Information processing device, information processing system, information processing method, and information processing program

Abstract

This control device comprises: a first acquisition unit that acquires an event in which there is a possibility of the occurrence of a defect mode, with regard to a semiconductor chip on which a pin mark has been formed due to contact by a probe pin; a second acquisition unit that acquires the event occurrence time; a third acquisition unit that acquires a first pin mark image, which is an image resulting from imaging the semiconductor chip prior to the event occurrence time, and a second pin mark image, which is an image resulting from imaging the semiconductor chip after the event occurrence time; an association unit that associates the event in which there is a possibility of the occurrence of the defect mode with the set of the first pin mark image and the second pin mark image; and an output unit that, for each event in which there is a possibility of the occurrence of the defect mode, outputs the set of the first pin mark image and the second pin mark image.

Description

Information processing device, information processing system, information processing method, and information processing program

[0001] This disclosure relates to an information processing device, an information processing system, an information processing method, and an information processing program.

[0002] In the inspection process of semiconductor chips, probe mark inspection (PMI) is sometimes performed to inspect for needle marks left by the contact of a probe needle. On the other hand, there are methods that use deep learning-based image classification technology, such as that used by CNN, to classify products as good or bad. This image classification technology is being applied to probe mark inspection to classify the condition of the needle marks as good or bad.

[0003] For example, Patent Document 1 (Japanese Patent Publication No. 2022-186039) discloses a technique for predicting the tip position of a probe needle using a prediction model that takes input data as input and the tip position of the probe needle as output. Patent Document 2 (Japanese Patent Publication No. 2023-052655) discloses a technique for acquiring multiple measurement values ​​by measuring a device under test via a jig, analyzing the multiple measurement values ​​to calculate fluctuation data showing the variation in measurement values ​​according to the number of times the jig has contacted the device under test, and managing the state of the jig based on the fluctuation data. Patent Document 3 (Japanese Patent Publication No. 2024-010713) discloses a technique for recognizing the needle region and needle tip region of a verification image by inputting a verification image taken by an imaging unit into a segmentation model trained using training data in which the range of the probe needle and the range of the needle tip are assigned to a training image of the tip of the probe needle.

[0004] Incidentally, when training by adding annotations to needle mark images, it is necessary to prepare a large number of images for which it is known whether the state of the needle mark is good or defective. The state of needle marks changes under the influence of variations in contact load, needle tip wear, and the like due to thermal fluctuation, contact, cleaning, and the like. It is difficult to add annotations that cover all of these changes. In particular, since the proportion of needle marks in a defective state is low, searching for needle mark images with defective needle mark states from a large amount of needle mark images places a heavy burden on the user. Furthermore, since it is difficult to intentionally cause a defective state of needle marks, it is possible to perform training using only good images having good needle mark states and treat images other than good images as defective images; however, in this case, there is a risk that the classification accuracy of the needle mark state will decrease.

[0005] An object of the present disclosure is to provide an information processing apparatus, an information processing system, an information processing method, and an information processing program capable of improving the classification accuracy of a needle mark state while reducing the burden on a user when adding annotations to a needle mark image.

[0006] The information processing apparatus according to the present disclosure comprises: a first acquisition unit that acquires an event that may cause a failure mode representing a defective needle mark state for a semiconductor chip on which a needle mark is formed by contact of a probe needle; a second acquisition unit that acquires an event occurrence time that is the time at which the event occurred; a third acquisition unit that acquires a first needle mark image that is an image of the semiconductor chip captured before the event occurrence time and a second needle mark image that is an image of the semiconductor chip captured after the event occurrence time; an association unit that associates the event that may cause the failure mode with the set of the first needle mark image and the second needle mark image; and an output unit that outputs the set of the first needle mark image and the second needle mark image for each of the events that may cause the failure mode.

[0007] The information processing system according to the present disclosure comprises the information processing apparatus described above, and a prober connected to the information processing apparatus.

[0008] The information processing method according to this disclosure involves a computer performing the following steps: acquiring an event that may occur in a semiconductor chip on which a needle mark has been formed by contact with a probe needle, which indicates that the condition of the needle mark is poor; acquiring an event occurrence time, which is the time when the event occurred; acquiring a first needle mark image, which is an image of the semiconductor chip taken before the event occurrence time, and a second needle mark image, which is an image of the semiconductor chip taken after the event occurrence time; associating the event that may occur in the condition of the poor mode with the pair of the first and second needle mark images; and outputting the pair of the first and second needle mark images for each event that may occur in the condition of the poor mode.

[0009] The information processing program according to this disclosure causes a computer to perform the following processes: acquire an event that may occur in a semiconductor chip on which a needle mark has been formed by contact with a probe needle, which indicates that the needle mark is in a poor condition; acquire an event occurrence time, which is the time when the event occurred; acquire a first needle mark image, which is an image of the semiconductor chip taken before the event occurrence time, and a second needle mark image, which is an image of the semiconductor chip taken after the event occurrence time; associate the event that may occur in the poor condition with the pair of the first and second needle mark images; and output the pair of the first and second needle mark images for each event that may occur in the poor condition.

[0010] According to this disclosure, it is possible to reduce the user's workload when adding annotations to needle mark images while improving the accuracy of needle mark classification.

