, specific maintenance recommendations are provided;

[0049] Using Python's Matplotlib, visualize the aging and condition information of semiconductors.

[0050] The results of each test are recorded in the database.

[0051] Secondly, the present invention provides a semiconductor aging test system, comprising,

[0052] The preprocessing module collects the performance parameters of the semiconductor and performs preprocessing.

[0053] The monitoring module establishes a quantum platform and deploys a network of nanosensors to monitor semiconductors and obtain semiconductor monitoring data.

[0054] The prediction module constructs an aging prediction model and obtains the aging trend of semiconductors based on the collected performance parameters.

[0055] The parsing module analyzes semiconductor monitoring data to obtain semiconductor status information;

[0056] The testing module combines the aging trend and status information of semiconductors to form complete semiconductor test results;

[0057] The maintenance module provides maintenance suggestions and visual icons based on semiconductor test results.

[0058] Thirdly, the present invention provides a computer device including a memory and a processor, wherein the memory stores a computer program, wherein the computer program, when executed by the processor, implements any step of the semiconductor aging test method as described in the first aspect of the present invention.

[0059] 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, implements any step of the semiconductor aging test method as described in the first aspect of the present invention.

[0060] The beneficial effects of this invention are as follows: An aging prediction model is established, and by defining a feature extraction function and utilizing integral transform to capture hidden patterns in the collected data, a deep mining of semiconductor aging characteristics is achieved. This method reveals hidden patterns that are difficult to capture using traditional methods, improving the accuracy of aging prediction, and is particularly significant for identifying early signs of aging. Furthermore, the use of short-time Fourier transform to generate time-frequency maps of semiconductor monitoring data in the time-frequency domain captures the trend of microstructural changes, helping to detect potential problems early and reducing the risk of failure, especially excelling in detecting early defect initiation. Attached Figure Description

[0061] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0062] Figure 1 This is a flowchart of the semiconductor aging test method in Example 1.

[0063] Figure 2 This is a schematic diagram of the test results generated in Example 1. Detailed Implementation

[0064] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0065] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0066] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0067] Example 1, referring to Figure 1 and Figure 2 This is the first embodiment of the present invention, which provides a semiconductor aging test method, including the following steps:

[0068] S1. Collect the performance parameters of the semiconductor and perform preprocessing.

[0069] Specifically, the steps include the following:

[0070] Install smart sensors with IIoT capabilities on the production line. These sensors will monitor parameters such as the operating voltage, current, temperature, and humidity of semiconductor devices as performance parameters.

[0071] Data acquisition is performed using ADAS, a unified automated data acquisition platform that collects sensor performance parameters in real time through the IIoT platform, which stands for Industrial Internet of Things.

[0072] Start ADAS, set a fixed time interval (e.g., once per minute), and use a Network Time Protocol (NTP) server to synchronize the time of all data acquisition nodes.

[0073] S1.1 Clean and format-convert the collected performance parameters. Remove outliers (such as measurement results outside the reasonable range) and fill in missing values ​​(e.g., using interpolation). Then, apply the Z-score standardization formula to unify the format of the cleaned data for easier subsequent analysis.

[0074] This further illustrates that it lays a solid foundation for building an efficient aging prediction model in the future.

[0075] S2. Build a quantum platform and deploy a network of nanosensors to monitor semiconductors and obtain semiconductor monitoring data.

[0076] Specifically, the steps include the following:

[0077] By selecting IBM Quantum as its quantum cloud service platform and leveraging its advanced quantum computing resources and technical support, classical computers can seamlessly access quantum processors through API interfaces.

[0078] Configure a communication interface between the classical computer and the quantum processor to ensure efficient data transfer between them.

[0079] Write a Python script to initialize the quantum circuit. This script will set the initial state vector (representing a uniform superposition state) and define a series of commonly used quantum gate operations, such as the Hadamard gate (used to create superposition states) and the Pauli-X gate (equivalent to the classical NOT gate, which flips the state of the qubits).

[0080] Graphene-based nanosensor arrays are arranged at key locations in semiconductors (such as gate, source, and drain), and the individual nanosensors are connected by wired or wireless means to form a densely distributed sensing network, enabling centralized management and real-time transmission of data.

[0081] To maintain the stable operation of the nanosensor network over a long period of time, a high-efficiency switching-mode DC-DC converter was selected and paired with a large-capacity supercapacitor to ensure the stable operation of the nanosensor network.

[0082] S3. Construct an aging prediction model to obtain the aging degree of the semiconductor based on the collected performance parameters.

