Method, system, and electronic device for semiconductor device consistency optimization

Abstract

The application provides a semiconductor equipment consistency optimization method, system and electronic equipment, relates to the technical field of semiconductor technology, and is used for solving the technical problem that the state consistency of semiconductor equipment is difficult to detect. The consistency optimization method comprises the following steps: determining an optimal equipment in a plurality of semiconductor equipment according to log data during wafer manufacturing in different semiconductor equipment performing the same process step, each log data comprising sampling data of a plurality of sensors; obtaining an offset distance between the group center of the log data corresponding to each of the other semiconductor equipment and the group center of the log data corresponding to the optimal equipment; judging the deviation degree of each of the other semiconductor equipment relative to the optimal equipment according to each offset distance, and determining an optimizable component. By judging the deviation degree of each of the other semiconductor equipment, and then adjusting the semiconductor equipment with a large deviation degree, each semiconductor equipment can be kept in an optimal state, thereby improving the quality and yield of the wafer.

Description

Technical Field

[0001] This application relates to the field of semiconductor technology, and in particular to a method, system, and electronic device for semiconductor device consistency optimization. Background Technology

[0002] As semiconductor manufacturing progresses towards automation, high precision, reliability, and intelligence, a major challenge facing the semiconductor industry is the need to continuously improve the overall quality of the final product. In the case of wafers, the overall quality is affected by numerous process variability factors. To improve overall wafer quality, it is necessary to match the state of the semiconductor equipment on each wafer. By maintaining optimal conditions for each semiconductor device, consistency in overall wafer quality can be achieved.

[0003] However, the different states of semiconductor equipment in existing equipment plants are difficult to detect, resulting in inconsistent wafer quality and low wafer yield. Summary of the Invention

[0004] In view of the above problems, embodiments of this application provide a method, system and electronic device for semiconductor device consistency optimization, so as to improve the detection of the status of each semiconductor device, so that each semiconductor device is in a better state, thereby improving the quality and yield of the wafer.

[0005] In a first aspect, embodiments of this application provide a method for optimizing semiconductor device consistency, comprising:

[0006] Based on log data from wafer fabrication in different semiconductor devices performing the same process step, the optimal device is determined among a plurality of said semiconductor devices, each of said log data including sampling data from multiple sensors;

[0007] Obtain the offset distance between the group center of the log data corresponding to each of the other semiconductor devices and the group center of the log data corresponding to the optimal device;

[0008] Based on the offset distances described, the degree of deviation of each of the other semiconductor devices relative to the optimal device is determined, and the components that can be optimized are identified.

[0009] In some possible embodiments, based on each of the said offset distances, the degree of deviation of the other said semiconductor devices relative to the optimal device is determined, and the corresponding optimizable components are identified, including:

[0010] Determine whether each of the offset distances is greater than or equal to a preset value; the semiconductor device whose offset distance is greater than or equal to the preset value is a device to be optimized.

[0011] The target sensor is determined based on the difference between the sampled data of each sensor in the log data of the device to be optimized and the sampled data of the corresponding sensor in the log data of the optimal device.

[0012] Identify the component corresponding to the target sensor.

[0013] In some possible embodiments, obtaining the offset distance between the group center of the log data corresponding to each of the other semiconductor devices and the group center of the log data corresponding to the optimal device includes:

[0014] Calculate the Euclidean distance and cosine distance between the group center of the log data corresponding to each of the other semiconductor devices and the group center of the log data corresponding to the optimal device;

[0015] The offset distance is determined based on the Euclidean distance and the cosine distance.

[0016] In some possible embodiments, the offset distance is equal to the sum of the Euclidean distance and the cosine distance.

[0017] In some possible embodiments, an optimal device is determined among a plurality of said semiconductor devices based on log data from wafer fabrication in different semiconductor devices, each of said log data including signal data from multiple sensors, including:

[0018] Obtain log data for each step of the same process in the wafers of different semiconductor devices;

[0019] Calculate the statistical values ​​of the log data in each step, and perform feature dimensionality reduction on the log data;

[0020] Principal component analysis was used to cluster all the statistical values ​​to determine the candidate equipment.

