Sampling device for acquiring multi-path image on surface of wafer

Abstract

The invention provides a sampling device for acquiring multiple paths of images on the surface of a wafer, which comprises an image detection sensing device and a high-performance video processing and data conversion system card, and the high-performance video processing and data conversion system card is connected with the image detection sensing device through an SDI (Serial Digital Interface) cable. The image detection sensing device comprises a CCD module and a precision micro-motion adsorption platform, the precision micro-motion adsorption platform is used for placing a to-be-detected wafer, and the CCD module is used for acquiring an image of the surface of the detected wafer on the adsorption platform. A linkage type image acquisition method is embedded in the high-performance video processing and data conversion system card, and the high-performance video processing and data conversion system card is used for receiving image data signals sampled by the CCD module, storing, splicing, processing and converting the image data signals and then transmitting control signals to control the movement of the precise micro-motion adsorption platform. According to the invention, the sampling rate of the surface image or the surface structure of the wafer can be improved, and the difficulty of subsequent image server and algorithm processing is reduced.

Description

Technical Field

[0001] This invention belongs to the field of multi-channel image or video sampling and image processing, and specifically relates to a multi-channel image sampling device for acquiring wafer surfaces. Background Technology

[0002] Currently, commonly used wafer surface sampling methods are mainly divided into X-ray light-based inspection systems and single-vision-based imaging systems. While X-ray light inspection equipment can detect wafer surface flatness and integrity, and delve into the internal structure of the chip, it is costly, occupies a large area, has high maintenance costs, and is quite difficult to repair. Single-vision imaging systems used in integrated circuit inspection mostly employ a single hyperspectral camera. Hyperspectral imaging technology organically combines imaging and spectroscopy, possessing the ability to fully acquire all information of the target's two-dimensional space and one-dimensional spectrum, with fast imaging speed and real-time monitoring capabilities. However, currently commonly used spectral imaging equipment still uses a single-lens mode, limiting the imaging area. This is somewhat insufficient for higher-density wafer imaging, where acquiring an image of the entire wafer in a very short time is challenging. Summary of the Invention

[0003] This invention addresses the problems of expensive X-RAY equipment, high usage and maintenance costs in existing technologies, as well as the shortcomings of single-lens inspection equipment, such as limited single-image area, low resolution, and slow sampling rate in wafer image acquisition. It provides a multi-channel image sampling device for acquiring wafer surfaces, which can further improve the image acquisition speed of the entire wafer while reducing equipment costs.

[0004] To solve the above technical problems, the present invention provides the following technical solution: a multi-channel image sampling device for acquiring wafer surface, comprising: an image detection sensing device and a high-performance video processing and data conversion system card, wherein the high-performance video processing and data conversion system card is connected to the image detection sensing device via an SDI cable;

[0005] The image detection sensing device includes a CCD module and a precision micro-motion adsorption platform. The precision micro-motion adsorption platform is used to place the wafer to be inspected, and the CCD module is used to acquire images of the surface of the wafer to be inspected on the adsorption platform and convert them into digital information.

[0006] The high-performance video processing and data conversion system card has an embedded linkage image acquisition method, which is used to receive image data signals sampled by the CCD module, store, preprocess, process image stitching, and convert the image data signals. The stitched image is then transmitted to digital signal processing, and the control signal controls the movement of the precision micro-motion adsorption platform.

[0007] Furthermore, the aforementioned CCD module includes four integrated CCD image sensors and optical components. The optical components have a focusing knob for magnifying the tiny patterns or surface structures on the wafer and sending the image to the CCD image sensors.

[0008] Furthermore, the aforementioned CCD image sensor includes a video output interface, and the components are arranged in a 2x2 configuration.

[0009] The optical components include: an optical magnifying lens barrel and an image input lens; the optical magnifying lens barrel is located below the CCD image sensor; the focusing knob is located on the optical magnifying lens barrel and is used to achieve different magnifications; the image input lens is located at the bottom of the optical components to enable the wafer image signal to be input from below; the video output interface is located above the CCD image sensor and is used to output 12GSDI standard digital image signals.

[0010] Furthermore, the aforementioned precision micro-motion adsorption platform is a flat plate containing a pneumatic adsorption device and a lateral and longitudinal motion auxiliary mechanism. Below the flat plate are auxiliary lateral and longitudinal mechanical motion devices used to fix the wafer and control the wafer to move in multiple directions.

