Method and apparatus for manufacturing semiconductor device

Abstract

Provided is a method for manufacturing a semiconductor device, in which a semiconductor element to be inspected is set as a first semiconductor element, a semiconductor element for acquiring teacher data is set as a second semiconductor element, and the following steps are performed: acquiring teacher data including a plurality of pieces of data relating to characteristics of the second semiconductor element; making a learned model by using the teacher data; and determining whether or not to perform an electrical test of the first semiconductor element on the basis of an output obtained by inputting a plurality of data relating to the characteristics of the first semiconductor element into the learned model.

Description

[0001] Cross-referencing of related applications

[0002] This application is based on Japanese Patent Application No. 2022-162850, filed on October 10, 2022, the contents of which are incorporated herein by reference. Technical Field

[0003] This disclosure relates to a method and apparatus for manufacturing a semiconductor device. Background Technology

[0004] Before semiconductor devices containing MOSFETs and other components leave the factory, they undergo high-voltage switching and withstand voltage tests. If components fail these tests due to short circuits, or if foreign matter adheres to the chiptester probes, worktables, collets, or other equipment, continuing testing as before will likely result in more measurement and appearance defects due to further damage to the chip. Therefore, in cases of component damage, the equipment is replaced to avoid further damage to the chip.

[0005] Replacing the equipment would increase chip costs due to decreased throughput and equipment expenses. Therefore, in the event of a short-circuit failure, overcurrent detection and current cut-off are used to protect the equipment. For example, if the drain current increases sharply beyond a reference value due to MOSFET failure, the current supply to the device is cut off, and the drain current stops.

[0006] However, according to this method, since a large current flows before the current is cut off, it is difficult to reduce the adhesion of foreign objects to the appliance due to component damage.

[0007] Regarding the manufacturing of semiconductor devices, for example, Patent Document 1 proposes determining whether aging tests are required for each batch. For batches determined to require aging tests, aging tests are performed before the final testing process. Regarding the necessity of aging tests, multiple chips are pre-tested using probe tests and aging tests, and the determination is based on the results against a pre-defined benchmark. Specifically, among the probe test items, items with high correlation to the aging test results are selected, and the inspection results of each chip in that test item are plotted as point data on a two-dimensional graph. A determination benchmark is set based on the distribution of the point data.

[0008] Existing technical documents

[0009] Patent documents

[0010] Patent Document 1: Japanese Patent No. 6310782 Summary of the Invention

[0011] For example, aging tests can be used to filter out defective chips before electrical inspection, thereby preventing component damage during electrical inspection and protecting the appliance. However, in the method described in Patent Document 1, the test results are converted into one point of data for each chip, resulting in significant information loss. Consequently, sufficient judgment accuracy cannot be obtained, and component damage is possible.

[0012] In view of the above-mentioned problems, the present disclosure aims to provide a method and apparatus for manufacturing a semiconductor device capable of suppressing damage to semiconductor elements.

[0013] According to one technical solution of this disclosure, a method for manufacturing a semiconductor device is provided, wherein a semiconductor element to be inspected is designated as a first semiconductor element, a semiconductor element used to obtain teacher data is designated as a second semiconductor element, and the following steps are performed: obtaining teacher data containing multiple data related to the characteristics of the second semiconductor element; creating a learned model using the teacher data; and determining whether to perform an electrical inspection of the first semiconductor element based on the output obtained by inputting multiple data related to the characteristics of the first semiconductor element into the learned model.

[0014] Therefore, the teacher data contains multiple data related to the characteristics of the second semiconductor element, and the determination is made using multiple data related to the characteristics of the first semiconductor element, thus improving the determination accuracy and suppressing damage to the semiconductor element.

[0015] Furthermore, according to another technical solution, a semiconductor device manufacturing apparatus includes a control unit that determines whether to perform electrical checks on semiconductor devices based on the output obtained by inputting multiple data related to the characteristics of semiconductor elements into a learned model.

[0016] Therefore, since multiple data related to the characteristics of the first semiconductor element are used for determination, the determination accuracy is improved, and damage to the semiconductor element can be suppressed.

[0017] In addition, the parenthesized reference marks assigned to each constituent element are used to indicate an example of the correspondence between the constituent element and the specific constituent element described in the embodiments described later. Attached Figure Description

[0018] Figure 1 This is a diagram showing the structure of the inspection device in the first embodiment.

[0019] Figure 2 This is a circuit diagram for checking the circuit.

[0020] Figure 3 This is a flowchart of the manufacturing process of a semiconductor device.

[0021] Figure 4 This is a diagram showing the withstand voltage of a semiconductor device.

[0022] Figure 5 This is a graph representing the threshold voltage of a semiconductor device.

[0023] Figure 6 This is a flowchart of the manufacturing process of the semiconductor device in the second embodiment.

[0024] Figure 7 This is a flowchart of a high-voltage test.

