Two-dimensional material identification positioning and morphology analysis system based on artificial intelligence

Abstract

The invention relates to a two-dimensional material identification positioning and morphology analysis system based on artificial intelligence, and the system comprises a microscopic imaging system which is used for observing and confirming the position of a detected sample; the plane scanning displacement table has an X motion direction and a Y motion direction and is used for completing sample positioning and moving the identified to-be-detected sample to the position below the probe; the Z-axis approaching displacement table is used for fixing a signal acquisition circuit and realizing approaching of the probe to the sample; the nanometer displacement table has X-axis movement, Y-axis movement and Z-axis movement, the X-axis movement and the Y-axis movement are used for completing nanometer-level scanning on the plane, and the Z-axis movement is used for controlling the distance between the probe and the sample within the nanometer range after the scanning probe approaches; the signal acquisition circuit is used for acquiring atomic force signals and transmitting the atomic force signals to the signal processing equipment; the motion control unit is used for receiving signals from the calculation and analysis equipment and controlling the plane scanning displacement table and the nanometer displacement table to move, the signal processing unit is used for generating driving signals and collecting morphology information, and the calculation and analysis equipment carries an artificial intelligence algorithm for identifying needed two-dimensional materials, controlling movement and processing the signals. The two-dimensional material is identified in an artificial intelligence mode, the labor cost of searching for a single layer and a thin layer under a microscope is reduced, meanwhile, the device is automatically moved to the position below a probe for surface appearance detection, the trouble that searching and positioning need to be carried out again when samples are transferred among multiple devices is avoided, and the detection efficiency is improved. Meanwhile, the risk of pollution and damage in the sample transfer process is reduced.

Description

Technical Field

[0001] The present invention relates to an automated morphology analysis instrument, and belongs to the field of artificial intelligence computer vision automatic control and microscopic surface imaging, and specifically to an artificial intelligence-based two-dimensional material recognition, positioning and morphology analysis system. Background Art

[0002] Thin-layer two-dimensional materials possess excellent electromagnetic and optoelectronic properties, and hold broad application prospects in the fabrication of sensors and other micro-nanoelectronic devices. Currently, the fabrication of single-layer and thin-layer 2D material devices typically requires manual identification and searching under a microscope, which is time-consuming and laborious. Traditional computer vision methods, however, require re-adjustment of recognition parameters when the environment changes, making them cumbersome. Artificial intelligence image recognition methods, which are insensitive to environmental changes, offer significant advantages in identifying 2D materials.

[0003] To study the properties of thin layers of two-dimensional materials, atomic force microscopy (AFM) is often used to collect topographic information. This requires transferring the thin layer of material to the topography analysis equipment and repositioning the sample, which not only consumes a lot of labor costs but also increases the probability of damage and contamination during sample transfer. Therefore, after identifying the two-dimensional material, automatically locating and analyzing its topography will greatly help improve the yield of two-dimensional material devices. Summary of the Invention

[0004] In view of the shortcomings of the prior art described above, the purpose of the present invention is to propose an artificial intelligence-based two-dimensional material identification, positioning and morphology analysis system. The present invention adopts an artificial intelligence-based image segmentation algorithm to realize the segmentation and identification of two-dimensional materials, single thin layers and thick layers, and can determine their position coordinates on the substrate. Through the plane scanning displacement stage, the sample to be tested is positioned under the atomic force microscope probe, and the surface morphology analysis of the material is realized in situ. The present invention can automatically identify the sample, avoiding the manual search for the sample under the microscope, and automatically locate and perform in-situ morphology analysis, which not only prevents potential damage and contamination to the material during the transfer process of multiple devices, but also reduces the labor cost of searching and aligning the material again, thereby improving the efficiency of two-dimensional material device preparation.

[0005] To achieve the above objectives, the present invention provides an artificial intelligence-based system for identifying, positioning, and analyzing two-dimensional materials. According to an embodiment of the invention, the system includes a microscopic imaging system; a planar scanning stage; a Z-axis approach stage; a nanometer stage; a signal acquisition circuit; a motion control device; a signal processing device; a computational analysis device; and an artificial intelligence image segmentation algorithm.

