Detection device, system and method

Abstract

The embodiment of the invention provides a detection device, system and method, relates to the technical field of integrated circuits, and is used for improving the detection precision and detection efficiency of detection equipment. The detection device comprises a machine table; a controller, an optical detector and a charged particle detector are arranged on the machine table; the controller receives detection data generated when the optical detector performs optical detection on the semiconductor sample, wherein the detection data comprises position information of at least one defect on the semiconductor sample; the controller outputs the first position information to the charged particle detector, and controls the charged particle detector to detect the defect indicated by the first position information on the semiconductor sample through charged particle beam detection; wherein the defect indicated by the first position information comprises at least one defect in the detection data.

Description

Technical Field

[0001] This application relates to the field of integrated circuit technology, and in particular to a detection device, system, and method. Background Technology

[0002] In the manufacturing process of integrated circuits (ICs), both unfinished and manufactured circuit components need to be inspected to ensure they are free of defects. Furthermore, bare wafers or unpatterned wafers must be inspected before IC manufacturing to ensure they are defect-free or meet required specifications. However, as semiconductor processes continue to evolve and wafer sizes shrink, the precision and efficiency of inspection equipment are crucial to shorten development cycles and improve product yield. Summary of the Invention

[0003] Embodiments of this application provide a detection apparatus, system, and method for improving the detection accuracy and efficiency of detection equipment.

[0004] To achieve the above objectives, the embodiments of this application adopt the following technical solutions:

[0005] In a first aspect, a detection apparatus is provided, comprising: a machine base; a controller, an optical detector, and a charged particle detector are disposed on the machine base; the controller receives detection data generated by the optical detector performing optical detection on a semiconductor sample, the detection data including location information of at least one defect on the semiconductor sample; the controller outputs first location information to the charged particle detector, and controls the charged particle detector to detect the defect indicated by the first location information on the semiconductor sample by detecting a charged particle beam; wherein the defect indicated by the first location information includes at least one defect in the detection data.

[0006] The detection apparatus provided in this application integrates an optical inspection instrument and a charged particle detector on a single machine. This allows the high-efficiency optical inspection instrument to first perform a large-area inspection of the sample to comprehensively detect defects in the semiconductor sample. Simultaneously, to avoid false detections by the optical inspection instrument, after the optical inspection is completed, the semiconductor sample needs to be transferred to the charged particle detector for further inspection. The charged particle detector can perform a secondary inspection only on the areas where defects were detected by the optical inspection instrument, thereby improving detection accuracy. Furthermore, by reducing the detection area of ​​the charged ion beam detector, the detection speed and efficiency of the sample are improved.

[0007] In one possible implementation, the controller is also configured to send detection data to the data center and receive first location information from the data center, wherein the defect indicated by the first location information includes a portion of the defect indicated by the detection data.

[0008] In this embodiment, considering the low resolution of the optical inspection instrument, which may produce false alarms, the data center needs to screen the detection data detected by the optical inspection instrument, screen out the defects that may be false alarms, and generate first location information so as to instruct the charged particle detector to perform secondary detection on the defects that may be false alarms.

[0009] In one possible implementation, the confidence level of the defect indicated by the first location information is lower than a first threshold.

[0010] In this embodiment, the data center can perform confidence calculations on the detection data detected by the optical detector, thereby filtering out defects that may be false alarms and generating first location information to instruct the charged particle detector to perform secondary detection on defects that may be false alarms.

[0011] In one possible implementation, the detection data and the first position information correspond to the same coordinate system. This eliminates the need for the charged particle detector to re-establish the coordinate system, thus reducing detection time.

[0012] In one possible implementation, positioning points are provided on the semiconductor sample; the controller is also used to construct a coordinate system based on the positioning points, and the detection data is the coordinates of the defects on the semiconductor sample in the coordinate system.

[0013] In one possible implementation, the controller is also used to align the semiconductor sample with the charged particle detector based on the correspondence between the positioning point and the coordinate system. This eliminates the need for the charged particle detector to re-establish the coordinate system, thus reducing detection time.

[0014] In one possible implementation, the first location information is the coordinates of at least one defect in a coordinate system; the controller is used to control the charged particle detector to acquire a first image based on the first location information, the first image being an image of the semiconductor sample detected by the charged particle detector at the coordinates indicated by the first location information.

[0015] In one possible implementation, the instrument includes a stage, a first detection position, and a second detection position. The stage is used to place a semiconductor sample. The first detection position corresponds to an optical detector, and the second detection position corresponds to a charged particle detector. The controller is also used to control the stage to move from the first detection position to the second detection position.

[0016] In one possible implementation, the field of view of the charged particle detector is smaller than that of the optical detector.

