Atomic-scale laser optimization method and device for carbon-based integrated circuit

Abstract

The invention relates to an atomic-scale laser optimization method and device for a carbon-based integrated circuit, belongs to the technical field of integrated circuit manufacturing, and solves the problems of how to effectively remove polymerization residues and ensure the transportation performance of carbon nanotubes at the same time. The method comprises the following steps: scanning the film surface of a carbon nanotube wafer to obtain a group of surface topography images; identifying the parameter information of the residual polymer on the surface of the carbon nanotube wafer and identifying the defect information of the carbon nanotube by using a machine vision and image processing algorithm; determining an optimal laser working parameter based on the parameter and / or defect information; carrying out fixed-point irradiation on the residual polymer and the defects of the carbon nano tube, so that the residual polymer is subjected to photolysis, and the in-situ defects of the carbon nano tube are discharged and repaired; and repeatedly acquiring surface topography images of the carbon nanotube wafer after fixed-point irradiation, and verifying the residual amount of the residual polymer and the defect repair state until the residual polymer is completely removed and the defect is repaired. And surface detection and quality improvement of the carbon nanotube wafer are realized.

Description

Technical Field

[0001] The present invention relates to the technical field of integrated circuit manufacturing, and in particular, to a method and device for optimizing carbon-based integrated circuits at the atomic level by laser. Background Art

[0002] With the wide application of carbon nanotubes (CNTs) materials in nanoelectronics and optoelectronic devices, the development of large-scale wafer-level carbon nanotube thin film preparation and quality control technologies has become the forefront and focus of current scientific and technological development. Especially in the manufacture of a new generation of carbon-based information processing chips, namely carbon-based integrated circuits, obtaining high-quality wafer-level carbon nanotube thin films is the key to achieving high performance and high reliability.

[0003] Currently, in the process of mass-producing wafer-level carbon nanotube array thin films by solution method, polymers or other carbon-based additives need to be introduced to obtain densely arranged and well-oriented carbon nanotubes. However, these polymers and additives aggregate and remain on the surface and inside of the carbon nanotubes, seriously hindering the smooth progress of subsequent wafer-level carbon nanotube thin film preparation processes, such as metal-carbon nanotube contact preparation, gate dielectric deposition, epitaxial heavy doping, etc. In particular, it will cause structural defects such as dislocations and grain boundaries, reduce the carrier mobility of the device, and increase the leakage current.

[0004] Although some physical or chemical methods can be used to remove polymer residues, the selection is relatively poor and it is easy to cause damage to the carbon nanotube structure and lead to performance degradation. Generally speaking, the existing technology has not yet realized a special process and device that can effectively remove residual polymers while ensuring the transport performance of carbon nanotubes. This has become a bottleneck restricting the progress and performance of carbon-based integrated circuits. Summary of the Invention

[0005] In view of the above analysis, the embodiments of the present invention aim to provide a method and device for optimizing carbon-based integrated circuits at the atomic level by laser, so as to solve problems such as how to effectively remove polymerization residues while ensuring the transport performance of carbon nanotubes.

[0006] On the one hand, an embodiment of the present invention provides a method for optimizing atomic-level lasers for carbon-based integrated circuits, including: scanning the thin film surface of a carbon nanotube wafer to obtain a set of surface topography images; using machine vision and image processing algorithms to identify the parameter information of residual polymers on the surface of the carbon nanotube wafer and the defect information of carbon nanotubes according to the set of surface topography images, wherein the parameter information of the residual polymers includes the target position of the residual polymers and the defect information of the carbon nanotubes includes the defect position; determining the optimal laser operating parameters of femtosecond lasers based on the parameter information and / or the defect information in combination with an optical property database; performing fixed-point irradiation on the residual polymers at the target position and the carbon nanotubes at the defect position respectively based on the optimal laser operating parameters, so that the residual polymers are photolyzed and discharged and the in-situ defects of the carbon nanotubes are repaired; and repeatedly collecting surface topography images of the carbon nanotube wafer after fixed-point irradiation, and verifying the remaining amount of the residual polymers and the defect repair status according to the repeatedly collected surface topography images until the residual polymers are completely removed and the defects are repaired.

