Failure unit repairing method, repairing device, repairing system and storage medium
By using the attribute information of failed units for cluster analysis and enhanced repair during the chip manufacturing process, the problem of low chip manufacturing yield is solved, higher chip yield and lower repair costs are achieved, and the use of the chip is extended. life.
Patent Information
- Application Number
- CN202310868063.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-07-14
- Publication Date
- 2025-01-24
AI Technical Summary
At this stage, there are limitations in the chip repair method, which leads to the chip manufacturing yield still need to be improved, especially in the later testing stage, which may appear, which increases the repair cost and reduces the service life of the chip.
By obtaining the attribute information of all the failure units in the target testing stage, at least one core failure unit is determined, and it is used as a cluster center to attribute the remaining failure units to the corresponding failure unit cluster. According to the cluster of each failure unit, the strategy of enhanced restoration around it is determined, and the potential failure unit is repaired by using a spare circuit.
It reduces the possibility of new failure units during the test phase, increases the yield of chip manufacturing, reduces the cost of repair, thereby extending the service life of the chip and enhancing the corporate image.
Smart Images
Figure CN119360929A_ABST
Abstract
Description
Technical Field
[0001] The present disclosure relates to the field of semiconductor technology, and in particular, to a method, a device, a system and a storage medium for repairing a failed unit. Background Art
[0002] In semiconductor manufacturing technology, a wafer with multiple dies can be formed first, and then the dies are sliced and packaged to form a chip. During the manufacturing process, the failed bits (FB) on the wafer need to be repaired to improve the chip manufacturing yield. However, the current chip repair method needs to be further improved. Summary of the invention
[0003] In view of this, the embodiments of the present disclosure provide a method, a device, a system and a storage medium for repairing a failed unit. The method of repairing a failed unit provided by the embodiments of the present disclosure includes:
[0004] Acquire attribute information of all failed units of the object to be analyzed in the target test phase, wherein the attribute information includes location information and information of failed test items;
[0005] Determine at least one core failure unit according to the attribute information of the failure unit;
[0006] The core failure unit is used as the cluster center, and the remaining failure units are assigned to the corresponding failure unit clusters;
[0007] According to each failed unit cluster, a strategy for performing enhanced repair on the periphery of the failed unit cluster is determined, wherein the enhanced repair is to repair the potential failed units by using a spare circuit.
[0008] In some embodiments, the determining of at least one core failure unit according to the attribute information of the failure unit includes:
[0009] Calculating the weighted Euclidean distance between any two failed units based on the attribute information;
[0010] Acquire a search radius, where the search radius is related to the size of the minimum repair area of the object to be analyzed;
[0011] For each of the failed units among all the failed units, using the search radius and the weighted Euclidean distance between the corresponding failed unit and each remaining failed unit, calculate the local area absolute density of the corresponding failed unit and the minimum weighted Euclidean distance between the corresponding failed unit and the remaining failed units;
[0012] The at least core failure unit is determined according to the local area absolute density and the minimum weighted Euclidean distance.
[0013] In some embodiments, determining the at least core failure unit according to the local area absolute density and the minimum weighted Euclidean distance includes:
[0014] The local area absolute density and the minimum weighted Euclidean distance of each failure unit in all the failure units are compared, and the failure unit corresponding to which the local area absolute density and the minimum weighted Euclidean distance are both greater than preset values is taken as the core failure unit.
[0015] In some embodiments, the step of determining, based on each failed unit cluster, a strategy for performing enhanced repair on the area around the failed unit cluster includes:
[0016] Determining whether to perform enhanced repair around each failed unit cluster according to the number of failed units contained in each failed unit cluster;
[0017] When determining to perform enhanced repair around the failed unit clusters, the direction and quantity of enhanced repair around the failed unit clusters are determined according to the morphology of each failed unit in each failed unit cluster.
[0018] In some embodiments, the determining whether to perform enhanced repair around each failed unit cluster according to the number of failed units contained in the failed unit cluster includes:
[0019] When the number of failed units included in the failed unit cluster is greater than a first threshold, it is determined to perform enhanced repair on the area around the failed unit cluster.
[0020] In some embodiments, the location information includes coordinate parameters, and the failed test item information includes the number of the failed test item;
[0021] According to the attribute information, the weighted Euclidean distance between any two failed units is calculated, including:
[0022] Respectively obtaining weighted values corresponding to the position information and the failed test item information;
[0023] The weighted Euclidean distance between any two failed units is calculated according to the weighted value, the coordinate parameters corresponding to the any two failed units, and the numbers of the test items that the any two failed units have failed.
[0024] In some embodiments, the weighted Euclidean distance wd between any two failed units is AB The calculation formula is:
[0025]
[0026] Among them, g A and g B is the number of the test items of any two failed units; A and x B is the horizontal coordinate of any two failed units; A and B are the ordinates of the two arbitrary failed units; w1, w2, w3 are the weighted values of the abscissa, ordinate, and test item number of the two arbitrary failed units, respectively, w1+w2+w3=1.
[0027] In some embodiments, the local area absolute density of each of the failure units is i The calculation formula is:
[0028]
[0029] Among them, d ij represents the weighted Euclidean distance between two failed units, d c is the search radius; The judgment function is,
[0030] In some embodiments, the test items include: functional test, current and voltage test, and margin test.
[0031] In some embodiments, the object to be analyzed is a die on a dynamic random access memory chip.
[0032] The embodiment of the present disclosure also provides a repair device for a failed unit, comprising:
[0033] An acquisition unit, used to acquire attribute information of all failed units of the object to be analyzed in the target test phase, wherein the attribute information includes location information and information of failed test items;
[0034] A determination unit, configured to determine at least one core failure unit according to the attribute information of the failure unit;
[0035] A processing unit, used to take the core failure unit as a cluster center and assign the remaining failure units to corresponding failure unit clusters;
[0036] The determination unit is further used to determine, according to each failed unit cluster, a strategy for performing enhanced repair on the periphery of the failed unit cluster, wherein the enhanced repair is to repair the potential failed unit by using a spare circuit.