[0017] In one possible implementation, the charged particle detector is a single-electron-beam detector or a multi-electron-beam detector.

[0018] In one possible implementation, the controller is configured to receive detection data, the detection data including location information of at least one defect on a semiconductor sample acquired by optical detection; the controller is also configured to output first location information, wherein the defect indicated by the first location information includes at least one defect in the detection data.

[0019] In one possible implementation, the controller is further configured to receive a second image, which is an image of a defect detected by an optical inspection instrument on a semiconductor sample; and to classify at least one defect in the detection data according to the second image and a classification model to obtain first location information.

[0020] In one possible implementation, the controller is further configured to receive a first image, which is an image of a semiconductor sample detected by a charged particle beam at coordinates indicated by first position information; and to optimize a classification model using the first and second images.

[0021] In one possible implementation, the controller is further configured to receive a second image, which is an image of a defect on a semiconductor sample obtained by optical detection; and to calculate the confidence of at least one defect in the detection data based on the second image and a confidence model to obtain first location information, wherein the confidence of the defect indicated by the first location information is lower than a first threshold.

[0022] In a second aspect, a detection system is provided, comprising: a data center and a detection device provided in the first aspect and any possible implementation thereof; the data center is configured to receive detection data, the detection data including location information of at least one defect on a semiconductor sample acquired by optical detection; the data center is further configured to output first location information, wherein the defect indicated by the first location information includes at least one defect in the detection data.

[0023] In one possible implementation, the data center is further configured to receive a second image, which is an image of a defect detected by an optical inspection instrument on a semiconductor sample; and to classify at least one defect in the detection data according to the second image and a classification model to obtain first location information.

[0024] In one possible implementation, the data center is also used to receive a first image, which is an image of a semiconductor sample detected by a charged particle beam at coordinates indicated by first location information; and to optimize a classification model using the first and second images.

[0025] In one possible implementation, the data center is further configured to receive a second image, which is an image of a defect on a semiconductor sample obtained by optical detection; and to calculate the confidence of at least one defect in the detection data based on the second image and a confidence model to obtain first location information, wherein the confidence of the defect indicated by the first location information is lower than a first threshold.

[0026] Thirdly, a detection method is provided, the method comprising: receiving at least one detection data, wherein the detection data is used to indicate the position information of a defect on a semiconductor sample obtained by optical detection; generating first position information, wherein the defect indicated by the first position information includes at least one defect in the detection data; and performing charged particle beam detection on the defect on the semiconductor sample based on the first position information.

[0027] In one possible implementation, the detection method further includes: receiving a second image, which is an image of a defect on a semiconductor sample obtained by optical detection; and classifying at least one defect in the detection data according to the second image and a classification model to obtain first location information.

[0028] In one possible implementation, the detection method further includes: receiving a first image, which is an image of a semiconductor sample detected by a charged particle beam at coordinates indicated by first position information, and optimizing a classification model using the first image and a second image.

[0029] In one possible implementation, the detection method further includes: receiving a second image, the second image being an image of a defect on a semiconductor sample obtained by optical detection; and calculating the confidence of at least one defect in the detection data based on the second image and a confidence model to obtain first location information, wherein the confidence of the defect indicated by the first location information is lower than a first threshold.

[0030] The technical effects of any of the possible implementations of the second and third aspects can be found in the technical effects of the different implementations of the first aspect described above, and will not be repeated here. Attached Figure Description

[0031] Figure 1 This is a schematic diagram of the detection device provided in the embodiments of this application;

[0032] Figure 2 This is a schematic diagram of the detection system provided in the embodiments of this application;

[0033] Figure 3 This is a schematic diagram of the structure of a detection device provided in another embodiment of this application;

[0034] Figure 4This is a schematic diagram of the structure of the optical inspection instrument provided in the embodiments of this application;

[0035] Figure 5 This is a schematic diagram of the structure of a semiconductor sample provided in an embodiment of this application;

[0036] Figure 6 This is a schematic diagram of the structure of the charged particle detector provided in the embodiments of this application;

[0037] Figure 7 This is a schematic diagram of the structure of a detection device provided in another embodiment of this application;

[0038] Figure 8 A schematic flowchart of the detection method provided in the embodiments of this application;

[0039] Figure 9 This is a schematic diagram of the working logic of the detection method provided in the embodiments of this application;

[0040] Figure 10 A schematic flowchart of a detection method provided in another embodiment of this application;

[0041] Figure 11 This is a schematic diagram of the working logic of a detection method provided in another embodiment of this application. Detailed Implementation