[0083] Specifically, the steps include the following:

[0084] Define a feature extraction function that uses integral transform to capture hidden patterns in the collected performance parameters. Hidden patterns refer to potential structures or regularities that are not directly apparent in the original data but can be revealed through calculation, such as changes in aging rate and periodic behavior. Its expression is:

[0085]

[0086] Where F(x) represents the result of capturing the latent pattern, x represents the performance parameter vector, α represents the regularization coefficient, t represents time, n represents the number of performance parameters, and x' h Let dt represent the h-th standardized performance parameter, and dt represent the differential symbol.

[0087] To further illustrate, by using Gaussian kernels for integral transformation, implicit patterns in the time and frequency domains can be effectively captured, revealing potential aging characteristics.

[0088] Mapping the results of hidden patterns to quantum states φ(F(x)) using a mapping function allows us to leverage the powerful parallel processing capabilities and sophisticated pattern recognition capabilities of quantum computing, significantly improving data analysis efficiency.

[0089] Based on the results mapped to the hidden modes of quantum states, the results of hidden modes of multiple quantum states are introduced. The inner product similarity of the results of any two hidden modes of quantum states in the quantum state space is calculated using a quantum kernel function. The inner product similarity between the results of all hidden modes of quantum states is then used to construct a quantum kernel matrix, the expression of which is:

[0090] K ij =|<φ(F(x) i ))|φ(F(x j ))>| 2 ;

[0091] Among them, K ij F(x) represents the inner product similarity between the i-th and j-th hidden mode results in the quantum state space. i F(x) represents the captured i-th latent pattern result. j ) represents the j-th hidden pattern result captured, and φ() represents the mapping function.

[0092] To further explain, by calculating inner product similarity, the degree of similarity between different samples can be quantified, providing a basis for subsequent information filtering. Furthermore, the quantum kernel matrix provides a global perspective, reflecting the interrelationships between all samples, which helps to discover potential patterns and associations.

[0093] Define an information filtering function to extract feature similarity information related to aging features from the quantum kernel matrix, and emphasize high-similarity feature pairs (feature pairs that show high similarity between different samples are given higher weights because they are likely to capture key patterns in the data). Its expression is:

[0094]

[0095] Where G(K) represents the feature similarity information related to aging characteristics extracted from the quantum kernel matrix K, and K represents the quantum kernel matrix. ij The inner product similarity of the results of the i-th and j-th quantum state hidden modes in the quantum state space is represented by β, which represents the temperature parameter K. ii K represents the similarity between the result of the hidden mode of the i-th quantum state and itself. jj This represents the similarity between the result of the hidden mode of the j-th quantum state and itself;

[0096] To further explain, the exponential decay factor can highlight high-similarity feature pairs and reduce the impact of low-similarity features, making the aging prediction model pay more attention to important features.

[0097] Three aging states are defined: normal, mild aging, and severe aging.

[0098] Feature similarity information related to aging characteristics is extracted from the quantum kernel matrix and combined into a feature similarity information vector z, which is then input into the fully connected layer. The softmax function is applied to obtain the probability of the semiconductor being in different aging states, and its expression is:

[0099]

[0100] Where, P(z) a Let G(K) represent the probability that a semiconductor is in the a-th aging state based on the feature similarity information vector z. a This represents the feature similarity information related to the a-th aging state extracted from the feature similarity information related to aging characteristics from the quantum kernel matrix K, where L represents the total number of aging states and l represents the index variable;

[0101] Based on the probabilities of the semiconductor being in different aging states, the highest probability can be used to make a judgment. For example, if the output probabilities are [0.1, 0.6, 0.3], then the semiconductor is currently in a mild aging state.

[0102] Similarly, a weighted average combined with the probabilities of all states can be used to generate a comprehensive score, the expression of which is:

[0103]

[0104] Among them, w a This represents the weight corresponding to the a-th aging state.

[0105] First, the probability P(z) for each aging state. a A weight is set to reflect the importance and severity of different aging states. For example, in the normal state, the weight w1 equals 0, which means that if the device is in a normal state, it will not contribute to the aging score; in the mild aging state, the weight w2 equals 0.5, which means that if the device is in a mild aging state, the aging score will increase by half; in the mild aging state, the weight w3 equals 1, which means that if the device is in a severe aging state, the aging score will double.

[0106] These weights can be set according to actual needs to reflect the importance of different aging states to the overall assessment.