[0025] Figure 8 This is the waveform of the drain current during a high-voltage test.

[0026] Figure 9 It is the waveform of the gate-source voltage during a high-voltage test.

[0027] Figure 10 It is the waveform of the drain-source voltage during a high-voltage test. Detailed Implementation

[0028] Hereinafter, embodiments of the present disclosure will be described based on the accompanying drawings. Furthermore, in each of the following embodiments, the same or equivalent parts will be given the same reference numerals for description.

[0029] (First Embodiment)

[0030] The first embodiment will be described. The semiconductor device manufacturing apparatus of this embodiment includes... Figure 1 The inspection apparatus shown includes a chip tester 1, a worktable 2, a digitizer 3, and a control unit 4.

[0031] The chip tester 1 is used to apply voltage to the semiconductor element under inspection, namely semiconductor element S1, and to measure the voltage, current, etc., output from semiconductor element S1. In this embodiment, semiconductor element S1 is a chip-based MOSFET element used in vehicle inverters, etc. Semiconductor element S1 corresponds to the first semiconductor element.

[0032] like Figure 1 As shown, a semiconductor device S1 is placed on the upper surface of a worktable 2. The upper surface of the semiconductor device S1 is connected to the chip testing machine 1 via a plurality of probes 1a. The chip testing machine 1 applies a voltage to the semiconductor device S1 via the plurality of probes 1a and measures the output of the semiconductor device S1.

[0033] The voltage and current measured by the chip tester 1 are input into the digitizer 3. The digitizer 3 converts the input signals into digital signals and sends them to the control unit 4.

[0034] The control unit 4 is used to control the chip tester 1 based on data sent from the chip tester 1 via the digitizer 3. The control unit 4 is composed of a microcomputer, which includes a CPU (not shown) and a storage unit composed of non-movable physical storage media such as ROM, RAM, flash memory, and HDD. CPU stands for Central Processing Unit, ROM for Read Only Memory, RAM for Random Access Memory, and HDD for Hard Disk Drive.

[0035] The control unit 4 stores the learned model created in step S102 (described later) and controls the chip tester 1 based on the output obtained by inputting the data sent from the chip tester 1 into the learned model.

[0036] Multiple circuit elements are arranged inside the chip tester 1. By touching the semiconductor element S1 with the probe 1a, the tester can form a circuit consisting of the semiconductor element S1 and the circuit elements arranged inside the chip tester 1. Figure 2 The inspection circuit shown.

[0037] In the inspection circuit, MOSFET elements 12 and 13 are connected to power supply 11. Specifically, the drain electrode of MOSFET element 12 is connected to the positive terminal of power supply 11, and the source electrode is connected to the drain electrode of MOSFET element 13. The source electrode of MOSFET element 13 is grounded. MOSFET elements 12 and 13 switch between on and off states according to the voltage supplied to their gate electrodes from power supplies 14 and 15.

[0038] Furthermore, MOSFET element 16, MOSFET element 17, IGBT element 18, and semiconductor element S1 are connected in series between the positive and negative terminals of power supply 11. Specifically, the drain electrode of MOSFET element 16 is connected to the positive terminal of power supply 11, and its source electrode is connected to the source electrode of MOSFET element 17. The drain electrode of MOSFET element 17 is connected to the collector electrode of IGBT element 18, and the emitter electrode of IGBT element 18 is connected to the drain electrode of semiconductor element S1. The source electrode of semiconductor element S1 is grounded.

[0039] MOSFET element 17 and IGBT element 18 switch on and off according to the voltage supplied to their gate electrodes from power supply 19 and power supply 20, respectively. The connection points of MOSFET elements 12 and MOSFET element 13 are connected to the connection points of MOSFET elements 17 and IGBT element 18 via coil 21.

[0040] In addition to the drain electrode, gate electrode, and source electrode, the semiconductor element S1 also has a Kelvin source electrode. The power supply 22 that applies the gate voltage to the semiconductor element S1 is located between the gate electrode and the Kelvin source electrode of the semiconductor element S1.

[0041] A gate drive circuit 23 is disposed between the power supply 22 and the gate electrode of the semiconductor element S1. The gate drive circuit 23 consists of a diode 24, a resistor 25, and a diode 26 and a resistor 27 connected in parallel with respect to the diode 24 and the resistor 25. The anode electrode of the diode 24 is connected to the power supply 22, and the cathode electrode of the diode 26 is connected to the power supply 22.

[0042] A voltmeter 28 is disposed between the drain electrode and the Kelvin source electrode of the semiconductor element S1. Furthermore, a voltmeter 29 is disposed between the gate electrode and the Kelvin source electrode of the semiconductor element S1. Voltmeters 28 and 29 are used to measure the drain-source voltage Vds and the gate-source voltage Vgs of the semiconductor element S1.