[0011] This is a side view showing an example of the appearance of a prober in a wafer test system. This is a perspective view showing an example of the appearance of a prober. This is a block diagram showing an example of the electrical configuration of a control device according to an embodiment. This is a diagram showing an example of needle mark images for each mode to be classified. This is a block diagram showing an example of the functional configuration of a control device according to an embodiment. This is a diagram showing an example of the change in needle mark state over time. This is a diagram showing an example of the change in needle mark state due to an event. This is a flowchart showing an example of the flow of event-specific needle mark image output processing by an information processing program according to an embodiment. This is a flowchart showing an example of the flow of probing processing by an information processing program according to an embodiment. This is a diagram showing an example of various information registered in the DB. This is a diagram used to explain the process of outputting a needle mark image for each event based on various information registered in the DB. This is a diagram used to explain the process of performing annotation from various information registered in the DB. This is a diagram used to explain the process of performing annotation from various information registered in the DB. This is a flowchart showing an example of the flow of learning processing by an information processing program according to an embodiment. This is a diagram used to explain the process of automatically determining the next action from the previous event and classification result. This is a flowchart showing an example of the flow of classification processing by an information processing program according to an embodiment.

[0012] Hereinafter, an example of an embodiment for carrying out the technology of this disclosure will be described in detail with reference to the drawings. Components and processes that perform the same operation, action, or function are given the same reference numerals throughout the drawings, and redundant explanations may be omitted as appropriate. Each drawing is only a schematic representation to the extent that the technology of this disclosure can be fully understood. Therefore, the technology of this disclosure is not limited to the illustrated examples. Furthermore, in this embodiment, explanations of configurations not directly related to the technology of this disclosure or well-known configurations may be omitted.

[0013] Figure 1A is a side view showing an example of the appearance of the prober 10 included in the wafer test system 1. Figure 1B is a perspective view showing an example of the appearance of the prober 10. The prober 10 is used in the wafer test system 1 to inspect the electrical characteristics of multiple semiconductor chips formed on a wafer W. Multiple semiconductor chips are formed on the wafer W, and multiple electrode pads are formed on each semiconductor chip. The wafer test system 1 is an example of an information processing system and includes the prober 10 and a control device 40. The control device 40 is an example of an information processing device.

[0014] As shown in Figures 1A and 1B, the prober 10 comprises a level pad 11, a base 12, a Y stage 13, a Y movement unit 14, an X stage 15, an X movement unit 16, a Zθ stage 17, a Zθ movement unit 18, a wafer chuck 20, a support column 23, a head stage 24, a card holder 25, a probe card 26, a wafer alignment camera 29, an upper and lower stage 30, a needle alignment camera 31, and a cleaning plate 32. Note that the external configuration of the prober 10 is not limited to the examples shown in Figures 1A and 1B and can be modified as appropriate.

[0015] A Y-stage 13 is supported on the upper surface of the base 12 via a Y-movement unit 14 so as to be movable in the Y-axis direction. The Y-axis direction is an example of a first direction. In addition, four level pads 11 are arranged on the lower surface of the base 12, which are adjustable in the Z-axis direction (vertical direction) of the base 12.

[0016] The Y-movement unit 14 includes, for example, a guide rail provided on the upper surface of the base 12 and parallel to the Y-axis, a slider provided on the lower surface of the Y-stage 13 and engaging with the guide rail, and an actuator such as a motor that moves the Y-stage 13 in the Y-axis direction. This Y-movement unit 14 moves the Y-stage 13 in the Y-axis direction on the base 12.

[0017] The X-stage 15 is supported on the upper surface of the Y-stage 13 via the X-movement unit 16 so as to be movable in the X-axis direction. The X-axis direction is an example of a second direction. The X-axis direction is perpendicular to the Y-axis direction. Here, "perpendicular" means that it is perpendicular including a predetermined error. The X-movement unit 16 includes, for example, a guide rail provided on the upper surface of the Y-stage 13 and parallel to the X-axis, a slider provided on the lower surface of the X-stage 15 and engaging with the guide rail, and an actuator such as a motor that moves the X-stage 15 in the X-axis direction. This X-movement unit 16 moves the X-stage 15 on the Y-stage 13 in the X-axis direction.

[0018] The upper surface of the X-stage 15 is provided with a Zθ stage 17 and upper and lower stages 30. The Zθ stage 17 is provided with a Zθ moving part 18. A wafer chuck 20 is held on the upper surface of the Zθ stage 17 via the Zθ moving part 18.

[0019] The Zθ moving unit 18 includes, for example, a lifting mechanism that moves the Zθ stage 17 in the Z-axis direction (vertical direction) and a rotation mechanism that rotates the Zθ stage 17 around the Z-axis. Therefore, the Zθ moving unit 18 moves the wafer chuck 20, which is held on the upper surface of the Zθ stage 17, in the Z-axis direction and rotates it around the Z-axis.

[0020] A wafer W is held on the upper surface of the wafer chuck 20 by various holding methods such as vacuum suction. The wafer chuck 20 is supported so as to be movable in the XYZ direction and so as to be rotatable around the Z axis via the Y stage 13, Y moving part 14, X stage 15, X moving part 16, Zθ stage 17, and Zθ moving part 18 described above. This allows the wafer W held in the wafer chuck 20 and the probe needle 35 described later to move relative to each other.