[0011] A wafer size detection sensor is embedded in the adsorption hole of the adsorption device to confirm the size of the placed wafer, as well as the accuracy and area of ​​subsequent motion control. The adsorption platform is precisely controlled by a high-performance video processing and data conversion system card.

[0012] Furthermore, the aforementioned linked image acquisition method is embedded in a system board, which includes a core FPGA, four 12GSDI chips and interfaces, four 10G optical fibers with SFP interfaces, one 40G optical fiber with a QSFP interface, and one PCIe interface; the 12GSDI chips, 10G optical fibers, 40G optical fibers, and PCIe interface are all connected to the FPGA.

[0013] Furthermore, the aforementioned linked image acquisition method specifically includes: a motion control process algorithm and a wafer image stitching algorithm.

[0014] Furthermore, the aforementioned wafer image stitching algorithm is to completely stitch together the images acquired by the CCD image sensing module under different horizontal / vertical micro-motion cycles and large horizontal / vertical cycles.

[0015] Furthermore, the aforementioned high-performance video processing and data conversion system card first encapsulates the obtained complete stitched image into communication transmission packets, converting it into PCIE data packets, QSFP data packets, or SFP data packets, and then transmits the signal to the host computer.

[0016] Compared with the prior art, the beneficial technical effects of the present invention using the above technical solution are as follows:

[0017] (1) For the commonly used single-vision imaging wafer inspection devices, the single imaging area is limited and the detection rate is slow. The present invention uses a combination of 4 CCD image sensors and a high-precision motion platform, adapted to the linkage image acquisition method and the real-time image stitching algorithm, to realize a multi-channel image sampling device for rapid acquisition of wafer surface, and further improve the sampling rate of wafer surface image or surface structure.

[0018] (2) To further improve image resolution and real-time performance, a 12G-SDI interface is used to transmit image data. The data transmission capacity of a single 12G-SDI interface reaches 11.88Gbps, and four channels approach 48Gbps. This data volume is insufficient for transmission using 10G fiber optic or 10GbE network cards. This invention uses an FPGA board with image processing capabilities to realize hardware filtering and processing of image data, pre-stitching of multi-frame, multi-scene images, and data conversion and transmission adapted to the capacity of fiber optic or PCIe interfaces. This achieves relay conversion and transmission of multi-channel ultra-high-capacity image data, further reducing the difficulty of subsequent image server and algorithm processing.

[0019] (3) Traditional single-lens image acquisition methods involve acquiring images row by row and column by column. This invention, considering the characteristics of a four-camera setup, divides the image acquisition process into two stages: a micro-motion cycle and a large-cycle cycle. Within the micro-motion cycle, images acquired through horizontal and vertical micro-motion are directly pre-stitched, allowing for the acquisition of a higher-resolution complete image in a shorter time. The images from the micro-motion cycle are then directly output for subsequent processing. Depending on the wafer size, different large-cycle cycles exist, resulting in varying numbers of images acquired during the micro-motion cycle. Compared to the image resolution achievable with a single lens, the image resolution acquired by this invention is significantly increased. Due to the increased area of ​​image acquisition within the micro-motion cycle, the final number of output images is reduced compared to a wafer of the same size, thus lowering the processing difficulty and capacity of the subsequent image server. Attached Figure Description

[0020] Figure 1 This is the overall system architecture diagram.

[0021] Figure 2 This is a diagram of an image detection sensing device.

[0022] Figure 3 This diagram shows the interface and resources of a high-performance video processing and data conversion system board.

[0023] Figure 4 This is a structural diagram of a precision micro-motion adsorption platform.

[0024] Figure 5This is a schematic diagram of wafer inspection and video partitioning stitching. Detailed Implementation

[0025] To better understand the technical content of the present invention, specific embodiments are described below in conjunction with the accompanying drawings.

[0026] In this invention, various aspects of the invention are described with reference to the accompanying drawings, in which numerous illustrative embodiments are shown. Embodiments of the invention are not limited to those depicted in the drawings. It should be understood that the invention is implemented through any of the various concepts and embodiments described above, as well as the concepts and embodiments described in detail below, because the concepts and embodiments disclosed herein are not limited to any particular implementation. Furthermore, some aspects of the invention disclosed may be used alone or in any suitable combination with other aspects of the invention disclosed.

[0027] like Figure 1 The diagram shown is an overall architecture diagram of the multi-channel image sampling device for obtaining wafer surfaces according to the present invention. It includes: an image detection sensing device and a high-performance video processing and data conversion system card. The high-performance video processing and data conversion system card is connected to the image detection sensing device via an SDI cable.