[0043] A galvanometer 30 is disposed between the IGBT element 18 and the semiconductor element S1. The galvanometer 30 is used to measure the drain current Id of the semiconductor element S1. The measurement results of the drain-source voltage Vds, gate-source voltage Vgs, and drain current Id obtained by the voltmeter 28, voltmeter 29, and galvanometer 30 are input into the digitizer 3 and sent to the control unit 4.

[0044] A power supply 31 and a gate drive circuit 32 are disposed between the gate electrode and the source electrode of the MOSFET element 16. The gate drive circuit 32 consists of a diode 33, a resistor 34, and a diode 35 and a resistor 36 connected in parallel with respect to the diode 33 and the resistor 34. The anode electrode of the diode 33 is connected to the power supply 31, and the cathode electrode of the diode 35 is connected to the power supply 31.

[0045] A voltmeter 37 and a voltmeter 38 are disposed between the drain and source electrodes and between the gate and source electrodes of the MOSFET element 16. The voltmeters 37 and 38 are used to measure the drain-source voltage and the gate-source voltage of the MOSFET element 16. A current meter 39 is disposed between the MOSFET element 16 and the MOSFET element 17. The current meter 39 is used to measure the drain current of the MOSFET element 16. The measurement results of the drain-source voltage, gate-source voltage, and drain current of the MOSFET element 16 obtained by the voltmeters 37, 38, and 39 are input into the digitizer 3 and sent to the control unit 4.

[0046] In this testing circuit, when current flows through semiconductor element S1, the output voltages of power supplies 14, 19, 20, and 22 are set to a high level, causing MOSFET elements 12, 17, 18, and S1 to be in a conducting state. Conversely, the output voltages of power supplies 15 and 31 are set to a low level, causing MOSFET elements 13 and 16 to be in a cutoff state. Thus, with... Figure 2 The current flows through the path indicated by arrow A1.

[0047] In this state, when the current flowing in semiconductor element S1 is cut off, the output voltage of power supply 20 is set to a low level, causing IGBT element 18 to be in a cut-off state. As a result, the current flowing in semiconductor element S1 is cut off, and the current flowing in coil 21 returns along the path shown by arrow A2.

[0048] A method for manufacturing a semiconductor device will be described. In this embodiment, by... Figure 3 Steps S101 to S109 shown are used to inspect the semiconductor device.

[0049] In step S101, a MOSFET element with the same design as semiconductor element S1 is prepared as semiconductor element S2, and teacher data containing multiple data related to the characteristics of semiconductor element S2 is obtained. Semiconductor element S2 is the semiconductor element used to obtain teacher data, equivalent to a second semiconductor element. Furthermore, multiple semiconductor elements S2 are prepared, and multiple data are obtained for each semiconductor element S2.

[0050] In this embodiment, waveform data related to the electrical characteristics of the semiconductor element S2 is used as multiple data related to the characteristics of the semiconductor element S2. Specifically, waveform data related to the electrical characteristics of the semiconductor element S2 is used. Figure 1 The inspection apparatus shown performs the same inspections on the semiconductor element S2 as in steps S106 and S107 described later. Specifically, for multiple semiconductor elements S2, DC tests are performed to investigate gate leakage current, drain leakage current, threshold voltage Vth, on-state voltage Von, withstand voltage, etc., as well as low-voltage AC tests. Thus, for each semiconductor element S2, waveform data such as Id-Vgs characteristics, composed of multiple data points, are obtained.

[0051] For example, by conducting a withstand voltage test on semiconductor element S2 at around 1200V, the following results can be obtained. Figure 4 The waveform data of the Id-Vgs characteristic is shown as a solid or dashed line. Furthermore, by varying the gate voltage of semiconductor element S2 within the range of 0 to 5V, the waveform data is obtained. Figure 5 The waveform data of the Id-Vgs characteristic is shown as solid or dashed lines.

[0052] For semiconductor element S2, after the above-described checks, the same checks as in step S109 described later are performed. That is, for semiconductor element S2, a high-voltage AC test is performed, which increases the likelihood of element damage compared to the checks described above. The results of this test are also included in the teacher data. Figure 4 , Figure 5 The solid line represents the waveform data of semiconductor element S2 that was not damaged during the high-voltage test, while the dashed line represents the waveform data of semiconductor element S2 that was damaged during the high-voltage test.

[0053] In step S102, a learned model is created using machine learning based on the teacher data obtained in step S101. For example, a learned model can be created using neural networks, deep learning, etc.

[0054] A learned model is created to output a value related to the probability of component failure when multiple data related to the characteristics of semiconductor component S1 are input. In this embodiment, the multiple data related to the characteristics of semiconductor component S1 are set as waveform data related to the electrical characteristics of semiconductor component S1. For example, a learned model is created so that when waveform data such as Id-Vgs characteristics obtained in tests such as withstand voltage tests where the probability of component failure of semiconductor component S1 is low are input, the probability of component failure of semiconductor component S1 is output in a high-voltage AC test where the probability of component failure is high. The control unit 4 stores the created learned model.