[0021] Based on the sampling data from the plurality of sensors corresponding to each semiconductor device, the average stability of all the sensors corresponding to each semiconductor device is determined;

[0022] The average stability values ​​are sorted, and the optimal device is determined from the candidate devices according to the order of the average stability values.

[0023] In some possible embodiments, principal component analysis is used to cluster all the statistical values ​​to determine candidate devices, including:

[0024] Obtain an original matrix consisting of all the statistical values, wherein each column of the original matrix corresponds to the statistical value for each of the semiconductor devices;

[0025] Subtract the mean of each column from the original matrix to decenter each column of the original matrix;

[0026] Calculate the covariance matrix of the original matrix after decentralization, as well as the eigenvalues ​​and eigenvectors of the covariance matrix;

[0027] Sort the feature values ​​from largest to smallest, and retain the feature vectors corresponding to the first k feature values, where k is a positive integer greater than or equal to 2;

[0028] The original matrix is ​​transformed into a new space constructed by the k eigenvectors to obtain the principal component matrix;

[0029] The candidate devices are determined based on the distribution of the first two principal components in the principal component matrix.

[0030] In some possible embodiments, determining the average stability of all the sensors corresponding to each semiconductor device based on the sampling data of the plurality of sensors corresponding to each semiconductor device includes:

[0031] The data length and time of the sampling data of the sensors corresponding to each semiconductor device are aligned using a cross-correlation function;

[0032] Calculate the average variance of the sampled data after alignment of the plurality of sensors corresponding to each semiconductor device, and determine the average stability of all the sensors of each semiconductor device.

[0033] In some possible embodiments, ranking the average stability values ​​and determining the optimal device from the candidate devices according to the order of the average stability values ​​includes:

[0034] Based on a prior knowledge base constructed from expert experience, at least one optimal device is selected from the top N candidate devices with the highest average stability, where N is a positive integer greater than or equal to 2.

[0035] The semiconductor device consistency optimization method provided in this application has at least the following advantages:

[0036] The semiconductor device consistency optimization method in this application embodiment determines the optimal device based on the log data corresponding to different semiconductor devices performing the same process step. It also determines the degree of deviation of other semiconductor devices relative to the optimal device based on the offset distance between the group center of the log data corresponding to each other semiconductor device and the group center of the log data corresponding to the optimal device, and identifies the components to be optimized. This obtains the state of each semiconductor device. By adjusting the semiconductor devices with large deviations, it is possible to keep each semiconductor device in a relatively consistent state, thereby improving the quality and yield of the wafer.

[0037] Secondly, embodiments of this application provide a system for optimizing semiconductor device consistency, comprising:

[0038] The first acquisition module is used to determine the optimal device among a plurality of semiconductor devices based on log data from wafer fabrication in different semiconductor devices performing the same process step, wherein each log data includes sampling data from a plurality of sensors.

[0039] The second acquisition module is used to obtain the group center of the log data corresponding to each of the other semiconductor devices and the offset distance between the group center of the log data corresponding to the optimal device.

[0040] The judgment module is used to determine the degree of deviation of each of the other semiconductor devices relative to the optimal device based on the offset distance, and to determine the components that can be optimized.

[0041] The semiconductor device consistency optimization system in this application embodiment can execute the technical solution shown in the above method embodiment, and its principle and beneficial effects are similar, so they will not be repeated here.

[0042] Thirdly, embodiments of this application provide an electronic device, including: at least one processor and a memory, the memory storing computer-executable instructions; the at least one processor executes the computer-executable instructions stored in the memory, causing the at least one processor to perform the semiconductor device consistency optimization method as described above. Attached Figure Description

[0043] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0044] Figure 1 This is a flowchart of a semiconductor device consistency optimization method in an embodiment of this application;

[0045] Figure 2 This is a state distribution diagram of the semiconductor device in the embodiments of this application. Detailed Implementation

[0046] This application provides a method for optimizing the consistency of semiconductor devices. Based on the log data corresponding to different semiconductor devices, the optimal device is determined. Based on the offset distance between the group center of the log data corresponding to each other semiconductor device and the group center of the log data corresponding to the optimal device, the degree of deviation of each other semiconductor device relative to the optimal device is determined, and the component to be optimized is determined. This obtains the state of each semiconductor device. By adjusting the semiconductor devices with large deviations, the state of each semiconductor device is improved, thereby improving the quality and yield of the wafer.