[0028] The image detection sensing device includes: four CCD modules and a precision micro-motion adsorption platform. The precision micro-motion adsorption platform is used to place the wafer to be inspected, and the CCD modules are used to acquire images of the surface of the wafer to be inspected on the adsorption platform and convert them into digital information.

[0029] The high-performance video processing and data conversion system card has an embedded linkage image acquisition method, which is used to receive image data signals sampled by the CCD module, store, preprocess, process image stitching, and convert the image data signals. The stitched image is then transmitted to digital signal processing, and the control signal controls the movement of the precision micro-motion adsorption platform.

[0030] The CCD module includes an integrated CCD image sensor and optical components. The optical components have a focusing knob for magnifying tiny patterns or surface structures on the wafer and sending the image to the CCD image sensor.

[0031] refer to Figure 2 A single CCD image detection sensor. The optical magnifying objective is located below the CCD image sensor; the focusing knob is located on the optical magnifying lens barrel to achieve different magnifications; the image input lens is at the bottom of the entire device, meaning the wafer image signal is input from below; the video output interface is located in the CCD image sensor module, at the top of the entire device, for outputting 12GSDI standard digital image signals.

[0032] The optical components include: an optical magnifying lens barrel and an image input lens; the optical magnifying lens barrel is located below the CCD image sensor; the focusing knob is located on the optical magnifying lens barrel and is used to achieve different magnifications; the image input lens is located at the bottom of the optical components to enable the wafer image signal to be input from below; the video output interface is located above the CCD image sensor and is used to output 12GSDI standard digital image signals.

[0033] Figure 3 The image shows a high-performance video processing and data conversion system board. The system board includes a core FPGA, four 12GSDI chips and interfaces, four 10G fiber optic interfaces with SFP connectors, one 40G fiber optic interface with a QSFP connector, and one PCIe interface. The 12GSDI, 10G, 40G, and PCIe interfaces are all connected to the FPGA. The system board primarily performs digital signal processing functions such as video data reception and preprocessing, feature extraction, image stitching, signal conversion, and transmission of the stitched image, as well as controlling the movement of a precision micro-motion adsorption platform.

[0034] The reception and preprocessing of video data involves identifying and filtering the image data transmitted from the SDI interface, determining its integrity, and preparing for subsequent feature extraction.

[0035] Feature extraction involves analyzing similarity points at image edges to provide foundational data for subsequent image stitching.

[0036] Image stitching processing is the complete stitching of images acquired by the CCD image sensing module in different horizontal / vertical micro-motion cycles and large horizontal / vertical cycles.

[0037] The digital signal processing involves two types of processes: signal conversion and image transmission after stitching. The complete stitched image acquired over a long period is first encapsulated into communication transmission packets, which are then converted into PCIE data packets, QSFP data packets, or SFP data packets. Finally, the images are transmitted to the host computer via the corresponding interface (SFP, QSFP, or PCIE).

[0038] See Figure 4 The diagram shows the structure of a precision micro-motion adsorption platform. The platform is a flat plate with adsorption holes and positioning marks, with auxiliary horizontal and vertical mechanical movement devices below the plate. The diagram shows a wafer size detection sensor embedded within the adsorption holes, used to confirm the size of the placed wafer and the accuracy and area of ​​subsequent motion control. Under the control of a high-performance video processing and data conversion system board, the adsorption platform performs precise motion control, including micro-motion cycles, large cycles, and reset processes.

[0039] See Figure 5The diagram illustrates wafer detection and video partitioning stitching under the control of a linked image acquisition method. It shows the motion control process of the linked image acquisition method and the output area after the wafer image stitching algorithm is implemented.

[0040] Based on the wafer image acquisition and motion control processes, the process is divided into horizontal / vertical micro-motion cycles and horizontal / vertical large cycles, which work in conjunction with the CCD image sensing module to acquire an image of the entire wafer. After a complete micro-motion cycle is completed, a high-resolution stitched image is output. After all large cycles are completed, the image of the entire wafer is output.

[0041] This invention overcomes the problems of expensive X-RAY equipment, high usage and maintenance costs, as well as the shortcomings of single-lens inspection equipment, such as limited single-image area, low resolution, and slow sampling rate.

[0042] This invention employs a combination of four CCD image sensors and a high-precision motion platform, adapted to a linked image acquisition method and a real-time image stitching algorithm, to realize a multi-channel image sampling device for rapidly acquiring wafer surface images, thereby further improving the sampling rate of wafer surface images or surface structures.