[0006] Furthermore, the microscopic imaging system includes an adjustment bracket, an image acquisition camera, a microscope tube, visible light illumination, and objective lenses of various magnifications, which are used to capture microscopic images, transmit them to a computing and analysis device, identify single layers of two-dimensional materials, and return coordinates. The planar scanning stage includes a slide rail, a slider, a power supply, and a transmission device, which receives motion control commands and performs scanning and positioning. The Z-axis approach stage includes a stage with fixing bolt holes, a power supply, a push rod, and a connecting rod fixing block, which is used to fix the signal acquisition circuit and perform probe approach.

[0007] Furthermore, the nanometer displacement stage comprises two vertically fixed linear displacement stages for scanning in the XY directions and a Z-axis displacement stage for vertical nanometer-scale displacement. The feed mechanism, which uses a piezoelectric crystal to drive a connecting rod, is used to detect probe approach, control the distance between the probe and the sample, and perform nanometer scanning to acquire topographic information. The signal acquisition circuit has fixing holes that can be fastened to the Z-axis approach stage. The circuit's main components include four operational amplifiers, which can be used to drive the atomic force probe and collect atomic force signals. The motion control device receives instructions from the computational analysis device, drives the planar scanning stage for scanning and addressing, drives the Z-axis approach stage probe for approach, and drives the piezoelectric displacement stage for microscopic scanning.

[0008] Furthermore, the signal processing device is used to output a resonant signal driving the atomic force probe, collect the atomic force signal, and transmit it to the computational analysis device. The computational analysis device issues a plane scanning instruction, collects image information from the image acquisition camera, determines the sample position through algorithmic recognition, issues an addressing instruction, receives information from the signal acquisition device, and uses feedback to adjust the Z-axis approach stage and nanometer stage to perform surface morphology analysis. The artificial intelligence image segmentation algorithm, which utilizes an artificial intelligence neural network algorithm and is deployed in the computational analysis device, segments the two-dimensional material image, identifies the number of two-dimensional material layers, and returns the coordinates of the sample location. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1 This is a schematic diagram of the overall three-dimensional structure of an artificial intelligence-based two-dimensional material identification, positioning and morphology analysis system in the present invention;

[0010] Figure 2 Schematic diagram of the three-dimensional structure of the microscopic imaging system and the Z approach translation stage bracket of the patented invention;

[0011] Figure 3 This is a schematic diagram of the three-dimensional structure of the patented microscopic imaging system of the present invention;

[0012] Figure 4 This is a schematic diagram of the three-dimensional structure of the Z approach translation stage of the present invention;

[0013] Figure 5 This is a schematic diagram of the three-dimensional structure of the plane scanning translation stage of the present invention;

[0014] Figure 6 This is a schematic diagram of the overall three-dimensional structure of the patented nano-displacement stage of the present invention; DETAILED DESCRIPTION

[0015] In order for those skilled in the art to better understand the scheme of this application, the technical scheme in this application will be described in detail and completely in conjunction with the drawings in the embodiments of this application. The described embodiments are only embodiments of a part of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by ordinary technicians in this field without making creative work should fall within the scope of protection of this application. The directional terms mentioned in the embodiments, such as "up", "down", "front", "back", "left", "right", etc., are only reference to the directions of the drawings and are not used to limit the scope of protection of this disclosure.

[0016] This patent provides an artificial intelligence-based system for identifying, positioning, and analyzing two-dimensional materials. The present invention comprises a microscopic imaging system 1; a planar scanning stage 3; a Z-axis approach stage 2; a nanometer stage 4; a signal acquisition circuit 21; a motion control device 5; a signal processing device 6; a computational analysis device 7; and an artificial intelligence image segmentation algorithm.

[0017] See also Figure 2 、 3 The microscopic imaging system and the support include a focus adjustment hand wheel 8, a microscope fixing support 9, a translation stage fixing support 10, an image acquisition camera 11, a microscope tube 12, a visible light illumination 13, an objective lens connecting part 14, and a magnifying objective lens 15.

[0018] See also Figure 4 The Z-axis approach stage 2 includes a signal acquisition circuit 21 and a stage 19. The stage 19 includes a power supply 16, a transmission push rod 17, and a push rod limit block 18. The signal acquisition circuit 21 includes a scanning probe 22 and a signal acquisition interface 20.

[0019] See also Figure 5 The plane scanning stage 3 includes a stage rail 26, a power feed 27, a transmission device 25, a slider 23, and a support base 24. Figure 6 The nano-displacement stage 4 includes a piezoelectric crystal 28 , a Z-axis displacement stage 29 , and a linear displacement stage 30 .