[0007] The beneficial effects of the above technical solutions are as follows: By integrating precise imaging at the atomic level, precise processing of picosecond lasers with specific wavelengths, and intelligent dynamic control and optimization technologies, a technical platform for automatically detecting and controlling the micro-nano structure and quality of the carbon nanotube surface in-situ is constructed. The non-destructive identification and efficient removal of various residual polymers are realized, and the atomic-level repair of carbon nanotube defects is carried out at the same time. This series of precise in-line characterization, quality evaluation and laser optimization technologies can realize the intelligent analysis and accurate improvement of the surface quality of carbon-based chip materials, making it applicable to automated large-scale processes without obstacles, simplifying operations, reducing costs, increasing production and yield; and is expected to accelerate the wide application of two-dimensional materials such as carbon nanotubes in new-generation micro-nano electronic devices and other fields.

[0008] Based on a further improvement of the above method, the parameter information of the residual polymers further includes the quantity of the residual polymers, the area ratio of the residual polymers, the spatial distribution of the residual polymers on the carbon nanotube surface, and the type of the residual polymers, wherein the type of the residual polymers includes cosolvents, surfactants, crosslinking agents, and / or sensitizers introduced during the preparation of the carbon nanotube wafer; and the defect information of the carbon nanotubes includes the number of carbon nanotube defects and the defect size.

[0009] Based on further improvements to the above method, using machine vision and image processing algorithms, further including identifying parameter information of residual polymers on the surface of carbon nanotube wafers and identifying parameter information of carbon nanotubes according to the set of surface topography images: constructing a convolutional neural network model, and training the convolutional neural network model using historical surface topography images of carbon nanotube wafers to generate a classification model, wherein the historical surface topography images of the carbon nanotube wafers are marked with the parameter information of the residual polymers and / or the defect information of the carbon nanotubes; directly inputting a set of surface topography images of the carbon nanotube wafer to be processed into the classification model to utilize the classification model to automatically detect the residual polymers and determine the types of the residual polymers and automatically detect the defect information, so as to output the parameter information of the residual polymers and the defect information of the carbon nanotubes through the classification model; and combining the region growing algorithm and the template matching method to automatically locate individual residual polymer particles to obtain the central coordinate data of the residual polymers and automatically locate the defect positions of the carbon nanotubes to obtain the central coordinate data of the carbon nanotubes.

[0010] Based on further improvements to the above method, determining the optimal laser operating parameters of femtosecond laser based on the parameter information and / or the defect information in combination with the optical property database includes: based on the parameter information of the residual polymers and / or the defect information of the carbon nanotubes, combining the optical property parameters in the optical property database, and inversely calculating the optimal laser operating parameters for photodissociating and discharging the residual polymers and / or repairing the in-situ defects of the carbon nanotubes, wherein the optimal laser operating parameters include: optimal wavelength, optimal pulse width, optimal peak power, optimal pulse repetition frequency, optimal spot scanning speed; the optical property database includes single-photon and multi-photon absorption coefficients, laser-induced thermal chemical reaction kinetic parameters, and photo-generated carrier quantization transport models.

[0011] Based on further improvements to the above method, further including that by performing fixed-point irradiation on the residual polymers at the target position, the residual polymers are photodissociated and discharged: when irradiating the residual polymers at the target position with femtosecond laser having the optimal laser operating parameters, the residual polymer molecules absorb multi-energy photons, the covalent bonds inside the residual polymer molecules are broken, the main chain and side chains of the residual polymers are broken to promote the decomposition of the residual polymers and the residual polymer molecular clusters remaining on the surface of the carbon nanotubes are effectively discharged.

[0012] Based on the further improvement of the above method, by performing fixed-point irradiation on the carbon nanotubes at the defect positions, the in-situ defect repair of the carbon nanotubes further includes: when femtosecond laser with the optimal laser working parameters irradiates the carbon nanotubes at the defect positions, part of the laser energy is directly absorbed by the carbon nanotubes, selectively exciting the local vibration of the defect positions existing in the carbon nanotubes; in the case of the local vibration, carbon atoms are rearranged and reconstructed to perform in-situ repair on the defects of the carbon nanotubes, where the defects of the carbon nanotubes include dislocations, defects or fractures.