[0047] To make the above-mentioned objectives, features, and advantages of the embodiments of this application more apparent and understandable, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0048] refer to Figure 1 This application provides a method for semiconductor device consistency optimization, which aims to make the state of semiconductor devices more consistent. The consistency optimization method includes the following steps:

[0049] Step S100: Based on the log data of wafer fabrication in different semiconductor devices performing the same process step, determine the optimal device among multiple semiconductor devices, where each log data includes sampling data from multiple sensors.

[0050] When multiple wafers undergo the same process step in different semiconductor devices, the states of these devices can vary, resulting in differences in wafer quality, with at least some wafers exhibiting lower quality and yield. Analyzing log data from the wafer fabrication process in each semiconductor device performing the same step reveals the state of the respective devices. This allows for adjustments to the devices that need improvement, bringing them to a better, or even optimal, state, ultimately leading to higher quality and yield wafers produced by different devices.

[0051] Performing the same process step means performing the same process steps on the same target film layer in each wafer, with identical parameters in each process step. The optimal chamber refers to one or more semiconductor devices in good condition among different semiconductor devices performing the same process step. Using this chamber as a benchmark and target, other semiconductor devices are adjusted to ensure that the condition of all semiconductor devices is relatively consistent, thereby ensuring that all semiconductor devices are in a superior state.

[0052] Each log data point includes sampling data from multiple sensors during wafer fabrication within the semiconductor equipment. These sensors include temperature sensors, pressure sensors, and other sensors; that is, each log data point contains multiple components. The sampling data from these multiple sensors during wafer fabrication can be extracted from the datalog file generated during the wafer fabrication process within the semiconductor equipment. This datalog file contains time-series data from multiple sensors.

[0053] Optionally, the log data may also include equipment information for the corresponding semiconductor device, such as batch number, model number, and other hardware information. The more types of data included in the log data, the more accurate the determination of the status of each semiconductor device.

[0054] Based on the log data corresponding to each semiconductor device, the semiconductor device with the best or even the best current state among all semiconductor devices is selected as the optimal device. Using the optimal device as the benchmark, the deviation of each other semiconductor device from the optimal device is determined. The smaller the deviation of any other semiconductor device from the optimal device, the closer its state is to that of the optimal device, and the better its state.

[0055] Understandably, there can be one, two, or more optimal devices. When only one optimal device is selected, subsequent comparisons of all other semiconductor devices with this single optimal device are straightforward. When multiple optimal devices are selected, subsequent comparisons of all other semiconductor devices with these optimal devices are necessary, making the comparison process more flexible.

[0056] Step S200: Obtain the group center of the log data corresponding to each of the other semiconductor devices and the offset distance between the group center of the log data corresponding to the optimal device.

[0057] Each semiconductor device corresponds to a log data. The group center of each log data is obtained to determine the offset distance between the group center corresponding to other semiconductor devices and the group center corresponding to the optimal device, thereby determining the degree of deviation between other semiconductor devices and the optimal device.

[0058] In some possible implementations, the offset between the group center of the log data corresponding to each of the other semiconductor devices and the group center of the log data corresponding to the optimal device is obtained, including:

[0059] Calculate the Euclidean and cosine distances between the group center of the log data corresponding to each of the other semiconductor devices and the group center of the log data corresponding to the optimal device; determine the offset distance based on the Euclidean and cosine distances.

[0060] Euclidean distance measures the absolute distance between two points in a multidimensional space. It is equal to the square root of the sum of the squares of the differences between the corresponding components in the group center corresponding to a semiconductor device and the group center corresponding to the optimal device. Euclidean distance characterizes the numerical difference between these two group centers.