[0055] In step S103, an inspection device (not shown) is used to inspect the wafer forming the semiconductor element S1 for defects. For example, light is shone onto the wafer, and surface defects and internal defects are detected based on the intensity of the reflected light.

[0056] In step S104, a semiconductor process is performed on the wafer to form multiple MOSFET elements, and the wafer is diced to divide it into chip units. This results in the formation of multiple chip-shaped semiconductor elements S1.

[0057] In step S105, the same defect inspection as in step S103 is performed on the multiple semiconductor elements S1. Furthermore, a visual inspection is performed on the multiple semiconductor elements S1 to check for the presence of cracks, etc.

[0058] After step S105, an electrical check is performed on the semiconductor element S1. This electrical check includes a first electrical check and a second electrical check performed after the first electrical check. The second electrical check is one that makes the semiconductor element S1 more susceptible to damage compared to the first electrical check. Specifically, the second electrical check involves applying a higher voltage to the semiconductor element S1 than the first electrical check. In this embodiment, steps S106 and S107 correspond to the first electrical check, and step S109 corresponds to the second electrical check.

[0059] In step S106, using Figure 1 The inspection device shown performs a DC test on the semiconductor element S1, obtains waveform data such as gate leakage current, drain leakage current, threshold voltage Vth, conduction voltage Von, and withstand voltage, and sends them to the control unit 4.

[0060] In step S107, using Figure 1 The inspection device shown performs an AC test on the semiconductor element S1, obtains waveform data related to the characteristics of the semiconductor element S1, and sends it to the control unit 4. The AC test performed in step S107 is a low-voltage switching test, etc.

[0061] In step S108, the control unit 4 determines whether to perform step S109 for the semiconductor device S1 after steps S106 and S107. Specifically, the control unit 4 inputs multiple data related to the characteristics of the semiconductor device S1 to the learned model created in step S102. As described above, in this embodiment, these multiple data are waveform data related to the electrical characteristics of the semiconductor device S1, and the control unit 4 inputs the waveform data obtained in steps S106 and S107 into the learned model. Furthermore, based on the output obtained therefrom, the control unit 4 determines whether to perform step S109 for the semiconductor device S1. For example, if the value representing the probability of device failure output from the learned model is less than a predetermined value, the control unit 4 determines to perform step S109; if the value is greater than or equal to the predetermined value, it determines not to perform step S109.

[0062] In step S109, for the semiconductor element S1 that was determined to be subject to step S109 in step S108, a high-voltage switching test that is prone to damage to the element is performed, and the switching withstand capacity, avalanche withstand capacity, etc. are investigated.

[0063] Then, the semiconductor element S1, which has been determined to be a qualified product in the inspections including steps S103, S105, S106, S107, and S109, is packaged. In this way, a semiconductor device containing the semiconductor element S1 is manufactured.

[0064] The effects of this embodiment will be explained. If the results of DC tests, etc., performed on multiple semiconductor elements S2 are converted into one point of data for each semiconductor element S2 to create teacher data, sufficient judgment accuracy cannot be obtained in step S108 due to the large amount of missing information. For example, if when obtaining teacher data, select... Figure 4 If the data at point P1 or point P2 is such that the difference between the semiconductor element S2 that has suffered damage and the semiconductor element S2 that has not suffered damage becomes smaller, it becomes difficult to reflect the difference between the two types of data in the learned model, resulting in a decrease in judgment accuracy.

[0065] In contrast, in this embodiment, teacher data is created to include... Figure 4 Waveform data, etc. Therefore, for example, besides... Figure 4 In addition to points P1 and P2, the teacher data also includes data from points P3, P4, and P5 where differences were generated in the two waveforms. Therefore, the difference between the two types of data is easily reflected in the learned model, thus improving the accuracy of the judgment.

[0066] As explained above, in this embodiment, multiple data related to the characteristics of semiconductor element S2 are included in the teacher data, and the implementation determination of the high-voltage test is performed using multiple data related to the characteristics of semiconductor element S1. Therefore, the determination accuracy is improved, and damage to semiconductor element S1 can be suppressed. As a result, damage to instruments such as probe 1a and foreign matter adhesion caused by component damage can be suppressed, and the increase in chip cost caused by decreased production efficiency and increased instrument costs can be reduced.

[0067] (Second Implementation)

[0068] The second embodiment will be described. This embodiment changes the timing of the electrical inspection determination compared to the first embodiment; otherwise, it is the same as the first embodiment. Therefore, only the parts that are different from the first embodiment will be described.

[0069] In this embodiment, the determination of whether to perform the electrical inspection is to be conducted midway through the second electrical inspection. Specifically, as follows: Figure 6 As shown, after step S107, step S108 is not performed; instead, step S109 begins. Furthermore, in step S109, as... Figure 7 As shown, steps S201 to S203 are performed sequentially.