[0042] The technical solutions of the embodiments of this application will be described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0043] In the following description, the terms "first," "second," etc., are used for descriptive convenience only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined with "first," "second," etc., may explicitly or implicitly include one or more of that feature. In the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0044] In the embodiments of this application, the terms "exemplary" or "for example" are used to indicate that something is an example, illustration, or description. Any embodiment or design that is described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the terms "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0045] Before being used in IC manufacturing, bare wafers (or "unpatterned" wafers, "blank" wafers) need to be inspected to ensure they are free of defects. For example, the wafers need to be inspected for contamination (e.g., particles, metallic contaminants) and surface quality (e.g., pits or dents, scratches, crystal defects), which can adversely affect wafer yield (e.g., how many high-quality circuit cells can be manufactured from the wafer) or the performance of the manufactured circuits (e.g., short circuits, poor contacts, etc., which may impair normal circuit operation). Furthermore, the results of bare wafer inspection are a good indicator of the cleanliness of the manufacturing or measurement equipment. If the bare wafers have a high defect density, the manufacturing or measurement equipment must be cleaned before starting the manufacturing process.

[0046] For example, for inspection methods with high speed / efficiency, such as optical inspection, full coverage of bare wafers can be performed using optical instruments such as optical microscopes. However, as described above, as the semiconductor industry strives to manufacture smaller circuits, the size of defects that can affect circuit manufacturing and operation and therefore need to be detected also becomes smaller (e.g., reduced to about 10 nm, which is below typical optical wavelengths). Therefore, inspection results reported by optical microscopes may contain a significant amount of inaccuracy, such as false alarms.

[0047] Charged particle (e.g., electron) beam microscopy, such as scanning electron microscopy (SEM), can be used to review defects identified by optical microscopy because electron beams have shorter wavelengths compared to photon beams and thus provide superior spatial resolution. In some specific embodiments, a bare wafer can first be placed under an optical microscope to identify potential defect sites. The bare wafer is then transferred to the SEM, where the potential defect sites can be further examined to determine if they correspond to actual defects. Therefore, it is necessary for the optical microscope to transmit the location of potential defects to the SEM.

[0048] Because some semiconductor samples (e.g., bare wafers) lack patterns, the location of defects cannot be indicated by referring to the wafer itself. Instead, they need to be represented by mathematical coordinates on images produced by optical microscopes and SEMs. However, because the wafer is not aligned correctly on the optical microscope and SEM, the same defect may have different coordinates on the optical and SEM images (hereinafter referred to as "optical plotting coordinates" and "electronic plotting coordinates," respectively). That is, within the SEM's field of view (FOV), the defect may not be precisely located at its optical plotting coordinates. Therefore, to locate the defect, the SEM needs to search around its optical plotting coordinates until it is found. This is very time-consuming, especially because SEM scanning speeds are low.

[0049] Therefore, this application embodiment also provides a detection device 100, such as Figure 1 As shown, the detection device 100 includes: a machine base 110; a controller 120, an optical detector 130, and a charged particle detector 140 are mounted on the machine base 110. The controller 120 is connected to the optical detector 130 and the charged particle detector 140 respectively.

[0050] The controller 120 receives detection data generated by the optical inspection instrument 130 performing optical inspection on the semiconductor sample 150. The detection data includes the location information of at least one defect on the semiconductor sample 150. The controller 120 outputs first location information to the charged particle detector 140 and controls the charged particle detector 140 to detect the defect indicated by the first location information on the semiconductor sample 150 through charged particle detection. The defect indicated by the first location information includes at least one defect in the detection data.

[0051] The detection apparatus 100 provided in this application integrates an optical inspection instrument 130 and a charged particle detector 140 on a single machine base 110. This allows the high-efficiency optical inspection instrument 130 to first perform a large-area inspection of the sample to be tested, comprehensively detecting any defects in the semiconductor sample 150. Simultaneously, to avoid false detections by the optical inspection instrument 130, after the optical inspection instrument 130 completes its inspection, the semiconductor sample 150 needs to be transferred to the charged particle detector 140 for further inspection. The charged particle detector 140 can perform a secondary inspection only on the area where the defects detected by the optical inspection instrument 130 are located, thereby improving the accuracy of the inspection. Furthermore, by reducing the detection area of ​​the charged ion beam detector, the detection speed and efficiency of the sample to be tested are improved.

[0052] This application also provides a detection system, such as... Figure 2 As shown, the detection system includes a data center and a detection device 100. The data center can filter the detection data generated by the optical detector 130 performing optical detection on the semiconductor sample 150, thereby further reducing the detection area of ​​the charged particle detector 140 and improving detection efficiency and accuracy.

[0053] The following section uses semiconductor sample 150 as an example to explain each part of the detection system in detail.

[0054] A data center (DC) is an information infrastructure that utilizes internet communication lines and bandwidth resources to transmit information, accelerate data, display content, compute data, and store resources between different regions around the world. Data centers can deploy power equipment and various types of computing devices, such as servers, management equipment, and security equipment.