[0021] The support column 23 is provided on the upper surface of the base 12 and supports the head stage 24 above the Y stage 13, X stage 15, and Zθ stage 17 (hereinafter simply referred to as stages 13, 15, and 17). As a result, the head stage 24 is fixed onto the base 12 via the support column 23.

[0022] A card holder 25 is held in the center of the head stage 24. The card holder 25 has a holding hole 25a that holds the outer circumference of the probe card 26, and the probe card 26 is held in this holding hole 25a. As a result, the probe card 26 is held in a position facing the wafer W via the head stage 24 and the card holder 25.

[0023] The probe card 26 has probe needles 35 arranged according to the arrangement of electrode pads on the semiconductor chip to be tested. These card holders 25 and probe cards 26 are replaced depending on the type of semiconductor chip.

[0024] The probe card 26 is provided with connection terminals (not shown) electrically connected to the probe needles 35, and a tester (not shown) is connected to these connection terminals. The tester supplies various test signals to the electrode pads of the semiconductor chip via the connection terminals of the probe card 26 and the probe needles 35, and also receives and analyzes the signals output from the electrode pads to test whether the semiconductor chip is functioning correctly. Note that the configuration of the tester and the test method are known technologies, so a detailed explanation is omitted.

[0025] The wafer alignment camera 29 photographs the semiconductor chips on the wafer W held in the wafer chuck 20. Based on the images captured by the wafer alignment camera 29, the positions of the electrode pads on the semiconductor chip to be inspected can be detected. The installation location and structure of the wafer alignment camera 29 are not particularly limited, but for example, it may be installed on the head stage 24.

[0026] The upper and lower stages 30 are equipped with a needle alignment camera 31 and a cleaning plate 32 at positions substantially opposite to the head stage 24, etc. The upper and lower stages 30 also have a lifting mechanism (not shown) that is movable in the Z-axis direction, allowing adjustment of the Z-axis position of the needle alignment camera 31 and the cleaning plate 32. The needle alignment camera 31 and the cleaning plate 32 are supported so as to be movable in the XYZ axes via the Y-stage 13 and Y-movement unit 14, the X-stage 15 and X-movement unit 16, and the upper and lower stages 30. This allows relative movement between the needle alignment camera 31 and the cleaning plate 32 and the probe needle 35. The needle alignment camera 31 may also be located on the Zθ-stage 17. In this case, the upper and lower stages 30 and the cleaning plate 32 may be omitted.

[0027] The needle alignment camera 31 photographs the probe needle 35 of the probe card 26. Based on the image of the probe needle 35 captured by this needle alignment camera 31, the position of the probe needle 35 can be detected. Specifically, the XY coordinates of the tip position of the probe needle 35 are detected from the position coordinates of the needle alignment camera 31, and the Z coordinate of the tip position of the probe needle 35 is detected from the focal point position of the needle alignment camera 31.

[0028] When inspecting semiconductor chips on wafer W with the prober 10 configured as described above, each time the probe card 26 is replaced, or each time a predetermined number of semiconductor chips are inspected, the stages 13, 15, and 17 are driven to move the needle alignment camera 31 relative to the position where the probe needle 35 will be photographed, and then the probe needle 35 is photographed by the needle alignment camera 31. Based on the image captured by the needle alignment camera 31, the tip position of the probe needle 35 is detected.

[0029] Furthermore, with the wafer W to be inspected held in the wafer chuck 20, each stage 13, 15, and 17 is driven to move the wafer alignment camera 29 relative to the shooting position of the wafer W, and then the semiconductor chip of the wafer W is photographed with the wafer alignment camera 29. Based on the image taken by the wafer alignment camera 29, the position of the electrode pads of the semiconductor chip to be inspected is detected.

[0030] Then, stages 13, 15, and 17 are driven to electrically contact the probe needle 35 with the electrode pad of the semiconductor chip to be inspected first. In this state, the tester performs inspection on the semiconductor chip to be inspected first. The remaining semiconductor chips to be inspected are then inspected in the same manner. Note that the specific inspection method for semiconductor chips is publicly known, so a detailed explanation is omitted here.

[0031] The control device 40 is a controller that controls various parts of the prober 10. The control device 40 may be built into the main body of the prober 10, or it may be provided separately from the prober body. The control device 40 is composed of, for example, a computing device such as a personal computer, and includes a computing circuit composed of various processors and memory. The control device 40 may be connected to multiple probers 10.

[0032] Figure 2 is a block diagram showing an example of the electrical configuration of the control device 40 according to this embodiment. As shown in Figure 2, the control device 40 according to this embodiment includes a CPU (Central Processing Unit) 41, a ROM (Read Only Memory) 42, a RAM (Random Access Memory) 43, an input / output interface (I / O) 44, a storage unit 45, a display unit 46, an operation unit 47, and a communication unit 48.

[0033] The CPU 41, ROM 42, RAM 43, and I / O 44 are connected to each other via a bus. The I / O 44 is connected to various functional units, including a storage unit 45, a display unit 46, an operation unit 47, and a communication unit 48. These functional units are capable of communicating with the CPU 41 via the I / O 44.