[0070] In step S101 of this embodiment, using Figure 1 The inspection apparatus shown performs the same high-voltage test on multiple semiconductor elements S2 as in step S109 of the first embodiment. Next, it obtains... Figures 8-10The waveform data of drain current Id, gate-source voltage Vgs, and drain-source voltage Vds are shown. Additionally, data on whether the device was damaged by the high-voltage test and the time elapsed until the device was damaged are obtained.

[0071] Figures 8-10 The solid line represents the waveform data of semiconductor element S2 that was not damaged during the high-voltage test, while the dashed line represents the waveform data of semiconductor element S2 that was damaged during the high-voltage test. Figures 8-10 During the test, the component failed at time t1 after the start of the test, and the drain current Id increased sharply. Then, at time t2 after the start of the test, the drain current Id exceeded the specified value. Figure 2 The IGBT element 18 is set to the off state, the current flowing in the semiconductor element S2 is cut off, and the drain current Id is reduced.

[0072] like Figure 8 , Figure 9 As shown, in the waveform data of the damaged semiconductor device S2 during the high-voltage test, compared with the waveform data of the undamaged semiconductor device S2, the values ​​of the drain current Id and the gate-source voltage Vgs are smaller at the beginning of the test. This is believed to be because, for example, cracks begin to appear in the gate oxide film after the start of the test, resulting in a decrease in the effective gate voltage.

[0073] In step S102, a learned model is created using machine learning with teacher data that includes the data obtained in step S101. The learned model is created such that when waveform data from a high-voltage test of semiconductor element S1 is input, it outputs the probability that semiconductor element S1 will be damaged under continuous high-voltage testing.

[0074] In step S201, a high-voltage test is performed on semiconductor element S1. In step S201, the test is ended in a shorter time than in step S109 of the first embodiment. The time t3 until the end of the test is set to be shorter than the time t1 during which semiconductor element S2 was damaged in step S101. Furthermore, in step S201, the waveform data of the drain-source voltage Vds, gate-source voltage Vgs, and drain current Id during the high-voltage test are transmitted to the control unit 4 by the digitizer 4.

[0075] In step S202, the control unit 4 determines whether to proceed to step S203 for the semiconductor element S1 after step S201, based on the waveform data sent from the digitizer 4. Specifically, the control unit 4 inputs the waveform data obtained in step S201 for the semiconductor element S1 to be determined into the learned model created in step S102. Furthermore, the control unit 4 determines whether to implement step S203 for the semiconductor element S1 based on the output obtained therefrom. For example, if the value representing the probability of element failure output from the learned model is less than a predetermined value, the control unit 4 determines to implement step S203; if the value is greater than or equal to the predetermined value, it determines not to implement step S203.

[0076] In step S203, for the semiconductor element S1 that was determined to be subject to step S203 in step S202, a high-voltage test is continued.

[0077] This embodiment achieves the same effect as the first embodiment through the same structure and operation.

[0078] (Other implementation methods)

[0079] Furthermore, this disclosure is not limited to the embodiments described above and can be appropriately modified. In addition, in each of the above embodiments, the elements constituting the embodiment are not necessarily essential, except where they are specifically stated to be necessary or where they are clearly considered necessary in principle. Furthermore, in each of the above embodiments, when referring to the number, value, quantity, range, etc., of the constituent elements of the embodiment, they are not limited to that specific quantity, except where they are specifically stated to be necessary or where they are clearly considered to be limited to that specific quantity in principle.

[0080] For example, in WAT (Wafer Acceptance Test) and packaging inspection, electrical inspection can be performed as in the first and second embodiments. This helps to suppress damage to the equipment caused by damage to the wafer or package.

[0081] Furthermore, in the second embodiment, step S108 can also be performed as in the first embodiment. That is, step S108 can be performed after step S107, with steps S201 and S202 performed for semiconductor element S1 determined to be subject to step S109, and step S203 performed for semiconductor element S1 determined to be subject to step S203. In this case, the semiconductor element S2 is also checked in the same way as in steps S106 and S107 in step S101, and teacher data containing the data obtained therefrom is created. Furthermore, in step S102, a learned model of the data obtained in steps S106 and S107 inputting about semiconductor element S1, and a learned model of the data obtained in step S201 inputting about about semiconductor element S1 are created for determination in steps S108 and S202.

[0082] Alternatively, steps S101 and S102 can be performed at a different time point than in the first and second embodiments. For example, steps S101 and S102 can be performed after steps S103 to S105.