[0055] In this embodiment, the data center can process various types of data, such as filtering the data detected by the optical detector 130 and comparing the data detected by the optical detector 130 with the data detected by the charged particle detector 140.

[0056] It is understood that the data center can be deployed in the same space as the detection device 100, or the data center can be set up remotely; this application embodiment does not impose any restrictions.

[0057] The testing device 100 includes a machine base 110, such as Figure 3 As shown, the machine 110 includes a stage 151, a first detection position 131 and a second detection position 141. A semiconductor sample 150 can be placed on the stage 151. The first detection position 131 corresponds to the optical detector 130, and the second detection position 141 corresponds to the charged particle detector 140. A transmission track can be provided between the first detection position 131 and the second detection position 141, so that the controller 120 can control the stage 151 to move from the first detection position 131 to the second detection position 141.

[0058] refer to Figure 4 The optical inspection instrument 130 includes a laser 132 for projecting an incident laser beam 133 onto a bare wafer. The laser beam 133 is scattered by the bare wafer, and the scattered light is detected by a photodetector 134. When the incident laser beam 133 hits a defect on the wafer, the intensity of the resulting scattered light changes. Therefore, potential defects can be detected by analyzing the change in the intensity of the scattered light.

[0059] In some embodiments, considering that the bare wafer does not have a printed pattern that can serve as a reference mark for determining its orientation, the edge of the bare wafer may be provided with positioning points 152 (e.g., Figure 5 The orientation of the wafer is marked by the positioning point 152. Both the optical inspection instrument 130 and the charged particle inspection instrument 140 can be aligned with the bare wafer based on the notch. In this way, there is no need for the charged particle inspection instrument 140 to re-establish a coordinate system. As long as the alignment between the bare wafer and the charged particle inspection instrument 140 is controlled according to the correspondence between the positioning point 152 and the coordinate system, the coordinate system established when the optical inspection instrument 130 inspects the bare wafer can be used.

[0060] When the optical inspection instrument 130 inspects the bare wafer, the controller 120 can control the stage 151 to rotate the bare wafer tangentially and move the wafer radially. In this way, the incident laser beam can illuminate the entire surface of the wafer to detect potential defects. Based on the rotation angle of the bare wafer and the radius of the laser beam, the position coordinates of the defects are calculated and recorded.

[0061] Because the optical inspection instrument 130 has low resolution and is prone to false alarms, a secondary inspection is required by the charged particle detector 140. Therefore, after inspection by the optical inspection instrument 130, the bare wafer is transferred to the second detection position 141 so that the charged particle detector 140 can inspect the bare wafer.

[0062] refer to Figure 6 The charged particle detector 140 may include an electron source 142 configured to generate a charged particle beam 143; at least one condenser lens 144 for converging the charged particle beam 143; a composite lens 145 for focusing the charged particle beam 143 onto a bare wafer; and a detection system 146 for detecting charged particles or X-rays emitted from the bare wafer.

[0063] The electron source 142 can be an electron source capable of generating, accelerating, and converging high-energy electrons, and the electron source 142 includes at least a cathode for emitting electrons. After the electron source 142 is energized, a large number of electrons accumulate on the cathode surface. Under the action of the electric field, the electrons escape from the cathode surface, thus acquiring ionized electrons. The ionized electrons can acquire higher kinetic energy under the action of the accelerating electrode, thereby forming a charged particle beam 143. The condenser lens 144 focuses the charged particle beam 143 to adjust the current of the charged particle beam before it enters the composite mirror 145. After applying a voltage to the composite mirror 145, a magnetic field and an electrostatic field are formed, which can focus the charged particle beam 143 onto the bare wafer and form a detection spot on the surface of the bare wafer. The detection system 146 detects the electrons emitted from the surface of the bare wafer and sends them to the controller 120, thereby realizing imaging.

[0064] although Figure 6 The charged particle detector 140 is shown as a single-electron-beam detector; in some other examples, it may also be a multi-electron-beam detector using multiple charged particle beams. As described above, the charged particle detector 140 can be configured to generate multiple charged particle beams for simultaneously probing multiple regions of the semiconductor sample 150. Correspondingly, the charged particle detector 140 may also include multiple sets of compound mirrors 116 (e.g., multiple sets of conical magnetic lenses and immersion magnetic lenses) for focusing the multiple charged particle beams. The multiple sets of compound mirrors 116 can be controlled jointly or individually to jointly or individually adjust the FoV size of the multiple charged particle beams.