[0037] The present disclosure also provides a repair system for a failed unit, including:
[0038] A repair unit, used to detect failed units of the object to be analyzed and feed back the detection results to the controller;
[0039] The controller analyzes the detection results and determines the enhanced repair strategy by executing the repair method described in the above embodiment of the present disclosure, and sends instructions to the repair unit, instructing the repair unit to repair the potential failure unit according to the enhanced repair strategy.
[0040] The embodiments of the present disclosure also provide a computer-readable storage medium, and when the computer program is executed by a processor, the steps of the repair method described in the above embodiments of the present disclosure are implemented.
[0041] The method for repairing a failed unit provided by the embodiment of the present disclosure obtains the failed unit information of the object to be analyzed in the target test phase, and then uses the attribute information of the failed unit of the object to be analyzed to determine at least one core failed unit; then uses the determined at least one core failed unit as the cluster center, and assigns the remaining failed units to the corresponding failed unit cluster cluster; finally, based on each failed unit cluster cluster, determines a strategy for enhanced repair of the surroundings of the failed unit cluster cluster, wherein the enhanced repair is to repair the potential failed units using a spare circuit; in this way, the occurrence of new failed units in the test process of the later test phase can be reduced, the manufacturing yield of the chip can be improved, and the cost of repair can be reduced, thereby increasing the service life of the chip and improving the corporate image. BRIEF DESCRIPTION OF THE DRAWINGS
[0042] Figure 1 A schematic diagram of the structure of a wafer provided in an embodiment of the present disclosure;
[0043] Figure 2 A schematic diagram of the structure of a repair system for a failed unit provided in an embodiment of the present disclosure;
[0044] Figure 3 A schematic diagram of an implementation flow of a method for repairing a failed unit provided in an embodiment of the present disclosure;
[0045] Figure 4 A schematic diagram of an implementation flow of determining a core failure unit provided in an embodiment of the present disclosure;
[0046] Figure 5 A schematic diagram of classifying failed units through cluster analysis provided in an embodiment of the present disclosure;
[0047] Figure 6 A schematic diagram of a grain structure including at least two failure units provided in an embodiment of the present disclosure;
[0048] Fig. 7A A schematic diagram of a grain structure corresponding to a search radius provided in an embodiment of the present disclosure;
[0049] Figure 7B A schematic diagram of a grain structure corresponding to another search radius provided in an embodiment of the present disclosure;
[0050] Figure 8 A schematic diagram of a grain structure including a plurality of clusters provided in an embodiment of the present disclosure;
[0051] Fig. 9 A schematic diagram of a grain structure with different core failure units provided in an embodiment of the present disclosure;
[0052] Fig.10 A schematic diagram of an implementation flow of a strategy for determining enhanced repair provided in an embodiment of the present disclosure;
[0053] Fig.11 A schematic diagram of the relationship between the number of different grains and failed units provided in an embodiment of the present disclosure;
[0054] Fig.12 A schematic diagram of a grain with a standard repair area and an enhanced repair area provided in an embodiment of the present disclosure;
[0055] Fig.13A A schematic diagram of repairing a failed unit on a die after standard repair provided by an embodiment of the present disclosure;
[0056] Fig. 13B A schematic diagram of repairing a failed unit on a die after enhanced repair provided by an embodiment of the present disclosure;
[0057] Fig.14 A schematic diagram of the composition structure of a failed unit repair device provided in an embodiment of the present disclosure.
[0058] In the above drawings (which are not necessarily drawn to scale), similar reference numerals may describe similar components in different views. Similar reference numerals with different letter suffixes may represent different examples of similar components. The accompanying drawings generally illustrate various embodiments discussed herein by way of example and not limitation. DETAILED DESCRIPTION
[0059] In order to make the technical solutions and advantages of the embodiments of the present disclosure clearer, the technical solutions of the present disclosure will be further described in detail below in conjunction with the accompanying drawings and embodiments. Although the exemplary implementation methods of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the implementation methods described herein. On the contrary, these implementation methods are provided in order to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.
[0060] The present disclosure is described in more detail in the following paragraphs by way of example with reference to the accompanying drawings. The advantages and features of the present disclosure will become more apparent from the following description and claims. It should be noted that the drawings are in very simplified form and in non-precise proportions, and are only used to conveniently and clearly assist in explaining the purpose of the embodiments of the present disclosure.
[0061] It will be understood that the meaning of “on,” “over,” and “over” in this disclosure should be interpreted in the broadest manner, so that “on” not only means that it is “on” something with no intervening features or layers (i.e., directly on something), but also includes the meaning that it is “on” something with intervening features or layers.
[0062] Additionally, for ease of description, spatially relative terms such as "on," "over," "above," "upper," and the like may be used herein to describe the relationship of one element or feature to another element or feature as illustrated in the figures. The spatially relative terms are intended to encompass different orientations of the device in use or operation in addition to the orientation depicted in the figures. The device may be otherwise oriented (rotated 90 degrees or at other orientations) and the spatially relative descriptors used herein may be interpreted accordingly.
[0063] In the disclosed embodiments, the term "substrate" refers to the material on which subsequent material layers are added. The substrate itself can be patterned. The material added on top of the substrate can be patterned or can remain unpatterned. In addition, the substrate can include a variety of semiconductor materials, such as silicon, silicon germanium, germanium, gallium arsenide, indium phosphide, etc. Alternatively, the substrate can be made of non-conductive materials, such as glass, plastic, or sapphire wafers.