[0020] It is important to note that:

[0021] The aforementioned microscopic imaging system can be an optical microscope, a metallographic microscope, or other microscopes that utilize visible light for microscopic imaging. The image acquisition camera can be equivalently replaced with a CCD or CMOS camera. The plane scanning stage can be powered by a stepper motor, servo motor, or piezoelectric crystal. For less precise applications, a conventional brushed DC motor, brushless motor, or servo can also be used, simply by replacing the corresponding motion control element.

[0022] The above-mentioned Z-axis approach translation stage and nano-translation stage, in this embodiment, replace the power feed device of the manual translation stage with an electric motor and a piezoelectric crystal, and use other displacement schemes to achieve Z-axis approach of the probe and nano-scanning of the sample, which are all regarded as other embodiments of the present invention.

[0023] The above-mentioned signal acquisition circuit uses a four-channel high-precision operational amplifier chip to realize the acquisition of atomic force signals in this embodiment. Signal acquisition circuits that use four single-channel operational amplifiers or two dual-channel operational amplifiers or other chip selections to realize the same principle and function of the present invention should all be regarded as other embodiments of the present invention.

[0024] In this embodiment, the aforementioned signal processing device uses a lock-in amplifier to detect the resonant frequency of the atomic force probe, output a resonant signal that drives the atomic force probe, and collect the atomic force signal. A signal generator or oscilloscope can replace the lock-in amplifier function and should be considered as another embodiment of the present invention.

[0025] The computing and analysis device described above implements an AI-based image segmentation algorithm, receives images, and issues motion control commands. This embodiment uses a general-purpose Windows host as the hardware carrier. Using a Linux system, an industrial computer, a server, a single-board host, or other programmable logic device capable of supporting AI algorithms should be considered other embodiments of the present invention.

[0026] The aforementioned artificial intelligence image segmentation algorithm uses the DeepLabV3Plus semantic segmentation model in this embodiment. Semantic segmentation algorithms using FCN, U-Net, DeepLab, SegNet, Mask R-CNN, RefineNet, and their related improved versions also implement the functions of the present invention and should be considered other embodiments of the present invention.

[0027] The specific working steps of this invention are as follows:

[0028] Sample preparation and image scanning. Fix the substrate with the two-dimensional material on the upper surface of the nanometer displacement stage 4, adjust the hand wheel 8, and move the microscope up and down to obtain a clear image. Send instructions to the motion control device 5 through the computing and analysis device 7 to move the scanning starting point to the objective lens. Set the scanning origin, scanning step length and scanning range in the host computer program and execute the scanning program. Under the control of the scanning instruction, the planar scanning displacement stage 3 will drive the nanometer displacement stage 4 and the sample, and scan the entire thing in an "S" shape under the objective lens 15. At each scanning coordinate, the camera 11 captures the image, and the image and coordinates are sent to the computing and analysis device 7 together.

[0029] Identification and Positioning. The computing and analysis device 7 uses an image segmentation algorithm to identify the image captured at each coordinate, recording the coordinates of the single layer and thin layer. After reviewing the identification results, the coordinates of the thin layer in the image are entered into the host computer, and a command is sent to the motion control device 5, which controls the planar scanning stage 3 to move to the coordinates and reposition it under the objective lens 15.

[0030] Surface morphology information collection: During the equipment installation process, the probe 22 is placed under the objective lens 15, so when the sample is positioned under the objective lens, it is also positioned under the scanning probe.

[0031] Resonant frequency measurement. Signal processing device 6 outputs a 100mV sinusoidal signal and adjusts the potentiometer in the circuit to compensate for parasitic capacitance and make the frequency response curve symmetrical. The resonant frequency of the atomic force probe used in this embodiment is around 32.7kHz. Therefore, the output signal frequency is adjusted from 20kHz to 40kHz. 100 sampling points are sampled, and the range is gradually narrowed until the probe resonant frequency is found. The output signal is fixed at this frequency, and the amplitude of the circuit output signal at this point is recorded.