[0107] Assuming the output probabilities are [0.7, 0.2, 0.1], then:

[0108] Degree of aging = 0.7 × 0 + 0.2 × 0.5 + 0.1 × 1;

[0109] Aging level = 0 + 0.1 + 0.1;

[0110] Aging level = 0.2.

[0111] An aging threshold is set in the range [0,1]. The aging degree of the semiconductor is obtained based on the value of the aging degree falling within the range of the aging threshold.

[0112] For example, if the aging degree value is less than 0.3, the semiconductor is judged to be in a normal state; if the aging degree value is greater than or equal to 0.3 and less than 0.7, the semiconductor is judged to be in a slightly aged state; if the aging degree value is greater than or equal to 0.7 and less than or equal to 1, the semiconductor is judged to be in a severely aged state.

[0113] The specific setting of the aging threshold can be customized according to the usage scenario and actual situation.

[0114] S4. Analyze the semiconductor monitoring data to obtain the semiconductor status information.

[0115] Specifically, the steps include the following:

[0116] S4.1. Use the short-time Fourier transform to generate a time-frequency plot of semiconductor monitoring data in the time-frequency domain. Because the short-time Fourier transform can provide both time and frequency information simultaneously, it is particularly suitable for analyzing short-term phenomena in non-stationary signals, such as microstructural changes in semiconductor devices. Its expression is:

[0117]

[0118] Where S(t,f) represents the time-frequency plot, which represents the signal strength at time point t and frequency f, v(τ) represents the semiconductor monitoring data, τ represents the time variable, r() represents the window, usually Hanning window, Gaussian window, etc., to limit the range of each Fourier transform, and q represents the imaginary unit.

[0119] Obtaining time-frequency plot slices from a time-frequency plot is equivalent to extracting a series of vertical slices along the time axis on the time-frequency plot.

[0120] Calculate the amplitude at each frequency f, and find the maximum value of the amplitude and its corresponding frequency;

[0121] The frequency corresponding to the maximum value is the maximum peak frequency;

[0122] To further explain, the maximum peak frequency reflects the strongest frequency component at each point in time, which helps to identify transient events or changing trends.

[0123] The power spectral density is calculated based on the time-frequency plot. The power spectral density is the energy distribution of the signal at different frequencies.

[0124] For each frequency f, calculate the average power spectral density at all time points t, which is the average power spectral density.

[0125] To further explain, the average power spectral density provides the distribution of power over the entire time range, revealing the main frequency components of the signal and their relative intensities.

[0126] The maximum peak frequency and average power spectral density are combined to form a microscopic feature vector.

[0127] S4.2. Convolutional Neural Network (CNN) was chosen as the main model because it is good at processing two-dimensional data (such as time-frequency graphs) and can automatically learn complex patterns.

[0128] The cross-entropy loss function is used to measure the difference between the predicted value and the true label, and the backpropagation algorithm is used to update the CNN parameters, minimize the loss function, and obtain the semiconductor state information;

[0129] Status information, through continuous monitoring of microscopic features, can detect potential signs of aging at an early stage, enabling timely warnings and preventing failures.

[0130] S5. Combine the aging degree and status information of the semiconductor to form a complete semiconductor test result.

[0131] Specifically, the steps include the following:

[0132] Define a comprehensive evaluation function Z, which combines the aging degree and condition information of the semiconductor through a weighted summation to form a comprehensive semiconductor test score. Its expression is:

[0133] Z = λ1 × A(P(z)) a )+λ2×Y;

[0134] Where Z represents the overall semiconductor test score, λ1 represents the weight of the semiconductor aging degree, reflecting the importance of the aging degree in the final score, and A(P(z)) a ) represents the aging degree of the semiconductor, λ2 represents the weight of the semiconductor state information, and Y represents the semiconductor state information;

[0135] The weights for semiconductor aging degree and semiconductor state information can be customized according to actual needs.

[0136] The Sigmoid function is used to limit the overall semiconductor test score to the range [0,1].

[0137] Test thresholds are established based on historical semiconductor data, and three intervals are set between [0,1] to represent different health levels; for example, good, need attention, and urgent maintenance.

[0138] After obtaining a comprehensive semiconductor test score, it can be compared with the test threshold range to obtain a complete semiconductor test result.

[0139] To further explain, combining aging level and status information makes the assessment more comprehensive and reduces the risk of misjudgment that may result from relying on a single indicator.

[0140] S6. Based on semiconductor test results, provide maintenance suggestions and visual icons.

[0141] Specifically, the steps include the following:

[0142] If the semiconductor test results indicate a good condition, continue monitoring without taking immediate action.