[0065] like Figure 7As shown, controller 120 includes memory 121, image acquisition unit 122, and processor 123. Processor 123 may include a computer, server, mainframe, terminal, personal computer, any kind of mobile computing device, microprocessor-based system, microcontroller, embedded system (e.g., firmware), or any other suitable control circuitry or system. Processor 123 may be specifically configured with hardware or software modules for controlling the operation of charged particle detector 140. For example, processor 123 may change the voltage applied to the lens to adjust the FoV dimension of charged particle detector 140.

[0066] Image acquisition unit 122 may be a computer system similar to processor 123. Image acquisition unit 122 can be connected to photodetector 134 and detection system 146 via a medium such as an electrical conductor, fiber optic cable, portable storage medium, IR, Bluetooth, Internet, wireless network, radio, or a combination thereof. Image acquisition unit 122 can receive signals from photodetector 134 and detection system 146 and can construct an image of semiconductor sample 150. Image acquisition unit 122 can also perform various post-processing functions, such as generating contours, overlaying indicators on the acquired image, etc. Image acquisition unit 122 can be configured to perform adjustments such as brightness and contrast of the acquired image.

[0067] The memory 121 may be a storage medium such as random access memory (RAM), a hard disk, cloud storage, or other types of computer-readable storage. The memory 121 may be coupled to the image acquisition unit 122 and the processor 123. The memory 121 stores computer instructions or programs that can be accessed and executed by the image acquisition unit 122 and the processor 123 for performing functions consistent with this disclosure. The memory 121 may also be used to save scanned raw image data as raw images and post-processed images.

[0068] Figure 8 This is a flowchart illustrating a detection method provided in an embodiment of this application. Figure 9 yes Figure 8 A schematic diagram of the working logic of the detection method is provided. This method is used for bare wafer inspection. For example, this method can be applied to the aforementioned detection system. (Refer to...) Figure 8 The method may include the following steps S10 to S70.

[0069] S10: After the bare wafer is placed on the stage 151, the controller 120 controls the stage 151 to be in the first detection position 131.

[0070] After the bare wafer is placed on the stage 151, the controller 120 controls the position of the stage 151 to position the bare wafer at the center of the laser beam of the optical inspection instrument 130. Then, according to the requirements of the optical inspection instrument 130, the controller controls the optical inspection instrument 130 to calibrate the target lens to ensure that the imaging optical axis of the optical inspection instrument 130 is aligned with the plane of the semiconductor sample 150.

[0071] Then, a specific point is selected on the semiconductor sample 150 as the origin of the coordinate system. Since the bare wafer does not have a printed pattern that can serve as a reference mark for determining its orientation, in some embodiments, positioning points 152 are provided on the edge of the bare wafer (e.g., Figure 5 The positioning point 152 shown is used to mark its orientation. Therefore, the controller 120 can use the positioning point 152 as the origin of the coordinate system and define the coordinate axis orientation based on the origin of the coordinate system: for example, the X-axis and Y-axis are usually set to be horizontal and vertical, respectively, and the Z-axis is the direction perpendicular to the surface of the semiconductor sample 150.

[0072] S20: Controller 120 controls optical inspection instrument 130 to perform optical inspection on bare wafers.

[0073] When the laser beam emitted from the optical inspection unit 130 illuminates the bare wafer, it is scattered on the surface of the wafer, forming scattered light. This scattered light is detected by a photodetector to generate inspection data. In some embodiments, this inspection data can be light scattering data, which includes the intensity of the scattered light. This allows the controller 120 to detect potential defects in the bare wafer by analyzing changes in the intensity of the scattered light.

[0074] S30: The optical inspection instrument 130 sends the inspection data of the bare wafer to the controller 120, so that the controller 120 can send the inspection data to the data center.

[0075] In some embodiments, the controller 120 can also form a second image based on light scattering data, which is an image of the defect detected by the optical inspection instrument 130 on the semiconductor sample 150. Since the intensity of the scattered light changes when the laser beam hits a defect on the bare wafer, the brightness of the affected area in the image also changes accordingly. The controller 120 can detect potential defects on the bare wafer by analyzing the image to obtain the intensity change of the scattered light. Furthermore, the controller 120 can obtain the coordinates of at least one defect of the semiconductor sample 150 in the coordinate system established in the above steps.

[0076] S40: Data center 200 filters and classifies at least one defect in the detection data to obtain first location information.

[0077] Since the optical inspection instrument 130 has low resolution, there may be false alarms. Therefore, the data center 200 needs to screen the defects detected by the optical inspection instrument 130 based on the second image to screen out the defects that may be false alarms, so that the charged particle detector 140 can perform secondary detection on such defects that may be false alarms.