[0043] This invention uses an FPGA board with image processing capabilities to realize hardware filtering and processing of image data, pre-stitching of multi-frame and multi-scene images, and data conversion and transmission adapted to the capacity of fiber optic or PCIe interfaces. It realizes multi-channel ultra-large capacity image data relay conversion and transmission, further reducing the difficulty of subsequent image server and algorithm processing.

[0044] Compared to the image resolution achievable with a single lens, the image resolution obtained by this invention will be multiplied. Due to the increased area for image acquisition within the micro-motion cycle, the number of images output is reduced compared to a wafer of the same size, thus lowering the processing difficulty and capacity of the subsequent image server.

[0045] While the present invention has been described above with reference to preferred embodiments, it is not intended to limit the invention. Those skilled in the art can make various modifications and refinements without departing from the spirit and scope of the invention. Therefore, the scope of protection of the present invention shall be determined by the claims.

Claims

1. A multi-channel image sampling device for acquiring wafer surface, characterized in that, include: The image detection sensing device and the high-performance video processing and data conversion system card are connected to the image detection sensing device via an SDI cable. The image detection sensing device includes a CCD module and a precision micro-motion adsorption platform. The precision micro-motion adsorption platform is used to place the wafer to be inspected, and the CCD module is used to acquire images of the surface of the wafer to be inspected on the adsorption platform and convert them into digital information. The high-performance video processing and data conversion system card has an embedded linkage image acquisition method, which is used to receive image data signals sampled by the CCD module, store, preprocess, process image stitching, and convert the image data signals. The stitched image is then transmitted to digital signal processing, and the control signal controls the movement of the precision micro-motion adsorption platform.

2. The multi-channel image sampling device for acquiring wafer surfaces according to claim 1, characterized in that, The CCD module includes four integrated CCD image sensors and optical components. The optical components have a focusing knob, which is used to magnify the tiny patterns or surface structures on the wafer and send the image to the CCD image sensors.

3. The multi-channel image sampling device for acquiring a wafer surface according to claim 2, characterized in that, The CCD image sensor includes a video output interface and is arranged in a 2x2 configuration. The optical components include: an optical magnifying lens barrel and an image input lens; the optical magnifying lens barrel is located below the CCD image sensor; The focusing knob is located on the optical magnifying lens barrel and is used to achieve different magnifications; the image input lens is at the bottom of the optical components, allowing the wafer image signal to be input from below; the video output interface is located above the CCD image sensor and is used to output 12GSDI standard digital image signals.

4. The multi-channel image sampling device for acquiring wafer surfaces according to claim 3, characterized in that... The precision micro-motion adsorption platform is a flat plate containing a pneumatic adsorption device and a lateral and longitudinal motion auxiliary mechanism. There is an auxiliary lateral and longitudinal mechanical motion device under the flat plate, which is used to fix the wafer and control the wafer to move in multiple directions. A wafer size detection sensor is embedded in the adsorption hole of the adsorption device to confirm the size of the placed wafer, as well as the accuracy and area of ​​subsequent motion control. The adsorption platform is precisely controlled by a high-performance video processing and data conversion system card.

5. The multi-channel image sampling device for acquiring a wafer surface according to claim 3, characterized in that, The linked image acquisition method is embedded in the system board, which includes a core FPGA, four 12GSDI chips and interfaces, four 10G optical fibers with SFP interfaces, one 40G optical fiber with QSFP interface, and one PCIe interface. The 12GSDI chip, 10G fiber optic cable, 40G fiber optic cable, and PCIe interface are all connected to the FPGA.

6. The multi-channel image sampling device for acquiring wafer surface according to claim 1, characterized in that, The linked image acquisition method specifically consists of a motion control process algorithm and a wafer image stitching algorithm.

7. The multi-channel image sampling device for acquiring a wafer surface according to claim 6, characterized in that, The wafer image stitching algorithm is used to completely stitch together images acquired by the CCD image sensing module under different horizontal / vertical micro-motion cycles and large horizontal / vertical cycles.

8. The multi-channel image sampling device for acquiring wafer surfaces according to claim 1, characterized in that, The high-performance video processing and data conversion system card first encapsulates the obtained complete stitched image into communication transmission packets, converting them into PCIE data packets, QSFP data packets, or SFP data packets, and then transmits the signals to the host computer.

Images