[0143] If semiconductor test results indicate a need for attention, increase monitoring frequency, prepare a preventative maintenance plan, and consider replacing key components or optimizing operating conditions.

[0144] If the semiconductor test results indicate emergency maintenance, stop using the device, conduct a detailed diagnosis, develop a detailed maintenance plan, and ensure that operation is restored as soon as possible.

[0145] Using Python's Matplotlib, you can create visual charts illustrating the aging and condition information of semiconductors. Matplotlib supports various chart types, such as line charts, bar charts, scatter plots, pie charts, radar charts, and heatmaps, to meet different data display needs.

[0146] Choose a relational database (such as MySQL or PostgreSQL) or a NoSQL database (such as MongoDB), select the appropriate database type according to your needs, and record the results of each test in the database for future reference.

[0147] This embodiment also provides a semiconductor aging test system, including:

[0148] The preprocessing module collects the performance parameters of the semiconductor and performs preprocessing.

[0149] The monitoring module establishes a quantum platform and deploys a network of nanosensors to monitor semiconductors and obtain semiconductor monitoring data.

[0150] The prediction module constructs an aging prediction model and obtains the aging trend of semiconductors based on the collected performance parameters.

[0151] The parsing module analyzes semiconductor monitoring data to obtain semiconductor status information;

[0152] The testing module combines the aging trend and status information of semiconductors to form complete semiconductor test results;

[0153] The maintenance module provides maintenance suggestions and visual icons based on semiconductor test results.

[0154] This embodiment also provides a computer device applicable to semiconductor aging test methods, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the semiconductor aging test method proposed in the above embodiment.

[0155] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0156] This embodiment also provides a storage medium storing a computer program, which, when executed by a processor, implements the semiconductor aging test method proposed in the above embodiments. The 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.

[0157] In summary, this invention achieves in-depth mining of semiconductor aging characteristics by: establishing an aging prediction model, defining a feature extraction function, and using integral transform to capture hidden patterns in the collected data. This method reveals hidden patterns that are difficult to capture using traditional methods, improving the accuracy of aging prediction, and is particularly significant for identifying early signs of aging. Furthermore, by using short-time Fourier transform to generate time-frequency maps of semiconductor monitoring data in the time-frequency domain, this method captures the trend of microstructural changes, helps to detect potential problems early, reduces the risk of failure, and is especially effective in detecting early defect initiation.

[0158] Example 2, referring to Table 1, is the second embodiment of the present invention. To further verify the technical solution of the present invention, experimental simulation data of the semiconductor aging test method are given.

[0159] To verify the effectiveness of the semiconductor aging test method proposed in this invention, a series of experiments were designed to compare it with existing comparative methods. Four different semiconductor devices were used as test objects, denoted as Comparative Method Test Device 1, Comparative Method Test Device 2, Invention Method Test Device 1, and Invention Method Test Device 2. The experimental environment was set under standard laboratory conditions, with the temperature controlled at 25℃±1℃ and the humidity maintained at 40%±5%.

[0160] First, for each test object, its performance parameters such as operating voltage, current, temperature, and humidity are collected and preprocessed. Specific steps include:

[0161] Real-time data acquisition: Using high-precision data acquisition equipment, the above performance parameters are recorded every second to ensure the accuracy and integrity of the data.

[0162] Data cleaning: The collected data is cleaned to remove outliers and fill in missing values. For example, data points that are outside the normal range (such as voltage exceeding the rated value) are marked as outliers and removed; for missing values, linear interpolation is used to fill in the missing values.

[0163] Standardization processing: The cleaned data is standardized so that the values ​​of all performance parameters are distributed in the range of [0,1], which facilitates subsequent analysis.

[0164] Next, a quantum platform was built, and a network of nanosensors was deployed to monitor the semiconductor. IBM Quantum was chosen as the quantum cloud service platform. The hardware interface between the classical computer and the quantum processor was configured, and an initialization script was written to set the initial state vector. Nanosensor arrays were placed at key locations on the semiconductor to form a densely distributed sensor network, enabling real-time monitoring of the semiconductor's state and obtaining detailed monitoring data.

[0165] Constructing an aging prediction model is one of the core components of this invention. A feature extraction function is defined, and implicit patterns in the collected data are captured through integral transformation.

[0166] Furthermore, the results of the hidden patterns are mapped to quantum states, and based on the results of the hidden patterns, the inner product similarity of the results of any two hidden patterns in the quantum state space is calculated using a quantum kernel function to construct a quantum kernel matrix. Then, an information filtering function is defined to extract feature similarity information related to aging features from the quantum kernel matrix, and to emphasize high-similarity feature pairs.