[0034] The control unit is comprised of a CPU 41, ROM 42, RAM 43, and I / O 44. The control unit may be configured as a sub-control unit that controls the operation of a part of the control device 40, or as part of the main control unit that controls the operation of the entire control device 40. Integrated circuits or IC chipsets, such as LSIs (Large Scale Integrations), are used in part or all of each block of the control unit. Individual circuits may be used for each of the above blocks, or circuits that integrate part or all of them may be used. The above blocks may be provided as a single unit, or some of the blocks may be provided separately. Furthermore, parts of each of the above blocks may be provided separately. For the integration of the control unit, dedicated circuits or general-purpose processors may be used, not limited to LSIs.

[0035] For example, the storage unit 45 can be an HDD (Hard Disk Drive), an SSD (Solid State Drive), or flash memory. The storage unit 45 stores an information processing program 45A for executing the information processing according to this embodiment. This information processing program 45A may also be stored in ROM 42.

[0036] The information processing program 45A may, for example, be pre-installed on the control device 40. The information processing program 45A may also be implemented by storing it on a non-volatile storage medium or by distributing it via a network and installing it on the control device 40 as appropriate. Examples of non-volatile storage mediums include CD-ROM (Compact Disc Read Only Memory), magneto-optical disk, HDD, DVD-ROM (Digital Versatile Disc Read Only Memory), flash memory, memory card, etc.

[0037] The display unit 46 may use, for example, a liquid crystal display (LCD), an organic EL (Electroluminescence) display, or the like. The display unit 46 may also have an integrated touch panel. The operation unit 47 is equipped with, for example, a keyboard, mouse, or other device for operation input. The display unit 46 and the operation unit 47 receive various instructions from the user of the control device 40. The display unit 46 displays various information such as the results of processing performed in response to instructions received from the user, and notifications regarding processing.

[0038] The communication unit 48 is connected to a network such as the Internet, LAN (Local Area Network), or WAN (Wide Area Network), and can communicate with, for example, the wafer alignment camera 29 and the needle alignment camera 31 via the network. The communication unit 48 enables communication with each component of the prober 10. The communication unit 48 may be connected via a wired connection or a wireless connection.

[0039] In this needle mark inspection, the semiconductor chip on which a needle mark has been formed by contact with the probe needle 35 is photographed by the wafer alignment camera 29, and a needle mark image is acquired. The needle mark image is classified into various modes according to the shape, size, position, and other conditions of the needle mark. "Contact" refers to the contact between the probe needle 35 and the semiconductor chip.

[0040] Figure 3 shows an example of needle mark images for each classification mode. In Figure 3, examples of classification modes include "normal needle mark," "poor contact," "excessive contact," "foreign matter adhesion," "misalignment," "needle mark splash," and "needle mark slippage." However, the user may decide how to classify them as appropriate. For example, "poor contact," "excessive contact," and "misalignment" may be classified together as a poor mode (hereinafter referred to as "NG mode"), and "normal needle mark" and "foreign matter adhesion" may be classified together as a normal mode (hereinafter referred to as "OK mode"). However, "foreign matter adhesion" may also be classified as an NG mode. On the other hand, the classification of "normal needle mark" may also be further subdivided, for example, by distinguishing and classifying based on the size of the needle mark.

[0041] By the way, when using machine learning to classify needle mark images, it is necessary to assign appropriate annotations to needle mark images for each mode. In this case, as mentioned above, it is necessary to prepare many images in which the condition of the needle mark is known to be good or bad. The condition of the needle mark changes due to the influence of thermal fluctuations, contact, cleaning, etc., which affect contact load fluctuations and needle tip wear. It is difficult to assign annotations that cover all of these changes. In particular, since the proportion of needle marks in a bad state is low, it is a heavy burden on the user to search for needle mark images in a bad state from a large number of needle mark images. Also, since it is difficult to intentionally induce a bad state in the needle mark, it is possible to train only on good images with good needle marks and treat all other images as bad images, but in this case the classification accuracy of the needle mark condition will decrease.

[0042] Therefore, in this embodiment, by creating a database of needle mark images, events that may cause NG mode to occur, and the time of event occurrence, it is made easier to retrieve needle mark images of various modes necessary for annotation.

[0043] Specifically, the CPU 41 of the control device 40 according to this embodiment functions as the various parts shown in Figure 4 by writing the information processing program 45A stored in the ROM 42 or storage unit 45 to the RAM 43 and executing it.

[0044] FIG. 4 is a block diagram showing an example of a functional configuration of a control device 40 according to the present embodiment. As shown in FIG. 4, a CPU 41 of the control device 40 according to the present embodiment functions as a first acquisition unit 41A, a second acquisition unit 41B, a third acquisition unit 41C, an association unit 41D, an output unit 41E, a learning unit 41F, a classification unit 41G, and a warning unit 41H. Note that the first acquisition unit 41A, the second acquisition unit 41B, and the third acquisition unit 41C may be configured as a single acquisition unit.