[0061] The cosine distance is equal to the vector product of the group center corresponding to a semiconductor device and the group center corresponding to the optimal device, divided by the length of these two group centers. In other words, it is equal to the sum of the products of the corresponding components of the group center corresponding to the semiconductor device and the group center corresponding to the optimal device, divided by the square root of the sum of the squares of the components of the semiconductor device's group center, and then divided by the square root of the sum of the squares of the components of the optimal device's group center. The cosine distance represents the difference in direction between these two group centers, which is also an angular difference.

[0062] Optionally, the offset distance corresponding to each of the other semiconductor devices can be the sum of the Euclidean distance and the cosine distance, thereby comprehensively determining the numerical and angular differences between the semiconductor device and the optimal device.

[0063] Step S300: Based on each offset distance, determine the degree of deviation of each other semiconductor device from the optimal device, and identify the components that can be optimized.

[0064] Each of the other semiconductor devices can be assigned an offset distance. The magnitude of this offset distance determines the degree of deviation of the corresponding semiconductor device from the optimal device. A larger offset distance indicates a worse state of the semiconductor device relative to that offset distance, and a greater degree of deviation from the optimal device.

[0065] In some possible embodiments, based on each offset distance, the degree of deviation of other semiconductor devices relative to the optimal device is determined and the optimizable components are identified (step S300), including:

[0066] Step S310: Determine whether each offset distance is greater than or equal to a preset value. Semiconductor devices with offset distances greater than or equal to preset values ​​are devices to be optimized.

[0067] For each semiconductor device other than the optimal device, the offset distance is determined. It is then checked whether this offset distance is greater than or equal to a preset value to obtain the status of each other semiconductor device. If yes, meaning the offset distance between the group center of the semiconductor device and the group center of the optimal device is greater than or equal to the preset value, the device status can be further optimized, for example, by replacing components or adjusting parameters. If no, meaning the offset distance between the group center of the semiconductor device and the group center of the optimal device is less than the preset value, the device status is good and requires no optimization or adjustment.

[0068] Step S320: Determine the target sensor based on the difference between the sampling data of each sensor in the log data corresponding to the device to be optimized and the sampling data of the corresponding sensor in the log data corresponding to the optimal device.

[0069] After identifying the equipment to be optimized, it is necessary to further determine the target sensors within that equipment. This will help identify the sensors that cause the equipment to deviate from its optimal state, allowing for targeted replacement, repair, or parameter adjustment. This will improve the utilization rate of semiconductor equipment, the quality of produced wafers, and ultimately, the overall yield of the semiconductor factory.

[0070] During the judgment, the sampling data of the corresponding sensor in the device to be optimized can be compared with that in the optimal device to determine the degree of difference of the sampling data. The sensor corresponding to the sampling data with a large degree of difference is the target sensor, and the corresponding sensor is the same type of sensor.

[0071] For example, the sampling data of the temperature sensor corresponding to the device to be optimized is compared with the sampling data of the temperature sensor corresponding to the optimal device. The difference between these two sets of sampling data is determined to identify whether the temperature sensor is a target sensor. Similarly, the sampling data of the pressure sensor corresponding to the device to be optimized is compared with the sampling data of the pressure sensor corresponding to the optimal device. The difference between these two sets of sampling data is determined to identify whether the pressure sensor is a target sensor. Likewise, the sampling data of other sensors among the multiple sensors corresponding to the device to be optimized is compared with the sampling data of the corresponding sensor corresponding to the optimal device. The difference between each set of sampling data is determined, thereby identifying the target sensors one by one. There can be multiple target sensors, i.e., target sensors can form a target sensor group.

[0072] Understandably, the component corresponding to the target sensor is the component to be optimized. Because of this component, the sampling data from the target sensor differs significantly from the sampling data from the corresponding sensor in the optimal device. For example, if the target sensor is a pressure sensor, a significant difference between the pressure data obtained by the target sensor and the pressure data obtained by the pressure sensor in the optimal device indicates that the pressure control component (e.g., the vacuum pump) in the device to be optimized needs repair or replacement to improve the condition of the device.