[0013] Based on the further improvement of the above method, when the target position of the residual polymer is the same as the defect position of the carbon nanotubes, fixed-point irradiation is performed on the target position of the residual polymer by femtosecond laser, and the laser energy breaks the molecular chains of the residual polymer present through multi-energy photon absorption to remove the residual polymer; at the same time, when the femtosecond laser irradiates the defect position of the carbon nanotubes, the laser energy is absorbed by the carbon nanotubes, exciting the carbon nanotubes to perform local atomic vibration, causing the defect regions in the carbon nanotubes to be reconstructed, and reordering the carbon atoms to perform in-situ repair on the defects of the carbon nanotubes; when the target position of the residual polymer is different from the defect position of the carbon nanotubes, first, laser irradiation is performed on the target position of the residual polymer to remove the residual polymer, and then another path of laser scanning is performed separately on the defect position of the carbon nanotubes, controlling the laser parameters to excite the local atomic vibration of the defect regions of the carbon nanotubes to perform in-situ repair on the defects of the carbon nanotubes.

[0014] Based on the further improvement of the above method, repeatedly acquiring surface topography images of the carbon nanotube wafer after fixed-point irradiation, and verifying the remaining amount of the residual polymer and the defect repair status according to the repeatedly acquired surface topography images until the residual polymer is completely removed and the defects are repaired further includes: re-acquiring another surface topography image of the carbon nanotube wafer and comparing it with a set of surface topography images before the process to determine the surface damage and reconstruction degree of the carbon nanotube wafer; measuring the surface roughness parameters and elemental composition analysis results to verify the removal degree of the residual polymer and the surface reconstruction degree; performing capacitance current and carrier mobility measurements on the carbon nanotube wafer to evaluate the electrical performance parameters of the carbon nanotube film and indirectly verify the surface quality of the carbon nanotube film.

[0015] On the other hand, an embodiment of the present invention provides a carbon-based integrated circuit atomic-level laser optimization device, comprising: an image scanning module for scanning the thin film surface of a carbon nanotube wafer to obtain a set of surface topography images; an identification module for using machine vision and image processing algorithms to identify the parameter information of residual polymers on the surface of the carbon nanotube wafer and the defect information of carbon nanotubes according to the set of surface topography images, wherein the parameter information of the residual polymers includes the target position of the residual polymers and the defect information of the carbon nanotubes includes the defect position; a laser determination module for determining the optimal laser operating parameters of femtosecond laser based on the parameter information and / or the defect information in combination with an optical property database; a wafer processing module for respectively performing spot irradiation on the residual polymers at the target position and the carbon nanotubes at the defect position, causing the residual polymers to undergo photolysis and be discharged, and repairing the in-situ defects of the carbon nanotubes; and a verification module for repeatedly collecting another set of surface topography images of the carbon nanotube wafer after spot irradiation, and verifying the remaining amount of the residual polymers and the defect repair status according to the another set of surface topography images until the residual polymers are completely removed and the defects are repaired.

[0016] Based on a further improvement of the above device, the parameter information of the residual polymers further includes the quantity of the residual polymers, the area ratio of the residual polymers, the spatial distribution of the residual polymers on the surface of the carbon nanotubes, and the type of the residual polymers, wherein the type of the residual polymers includes cosolvents, surfactants, crosslinking agents, and / or sensitizers introduced during the preparation process of the carbon nanotube wafer; and the defect information of the carbon nanotubes includes the number of defects and the size of the defects of the carbon nanotubes.

[0017] Compared with the prior art, the present invention can at least achieve one of the following beneficial effects:

[0018] 1. It can achieve non-destructive identification and efficient removal of various residual polymers, and at the same time, precise parameters are used for structural repair at the atomic level on the surface of carbon nanotubes, which can ensure that the electronic structure of carbon nanotubes is not damaged, thereby ensuring their electron transport performance. It can integrate a variety of atomic-precision detection and analysis technologies to build a new platform for automatic, in-situ detection and control of the micro-nano structure and quality change on the surface of carbon nanotubes, and can monitor and evaluate the entire laser processing optimization process in real time.

[0019] 2. Based on cutting-edge algorithms such as deep learning, it is possible to achieve intelligent identification and analysis of the surface topography and defects of carbon-based integrated circuit chips, improve the speed and accuracy of surface quality diagnosis and control, and meet the requirements of industrial applications. Through precise matching and optimization of parameters, the operation can be simplified, labor costs can be reduced, automated processing can be achieved, and the production yield and good product rate of carbon-based chip materials can be increased. It is expected to accelerate the wide application and marketization of advanced electronic functional materials such as carbon nanotubes in the fields of new-generation information storage, computing, and other frontier high-tech fields.

[0020] 3. The types and components of the residual polymers on the surface of carbon nanotubes can be determined, laying a foundation for establishing a system for quantitatively evaluating the influence of the surface quality and electronic structure of carbon nanotubes. Automated high-speed detection and diagnosis of the surface topography and defects of carbon nanotubes can be realized, replacing traditional inefficient manual analysis, and generally promoting the progress of quality control technology.