[0045] The first acquisition unit 41A acquires an event that may cause an NG mode for a semiconductor chip on which a probe mark is formed by contact of a probe needle 35. The acquired event is registered in, for example, a database (hereinafter referred to as "DB") 45C. The DB 45C is stored in, for example, the storage unit 45, but may be stored in an external storage device. Here, the NG mode is a mode indicating that the state of a probe mark is defective. The NG mode includes, for example, at least one of positional deviation, excessive contact, poor contact, absence of probe mark, probe mark splashing, probe mark slipping, and foreign matter adhesion. The event includes, for example, at least one of temperature change, change in the overdrive amount (hereinafter referred to as "OD amount") of the probe needle 35, change in the contact position of the probe needle 35, measurement of the position of the probe needle 35, wafer replacement, and cleaning. However, the OD amount refers to an amount by which the probe needle 35 is fed in a direction in which it bites in after the probe needle 35 is brought into contact with the semiconductor chip.

[0046] The second acquisition unit 41B acquires an event occurrence time. The event occurrence time is the time at which the event acquired by the first acquisition unit 41A occurs. The acquired event occurrence time is registered in, for example, the DB 45C.

[0047] The third acquisition unit 41C acquires a first probe mark image and a second probe mark image. The first probe mark image is an image obtained by photographing the semiconductor chip before the event occurrence time, and the second probe mark image is an image obtained by photographing the semiconductor chip after the event occurrence time. The acquired first probe mark image and second probe mark image are registered in, for example, the DB 45C.

[0048] FIG. 5A is a diagram showing an example of a change in needle trace state over time. FIG. 5B is a diagram showing an example of a change in needle trace state due to an event. In the example of FIG. 5A, changes in the size of needle traces and changes in the position of needle traces over time, rather than due to event occurrence, are shown. On the other hand, in the example of FIG. 5B, changes in needle trace state based on the time when an event of OD amount change occurs, and changes in needle trace state based on the time when an event of cleaning occurs are shown. That is, the needle trace image before the OD amount change is the first needle trace image, and the needle trace image after the OD amount change is the second needle trace image. Similarly, the needle trace image before cleaning is the first needle trace image, and the needle trace image after cleaning is the second needle trace image.

[0049] The associating unit 41D associates, for example, an event that may cause an NG mode with a pair of the first needle trace image and the second needle trace image based on various types of information registered in the DB 45C.

[0050] The output unit 41E outputs a pair of the first needle trace image and the second needle trace image for each event that may cause an NG mode. Here, the output destination of the pair of the first needle trace image and the second needle trace image may be, for example, the display unit 46 or the storage unit 45.

[0051] The learning unit 41F performs machine learning using a needle trace image to which an annotation is added based on a pair of the first needle trace image and the second needle trace image as learning data, thereby generating a trained model 45B that receives a needle trace image as an input and outputs a classification result of a needle trace state. Note that "annotation" refers to a label added to target data (here, a needle trace image) or a process of adding a label in supervised learning, which is one type of machine learning. The label is added by, for example, a user. The learning unit 41F stores the generated trained model 45B in, for example, the storage unit 45. The machine learning method is not particularly limited, and for example, Deep Learning or the like is used.

[0052] The classification unit 41G uses the trained model 45B to classify the input needle mark image into a predetermined mode based on the state of the needle mark. Here, the predetermined modes include NG mode and OK mode. As described above, the NG mode is a mode that includes at least one of the following: misalignment, excessive contact, poor contact, no needle mark, needle mark splash, needle mark slippage, and foreign matter adhesion. The OK mode is a mode that indicates the state of the needle mark is normal.

[0053] The learning unit 41F and the classification unit 41G may be executed by an external server computer (not shown) instead of the control device 40. In this case, the control device 40 transmits learning data to the server computer, and the server computer generates and stores a trained model 45B based on the learning data. The server computer then uses the stored trained model 45B to classify the needle trace images input from the control device 40 into predetermined modes and transmits the classification results to the control device 40.

[0054] The warning unit 41H issues a warning if the needle mark condition is classified as a faulty mode. The warning may include, for example, a message warning that the needle mark condition is faulty, or a message prompting the user to stop the prober 10.

[0055] Next, the operation of the control device 40 according to this embodiment will be explained with reference to Figure 6.

[0056] Figure 6 is a flowchart showing an example of the flow of event-specific needle mark image output processing by the information processing program 45A according to this embodiment.

[0057] When an instruction is given to execute processing by the information processing program 45A, the CPU 41 of the control device 40 executes the program by writing the information processing program 45A, which is stored in the ROM 42 or the storage unit 45, to the RAM 43.

[0058] In step S101 of Figure 6, the CPU 41 acquires an event that may cause NG mode to occur and registers it in DB 45C as an example.

[0059] In step S102, the CPU 41 obtains the event occurrence time, which is the time when the event acquired in step S101 occurred, and registers it in DB 45C as an example.

[0060] In step S103, the CPU 41 acquires the first and second needle trace images and registers them in DB 45C as an example. As described above, the first needle trace image is an image of the semiconductor chip taken before the event occurred, and the second needle trace image is an image of the semiconductor chip taken after the event occurred.

[0061] In step S104, the CPU 41 associates events that may cause NG mode to occur with the pair of first needle mark image and second needle mark image, based on various information registered in DB 45C as an example.