[0078] Considering that even minor defects can affect the electrical performance of a bare wafer in practical applications, these defects can be classified into Category 1, or "critical defects." However, some areas, such as the dummy fill region, have little impact on the final performance of the bare wafer due to defects. Defects in these areas do not affect the chip's electrical performance; these are classified into Category 2, or "unimportant defects." Handling unimportant defects, including automatic defect classification, recording, and storage, is of limited significance. If every potentially falsely reported defect on the bare wafer were to undergo secondary detection using a charged particle beam device, it would waste detection resources and reduce detection efficiency.

[0079] Therefore, in some examples, the data center can input the second image detected by the optical inspection instrument into a trained classification model, which can filter and classify at least one defect displayed in the second image, select the defect that needs to be detected a second time, and generate first position information containing the coordinates of the defect, so that the controller 120 can control the charged particle detector 140 to perform charged particle beam detection on the defect indicated by the first position information on the semiconductor sample 150.

[0080] S50: Controller 120 controls the stage 151 to move to the second detection position 141.

[0081] After optical inspection of the bare wafer is completed, the controller 120 moves the stage 151 to the second detection position 141 to enable charged particle beam detection of the bare wafer. The controller 120 controls the alignment of the semiconductor sample 150 with the charged particle detector 140 according to the correspondence between the positioning point 152 and the coordinate system established above.

[0082] In this embodiment, since the optical inspection instrument 130 and the charged particle detector 140 are integrated on a single machine tool 110, the optical inspection instrument 130 and the charged particle detector 140 can have the same coordinate system (i.e., the same defects on the bare wafer have the same optical plotting coordinates and electronic plotting coordinates). Therefore, there is no need to recalibrate the transformation relationship between these two coordinate systems, which can further improve inspection efficiency.

[0083] S60: The controller 120 receives the first position information from the data center 200 and controls the charged particle detector 140 to detect the defect indicated by the first position information on the semiconductor sample 150 by using charged particle beam detection.

[0084] The charged particle beam in the charged particle detector 140 is focused after passing through a condenser lens. The focused beam is then focused onto a bare wafer through a composite lens, forming a detection spot on the wafer's surface. The controller 120 controls the formation of the detection spot at the coordinates indicated by the first position information. The detection system detects electrons emitted from the surface of the bare wafer, thereby forming a first image. That is, the first image is the image of the semiconductor sample 150 detected by the charged particle detector 140 at the coordinates indicated by the first position information.

[0085] In this embodiment, the charged particle detector 140 only performs secondary detection on the area where the defect is detected by the optical detector 130. Therefore, the field of view of the charged particle detector 140 should be smaller than that of the optical detector 130. This can improve the detection accuracy, as well as the detection speed and efficiency of the sample to be tested.

[0086] S70: Data Center 200 receives the first image.

[0087] Because the optical inspection instrument 130 is prone to false detections, the data center can analyze the first image to obtain the detection results of the charged particle detector 140 on the semiconductor sample. This allows for secondary verification of the defects detected by the optical inspection instrument 130 based on the results from the charged particle detector 140, filtering out false alarms and identifying the actual defects. Then, the defects detected simultaneously by both the optical inspection instrument 130 and the charged particle detector 140 are classified, recorded, and stored.

[0088] In some optional embodiments, the data center can further enhance the classification model using the first and second images. This classification model can classify at least one defect in the detection data detected by the optical inspection instrument 130. The detection device 100 can input the first and second images obtained from each detection into the classification model, thereby continuously iteratively optimizing the classification model and gradually improving its classification accuracy. This can improve the detection accuracy of the detection device 100.

[0089] Figure 10 This is a flowchart illustrating a detection method provided in an embodiment of this application. Figure 11 yes Figure 10 A schematic diagram of the working logic of the detection method is provided. This method is used for bare wafer inspection. For example, this method can be applied to the aforementioned detection system. (Refer to...) Figure 10 The method may include the following steps S11 to S71.

[0090] S11: After the bare wafer is placed on the stage 151, the controller 120 controls the stage 151 to be in the first detection position 131.

[0091] After the bare wafer is placed on the stage 151, the controller 120 controls the position of the stage 151 to position the bare wafer at the center of the laser beam of the optical inspection instrument 130. Then, according to the requirements of the optical inspection instrument 130, the controller controls the optical inspection instrument 130 to calibrate the target lens to ensure that the imaging optical axis of the optical inspection instrument 130 is aligned with the plane of the semiconductor sample 150.

[0092] Then, a specific point is selected on the semiconductor sample 150 as the origin of the coordinate system. Since the bare wafer does not have a printed pattern that can serve as a reference mark for determining its orientation, in some embodiments, positioning points 152 are provided on the edge of the bare wafer (e.g., Figure 5 The positioning point 152 shown is used to mark its orientation. Therefore, the controller 120 can use the positioning point 152 as the origin of the coordinate system and define the coordinate axis orientation based on the origin of the coordinate system: for example, the X-axis and Y-axis are usually set to be horizontal and vertical, respectively, and the Z-axis is the direction perpendicular to the surface of the semiconductor sample 150.