[0064] In the disclosed embodiments, the term "layer" refers to a material portion including an area with a thickness. A layer may extend over the entirety of a lower or upper structure, or may have a range that is smaller than the range of a lower or upper structure. In addition, a layer may be a region of a homogeneous or inhomogeneous continuous structure having a thickness that is less than the thickness of a continuous structure. For example, a layer may be located between the top and bottom surfaces of a continuous structure, or a layer may be between any horizontal faces at the top and bottom surfaces of a continuous structure. A layer may extend horizontally, vertically, and / or along an inclined surface. A layer may include multiple sublayers. For example, an interconnect layer may include one or more conductors and contact sublayers (where interconnects and / or via contacts are formed), and one or more dielectric sublayers.
[0065] In the embodiments of the present disclosure, the terms "first", "second", etc. are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence.
[0066] The repair method involved in the embodiments of the present disclosure can be used to repair a variety of memory chips or logic chips. In some embodiments, the memory chip can be a volatile memory chip, such as a dynamic random access memory (DRAM) chip and a static random access memory chip, or can be a non-volatile memory chip, such as a phase change random access memory chip, a magnetoresistive random access memory chip, a ferroelectric random access memory chip, and a resistive random access memory chip. The logic chip can be, for example, a microprocessor, a field programmable logic gate, or a digital signal processor. The type of chip is not limited in the embodiments of the present disclosure.
[0067] The chip manufacturing process can be divided into the front-end process and the back-end process. Among them, the front-end process is the process for the entire wafer, which may include wafer manufacturing and wafer testing (the test object is for each die in the entire wafer, the purpose is to ensure that each die in the entire wafer can basically meet the characteristics or design specifications of the device, which may include voltage, current, timing and function verification). Figure 1 , Figure 1 A schematic diagram of the structure of a wafer provided by an embodiment of the present disclosure is shown; Figure 1As shown, the wafer 100 may include multiple crystal grains 101. When the wafer 100 is in the circuit probe test (CP Test) stage, different failure units 102 (failure units are one or more failed storage units in the crystal grains) will be obtained according to different test conditions, and then the failed units will be repaired using a backup circuit. As to how to detect the failed units, the embodiments of the present disclosure do not limit it. For example, during wafer testing, automatic test equipment (ATE) can be used to perform probe testing on each crystal grain on the wafer under different test conditions. And, the back-end process is the process of slicing and packaging multiple crystal grains on the wafer to form chips, which may include various processes such as packaging, final testing and finished product warehousing.
[0068] In the manufacturing process of the above-mentioned chip, the method of repairing the failed units on the wafer can be used to improve the manufacturing yield of the chip. For example, each failed unit on the wafer can be repaired by a method such as a fuse process to adjust one or more performance parameters of the wafer, so that the performance parameters of each failed unit that has been repaired are within the corresponding specification range, thereby improving the manufacturing yield of the chip. However, the current chip repair method has certain limitations, and the chip yield still needs to be further improved.
[0069] Specifically, through research, it is found that after executing the circuit probe test, the wafer will still be subjected to the final test (FT), and new failure units (NFB) (also known as unknown failure units (UFB), potential failure units, etc.) will appear in the final test stage. For example, before shipment, the new failure units found in the endurance test (RDBI) of the final test stage need to be repaired again. At this time, the repair cost increases. If the number of new failure units found is too large, it may lead to an inability to repair, and finally cause the chip yield to drop significantly. For example, after shipment, the customer's product fails to pass the internal test or the chip fails during the normal service life cycle, that is, the product quality is poor, which increases the company's after-sales cost and corporate image. Serious consequences may lead to the loss of core customers.
[0070] Based on one or more of the above problems, the embodiments of the present disclosure provide a method, a device, a system and a storage medium for repairing a failed unit, wherein the repair method can be applied to chip manufacturing scenarios, illustratively, can be applied to the specific scenario of wafer circuit probe testing, or can also be the repair stage after wafer testing, without specific limitation. In the embodiments of the present disclosure, the potential failed units in the object to be analyzed are determined by obtaining the attribute information of the failed units in the object to be analyzed, that is, the new failed units are predicted in advance, and the predicted new failed units (that is, the aforementioned potential failed units) are repaired, thereby further improving the manufacturing yield of the chip.
[0071] In order to facilitate the overall understanding of the technical solution provided by the embodiment of the present disclosure, the embodiment of the present disclosure also provides a repair system for a failed unit. Next, the repair system for a failed unit provided by the embodiment of the present disclosure is described. Figure 2 As shown, Figure 2 The present invention provides a schematic diagram of the structure of a repair system for a failed unit provided in an embodiment of the present invention; the repair system for a failed unit includes:
[0072] The repair unit 201 is used to detect the failed unit of the object to be analyzed and feed back the detection results to the controller; the controller 202 analyzes the detection results and determines the enhanced repair strategy by executing the repair method described in each embodiment of the present disclosure, and issues instructions to the repair unit to instruct the repair unit to repair the potential failed unit according to the enhanced repair strategy. Among them, the repair unit may include a failed unit detection module on the machine, and may also include a module that specifically performs the repair operation. The controller can be a device on the machine for controlling and directing the input, operation and processing of calculation results of programs and data, and can also be other processing devices that are communicatively connected to the chip manufacturing equipment.
[0073] It should be noted that the position of the failed unit tested in the target test phase is repaired, and this disclosure refers to this type of repair as standard repair. During the repair process, the position of the potential failed unit that is tested in the non-target test phase but predicted by the repair method of the failed unit provided by the embodiment of the present disclosure will also be repaired. The reason is that the storage units at these positions are likely to be judged as failed units in subsequent endurance tests and other tests in the back-end test phase. This type of repair is called predictive repair (Predict Repair) or enhanced repair in the present disclosure. In the e-Fuse (one-time programmable memory) fuse stage, the actual position repair is performed according to the redundancy allocation (RA) result. Among them, redundancy allocation is the allocation of spare circuits using a repair algorithm, and the spare circuits are used to repair failed units and potential failed units. Standard repair is performed on failed units, and predictive repair is performed on potential failed units.