[0073] Step S330: Determine the component to be optimized corresponding to the target sensor.

[0074] Each target sensor can correspond to one or more components in the device to be optimized, thereby identifying the components that need to be replaced or have their parameters adjusted. For example, if the temperature sensor in the device to be optimized is a target sensor, and the temperature data it represents for the device to be optimized differs significantly from the temperature data for the optimal device, then the heating device in the device to be optimized can be replaced, and / or the cooling device in the device to be optimized can be replaced, and / or the parameters of the heating device in the device to be optimized can be adjusted, and / or the parameters of the cooling device in the device to be optimized can be adjusted.

[0075] In summary, in this embodiment, the optimal device is determined based on the log data corresponding to different semiconductor devices. The deviation of each semiconductor device from the optimal device is determined based on the offset distance between the group center of the log data corresponding to each other semiconductor device and the group center of the log data corresponding to the optimal device. The device to be optimized is then determined, thereby obtaining the state of each semiconductor device. By adjusting the semiconductor devices with large deviations, the state of each semiconductor device can be improved, thereby improving the quality and yield of the wafer.

[0076] In some possible embodiments, based on log data from wafer fabrication in different semiconductor devices, an optimal device is determined among multiple semiconductor devices, each log data including signal data from multiple sensors (step S100), including:

[0077] Step S110: Obtain log data of each step in the same process for wafers in different semiconductor devices.

[0078] A process is divided into multiple steps, and these steps can be retrieved from the data log file.

[0079] Step S120: Calculate the statistical values ​​of the log data in each step and perform feature dimensionality reduction on the log data.

[0080] To facilitate subsequent dimensionality reduction and processing of the log data, statistical values ​​can be calculated for each step. These statistical values ​​can include various parameters such as the maximum, minimum, or covariance of the log data at each step. After calculating the statistical values, the feature length of each log data point is n, where n is a positive integer.

[0081] Step S130: Use principal component analysis to cluster all statistical values ​​to determine candidate equipment.

[0082] Principal Component Analysis (PCA) is performed on the statistical values ​​corresponding to each semiconductor device. Based on the distribution of the principal components in the first few dimensions (e.g., the first two dimensions), the range of optimal devices is initially selected, i.e., the optimal device is selected from the candidate devices.

[0083] See Figure 2 After clustering the statistical values ​​of each semiconductor device using principal component analysis, two semiconductor devices, 4 and 5, deviated from the majority of the semiconductor devices, such as... Figure 2 As shown, semiconductor device 5 is located to the left of most of the semiconductor devices, and semiconductor device 4 is located to the right of most of the semiconductor devices. The other semiconductor devices besides these two are clustered in the middle area; these are candidate devices. Although the candidate devices are clustered, they are still somewhat dispersed, requiring further narrowing down the range of candidate devices to determine the optimal device.

[0084] In some possible implementations, principal component analysis is used to cluster all statistical values ​​to determine candidate devices (step S130), including:

[0085] Step S131: Obtain the original matrix composed of all statistical values, where each column of the original matrix represents the statistical value corresponding to each semiconductor device.

[0086] Specifically, the characteristic length of the statistical value corresponding to each semiconductor device is n. The m wafers from different semiconductor devices form an m*n original matrix X, which can be represented as:

[0087]

[0088] Step S132: Subtract the mean of each column from the original matrix to decenter each column of the original matrix.

[0089] Step S133: Calculate the covariance matrix of the original matrix after decentralization, as well as the eigenvalues ​​and eigenvectors of the covariance matrix.

[0090] Step S134: Sort the eigenvalues ​​from largest to smallest, and retain the eigenvectors corresponding to the first k eigenvalues, where k is a positive integer greater than or equal to 2.

[0091] Specifically, if k can be 2, then the transformation matrix P includes the eigenvectors corresponding to the first two eigenvalues, that is, the eigenvector corresponding to the largest eigenvalue and the eigenvector corresponding to the second largest eigenvalue.