[0062] In step S105, the CPU 41 outputs a pair of first and second needle mark images for each event that may cause an NG mode to occur, and terminates the event-specific needle mark image output processing by the information processing program 45A. Here, the output destination for the pair of first and second needle mark images may be, as described above, the display unit 46 or the storage unit 45.

[0063] Next, the probing process according to this embodiment will be described with reference to Figures 7 and 8. During this probing process, the needle trace image, events that may cause an NG mode, and the time of event occurrence are acquired and registered in DB45C.

[0064] Figure 7 is a flowchart showing an example of the probing process flow by the information processing program 45A according to this embodiment.

[0065] When an instruction is given to execute processing by the information processing program 45A, the CPU 41 of the control device 40 executes the program by writing the information processing program 45A, which is stored in the ROM 42 or the storage unit 45, to the RAM 43.

[0066] In step S111 of Figure 7, the CPU 41 determines whether an event has occurred in the prober 10 that could potentially cause the NG mode to occur. If it is determined that an event has occurred (positive determination), the process proceeds to step S112; if it is determined that no event has occurred (negative determination), the process proceeds to step S113.

[0067] In step S112, the CPU 41 registers the event that occurred and the time of the event in DB 45C, and then proceeds to step S113.

[0068] In step S113, the CPU 41 determines whether or not it is time to perform a needle mark inspection. If it determines that it is time to perform a needle mark inspection (positive determination), the process proceeds to step S114. If it determines that it is not time to perform a needle mark inspection (negative determination), the process proceeds to step S116.

[0069] In step S114, the CPU 41 controls the prober 10 to perform needle mark inspection.

[0070] In step S115, the CPU 41 registers the first needle trace image taken before the event occurred and the second needle trace image taken after the event occurred in DB 45C, and then proceeds to step S116.

[0071] In step S116, the CPU 41 determines whether the probing process has finished. If it determines that the probing process has not finished (negative determination), the process returns to step S111 and is repeated. If it determines that the probing process has finished (positive determination), the probing process by this information processing program 45A is terminated.

[0072] Figure 8 shows an example of the various types of information registered in DB45C. The needle trace image obtained by needle trace inspection, which is performed at predetermined intervals during the probing process, is registered in DB45C along with probe card information, device information, and time information. In addition, various events (e.g., alignment, cleaning, needle position adjustment, OD amount change, contact position change, measurement-related alarms, etc.) are registered in DB45C along with time information.

[0073] As shown in Figure 8, DB45C registers the first list 101, the second list 102, the third list 103, and the fourth list 104. The first list 101 registers the product ID (product identification information), device name (device identification information), and card name (probe card 26 identification information) of the semiconductor chip to be inspected. The second list 102 registers the product ID, needle tip ID (probe needle 35 identification information), needle tip coordinate X (coordinate of the probe needle 35 in the X-axis direction), and needle tip coordinate Y (coordinate of the probe needle 35 in the Y-axis direction). The third list 103 registers the product ID, time (event occurrence time), and event (occurring event). The fourth list 104 registers the product ID, needle tip ID, time (acquisition time of the needle trace image), and image (needle trace image).

[0074] Figure 9 is a diagram illustrating the process of outputting a needle mark image for each event based on various information registered in DB45C. In order to perform annotation necessary for machine learning, the DB45C, which has registered needle mark images and events in the probing process described in Figures 7 and 8 above, associates the event with the needle mark images obtained before and after the event occurrence time in the needle mark inspection, that is, the pair of the first needle mark image and the second needle mark image, and outputs the pair of the first and second needle mark images that may indicate a change in the needle mark state. The needle mark image list 105 shown in Figure 9 shows the case where a pair of before image (first needle mark image) and after image (second needle mark image) is output for each event. In the needle mark image list 105, the event is associated with the pair of the first and second needle mark images based on the variety ID and time information registered in the third list 103 shown in Figure 8 above, and the variety ID, needle tip ID, and time information registered in the fourth list 104. This makes it possible to secure a sufficient number of needle trace images for the various modes required for annotation.

[0075] In this embodiment, the process can be broadly divided into a data storage phase, an annotation and learning phase, and a needle trace image classification phase.

[0076] In the data accumulation phase, as mentioned above, since needle mark images need to be accumulated during annotation, event data and periodic needle mark images are acquired for each lot while normal operations are being carried out. Then, various information for multiple lots is registered in DB45C.

[0077] In the annotation and learning phase, the control device 40 itself performs the operation, or a server computer on the network accesses the control device 40 and performs the operation.

[0078] Figures 10 and 11 illustrate the process of performing annotation from various information registered in DB45C. As shown in Figure 10, events are listed in the third list 103, and the needle marks before and after the event can be compared for each needle tip of the corresponding probe card 26 in the needle mark image lists 106 and 107. This allows the user to perform annotation work while narrowing down the events of interest. In the example in Figure 10, when the event "wafer change" occurs, the needle mark image list 106 is created, and when the event "needle alignment" occurs, the needle mark image list 107 is created. The user can perform annotation work while looking at the needle mark image lists 106 and 107.