[0074] The present disclosure provides a method for repairing a failed unit, which can be performed by any electronic device with computing processing capabilities. For example, the method can be performed by, for example, semiconductor manufacturing equipment, or other processing equipment that is communicatively connected to the chip manufacturing equipment, without specific limitation.
[0075] Specifically, refer to Figure 3 , Figure 3 FIG. 1 is a schematic diagram of an implementation flow of a method for repairing a failed unit in an embodiment of the present disclosure, such as Figure 3 As shown, the method for repairing a failed unit provided in the embodiment of the present disclosure may include the following steps:
[0076] Step S301: Acquire attribute information of all failed units of the object to be analyzed in the target test phase, wherein the attribute information includes location information and information of failed test items;
[0077] Step S302: determining at least one core failure unit according to the attribute information of the failure unit;
[0078] Step S303: taking the core failure unit as the cluster center, and assigning the remaining failure units to corresponding failure unit clusters;
[0079] Step S304: according to each failed unit cluster, determine a strategy for performing enhanced repair on the periphery of the failed unit cluster, wherein the enhanced repair is to repair the potential failed units by using a spare circuit.
[0080] In step S301, the object to be analyzed may be a wafer, a crystal grain, or a certain area on a crystal grain / wafer. Exemplarily, the object to be analyzed may be multiple crystal grains of the same type of product. For example, in the following embodiments, the object to be analyzed is illustrated as a crystal grain on a dynamic random access memory chip, but the present disclosure is not limited thereto. Each crystal grain includes multiple storage units (cells).
[0081] In the disclosed embodiment, the target test phase may be any phase of testing the object to be analyzed, for example, the target test phase may be a circuit probe test phase, or may be one or more test items in the circuit probe test phase. In the circuit probe test phase, a plurality of test items (items) may be tested on the die. The failed unit may be a storage unit that has at least one test item that has not passed or failed the test during the test process of the circuit probe test phase. The attribute information of the failed unit may be relevant information of the failed unit, for example, the attribute information of the failed unit may be the location information of the failed unit (failed storage unit) on the die / wafer, and the test item information that caused the failed unit to be judged as failed in the test.
[0082] In step S302, at least one core failure unit is determined according to the acquired attribute information of the failure unit.
[0083] In some embodiments, reference Figure 4 , Figure 4 The present invention provides a schematic diagram of an implementation process of determining a core failure unit provided in an embodiment of the present invention; wherein, determining at least one core failure unit according to the attribute information of the failure unit comprises the following steps:
[0084] Step S401: Calculate the weighted Euclidean distance between any two failed units according to the attribute information;
[0085] Step S402: obtaining a search radius, where the search radius is related to the size of the minimum repair area of the object to be analyzed;
[0086] Step S403: for each of the failure units, using the search radius and the weighted Euclidean distance between the corresponding failure unit and each remaining failure unit, calculate the local area absolute density of the corresponding failure unit and the minimum weighted Euclidean distance between the corresponding failure unit and the remaining failure units;
[0087] Step S404: determining at least the core failure unit according to the local area absolute density and the minimum weighted Euclidean distance.
[0088] It should be noted that, in order to achieve the above content, in the embodiment of the present disclosure, the number, classification, similarity, etc. of the failed units can be obtained through cluster analysis; for example, Figure 5 As shown, Figure 5 A schematic diagram of classifying failure units through cluster analysis provided in an embodiment of the present disclosure; wherein squares, circles, and triangles represent the categories to which three types of failure units belong, respectively, and the failure units of the object to be analyzed are classified through cluster analysis. Exemplarily, through cluster analysis, the failure units of the object to be analyzed are calculated, for example, by calculating the Euclidean distance, cosine similarity, etc. of the failure units to determine the similarity of the failure units.
[0089] It should be noted that cluster analysis is an unsupervised learning method in machine learning, which can classify objects to be analyzed with similar properties into one category, and cluster analysis does not require prior knowledge for training. Cluster analysis can identify cluster structures of arbitrary shape distribution, and the number of clusters does not need to be determined in advance. The disclosed embodiment uses the clustering method in machine learning to classify failed units with similar properties, which can dig out complex rules and improve the accuracy of determining potential failed units, thereby further improving the manufacturing yield of the chip. Specifically, in step S401, according to the attribute information of the failed unit, cluster analysis is used to calculate the weighted Euclidean distance between any two failed units. Here, the weighted Euclidean distance is determined based on the coordinates and weighted values of the failed unit. Among them, the attribute information of the failed unit is the location information of the failed unit and the test item information that the failed unit has not passed.
[0090] In some embodiments, the location information includes coordinate parameters, and the failed test item information includes the number of the failed test item;
[0091] According to the attribute information, the weighted Euclidean distance between any two failed units is calculated, including:
[0092] Respectively obtaining weighted values corresponding to the position information and the failed test item information;
[0093] The weighted Euclidean distance between any two failed units is calculated according to the weighted value, the coordinate parameters corresponding to the any two failed units, and the numbers of the test items that the any two failed units have failed.
[0094] Exemplarily, the location information of the failed unit may be an X coordinate and a Y coordinate. In other embodiments, the location information may be represented by polar coordinates, or other distance measurement methods such as cosine distance may be used, which will not be described in detail here.
[0095] In some embodiments, the test items may include: functional test (such as read / write function test), current and voltage test (DC test), margin test (Margin test, wherein the margin test may be used to measure time-related information). It should be understood that if different failed units fail in the same category of test items, the similarity between them is higher.