[0083] Furthermore, image data of semiconductor elements S1 and S2 can also be used as multiple data related to the characteristics of semiconductor elements S1 and S2. For example, a learned model can be created using teacher data containing image data obtained through defect inspection of semiconductor element S2. In step S108, the image data obtained through defect inspection of semiconductor element S1 is input into the learned model, and the implementation determination in step S109 is performed. Alternatively, a learned model can be created using teacher data containing image data obtained through appearance inspection of semiconductor element S2. In step S108, the image data obtained through appearance inspection of semiconductor element S1 is input into the learned model, and the implementation determination in step S109 is performed. When using image data, both waveform data and image data related to the electrical characteristics of semiconductor elements S1 and S2 can be used as multiple data related to the characteristics of semiconductor elements S1 and S2, or only image data can be used without waveform data. Furthermore, when using image data, both image data obtained during defect inspection and image data obtained during appearance inspection can be used. Among the multiple data related to the characteristics of semiconductor elements S1 and S2, there may also be data that is different from the waveform data and image data mentioned above.

[0084] Furthermore, for semiconductor device S1 that is determined not to require steps S109 and S203 based on steps S108 and S202, the same checks as in steps S109 and S203 can be performed, and the resulting data can be included in the teacher data. That is, it is also possible to investigate whether semiconductor device S1, which is deemed to have a high probability of failure, actually failed during the high-voltage test, and reflect the results in the teacher data. This further improves the accuracy of the determinations made in steps S108 and S202.

[0085] The control unit and method described in this disclosure can also be implemented by a special-purpose computer consisting of a processor and memory programmed to perform one or more functions embodied in a computer program. Alternatively, the control unit and method described in this disclosure can also be implemented by a special-purpose computer consisting of a processor configured with one or more special-purpose hardware logic circuits. Alternatively, the control unit and method described in this disclosure can also be implemented by one or more special-purpose computers consisting of a processor and memory programmed to perform one or more functions and a processor configured with one or more hardware logic circuits. Furthermore, the computer program can also be stored as instructions to be executed by the computer in a computer-readable non-removable tangible recording medium.

[0086] (The technical solution disclosed herein)

[0087] [Technical Solution 1]

[0088] A method for manufacturing a semiconductor device, wherein the semiconductor element to be inspected is a first semiconductor element (S1), and the semiconductor element used to obtain teacher data is a second semiconductor element (S2), and the following steps are performed: obtaining the teacher data containing multiple data related to the characteristics of the second semiconductor element; using the teacher data to create a learned model; and determining whether to perform an electrical inspection of the first semiconductor element based on the output obtained by inputting multiple data related to the characteristics of the first semiconductor element into the learned model.

[0089] [Second Technical Solution]

[0090] The semiconductor device manufacturing method according to the first technical solution includes multiple data related to the characteristics of the first semiconductor element, including waveform data related to the electrical characteristics of the first semiconductor element, and multiple data related to the characteristics of the second semiconductor element, including waveform data related to the electrical characteristics of the second semiconductor element.

[0091] [Technical Solution 3]

[0092] In the semiconductor device manufacturing method described in the first or second technical solution, the electrical inspection includes a first electrical inspection and a second electrical inspection performed after the first electrical inspection. The teacher data includes data obtained through the first electrical inspection of the second semiconductor element. In the process of determining whether to perform the electrical inspection, the second electrical inspection is determined based on the output obtained by inputting the data obtained through the first electrical inspection of the first semiconductor element into the learned model.

[0093] [Technical Solution 4]

[0094] The semiconductor device manufacturing method as described in the first or second technical solution includes an electrical inspection comprising a first electrical inspection and a second electrical inspection performed after the first electrical inspection, wherein a higher voltage than that applied in the first electrical inspection is applied. The teacher data includes data obtained through the second electrical inspection of the second semiconductor element. The process of determining whether to perform the electrical inspection is performed midway through the second electrical inspection. In the process of determining whether to perform the electrical inspection, the process of determining whether to continue performing the second electrical inspection is based on the output obtained by inputting the data obtained up to the midway through the second electrical inspection of the first semiconductor element into the learned model.

[0095] [Fifth Technical Solution]

[0096] In the semiconductor device manufacturing method described in the first or second technical solution, the electrical inspection includes a first electrical inspection and a second electrical inspection performed after the first electrical inspection, applying a voltage higher than that of the first electrical inspection. The teacher data includes data obtained through the first electrical inspection of the second semiconductor element and data obtained through the second electrical inspection. The process of determining whether to perform the electrical inspection is performed before the second electrical inspection and during the second electrical inspection.

[0097] In the process of determining whether to perform the aforementioned electrical inspection before the second electrical inspection, the determination of whether to perform the second electrical inspection is based on the output obtained by inputting the data obtained from the first electrical inspection of the first semiconductor element into the learned model.

[0098] In the process of determining whether to perform the electrical inspection during the second electrical inspection, the decision on whether to continue performing the second electrical inspection is based on the output obtained by inputting the data obtained up to the middle of the second electrical inspection of the first semiconductor element into the learned model.