[0032] Probe approach. Start the upper computer approach program and send instructions to the motion control device 5 to drive the feed mechanism 16 of the Z-axis approach stage 2 to extend the transmission push rod 17 and push the push rod limit block 18 to drive the stage to move downward. The approach step length is 0.1-1μm, which can be set in the upper computer program. During this process, the signal processing device 6 continuously collects the amplitude of the signal in the circuit. When the signal is less than 90% of the amplitude of the output signal when it is far away from the sample surface, it is changed to a fine approach with a step length of 0.1μm. The approach is stopped when it is less than 80% of the amplitude signal away from the sample surface. At this time, the nano-stage 4 is controlled to move up and down in the range of 0.1μm with a step length of 1 to 10nm. It should be observed that the collected signal has a corresponding fluctuation with the stage, which proves that the needle is successfully advanced. Otherwise, the probe approach is executed again.

[0033] Acquire surface topography signals. The host computer determines the scanning range and number of sampling points and initiates the scanning process. The amplitude of each point is collected during the scanning process to represent the height and depth of the sample, thereby reflecting the sample's topography. After the scan is complete, the probe retraction process is automatically executed. The Z-axis approach stage 2 moves the probe upward away from the sample surface, and the nanostage 4 returns to the scanning starting point.

[0034] If you need to obtain the morphology information of other thin layers on the substrate, enter the coordinates of the image where the thin layer identification result is located. The plane translation stage will move the coordinates under the objective lens and the probe, and then execute the probe approach and surface morphology signal acquisition again.

[0035] The above are only preferred embodiments of the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A two-dimensional material identification, positioning and morphology analysis system based on artificial intelligence, characterized in that: It includes a microscopic imaging system; a plane scanning translation stage; a Z-axis approach translation stage; a nanometer translation stage; a signal acquisition circuit; a motion control device; a signal processing device; a computing and analysis device; and an artificial intelligence image segmentation algorithm.

2. The artificial intelligence-based two-dimensional material identification, positioning and morphology analysis system according to claim 1, characterized in that The microscopic imaging system includes an adjustment bracket, an image acquisition camera, a microscope tube, visible light illumination, and objective lenses with various magnifications.

3. The artificial intelligence-based two-dimensional material identification, positioning and morphology analysis system according to claim 1, characterized in that The plane scanning displacement stage includes a slide rail, a slider, a power supply device, and a transmission device.

4. The artificial intelligence-based two-dimensional material identification, positioning and morphology analysis system according to claim 1, characterized in that The Z-axis approach translation stage comprises a translation stage with fixing bolt holes, a power supply device, a pushing connecting rod, and a connecting rod fixing block.

5. The artificial intelligence-based two-dimensional material identification, positioning and morphology analysis system according to claim 1, characterized in that The nano-displacement stage includes two vertically fixed linear displacement stages to achieve scanning in the XY direction, and a Z-axis displacement stage to achieve nano-level displacement up and down, wherein the feeding mechanism uses a piezoelectric crystal to push a connecting rod.

6. The artificial intelligence-based two-dimensional material identification, positioning and morphology analysis system according to claim 1, characterized in that The signal acquisition circuit has a fixing hole that can be fastened to the Z-axis proximity stage. The main components of the circuit include four operational amplifiers, which can be used to drive the atomic force probe and collect atomic force signals.

7. The artificial intelligence-based two-dimensional material identification, positioning and morphology analysis system according to claim 1, characterized in that The motion control device receives instructions from the computing and analyzing device, drives the plane scanning displacement stage to scan and address, drives the Z-axis approach displacement stage probe to approach, and drives the piezoelectric displacement stage to perform microscopic scanning.

8. The artificial intelligence-based two-dimensional material identification, positioning and morphology analysis system according to claim 1, characterized in that The signal processing device is used to output a resonance signal for driving the atomic force probe, collect the atomic force signal, and send it to the calculation and analysis device.

9. The artificial intelligence-based two-dimensional material identification, positioning and morphology analysis system according to claim 1, characterized in that The computing and analysis device issues a plane scanning instruction, collects image information from the image acquisition camera, determines the sample position through algorithm recognition, issues an addressing instruction, receives information from the signal acquisition device, and feedback-adjusts the Z-axis approach stage and the nano-stage to perform surface morphology analysis.

10. The artificial intelligence-based two-dimensional material identification, positioning and morphology analysis system according to claim 1, characterized in that The artificial intelligence image segmentation algorithm adopts an artificial intelligence neural network algorithm, which is deployed on a computing and analysis device to segment a two-dimensional material image, identify the number of two-dimensional material layers, and return the coordinates of the sample's location.

Images