[0096] refer to Figure 6 , Figure 6 A schematic diagram of a grain structure including at least two failure units provided in an embodiment of the present disclosure; wherein: Figure 6 , a wafer including at least some of the crystal grains is shown, 601 is used to represent any crystal grain on the wafer, 602 is used to represent any failure unit on the crystal grain, and k represents the distance of a bit; wherein, the two failure units A and B shown in the figure are both isolated failure units, and there is no adjacent failure unit around the isolated failure unit. Any two failure units may also be adjacent non-isolated failure units, and only failure units A and failure units B are used as examples for illustration, which is not used to limit the scope of the present disclosure.
[0097] In some embodiments, the weighted Euclidean distance wd between any two failed units is AB The calculation formula is:
[0098]
[0099] Among them, g A and g B is the number of the test items of any two failed units; A and x B is the horizontal coordinate of any two failed units; A and B are the ordinates of the two arbitrary failed units; w1, w2, w3 are the weighted values of the abscissa, ordinate, and test item number of the two arbitrary failed units, respectively, w1+w2+w3=1.
[0100] Exemplary, reference Figure 6 , the coordinates of the failed unit A are (x A ,y A ), the coordinates of the failed unit B are (x B ,y B ), g A is the number of the test item of the failed unit A, g B is the number of the test item of the failed unit B, w1 is the weighted value corresponding to the difference between the horizontal coordinates of the failed unit A and the failed unit B, w2 is the weighted value corresponding to the difference between the vertical coordinates of the failed unit A and the failed unit B, and w3 is the weighted value corresponding to the number of the test item of the failed unit A and the failed unit B.
[0101] In some embodiments, the numbers of the test items of the failed unit A and the failed unit B (g A -g B ) can be obtained according to the following formula (2):
[0102]
[0103] Among them, D subdomain is the width of the minimum repair area designed for the chip, which can be set according to the actual situation, and k is an integer; for example, D subdomain Select 1024; k is generally selected as 16.
[0104] For example, the coordinates of the failed unit A are (3, 2), the coordinates of the failed unit B are (4, 4), and D subdomain =1024; k = 16; w1 = w2 = w3 = 1 / 3, then the weighted Euclidean distance wd between the failed unit A and the failed unit B AB The value of is about 37.
[0105] It should be noted that when the two selected failure units are the same, the conventional Euclidean distance calculation formula The value of the Euclidean distance between the two failed units (about 64) is greater than the value of the weighted Euclidean distance between the two failed units obtained by the weighted Euclidean distance calculation formula (formula (1)); it should be understood that if the value of the Euclidean distance is larger, the clustering effect is coarser and the accuracy of the clustering result is lower. The weighted Euclidean distance obtained by the weighted Euclidean distance calculation formula in the embodiment of the present disclosure is smaller, the clustering effect is fine, and the clustering result accuracy is higher, which is conducive to improving the accuracy of determining the potential failed units in the object to be analyzed.
[0106] In addition, in the embodiment of the present disclosure, the weighted Euclidean distance formula is used, that is, the weighted Euclidean distance between two failure units is calculated by adding a weighted value to the formula, which can better adjust the similarity between the failure units; at the same time, for different products, the three attribute information (such as the horizontal coordinate, vertical coordinate, and test items of the failed test) can be adjusted separately in the role of the failure unit in the cluster analysis to improve the accuracy of determining potential failure units.
[0107] It should be noted that subdomain is a concept in the repair process of dynamic random access memory chips. It is related to circuit design. The entire storage cell area of a die can be divided into multiple subdomains, that is, the large storage cell area is divided into multiple small storage cell areas (subdomains). The spare circuit allocated to each subdomain is related to the subdomain.A -g B The value setting is also related to the sub-region. A -g B If the value of g is set too large, the similarity between different failure units is too small, and abnormal values (such as noise points) will appear in the subsequent method search; A -g B If the value is set too small, the similarity between failed units that failed different test items is too high, and the clustering result is unreasonable.
[0108] In step S402, a search radius is obtained, where the search radius is related to the size of the minimum repair area of the object to be analyzed.
[0109] As mentioned above, reference Fig. 7A , Figure 7B , Fig. 7A is a schematic diagram of a grain structure corresponding to a search radius provided in an embodiment of the present disclosure, Figure 7B FIG7 is a schematic diagram of a grain structure corresponding to another search radius provided in an embodiment of the present disclosure, FIG7 and FIG Figure 7B The "X" shown in the figure represents a failed unit; if the selected search radius d c Too big (such as Fig. 7A As shown in the figure, all failed units are of the same category). When searching, more failed units will be found, the clustering effect will be rough, the clustering result accuracy will be low, and more categories of data cannot be obtained. c Too small (such as Figure 7B As shown in the figure, all failed units are divided into three categories, and the number of failed units in each category is small), there will be many failed units that will not be assigned to specific categories, that is, the similarity between failed units that failed different test items is too high, and the clustering result is unreasonable. Based on this, the search radius d c It is important to choose a reasonable setting.
[0110] The search radius d provided in the embodiment of the present disclosure c The calculation formula is as follows: Formula (3).
[0111] d c =a*D subdomain (3)
[0112] Among them, a is the search range control coefficient, D subdomain The width of the minimum repair area designed for the chip. In some specific embodiments, the search range control coefficient α ranges from [0, 1], for example, the value of a is 1 / 4. Exemplarily, D subdomain =1024; α = 1 / 4; then d c= 256. The search range control coefficient α in the embodiment of the present disclosure is to facilitate the control of the clustering effect in data mining, which is conducive to improving the accuracy of determining the potential failure units in the object to be analyzed.
[0113] In step S403, for each failed unit among all the failed units, the local area absolute density of the corresponding failed unit and the minimum weighted Euclidean distance between the corresponding failed unit and the remaining failed units are calculated using the search radius and the weighted Euclidean distance between the corresponding failed unit and each remaining failed unit.