[0167] Finally, feature similarity information related to aging characteristics is extracted from the quantum kernel matrix and combined into a feature similarity information vector, which is then input into the fully connected layer. The softmax function is applied to obtain the probability of the semiconductor being in different aging states, thereby determining the degree of semiconductor aging.

[0168] When analyzing semiconductor monitoring data, a short-time Fourier transform is used to generate a time-frequency graph, from which the maximum peak frequency and average power spectral density are extracted and combined into a microscopic feature vector. Based on this microscopic feature vector, a convolutional neural network (CNN) is introduced. Using supervised learning and cross-entropy as the loss function, the CNN is trained to obtain the semiconductor's state information, including microstructural changes and defect initiation.

[0169] When analyzing semiconductor monitoring data, a short-time Fourier transform is used to generate a time-frequency graph, from which the maximum peak frequency and average power spectral density are extracted and combined into a microscopic feature vector. Based on this microscopic feature vector, a convolutional neural network (CNN) is introduced. Using supervised learning and cross-entropy as the loss function, the CNN is trained to obtain the semiconductor's state information, including microstructural changes and defect initiation.

[0170] Finally, by combining the semiconductor's aging level and condition information, a comprehensive evaluation function is defined to obtain complete semiconductor test results, providing specific maintenance suggestions and visualization charts. Each test result is recorded in a database for long-term tracking and analysis.

[0171] The details are shown in Table 1 below:

[0172] Table 1 Test Comparison Table

[0173]

[0174] Analysis of the above table clearly shows that the method of this invention has significant advantages and innovative effects compared to the comparative method:

[0175] Aging degree rating: The aging degree ratings of the devices tested by the comparative method were 0.73 and 0.81, while the aging degree ratings of the devices tested by the method of the present invention were 0.59 and 0.63. This indicates that the method of the present invention can identify signs of semiconductor aging earlier, thereby allowing for preventative measures to be taken in advance and extending device lifespan.

[0176] State information scoring: The state information scores of the devices tested by the comparative methods were 0.68 and 0.72, while the state information scores of the devices tested by the method of the present invention were 0.86 and 0.89. This indicates that the method of the present invention can capture the state information of semiconductors more comprehensively and accurately, especially in terms of microstructure changes, providing a richer description of their health status.

[0177] Overall Score: Although the overall scores are similar (0.71, 0.77 and 0.72, 0.76 respectively), the method of this invention shows higher accuracy in scoring aging degree and status information, making the overall score more reliable. Furthermore, because the method of this invention can detect potential problems earlier, the overall score better reflects the actual health condition.

[0178] Maintenance Recommendations: The maintenance recommendations provided by the method of this invention are more precise. They not only identify situations requiring attention but also offer more targeted operational suggestions. For example, the recommendations to "continue monitoring" and "increase monitoring frequency" are based on detailed status information, which helps in developing reasonable maintenance plans and avoiding unnecessary downtime.

[0179] Visualization Chart Score: The method of this invention significantly outperforms the comparison methods in terms of visualization chart scores, at 0.94 and 0.91 respectively. This is mainly because the method of this invention can generate more intuitive and detailed charts, helping technical personnel quickly understand the meaning behind the data and improve decision-making efficiency.