[0092] Step S135: Transform the original matrix into a new space constructed by the k eigenvectors to obtain the principal component matrix. It can be understood that the principal component matrix Y = PX.

[0093] Step S136: Determine the candidate equipment based on the distribution of the first two principal components in the principal component matrix.

[0094] Specifically, candidate equipment is determined based on the distribution of the principal components in the first two dimensions.

[0095] Step S140: Determine the average stability of all sensors corresponding to each semiconductor device based on the sampling data of multiple sensors corresponding to each semiconductor device.

[0096] The sampling data from multiple sensors corresponding to each semiconductor device are grouped together. By assessing the stability of the sampling data from this group of sensors for each semiconductor device, the stability of the semiconductor device corresponding to that group of sensors can be measured, thus ensuring that the determined optimal device has good stability.

[0097] In some possible implementations, the average stability of all sensors corresponding to each semiconductor device is determined based on the sampling data of multiple sensors corresponding to each semiconductor device (step S140), including:

[0098] By using cross-correlation functions, the sampled data of the sensors corresponding to each semiconductor device are aligned in terms of data length and time.

[0099] Calculate the average variance of the aligned sampled data from multiple sensors corresponding to each semiconductor device, and determine the average stability of all sensors for each semiconductor device.

[0100] Specifically, because different wafers undergo slightly different processing times within the same process on various semiconductor devices, the data lengths of the sensors corresponding to different semiconductor devices vary, and there are time discrepancies. For example, the sampling data of pressure sensors corresponding to different semiconductor devices have different data lengths and are not synchronized in time. To ensure better stability of the optimal device, in this embodiment, the data lengths and times of the corresponding sensors are aligned. The specific steps are as follows:

[0101] The corresponding multiple sensors can be of the same type, forming a matrix of sensor sampling data of unequal length, sensor[j], which is represented as:

[0102]

[0103] Where wi_sj=(wi_sj[1],wi_sj[2],…,wi_sj[k]) represents the sampled data in the matrix sensor[j] corresponding to the i-th wafer, i∈[1,m], and the lengths k of different wi_sj are not completely equal.

[0104] Select the shortest data type, wi_sj, as the reference data;

[0105] Calculate the cross-correlation coefficient between the matrix sensor[j] and wi_sj for other data lengths.

[0106] For w1_sj and wi_sj, assume that the data length of w1_sj is z1 and the data length of wi_sj is z2. w1_sj=(w1_sj[1],w1_sj[2],...,w1_sj[k1]); wi_sj=(si_sj[1],wi_sj[2],...wi_sj[k2]), z2≥z1.

[0107] The linear cross-correlation result between w1_sj and wi_sj is expressed as C(w1_sj, wi_sj)={C p p = 0, 1, 2, ...}, where, Each C p It is obtained through an operation similar to "sliding dot product".

[0108] The cross-correlation coefficient represents the similarity between each subvector in w1_sj of the same length as wi_sj and wi_sj. The index with the largest value in C(w1_sj, wi_sj) is the starting index of the subvector in w1_sj that is most similar to wi_sj, which is the part of the data that is relatively homogeneous in time.

[0109] By aligning the sampling data from different wafers of the same sensor within the same device, the variance at each time point t is calculated using the aligned sampling data. The average variance over all time points is then calculated; a larger average variance indicates poorer sensor stability. The average variance of multiple sensors corresponding to each semiconductor device is then averaged to characterize the average stability of all sensors corresponding to each semiconductor device.

[0110] Step S150: Sort the average stability values ​​and determine the optimal device from the candidate devices according to the order of average stability values.

[0111] In some possible embodiments, the average stability is sorted and the optimal device is determined from the candidate devices according to the order of average stability, including: selecting at least one optimal device from the top N candidate devices with the highest average stability based on a prior knowledge base constructed from expert experience, where N is a positive integer greater than or equal to 2.

[0112] By combining historical data from a prior knowledge base, at least one optimal device is determined from the candidate devices with the highest average stability. This optimal device may or may not include the candidate device with the highest average stability.