[0114] Among them, after obtaining the attribute information of all failed units, the weighted Euclidean distance between the failed unit and other remaining failed units except the failed unit is determined according to the attribute information of each failed unit, and the local area absolute density of the failed unit is calculated according to the weighted Euclidean distance of the failed unit and the search radius; and then the minimum weighted Euclidean distance is determined from multiple weighted Euclidean distances through calculation and comparison.
[0115] In some embodiments, the local absolute density of each failed unit is i The calculation formula is:
[0116]
[0117] Among them, d ij represents the weighted Euclidean distance between two failed units, d c is the search radius; The judgment function is,
[0118] Here, d ij represents the weighted Euclidean distance between failed unit i and failed unit j; where failed unit i is the current failed unit, and failed unit j is any one of the remaining failed units except failed unit i. The local absolute density ρ of failed unit i i It can be obtained according to the above formula (4).
[0119] According to the local absolute density ρ of the failed unit i i , determine the minimum weighted Euclidean distance υ between the failure units with greater absolute density i ; Among them, the minimum weighted Euclidean distance υ i The calculation formula is as follows: Formula (5).
[0120]
[0121] In other words, the weighted Euclidean distance between the determined failed unit i and the failed unit j, and the search radius d of the current failed unit i cAfter that, the absolute density ρ of the local area of the failed unit i can be determined by calculating and comparing formula (4) and formula (5): i And the minimum weighted Euclidean distance υ among all weighted Euclidean distances corresponding to the failed unit i i It should be noted that the minimum weighted Euclidean distance υ of the failed unit i is calculated by formula (5): i When the failure unit j is the local absolute density ρ of the remaining failure units except the failure unit i j The absolute density of the local area of the failed unit i is ρ i Large failure unit.
[0122] In step S404, the at least core failure unit is determined according to the local area absolute density and the minimum weighted Euclidean distance.
[0123] In some embodiments, determining the at least core failure unit according to the local area absolute density and the minimum weighted Euclidean distance includes:
[0124] The local area absolute density and the minimum weighted Euclidean distance of each failure unit in all the failure units are compared, and the failure unit corresponding to which the local area absolute density and the minimum weighted Euclidean distance are both greater than preset values is taken as the core failure unit.
[0125] Here, the preset value includes a first preset value and a second preset value.
[0126] Specifically, refer to Figure 8 , Figure 8 is a schematic diagram of a grain structure including multiple clusters provided in an embodiment of the present disclosure; the absolute densities ρ of all local areas corresponding to all failed units are obtained i Compare and obtain the minimum weighted Euclidean distance υ of all failed units i For comparison, the absolute density ρ in all local areas i The failure unit corresponding to the local area absolute density greater than the first preset value R1 in the above example, and the minimum weighted Euclidean distance v of the failure unit i When the value is greater than the second preset value R2 of the failure unit, the failure unit is regarded as a core failure unit. The number of core failure units can be one or more. The first preset value R1 and the second preset value R2 can be selected and set according to actual conditions.
[0127] In step S303, the core failure unit is used as the cluster center, and the remaining failure units are assigned to corresponding failure unit clusters.
[0128] Among them, the core failure unit can be understood as the core of the failure unit block area at the clustering method level. Its actual physical condition may be the essential reason for the failure of this area, so the failure units in this area are classified into the same cluster (category) as the core failure units. Figure 8 , multiple failure units (such as 1, 2, 3, ..., 11) are shown, among which failure units 9, 10, and 11 are core failure units. Here, failure units 9, 10, and 11 are used as cluster centers, and failure units 1, 2, 3, 4, 5, 6, 7, and 8 are assigned to corresponding failure unit clusters. The clusters can be understood as classifying multiple failure units according to different conditions, and each category is a cluster.
[0129] refer to Fig. 9 , Fig. 9 A schematic diagram of a grain structure with different core failure units provided in an embodiment of the present disclosure; wherein 901 is used to represent a grain, 902 is used to represent a core failure unit E, 903 is used to represent a core failure unit F, 904 is used to represent a non-core failure unit, and 905 is used to represent an isolated failure unit. The core failure unit E and the core failure unit F belong to different failure unit clusters, respectively, and multiple non-core failure units 904 belong to corresponding failure unit clusters.
[0130] It should be noted that the clustering algorithm used in the embodiment of the present disclosure can identify clusters of any shape distribution, and the number of clusters does not need to be predetermined. The failure units of the same cluster are combined with the test information of the chip, so that the similarity measurement between the failure units is more accurate. In addition, the search radius proposed in the present disclosure can better mine potential clustered failure units.
[0131] In step S304, according to each failed unit cluster, a strategy for performing enhanced repair on the periphery of the failed unit cluster is determined, wherein the enhanced repair is to repair the potential failed units by using a spare circuit.
[0132] In some embodiments, reference Fig.10 , Fig.10 The present invention provides a schematic diagram of an implementation flow of a strategy for determining enhanced repair provided in an embodiment of the present invention; wherein, according to each failed unit cluster, determining a strategy for enhanced repair around the failed unit cluster includes the following steps:
[0133] Step S1001: determining whether to perform enhanced repair on the periphery of each failed unit cluster according to the number of failed units contained in the failed unit cluster;
[0134] Step S1002: when determining to perform enhanced repair on the periphery of the failed unit cluster, determine the direction and quantity of enhanced repair on the periphery of the failed unit cluster according to the morphology of each failed unit in each failed unit cluster.
[0135] In step S1001, the determining whether to perform enhanced repair on the periphery of each failed unit cluster according to the number of failed units contained in the failed unit cluster includes:
[0136] When the number of failed units included in the failed unit cluster is greater than a first threshold, it is determined to perform enhanced repair on the area around the failed unit cluster.