[0079] On the other hand, as shown in Figure 11, the needle mark images are listed in the fourth list 104, and the needle mark image list 108 allows for comparison of the needle marks before and after the event for each needle tip of the corresponding probe card 26. This allows the user to perform annotation work while narrowing down the events of interest. In the example in Figure 11, when a specific event occurs, the needle mark image list 108 is created from the needle mark images listed in the fourth list 104. The user can perform annotation work while looking at the needle mark image list 108.

[0080] Figure 12 is a flowchart showing an example of the learning process flow by the information processing program 45A according to this embodiment.

[0081] When an instruction is given to execute processing by the information processing program 45A, the CPU 41 of the control device 40 executes the program by writing the information processing program 45A, which is stored in the ROM 42 or the storage unit 45, to the RAM 43.

[0082] In step S121 of Figure 12, the CPU 41 acquires an annotated needle mark image, as shown in Figures 10 and 11 above, as an example.

[0083] In step S122, the CPU 41 performs machine learning using the annotated needle mark images as training data to generate a trained model 45B that takes the needle mark images as input and outputs the classification result of the needle mark state.

[0084] In step S123, the CPU 41 stores the trained model 45B generated in step S122 in, for example, the memory unit 45, and terminates the training process by the information processing program 45A.

[0085] In the needle mark image classification phase, the control device 40 itself performs the classification, or a server computer on the network accesses the control device 40 to perform the classification. In the needle mark image classification phase, the trained model 45B is used to classify the needle mark images. This classifies needle marks in NG modes, such as misalignment, excessive contact, poor contact, no needle mark, and foreign matter adhesion, and issues warnings, such as alarm shutdown. This helps to suppress measurement errors, damage to the probe card 26, damage to the wafer W, etc.

[0086] Figure 13 illustrates the process by which the prober 10 automatically determines the next action based on the previous event and classification result. The list shown in Figure 13 lists the previous event and the prober 10's actions for each NG mode of the needle trace. As shown in Figure 13, the prober 10 can automatically determine the next action or assist in the operation based on the previous event and classification result.

[0087] As shown in Figure 13, when the NG mode is "foreign matter adhesion," the preceding event is divided into "other than cleaning" and "cleaning." If it is "other than cleaning," the next action of the prober 10 is "cleaning," and if it is "cleaning," the next action of the prober 10 is "cleaning sheet replacement." When the NG mode is "excessive contact, poor contact, no needle marks," the preceding event is divided into "wafer alignment, wafer replacement, alarm stop" and "needle positioning, calibration for thermal fluctuation correction." If it is "wafer alignment, wafer replacement, alarm stop," the next action of the prober 10 is "calibration for thermal fluctuation correction," and if it is "needle positioning, calibration for thermal fluctuation correction," the next action of the prober 10 is "cleaning." When the NG mode is "misalignment," the preceding event is divided into "wafer alignment, wafer replacement, alarm stop" and "needle positioning, calibration for thermal fluctuation correction." If the event is "wafer alignment, wafer change, alarm stop," the next action of prober 10 will be "calibration for thermal fluctuation correction," and if the event is "needle positioning, calibration for thermal fluctuation correction," the next action of prober 10 will be "operator call, needle tip confirmation request." If the NG mode is "needle mark bounce" or "needle mark slip," there is no preceding event, and the next action of prober 10 will be "operator call, Z-axis acceleration change recommended."

[0088] Figure 14 is a flowchart showing an example of the classification process flow by the information processing program 45A according to this embodiment.

[0089] When an instruction is given to execute processing by the information processing program 45A, the CPU 41 of the control device 40 executes the program by writing the information processing program 45A, which is stored in the ROM 42 or the storage unit 45, to the RAM 43.

[0090] In step S131 of Figure 14, the CPU 41 acquires the needle trace image to be classified.

[0091] In step S132, the CPU 41 inputs the needle trace image acquired in step S131 to the trained model 45B, and the trained model 45B classifies the needle trace state of the input needle trace image into a predetermined mode.

[0092] In step S133, the CPU 41 determines whether the needle mark state is classified as NG mode. If it is determined that the needle mark state is classified as NG mode (positive determination), the process proceeds to step S134. If it is determined that the needle mark state is not classified as NG mode, that is, classified as OK mode (negative determination), the classification process by this information processing program 45A is terminated.

[0093] In step S134, the CPU 41 outputs a message to warn that the needle mark state has been classified as NG mode, and the classification process by the information processing program 45A is terminated.

[0094] Thus, according to this embodiment, it is possible to reduce the user's burden when adding annotations to needle mark images while improving the accuracy of needle mark classification.

[0095] Furthermore, at least one of the following events can be obtained: temperature change, change in the OD amount of the probe needle, change in the contact position of the probe needle, measurement of the probe needle position, wafer replacement, and cleaning.

[0096] Furthermore, at least one of the following can be obtained as an NG mode: misalignment, excessive contact, poor contact, no needle mark, needle mark splashing, needle mark slippage, and foreign matter adhesion.

[0097] Furthermore, it is possible to generate a pre-trained model that can accurately classify the state of needle marks.

[0098] Furthermore, a warning can be issued if the needle mark status indicates an NG mode.

[0099] In this embodiment, we have described an example where the images before and after the event occurrence time are designated as the first needle mark image and the second needle mark image, respectively. However, if an event execution schedule is set in advance, the needle mark images to be acquired before and after the event may also be specified in advance, and the event, the first needle mark image, and the second needle mark image may be linked and registered in DB45C.