[0021] 4. A closed-loop control process from imaging detection to calculation parameters to precise laser processing can be constructed, which can be gradually iteratively optimized to ensure continuous and automatic improvement of processing quality. The system function of the present invention is modular, with high integration, facilitating subsequent debugging, upgrading, and maintenance, and can provide a technical platform for the processing of a wider range of two-dimensional materials. The key performance indicators of the laser system are advanced, and precise, high-efficiency, stable, and controllable laser irradiation and quality improvement of carbon nanotubes can be achieved.

[0022] In the present invention, the above technical solutions can also be combined with each other to achieve more preferred combination solutions. Other features and advantages of the present invention will be described in the subsequent specification, and some advantages can be made obvious from the specification or understood by implementing the present invention. The objectives and other advantages of the present invention can be realized and obtained through the content specifically pointed out in the specification and the drawings. Description of the Drawings

[0023] The drawings are only for the purpose of showing specific embodiments and are not considered as limiting the present invention. Throughout the drawings, the same reference signs denote the same components.

[0024] Figure 1 It is a flowchart of the atomic-level laser optimization method for carbon-based integrated circuits according to an embodiment of the present invention;

[0025] Figure 2 It is a specific flowchart of the atomic-level laser optimization method for carbon-based integrated circuits according to an embodiment of the present invention;

[0026] Figure 3 It is a flowchart of the machine vision and image processing algorithm according to an embodiment of the present invention;

[0027] Figure 4Schematic structural diagram of an atomic-level laser optimization device for a carbon-based integrated circuit according to an embodiment of the present invention;

[0028] Figure 5 SEM image of carbon nanotubes before laser optimization processing according to an embodiment of the present invention;

[0029] Figure 6 SEM image of carbon nanotubes after laser optimization processing according to an embodiment of the present invention;

[0030] Figure 7 Raman test image of the surface polymer of carbon nanotubes before laser optimization processing according to an embodiment of the present invention;

[0031] Figure 8 Raman test image of the surface polymer of carbon nanotubes after laser optimization processing according to an embodiment of the present invention;

[0032] Figure 9 Flowchart of an atomic-level laser optimization device for a carbon-based integrated circuit according to an embodiment of the present invention. Detailed implementation manners

[0033] The following specifically describes the preferred embodiments of the present invention in conjunction with the accompanying drawings, where the accompanying drawings form a part of this application and are used together with the embodiments of the present invention to explain the principles of the present invention, and are not used to limit the scope of the present invention.

[0034] Refer to Figure 1 , a specific embodiment of the present invention discloses a method for atomic-level laser optimization of a carbon-based integrated circuit, including in step S101, scanning the thin film surface of a carbon nanotube wafer to obtain a set of surface topography images; in step S102, using machine vision and image processing algorithms to identify the parameter information of the residual polymer on the surface of the carbon nanotube wafer and the defect information of the carbon nanotubes according to a set of surface topography images, where the parameter information of the residual polymer includes the target position of the residual polymer and the defect information of the carbon nanotubes includes the defect position; in step S103, determining the optimal laser operating parameters of the femtosecond laser based on the parameter information and / or defect information in combination with the optical property database; in step S104, by respectively performing fixed-point irradiation on the residual polymer at the target position and the carbon nanotubes at the defect position, causing the residual polymer to undergo photolysis and be discharged and repairing the in-situ defects of the carbon nanotubes; and in step S105, repeatedly collecting the surface topography images of the carbon nanotube wafer after fixed-point irradiation, and verifying the remaining amount of the residual polymer and the defect repair status according to the repeatedly collected surface topography images until the residual polymer is completely removed and the defects are repaired.

[0035] Compared with the prior art, in the atomic-level laser optimization method for carbon-based integrated circuits provided in this embodiment, by integrating precise imaging at the atomic level, precise processing with picosecond lasers of specific wavelengths, and intelligent dynamic control and optimization technologies, a technical platform for automatically detecting and controlling the micro-nano structures and quality on the surface of carbon nanotubes in-situ is constructed. Non-destructive identification and efficient removal of various residual polymers are achieved, and at the same time, atomic-level repair of carbon nanotube defects is performed. This series of precise in-line characterization, quality evaluation, and laser optimization technologies can realize intelligent analysis and accurate improvement of the surface quality of carbon-based chip materials, enabling them to be applied in automated large-scale processes without obstacles, simplifying operations, reducing costs, increasing production and yield; and is expected to accelerate the wide application of two-dimensional materials such as carbon nanotubes in new-generation micro-nano electronic devices and other fields.