[0113] This application also provides a system for semiconductor device consistency optimization, including:

[0114] The first acquisition module is used to determine the optimal device among multiple semiconductor devices based on log data from wafer fabrication in different semiconductor devices performing the same process step. Each log data includes sampling data from multiple sensors.

[0115] The second acquisition module is used to obtain the group center of the log data corresponding to each other semiconductor device and the offset distance between the group center of the log data corresponding to the optimal device.

[0116] The judgment module is used to determine the degree of deviation of other semiconductor devices from the optimal device based on each offset distance, and to identify the corresponding optimizable components.

[0117] The semiconductor device consistency optimization system in this application embodiment can execute the technical solution shown in the above method embodiment, and its principle and beneficial effects are similar, so they will not be repeated here.

[0118] This application also provides an electronic device, including: at least one processor and a memory, the memory storing computer-executable instructions; the at least one processor executes the computer-executable instructions stored in the memory, causing the at least one processor to perform the semiconductor device conformance optimization method described above. The memory can be either independent or integrated with the processor. When the memory is independently configured, the device also includes a bus for connecting the memory and the processor.

[0119] This disclosure also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the above-described method for semiconductor device consistency optimization.

[0120] This disclosure also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method for semiconductor device consistency optimization.

[0121] It is understood that the above division of modules is merely a logical functional division. In actual implementation, they can be fully or partially integrated into a single physical entity, or they can be physically separated. These modules can be implemented entirely in software via processing element calls; they can be fully implemented in hardware; or some modules can be implemented by processing element calls to software, while others are implemented in hardware. They can be separate processing elements, integrated into a chip within the aforementioned device, or stored as program code in the device's memory, invoked and executed by a processing element. The implementation of other modules is similar. Furthermore, these modules can be fully or partially integrated together, or implemented independently. The processing element here can be an integrated circuit with signal processing capabilities. During implementation, each step of the above method or each of the above modules can be completed through integrated logic circuits in the hardware of the processor element or through software instructions.

[0122] For example, these modules can be one or more integrated circuits configured to implement the above methods, such as one or more Application Specific Integrated Circuits (ASICs), one or more Digital Signal Processors (DSPs), or one or more Field Programmable Gate Arrays (FPGAs). As another example, when a module is implemented using processing element scheduler code, the processing element can be a general-purpose processor, such as a Central Processing Unit (CPU) or other processor capable of calling program code. Furthermore, these modules can be integrated together to implement a System on Chip (SoC).

[0123] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product. A computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the flow or function according to the embodiments of this disclosure is generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., a solid-state disk (SSD)).

[0124] The processor mentioned above can be a Central Processing Unit (CPU), or other general-purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), etc. The memory may include high-speed RAM, and may also include non-volatile memory (NVM), such as at least one disk drive, or a USB flash drive, external hard drive, read-only memory, hard disk, or optical disk. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc.

[0125] The aforementioned storage medium can be implemented using any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disks, or optical disks. The storage medium can be any available medium accessible to general-purpose or special-purpose computers.

[0126] The various embodiments or implementation methods described in this specification are presented in a progressive manner. Each embodiment focuses on the differences from other embodiments, and the same or similar parts between the embodiments can be referred to each other.

[0127] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "illustrative embodiment," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with an embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0128] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