[0180] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A semiconductor aging test method, characterized in that: include, Collect semiconductor performance parameters and perform preprocessing; A quantum platform was built, and a network of nanosensors was deployed to monitor semiconductors and obtain semiconductor monitoring data. An aging prediction model is constructed to determine the aging degree of the semiconductor based on the collected performance parameters. The specific steps include the following: Define a feature extraction function that captures hidden patterns in the acquired performance parameters through integral transform; its expression is: ； in, This represents the result of capturing hidden patterns. Represents a performance parameter vector. Represents the regularization coefficient. Indicates time, Indicates the number of performance parameters. Indicates the first A standardized performance parameter Represents the differential symbol; Mapping the results of hidden modes to quantum states; Based on the implicit mode results mapped to quantum states, the implicit mode results of multiple quantum states are introduced. The inner product similarity of the results of any two quantum state implicit modes in the quantum state space is calculated using the quantum kernel function. The inner product similarity between the results of all quantum state implicit modes is used to form a quantum kernel matrix. Define an information filtering function to extract feature similarity information related to aging features from the quantum kernel matrix, and emphasize highly similar feature pairs. Its expression is: ； in, Represents the quantum kernel matrix Extract feature similarity information related to aging characteristics. Represents the quantum kernel matrix. Indicates the first The and the first The inner product similarity of the results of the hidden modes of a quantum state in the quantum state space. Indicates semiconductor temperature parameters, Indicates the first The similarity between the results of the hidden modes of a quantum state and itself. Indicates the first The similarity between the results of a quantum state hidden mode and itself; Feature similarity information related to aging characteristics is extracted from the quantum kernel matrix and combined into a feature similarity information vector, which is then input into the fully connected layer. The softmax function is applied to obtain the probability of the semiconductor being in different aging states, and its expression is as follows: ； in, Represents a vector based on feature similarity information. Semiconductors are in the first The probability of a certain aging state. Represents the quantum kernel matrix Extracting feature similarity information related to aging characteristics from the first... Similarity information of features related to various aging states This represents the total number of aging states. Indicates an index variable; The degree of semiconductor aging is obtained based on the probability of the semiconductor being in different aging states. Analyze semiconductor monitoring data to obtain semiconductor status information; By combining the aging degree and condition information of semiconductors, a complete semiconductor test result is formed; Based on semiconductor test results, maintenance recommendations and visual icons are provided.

2. The semiconductor aging test method as described in claim 1, characterized in that: The semiconductor performance parameters are collected and preprocessed, specifically including the following steps: Real-time acquisition of semiconductor operating voltage, current, temperature, and humidity as performance parameters; The collected performance parameters are cleaned, converted in format, outliers are removed, and missing values ​​are filled. The cleaned data is then standardized.

3. The semiconductor aging test method as described in claim 2, characterized in that: The process of building a quantum platform and deploying a network of nanosensors to monitor semiconductors and obtain semiconductor monitoring data includes the following steps. IBM Quantum was chosen as the quantum cloud service platform. Configure the hardware interface between the classical computer and the quantum processor, and write an initialization script to set the initial state vector; Nanoscale sensor arrays are arranged at key locations in semiconductors to form a densely distributed sensing network; Nanoscale sensor networks monitor the state of semiconductors in real time and obtain monitoring data.

4. The semiconductor aging test method as described in claim 3, characterized in that: Analyzing semiconductor monitoring data to obtain semiconductor status information involves the following steps: Use short-time Fourier transform to generate a time-frequency plot of semiconductor monitoring data in the time-frequency domain; Extract the maximum peak frequency and average power spectral density from the time-frequency plot and combine them into a microscopic feature vector; Based on microscopic feature vectors, a convolutional neural network is introduced; By using supervised learning and cross-entropy as the loss function, a convolutional neural network is trained to obtain the state information of the semiconductor. The state information of the semiconductor includes changes in the semiconductor's microstructure and the initiation of defects.

5. The semiconductor aging test method as described in claim 4, characterized in that: Combining the aging degree and condition information of semiconductors to form complete semiconductor test results includes the following steps: Define a comprehensive evaluation function that combines the aging degree and condition information of the semiconductor to form a comprehensive semiconductor test score. Its expression is: ； in, This represents the overall semiconductor test score. Weights representing the degree of semiconductor aging Indicates the degree of aging of semiconductors. Weights representing semiconductor state information This indicates the status information of the semiconductor; Test thresholds are set based on historical semiconductor data. Complete semiconductor test results are obtained by determining whether the comprehensive semiconductor test score falls within the range of the test thresholds.

6. The semiconductor aging test method as described in claim 5, characterized in that: Based on semiconductor test results, maintenance recommendations and visual icons are provided, specifically including the following steps. Based on the semiconductor test results, specific maintenance recommendations are provided; Using Python's Matplotlib, visualize the aging and condition information of semiconductors. The results of each test are recorded in the database.

7. A semiconductor aging test system, based on the semiconductor aging test method according to any one of claims 1 to 6, characterized in that: include, The preprocessing module collects the performance parameters of the semiconductor and performs preprocessing. The monitoring module establishes a quantum platform and deploys a network of nanosensors to monitor semiconductors and obtain semiconductor monitoring data. The prediction module constructs an aging prediction model and obtains the aging trend of semiconductors based on the collected performance parameters. The parsing module analyzes semiconductor monitoring data to obtain semiconductor status information; The testing module combines the aging trend and status information of semiconductors to form complete semiconductor test results; The maintenance module provides maintenance suggestions and visual icons based on semiconductor test results.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the semiconductor aging test method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the semiconductor aging test method according to any one of claims 1 to 6.

Images