[0100] In this embodiment, each process is executed on any computer. Furthermore, any computer may execute these processes using a processor as hardware, a program as software, or a combination thereof. In that case, the processor is configured to work in cooperation with the program to execute the various processes in this embodiment, and can function as a unit or means in this embodiment. Also, the execution order of the processes by the processor is not limited to the order described and may be changed as appropriate. Any computer may be a general-purpose computer, a computer designed for a specific purpose, a workstation, or any other system capable of executing each process.

[0101] A processor may consist of one or more hardware components, and the type of hardware is not limited. Examples of processors include CPUs, GPUs (Graphics Processing Units), ASICs (Application Specific Integrated Circuits), and programmable logic devices (e.g., SPLDs (Simple Programmable Logic Devices), CPLDs (Complex Programmable Logic Devices), and FPGAs (Field Programmable Gate Arrays)). Furthermore, the hardware components may be a combination of different types of hardware. When multiple hardware components are configured to execute one or more processes of a processor, these components may reside in physically separate devices or in the same device. Furthermore, in any embodiment, the order of the processes performed by the processor is not limited to the order described above and may be changed as appropriate. The hardware components are composed of electrical circuits (circuits) and the like, which are combinations of circuit elements such as semiconductor elements.

[0102] Furthermore, the program may be firmware or software such as microcode. Alternatively, the program may be, for example, a group of program modules, each function of which may be implemented by a processor configured to perform its respective function. The program may be program code or multiple code segments stored on one or more non-temporary computer-readable media (e.g., storage media or other storage). The program may be divided and stored on multiple non-temporary computer-readable media located in physically separate devices. The program code or code segments may represent any combination of procedures, functions, subprograms, routines, subroutines, modules, software packages, classes, or instructions, data structures, or program statements. The program code or code segments may be connected to other code segments or hardware circuits by sending and receiving information, data, arguments, parameters, or memory contents. Furthermore, the program according to the embodiment may be provided as a program product.

[0103] The disclosure of Japanese Patent Application No. 2025-054410, filed on 27 March 2025, is incorporated herein by reference in its entirety. All documents, patent applications, and technical standards described herein are incorporated herein by reference to the same extent as if each individual document, patent application, and technical standard were specifically and individually noted to be incorporated by reference.

Claims

1. An information processing device comprising: a first acquisition unit that acquires an event that may cause a defect mode to occur, indicating that the condition of the needle marks on a semiconductor chip formed by contact with a probe needle is poor; a second acquisition unit that acquires an event occurrence time, which is the time when the event occurred; a third acquisition unit that acquires a first needle mark image, which is an image of the semiconductor chip taken before the event occurrence time, and a second needle mark image, which is an image of the semiconductor chip taken after the event occurrence time; an association unit that associates the event that may cause the defect mode with the pair of the first needle mark image and the second needle mark image; and an output unit that outputs the pair of the first needle mark image and the second needle mark image for each event that may cause the defect mode.

2. The information processing apparatus according to claim 1, wherein the event includes at least one of a temperature change, a change in the amount of overdrive of the probe needle, a change in the contact position of the probe needle, a measurement of the position of the probe needle, wafer replacement, and cleaning.

3. The information processing apparatus according to claim 1, wherein the failure mode includes at least one of misalignment, excessive contact, poor contact, no needle mark, needle mark splashing, needle mark slippage, and foreign matter adhesion.

4. The information processing apparatus according to claim 1, further comprising a learning unit that generates a trained model that takes a needle mark image as input and outputs a classification result of the needle mark state by performing machine learning on a needle mark image to which annotations have been added based on the set of the first needle mark image and the second needle mark image as training data.

5. The information processing apparatus according to claim 4, further comprising: a classification unit that classifies the needle trace state of an input needle trace image into a predetermined mode using the trained model; and a warning unit that issues a warning when the needle trace state is classified into a poor mode.

6. An information processing system comprising: an information processing device according to any one of claims 1 to 5; and a prober connected to the information processing device.

7. An information processing method in which a computer performs the following steps: acquire an event that may occur in a semiconductor chip on which a needle mark has been formed by contact of a probe needle, indicating that the condition of the needle mark is poor; acquire an event occurrence time, which is the time when the event occurred; acquire a first needle mark image, which is an image of the semiconductor chip taken before the event occurrence time, and a second needle mark image, which is an image of the semiconductor chip taken after the event occurrence time; associate the event that may occur in the poor mode with the pair of the first and second needle mark images; and output the pair of the first and second needle mark images for each event that may occur in the poor mode.

8. An information processing program that causes a computer to perform the following steps: acquire an event that may cause a defect mode to occur, indicating that the condition of the needle marks on a semiconductor chip formed by contact with a probe needle is poor; acquire an event occurrence time, which is the time when the event occurred; acquire a first needle mark image, which is an image of the semiconductor chip taken before the event occurrence time, and a second needle mark image, which is an image of the semiconductor chip taken after the event occurrence time; associate the event that may cause the defect mode to occur with the pair of the first and second needle mark images; and output the pair of the first and second needle mark images for each event that may cause the defect mode to occur.

Images