[0093] S21: Controller 120 controls optical inspection instrument 130 to perform optical inspection on bare wafer.

[0094] When the laser beam emitted from the optical inspection unit 130 illuminates the bare wafer, it is scattered on the surface of the wafer, forming scattered light. This scattered light is detected by a photodetector to generate inspection data. In some embodiments, this inspection data can be light scattering data, which includes the intensity of the scattered light. This allows the controller 120 to detect potential defects in the bare wafer by analyzing changes in the intensity of the scattered light.

[0095] S31: The optical inspection instrument 130 sends the inspection data of the bare wafer to the controller 120, so that the controller 120 can send the inspection data to the data center.

[0096] In some embodiments, the controller 120 can also form a second image based on light scattering data, which is an image of the defect detected by the optical inspection instrument 130 on the semiconductor sample 150. Since the intensity of the scattered light changes when the laser beam hits a defect on the bare wafer, the brightness of the affected area in the image also changes accordingly. The controller 120 can detect potential defects on the bare wafer by analyzing the image to obtain the intensity change of the scattered light. Furthermore, the controller 120 can obtain the coordinates of at least one defect on the semiconductor sample 150 in the coordinate system established in the above steps.

[0097] S41: Data center 200 calculates the confidence level for each defect in the detection data and filters out defects with low confidence levels to obtain the first location information.

[0098] Because the optical inspection instrument 130 has low resolution, it may produce false alarms. Therefore, the data center needs to use a trained confidence model to calculate the confidence level of defects shown in the second image. Confidence level typically refers to the degree of certainty the confidence model has in predicting a certain category. For example, if the model predicts a defect in the second image as a "real defect" and gives a 90% confidence level, it means the model believes there is a 90% probability that the defect actually exists. If the model predicts a defect in the second image as a "real defect" and gives a 30% confidence level, it means the model believes there is a 30% probability that the defect actually exists, meaning the defect may be a false alarm. Thus, the data center can filter out defects with confidence levels below a first threshold, allowing for secondary detection of these low-confidence defects using the charged particle detector 140.

[0099] S51: Controller 120 controls the stage 151 to move to the second detection position 141.

[0100] After optical inspection of the bare wafer is completed, the controller 120 moves the stage 151 to the second detection position 141 to enable charged particle beam detection of the bare wafer. The controller 120 controls the alignment of the semiconductor sample 150 with the charged particle detector 140 according to the correspondence between the positioning point 152 and the coordinate system established above.

[0101] In this embodiment, since the optical inspection instrument 130 and the charged particle detector 140 are integrated on a single machine tool 110, the optical inspection instrument 130 and the charged particle detector 140 can have the same coordinate system (i.e., the same defects on the bare wafer have the same optical plotting coordinates and electronic plotting coordinates). Therefore, there is no need to recalibrate the transformation relationship between these two coordinate systems, which can further improve inspection efficiency.

[0102] S61: The controller 120 receives the first position information from the data center 200 and controls the charged particle detector 140 to detect the defect indicated by the first position information on the semiconductor sample 150 by using charged particle beam detection.

[0103] The charged particle beam in the charged particle detector 140 is focused after passing through a condenser lens. The focused beam is then focused onto a bare wafer through a composite lens, forming a detection spot on the wafer's surface. The controller 120 controls the formation of the detection spot at the coordinates indicated by the first position information. The detection system detects electrons emitted from the surface of the bare wafer, thereby forming a first image. That is, the first image is the image of the semiconductor sample 150 detected by the charged particle detector 140 at the coordinates indicated by the first position information.

[0104] In this embodiment, the charged particle detector 140 only performs secondary detection on the area where the defect is detected by the optical detector 130. Therefore, the field of view of the charged particle detector 140 should be smaller than that of the optical detector 130. This can improve the detection accuracy, as well as the detection speed and efficiency of the sample to be tested.

[0105] S71: Data center 200 receives the first and second images.

[0106] Because the optical inspection instrument 130 is prone to false detections, the data center 200 can perform secondary confirmation of the defects detected by the optical inspection instrument 130 based on the second image. Then, the defects detected simultaneously by the optical inspection instrument 130 and the charged particle detector 140 are classified, recorded, and stored.

[0107] In some optional embodiments, the data center 200 can also train a confidence model using the first and second images. This confidence model can calculate the confidence level of at least one defect in the detection data detected by the optical inspection instrument 130. Each time the detection device 100 detects a defect, the first and second images can be input into the confidence model, thereby continuously training the confidence model, gradually improving the calculation accuracy of the confidence model, and enhancing the detection accuracy of the detection device 100.