[0099] [Sixth Technical Solution]

[0100] The method for manufacturing a semiconductor device according to any one of the first to fifth technical solutions includes multiple data relating to the characteristics of the first semiconductor element, including image data of the first semiconductor element, and multiple data relating to the characteristics of the second semiconductor element, including image data of the second semiconductor element.

[0101] [Seventh Technical Solution]

[0102] The method for manufacturing a semiconductor device according to any one of the first to sixth technical solutions, wherein the teacher data includes data obtained by defect inspection of the second semiconductor element, and in the process of determining whether to perform the electrical inspection, the data obtained by defect inspection of the first semiconductor element is input into the learned model.

[0103] [Technical Solution No. 8]

[0104] The method for manufacturing a semiconductor device as described in any of the first to seventh technical solutions, wherein the teacher data includes data obtained through the visual inspection of the second semiconductor element, and in the process of determining whether to perform the electrical inspection, the data obtained through the visual inspection of the first semiconductor element is input into the learned model.

[0105] [Technical Solution No. 9]

[0106] The method for manufacturing a semiconductor device as described in any of the first to eighth technical solutions, wherein the aforementioned teacher data includes data obtained by performing the same electrical checks on the first semiconductor element that is determined not to have undergone the aforementioned electrical checks.

[0107] [Technical Solution No. 10]

[0108] A semiconductor device manufacturing apparatus includes a control unit (4) that determines whether to perform an electrical inspection of the semiconductor device based on the output obtained by inputting multiple data related to the characteristics of the semiconductor element (S1) into a learned model.

[0109] [Technical Solution No. 11]

[0110] The semiconductor device manufacturing apparatus as described in the 10th technical solution includes multiple data related to the characteristics of the semiconductor element, including waveform data related to the electrical characteristics of the semiconductor element.

[0111] [Technical Solution No. 12]

[0112] In the semiconductor device manufacturing apparatus described in the 10th or 11th technical solution, the electrical inspection includes a first electrical inspection and a second electrical inspection performed after the first electrical inspection. The control unit determines whether to perform the second electrical inspection based on the output obtained by inputting the data obtained from the first electrical inspection of the semiconductor element into the learned model.

[0113] [Technical Solution No. 13]

[0114] In the semiconductor device manufacturing apparatus described in the 10th or 11th technical solution, the electrical check includes a first electrical check and a second electrical check performed after the first electrical check by applying a voltage higher than that of the first electrical check. The control unit determines whether to continue performing the second electrical check based on the output obtained by inputting data obtained up to the middle of the second electrical check of the semiconductor element into the learned model.

[0115] [Technical Solution No. 14]

[0116] In the semiconductor device manufacturing apparatus described in the 10th or 11th technical solution, the electrical inspection includes a first electrical inspection and a second electrical inspection performed after the first electrical inspection, which applies a voltage higher than that of the first electrical inspection. The control unit determines whether to perform the second electrical inspection based on the output obtained by inputting data obtained through the first electrical inspection of the semiconductor element into the learned model, and determines whether to continue performing the second electrical inspection based on the output obtained by inputting data obtained up to the middle of the second electrical inspection of the semiconductor element into the learned model.

[0117] [Technical Solution No. 15]

[0118] The semiconductor device manufacturing apparatus according to any one of the 10th to 14th technical solutions includes multiple data related to the characteristics of the semiconductor element, including image data of the semiconductor element.

[0119] [Technical Solution No. 16]

[0120] In the semiconductor device manufacturing apparatus described in any of the 10th to 15th technical solutions, the control unit inputs data obtained through defect inspection of the semiconductor element into the learned model.

[0121] [Technical Solution No. 17]

[0122] In the semiconductor device manufacturing apparatus described in any of the 10th to 16th technical solutions, the control unit inputs data obtained through the appearance inspection of the semiconductor element into the learned model.

Claims

1. A method for manufacturing a semiconductor device, characterized in that: Assume that the semiconductor element to be inspected is a first semiconductor element (S1), Assume that the semiconductor element used to obtain the teacher data is the second semiconductor element (S2), Carry out the following process: acquiring the teacher data including a plurality of data related to the characteristics of the second semiconductor element; Using the above teacher data, produce a learned model; and Whether or not to perform an electrical inspection of the first semiconductor element is determined based on an output obtained by inputting a plurality of data related to characteristics of the first semiconductor element into the learned model.

2. The method for manufacturing a semiconductor device according to claim 1, wherein: the plurality of data related to the characteristics of the first semiconductor element include waveform data related to the electrical characteristics of the first semiconductor element, The plurality of data on the characteristics of the second semiconductor element include waveform data on the electrical characteristics of the second semiconductor element.

3. The method for manufacturing a semiconductor device according to claim 1 or 2, wherein: The electrical inspection includes a first electrical inspection and a second electrical inspection performed after the first electrical inspection. The teacher data includes data obtained through the first electrical inspection of the second semiconductor element. In the step of determining whether to perform the electrical inspection, it is determined whether to perform the second electrical inspection based on an output obtained by inputting data obtained through the first electrical inspection of the first semiconductor element into the learned model.