Claims

1. A method for optimizing the consistency of semiconductor devices, characterized in that, include: Based on log data from wafer fabrication in different semiconductor devices performing the same process step, the optimal device is determined among a plurality of said semiconductor devices, each of said log data including sampling data from multiple sensors; Obtain the offset distance between the group center of the log data corresponding to each of the other semiconductor devices and the group center of the log data corresponding to the optimal device; Based on the aforementioned offset distances, the degree of deviation of each of the other semiconductor devices relative to the optimal device is determined, and the components that can be optimized are identified. Specifically, based on log data from wafer fabrication in different semiconductor devices, the optimal device is determined among multiple semiconductor devices. Each log data includes signal data from multiple sensors, including: Obtain log data for each step of the same process in the wafers of different semiconductor devices; Calculate the statistical values ​​of the log data in each step, and perform feature dimensionality reduction on the log data; Principal component analysis was used to cluster all the statistical values ​​to determine the candidate equipment. Based on the sampling data from the plurality of sensors corresponding to each semiconductor device, the average stability of all the sensors corresponding to each semiconductor device is determined; The average stability of each device is sorted, and the optimal device is determined from the candidate devices according to the order of the average stability. Based on the sampling data from the plurality of sensors corresponding to each semiconductor device, the average stability of all the sensors corresponding to each semiconductor device is determined, including: The data length and time of the sampling data of the sensors corresponding to each semiconductor device are aligned using a cross-correlation function; Calculate the average variance of the sampled data after alignment of the plurality of sensors corresponding to each semiconductor device, and determine the average stability of all the sensors of each semiconductor device.

2. The method for semiconductor device consistency optimization according to claim 1, characterized in that, Based on the aforementioned offset distances, the degree of deviation of each of the other semiconductor devices relative to the optimal device is determined, and the corresponding optimizable components are identified, including: Determine whether each of the offset distances is greater than or equal to a preset value; the semiconductor device whose offset distance is greater than or equal to the preset value is a device to be optimized. The target sensor is determined based on the difference between the sampled data of each sensor in the log data of the device to be optimized and the sampled data of the corresponding sensor in the log data of the optimal device. Identify the component corresponding to the target sensor.

3. The method for semiconductor device consistency optimization according to claim 2, characterized in that, Obtaining the offset distance between the group center of the log data corresponding to each of the other semiconductor devices and the group center of the log data corresponding to the optimal device includes: Calculate the Euclidean distance and cosine distance between the group center of the log data corresponding to each of the other semiconductor devices and the group center of the log data corresponding to the optimal device; The offset distance is determined based on the Euclidean distance and the cosine distance.

4. The method for semiconductor device consistency optimization according to claim 3, characterized in that, The offset distance is equal to the sum of the Euclidean distance and the cosine distance.

5. The method for semiconductor device consistency optimization according to claim 1, characterized in that, Principal component analysis was used to cluster all the statistical values ​​to determine candidate equipment, including: Obtain an original matrix consisting of all the statistical values, wherein each column of the original matrix corresponds to the statistical value for each of the semiconductor devices; Subtract the mean of each column from the original matrix to decenter each column of the original matrix; Calculate the covariance matrix of the original matrix after decentralization, as well as the eigenvalues ​​and eigenvectors of the covariance matrix; Sort the feature values ​​from largest to smallest, and retain the feature vectors corresponding to the first k feature values, where k is a positive integer greater than or equal to 2; The original matrix is ​​transformed into a new space constructed by the k eigenvectors to obtain the principal component matrix; The candidate devices are determined based on the distribution of the first two principal components in the principal component matrix.

6. The method for semiconductor device consistency optimization according to claim 1, characterized in that, The average stability of each device is ranked, and the optimal device is determined from the candidate devices according to the order of the average stability, including: Based on a prior knowledge base constructed from expert experience, at least one optimal device is selected from the top N candidate devices with the highest average stability, where N is a positive integer greater than or equal to 2.

7. A system for optimizing semiconductor device consistency, characterized in that, include: The first acquisition module is used to determine the optimal device among a plurality of semiconductor devices based on log data from wafer fabrication in different semiconductor devices performing the same process step, wherein each log data includes sampling data from a plurality of sensors. The second acquisition module is used to obtain the group center of the log data corresponding to each of the other semiconductor devices and the offset distance between the group center of the log data corresponding to the optimal device. The judgment module is used to determine the degree of deviation of each of the other semiconductor devices relative to the optimal device based on the offset distance, and to determine the components that can be optimized.

8. An electronic device, characterized in that, include: At least one processor and a memory, the memory storing instructions to be executed by the computer; The at least one processor executes computer execution instructions stored in the memory, causing the at least one processor to perform the semiconductor device conformance optimization method according to any one of claims 1-6.

Images