[0108] It should be noted that the data center 200 in this application embodiment can be set up in a remote cloud or integrated into the controller. That is, the controller receives the first image and the second image and uses the first image and the second image to train at least one of the classification model and the confidence model, or uses the second image to classify and screen the defects detected by the optical detector.

[0109] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A detection device, characterized in that, The application relates to a detection device and a data center. The device comprises: a machine table; a controller, an optical detector and a charged particle detector arranged on the machine table; the controller receives detection data generated by the optical detector when the optical detector performs optical detection on a semiconductor sample, wherein the detection data comprises position information of at least one defect on the semiconductor sample; 2. The detection device of claim 1, wherein, the controller outputs first position information to the charged particle detector, and controls the charged particle detector to detect a defect indicated by the first position information on the semiconductor sample by using a charged particle beam; wherein the defect indicated by the first position information comprises at least one defect in the detection data.

3. The detection device of claim 2, wherein, The controller is further configured to send the detection data to a data center and receive the first position information from the data center, wherein the defect indicated by the first position information comprises part of the defects indicated by the detection data.

4. The detection device according to any one of claims 1 to 3, characterized in that The confidence of the defect indicated by the first position information is lower than a first threshold.

5. The detection device of claim 4, wherein, The detection data and the first position information correspond to the same coordinate system. The semiconductor sample is provided with a positioning point; 6. The detection device of claim 5, wherein, The controller is further configured to construct a coordinate system according to the positioning point, and the detection data is the coordinates of the defects on the semiconductor sample in the coordinate system.

7. The detection device of claim 6, wherein, The controller is further configured to control the semiconductor sample to align with the charged particle detector according to the correspondence between the positioning point and the coordinate system. The first position information is the coordinates of the at least one defect in the coordinate system.

8. The detection device according to any one of claims 1 to 7, characterized in that, The controller is configured to control the charged particle detector to obtain a first image according to the first position information, wherein the first image is an image of the semiconductor sample at the coordinates indicated by the first position information detected by the charged particle detector. The machine table comprises a sample stage for placing the semiconductor sample; 9. The detection device according to any one of claims 1 to 8, characterized in that, The controller is further configured to control the sample stage to move from a first detection position to a second detection position, wherein the first detection position corresponds to the optical detector, and the second detection position corresponds to the charged particle detector.

10. The detection device according to any one of claims 1 to 9, characterized in that, The field of view of the charged particle detector is smaller than that of the optical detector.

11. A detection system characterized by, The charged particle detector is a single-electron-beam detection instrument or a multi-electron-beam detection instrument. The application relates to a detection device and a data center. The data center is configured to receive detection data, wherein the detection data comprises position information of at least one defect on a semiconductor sample obtained by an optical detection method; The data center is further configured to output first position information, wherein the defect indicated by the first position information comprises at least one defect in the detection data.

12. The detection system of claim 11, wherein, The data center is further configured to receive a second image, wherein the second image is an image of the defect on the semiconductor sample detected by the optical detector; At least one defect in the detection data is classified according to the second image and a classification model to obtain the first position information.

13. The detection system of claim 12, wherein, The data center is further configured to receive a first image, wherein the first image is an image of the semiconductor sample at the coordinates indicated by the first position information detected by the charged particle beam. optimizing the classification model by the first image and the second image.

14. The detection system of claim 11, wherein, The data center is also configured to receive a second image, the second image being an image of the defect on the semiconductor sample acquired by the optical detection method; and calculate a confidence of at least one defect in the detection data according to the second image and a confidence model to obtain the first position information, the first position information indicating that the confidence of the defect is lower than a first threshold.

15. A method of detection, characterized in that, comprising: receiving at least one detection data, wherein the detection data is used to indicate position information of a defect on a semiconductor sample acquired by an optical detection method; generating first position information, wherein the first position information indicates that the defect includes at least one defect in the detection data; performing charged particle beam detection on the defect on the semiconductor sample based on the first position information.

16. The detection method of claim 15, wherein, The detection method further comprises: receiving a second image, the second image being an image of the defect on the semiconductor sample acquired by the optical detection method; classifying at least one defect in the detection data according to the second image and a classification model to obtain the first position information.

17. The detection method of claim 16, wherein, The detection method further comprises: receiving a first image, the first image being an image of the semiconductor sample at a coordinate indicated by the first position information acquired by the charged particle beam detection, optimizing the classification model by the first image and the second image.

18. The detection method according to claim 16 or 17, characterized in that, The detection method further comprises: receiving a second image, the second image being an image of the defect on the semiconductor sample acquired by the optical detection method; calculating a confidence of at least one defect in the detection data according to the second image and a confidence model to obtain the first position information, the first position information indicating that the confidence of the defect is lower than a first threshold.

Images