[0137] In other words, reference Fig.11 , Fig.11 A schematic diagram of the relationship between the number of different grains and failed units provided in an embodiment of the present disclosure; wherein, in different grains, after determining the core failed unit and the failed unit cluster to which it belongs, the number of failed units in each failed unit cluster is determined, and when the number of failed units in each failed unit cluster is greater than a first threshold value H, it is determined to perform enhanced repair around the failed unit cluster; in some specific embodiments, the value range of the first threshold value H is 5-10. When the number of failed units in each failed unit cluster is less than the first threshold value H, it is determined not to perform enhanced repair around the failed unit cluster, such as isolating the failed unit.
[0138] It should be noted that when the number of failed units in a die is greater than the first threshold value H, other defects may easily be caused, for example, the die may fail to pass other test items.
[0139] In addition, it should be noted that the reference Fig. 9 , the distance between the isolated failure unit 905 and its nearest core failure unit is too large (or greater than a certain threshold), and in the clustering process, the isolated failure unit 905 will not be classified into a specific category.
[0140] In step S1002, the failed units in the failed unit clusters determined to be enhanced and repaired are divided, and the morphology formed by each failed unit in each failed unit cluster is different, and the corresponding enhanced repair methods are different, that is, the enhanced repair strategies are different; for example, in the morphology formed by each failed unit in the failed unit cluster, the unit distance in the row direction is greater than the unit distance in the column direction, then the enhanced repair method widens the boundary of the failed unit cluster in the row direction by several unit distances in the column direction; similarly, in the morphology formed by each failed unit in the failed unit cluster, the bit distance in the column direction is greater than the unit distance in the row direction, then the enhanced repair method widens the boundary of the failed unit cluster in the column direction by several unit distances in the row direction. Here, 1 unit distance can be equal to 1 bit distance. The specific widening amount can be determined based on the difference between the total number of failed units and the first threshold, or it can be widened according to the inherent specified widening amount based on experience.
[0141] It should be noted that repairing failed units is standard repair, and repairing the area widened beyond the boundaries of failed units is enhanced repair.
[0142] Next, after determining the enhanced repair strategy, the potential failure unit is repaired using the spare circuit. The spare circuit in the disclosed embodiment includes a row-direction spare circuit and a column-direction spare circuit. If the row-direction spare circuit is used, two units (bits) can be added above and below the standard repair area for repair; similarly, if the column-direction spare circuit is used, two units (bits) can be added to the left and right of the standard repair area for repair.
[0143] Exemplary, reference Fig.12 , Fig.12 FIG. 1 shows a schematic diagram of a grain of a standard repair area and an enhanced repair area provided by an embodiment of the present disclosure; Fig.12 It can be seen that the standard repair area 1201 of the die 1200 is enhanced by two rows on the left and right in the row direction, that is, the enhanced repair area 1202. It should be understood that if the characteristic information of the failed unit in the object to be analyzed does not meet the enhanced repair rule, such as an isolated failed unit, only standard repair is required.
[0144] The disclosed embodiment determines the potential failure units of the object to be analyzed by enhancing the repair strategy, which is conducive to repairing more potential failure units, can reduce the occurrence of potential failure units during the testing process in the later testing stage, improve the manufacturing yield of the chip, and then reduce the cost of repair, thereby increasing the service life of the chip and improving the corporate image.
[0145] Exemplary, reference Fig.13A and Fig. 13B , Fig.13AA schematic diagram of repairing a failed unit on a die after standard repair provided by an embodiment of the present disclosure, Fig. 13B A schematic diagram of repairing a failed unit on a die after enhanced repair is provided in an embodiment of the present disclosure; wherein 1301 is used to represent all failed units detected by the object to be analyzed in the target test phase, 1302, 1303, and 1304 are all used to represent potential failed units detected by the object to be analyzed in the durability test phase or other test phases, 1201 is used to represent a standard repair area, and 1202 is used to represent an enhanced repair area. Fig.13A and Fig. 13B It can be seen that by performing enhanced repair on the object to be analyzed, the potential failure units detected in the durability test phase or other test phases of the object to be analyzed can be predicted and repaired in advance. In this way, the yield of the chip can be improved, the cost of repair can be reduced, and the service life of the chip can be increased.
[0146] In some embodiments, the object to be analyzed is a die on a dynamic random access memory chip.
[0147] Based on the above repair method, the present disclosure provides a repair device for a failed unit, referring to Fig.14 , Fig.14 The present invention is a schematic diagram of the composition structure of a repair device for a failed unit provided in an embodiment of the present invention; wherein the repair device for a failed unit comprises:
[0148] The obtaining unit 1401 is used to obtain the attribute information of all failed units of the object to be analyzed in the target test phase, wherein the attribute information includes location information and information of failed test items;
[0149] A determination unit 1402, configured to determine at least one core failure unit according to the attribute information of the failure unit;
[0150] The processing unit 1403 is used to use the core failure unit as a cluster center and assign the remaining failure units to corresponding failure unit clusters;
[0151] The determining unit 1402 is further configured to determine, based on each failed unit cluster, a strategy for performing enhanced repair on the periphery of the failed unit cluster, wherein the enhanced repair is to repair potential failed units using a spare circuit.
[0152] In some embodiments, the determining unit is specifically configured to:
[0153] Calculating the weighted Euclidean distance between any two failed units based on the attribute information;
[0154] Acquire a search radius, where the search radius is related to the size of the minimum repair area of the object to be analyzed;
[0155] For each of the failed units among all the failed units, using the search radius and the weighted Euclidean distance between the corresponding failed unit and each remaining failed unit, calculate the local area absolute density of the corresponding failed unit and the minimum weighted Euclidean distance between the corresponding failed unit and the remaining failed units;
[0156] The at least core failure unit is determined according to the local area absolute density and the minimum weighted Euclidean distance.
[0157] In some embodiments, the determining unit is further configured to:
[0158] The local area absolute density and the minimum weighted Euclidean distance of each failure unit in all the failure units are compared, and the failure unit corresponding to which the local area absolute density and the minimum weighted Euclidean distance are both greater than preset values is taken as the core failure unit.