4. The method for manufacturing a semiconductor device according to claim 1 or 2, wherein: The electrical inspection includes a first electrical inspection and a second electrical inspection performed after the first electrical inspection and applying a higher voltage than the first electrical inspection. The teacher data includes data obtained through the second electrical inspection of the second semiconductor element. The step of determining whether to implement the electrical inspection is performed during the second electrical inspection. In the step of determining whether to perform the electrical inspection, it is determined whether to continue performing the second electrical inspection based on an output obtained by inputting data obtained up to the middle of the second electrical inspection of the first semiconductor element into the learned model.

5. The method for manufacturing a semiconductor device according to claim 1 or 2, wherein: The electrical inspection includes a first electrical inspection and a second electrical inspection performed after the first electrical inspection and applying a higher voltage than the first electrical inspection. The teacher data includes data obtained through the first electrical inspection of the second semiconductor element and data obtained through the second electrical inspection. The step of determining whether to implement the electrical inspection is performed before the second electrical inspection and during the second electrical inspection. In the step of determining whether to perform the electrical inspection before the second electrical inspection, determining whether to perform the second electrical inspection is based on an output obtained by inputting data obtained through the first electrical inspection of the first semiconductor element into the learned model, In the process of determining whether to implement the electrical inspection during the second electrical inspection, whether to continue the second electrical inspection is determined based on the output obtained by inputting data obtained up to the middle of the second electrical inspection of the first semiconductor element into the learned model.

6. The method for manufacturing a semiconductor device according to claim 1 or 2, wherein: the plurality of data related to the characteristics of the first semiconductor element include image data of the first semiconductor element, The plurality of data related to the characteristics of the second semiconductor element include image data of the second semiconductor element.

7. The method for manufacturing a semiconductor device according to claim 1 or 2, wherein: The teacher data includes data obtained by defect inspection of the second semiconductor element. In the step of determining whether to perform the electrical inspection, data obtained through the defect inspection of the first semiconductor element is input into the learned model.

8. The method for manufacturing a semiconductor device according to claim 1 or 2, wherein: The teacher data includes data obtained through appearance inspection of the second semiconductor element. In the step of determining whether to perform the electrical inspection, data obtained through the appearance inspection of the first semiconductor element is input into the learned model.

9. The method for manufacturing a semiconductor device according to claim 1 or 2, wherein: The teacher data includes data obtained by performing an inspection similar to the electrical inspection on the first semiconductor element determined not to be subjected to the electrical inspection.

10. A semiconductor device manufacturing apparatus, characterized in that: A control unit (4) is provided for determining whether to perform an electrical inspection of the semiconductor element (S1) based on an output obtained by inputting a plurality of data related to characteristics of the semiconductor element (S1) into a learned model.

11. The semiconductor device manufacturing apparatus according to claim 10, wherein: The plurality of data related to the characteristics of the semiconductor element include waveform data related to the electrical characteristics of the semiconductor element.

12. The semiconductor device manufacturing apparatus according to claim 10 or 11, characterized in that: The electrical inspection includes a first electrical inspection and a second electrical inspection performed after the first electrical inspection. The control unit determines whether to perform the second electrical inspection based on an output obtained by inputting data obtained through the first electrical inspection of the semiconductor element into the learned model.

13. The semiconductor device manufacturing apparatus according to claim 10 or 11, characterized in that: The electrical inspection includes a first electrical inspection and a second electrical inspection performed after the first electrical inspection and applying a higher voltage than the first electrical inspection. The control unit determines whether to continue the second electrical inspection of the semiconductor element based on an output obtained by inputting data obtained up to the middle of the second electrical inspection into the learned model.

14. The semiconductor device manufacturing apparatus according to claim 10 or 11, characterized in that: The electrical inspection includes a first electrical inspection and a second electrical inspection performed after the first electrical inspection and applying a higher voltage than the first electrical inspection. The above control unit, determining whether to perform the second electrical inspection based on an output obtained by inputting data obtained by the first electrical inspection of the semiconductor element into the learned model, Whether or not to continue the second electrical inspection of the semiconductor element is determined based on an output obtained by inputting data obtained up to the middle of the second electrical inspection of the semiconductor element into the learned model.

15. The semiconductor device manufacturing apparatus according to claim 10 or 11, characterized in that: The plurality of data related to the characteristics of the semiconductor element include image data of the semiconductor element.

16. The semiconductor device manufacturing apparatus according to claim 10 or 11, characterized in that: The control unit inputs data obtained through defect inspection of the semiconductor element into the learned model.

17. The semiconductor device manufacturing apparatus according to claim 10 or 11, characterized in that: The control unit inputs data obtained through the appearance inspection of the semiconductor device into the learned model.

Images