[0159] In some embodiments, the determining unit is further configured to:
[0160] Determining whether to perform enhanced repair around each failed unit cluster according to the number of failed units contained in each failed unit cluster;
[0161] When determining to perform enhanced repair on the periphery of the failed unit clusters, a specific method of performing enhanced repair on the periphery of the failed unit clusters is determined according to the morphology formed by each failed unit in each failed unit cluster.
[0162] In some embodiments, the determining unit is further configured to:
[0163] When the number of failed units included in the failed unit cluster is greater than a first threshold, it is determined to perform enhanced repair on the area around the failed unit cluster.
[0164] The embodiment of the present disclosure also provides a computer-readable storage medium, and when the computer program is executed by a processor, the steps in the repair method described in the above embodiment of the present disclosure are implemented.
[0165] The protection scope of the present disclosure is not limited thereto, and any person skilled in the art who is familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present disclosure, which should be included in the protection scope of the present disclosure. Therefore, the protection scope of the present disclosure should be based on the protection scope of the claims.
Claims
1. A method for repairing a failed unit, characterized in that: include: Acquire attribute information of all failed units of the object to be analyzed in the target test phase, wherein the attribute information includes location information and information of failed test items; Determine at least one core failure unit according to the attribute information of the failure unit; The core failure unit is used as the cluster center, and the remaining failure units are assigned to the corresponding failure unit clusters; According to each failed unit cluster, a strategy for performing enhanced repair on the periphery of the failed unit cluster is determined, wherein the enhanced repair is to repair the potential failed units by using a spare circuit.
2. The repair method according to claim 1, characterized in that: The step of determining at least one core failure unit according to the attribute information of the failure unit includes: Calculating the weighted Euclidean distance between any two failed units based on the attribute information; Acquire a search radius, where the search radius is related to the size of the minimum repair area of the object to be analyzed; For each of the failed units among all the failed units, using the search radius and the weighted Euclidean distance between the corresponding failed unit and each remaining failed unit, calculate the local area absolute density of the corresponding failed unit and the minimum weighted Euclidean distance between the corresponding failed unit and the remaining failed units; The at least core failure unit is determined according to the local area absolute density and the minimum weighted Euclidean distance.
3. The repair method according to claim 2, characterized in that: Determining at least the core failure unit according to the local area absolute density and the minimum weighted Euclidean distance includes: The local area absolute density and the minimum weighted Euclidean distance of each failure unit in all the failure units are compared, and the failure unit corresponding to which the local area absolute density and the minimum weighted Euclidean distance are both greater than preset values is taken as the core failure unit.
4. The repair method according to claim 1, characterized in that: The step of determining, based on each failed unit cluster, a strategy for enhancing and repairing the area around the failed unit cluster, includes: Determining whether to perform enhanced repair around each failed unit cluster according to the number of failed units contained in each failed unit cluster; When determining to perform enhanced repair around the failed unit clusters, the direction and quantity of enhanced repair around the failed unit clusters are determined according to the morphology of each failed unit in each failed unit cluster.
5. The repair method according to claim 4, characterized in that: The determining whether to perform enhanced repair around each failed unit cluster according to the number of failed units contained in the failed unit cluster includes: When the number of failed units included in the failed unit cluster is greater than a first threshold, it is determined to perform enhanced repair on the area around the failed unit cluster.
6. The repair method according to claim 2, characterized in that: The location information includes coordinate parameters, and the failed test item information includes the number of the failed test item; According to the attribute information, the weighted Euclidean distance between any two failed units is calculated, including: Respectively obtaining weighted values corresponding to the position information and the failed test item information; The weighted Euclidean distance between any two failed units is calculated according to the weighted value, the coordinate parameters corresponding to the any two failed units, and the numbers of the test items that the any two failed units have failed.
7. The repair method according to claim 6, characterized in that: The weighted Euclidean distance wd between any two failed units AB The calculation formula is: Among them, g A and g B is the number of the test items of any two failed units; A and x B is the horizontal coordinate of any two failed units; A and B are the ordinates of the two arbitrary failed units; w1, w2, and w3 are the weighted values of the abscissa, ordinate, and number of the test items of the two arbitrary failed units, respectively, and w1+w2+w3=1.
8. The repair method according to claim 2, characterized in that: The local absolute density of each failure unit is i The calculation formula is: Among them, d ij represents the weighted Euclidean distance between two failed units, d c is the search radius; The judgment function is, 9. The repair method according to claim 1, characterized in that: The test items include: functional test, current and voltage test, and margin test.
10. The repair method according to any one of claims 1 to 9, characterized in that: The object to be analyzed is a die on a dynamic random access memory chip.
11. A repair device for a failed unit, characterized in that: include: An acquisition unit, used to acquire attribute information of all failed units of the object to be analyzed in the target test phase, wherein the attribute information includes location information and information of failed test items; A determination unit, configured to determine at least one core failure unit according to the attribute information of the failure unit; A processing unit, used to take the core failure unit as a cluster center and assign the remaining failure units to corresponding failure unit clusters; The determination unit is further used to determine, according to each failed unit cluster, a strategy for performing enhanced repair on the periphery of the failed unit cluster, wherein the enhanced repair is to repair the potential failed unit by using a spare circuit.
12. A repair system for a failed unit, characterized in that: include: A repair unit, used to detect failed units of the object to be analyzed and feed back the detection results to the controller; The controller analyzes the detection results and determines the enhanced repair strategy by executing the repair method described in any one of claims 1 to 10, and sends instructions to the repair unit, instructing the repair unit to repair the potential failure unit according to the enhanced repair strategy.
13. A computer-readable storage medium, characterized in that: When the computer program is executed by a processor, the steps of the repair method according to any one of claims 1 to 10 are implemented.