[0036] In the following, reference Figure 1 is made to detail each step in the atomic-level laser optimization method for carbon-based integrated circuits according to the embodiments of the present invention.

[0037] In step S101, the thin film surface of a carbon nanotube wafer is scanned to obtain a set of surface topography images (i.e., initial topography images or first topography images). The thin film surface of the carbon nanotube wafer is scanned using scanning tunneling microscopy, atomic force microscopy, or scanning transmission microscopy based on transmission electron microscopy to obtain a set of surface topography images, and the surface topography images refer to surface topography images with a resolution at the molecular level or even the atomic level. The visual range achievable by scanning tunneling microscopy, atomic force microscopy, or scanning transmission microscopy based on transmission electron microscopy is small (for example, the visual range is less than 1 μm × 1 μm, such as 100 nm × 100 nm), and this small visual range cannot meet the detection area requirements of the carbon nanotube wafer. However, scanning the thin film surface of the carbon nanotube wafer using scanning tunneling microscopy, atomic force microscopy, or scanning transmission microscopy based on transmission electron microscopy can obtain a set of surface topography images of the entire thin film surface of the carbon nanotube wafer.

[0038] For example, scanning tunneling microscopy obtains an electron cloud distribution image of individual atoms and molecules by detecting the local quantum tunneling current on the surface of the sample; atomic force microscopy uses a non-contact probe with atomic-level resolution to map the precise three-dimensional distribution of atoms on the surface of the sample; scanning transmission microscopy based on transmission electron microscopy can visually reproduce the atomic structure arrangement of thin film materials by detecting quantized backscattered electrons and transmitted electron signals.

[0039] In step S102, using machine vision and image processing algorithms, parameter information of residual polymers on the surface of the carbon nanotube wafer and / or defect information of carbon nanotubes is recognized according to a set of surface topography images. Among them, the parameter information of the residual polymers includes: the target position of the residual polymers, the number of residual polymers, the area ratio of the residual polymers, the spatial distribution of the residual polymers on the surface of the carbon nanotubes, and the type of the residual polymers. Among them, the type of the residual polymers includes cosolvents, surfactants, crosslinking agents, and / or sensitizers introduced during the preparation of the carbon nanotube wafer. The defect information of the carbon nanotubes includes the number of defects of the carbon nanotubes, the defect positions, and the defect sizes.

[0040] Specifically, using machine vision and image processing algorithms to recognize the parameter information of the residual polymers on the surface of the carbon nanotube wafer and recognize the parameter information of the carbon nanotubes according to a set of surface topography images further includes: constructing a convolutional neural network model, and training the convolutional neural network model with historical surface topography images of the carbon nanotube wafer to generate a classification model, where the historical surface topography images of the carbon nanotube wafer are marked with the parameter information of the residual polymers and / or the defect information of the carbon nanotubes; directly inputting a set of surface topography images of the carbon nanotube wafer to be processed into the classification model to utilize the classification model to automatically detect the residual polymers and determine the type of the residual polymers, and automatically detect the defect information, so as to output the parameter information of the residual polymers and the defect information of the carbon nanotubes through the classification model; and combining the region growing algorithm and the template matching method to automatically locate individual residual polymer particles to obtain the center coordinate data of the residual polymers and automatically locate the defect positions of the carbon nanotubes to obtain the center coordinate data of the carbon nanotubes.

[0041] In step S103, based on the parameter information and / or the defect information, the optimal laser operating parameters of the femtosecond laser are determined in combination with the optical property database. Specifically, determining the optimal laser operating parameters of the femtosecond laser based on the parameter information and / or the defect information in combination with the optical property database further includes: based on the parameter information of the residual polymers and / or the defect information of the carbon nanotubes, combining the optical property parameters in the optical property database, and using the variational method for numerical simulation to inversely calculate the optimal laser operating parameters for photodecomposing and discharging the residual polymers and / or repairing the in-situ defects of the carbon nanotubes. The optimal laser operating parameters include: the optimal wavelength, the optimal pulse width, the optimal peak power, the optimal pulse repetition frequency, and the optimal spot scanning speed

[0042] Inverse calculation is carried out by means of variational method numerical simulation. This variational method numerical simulation usually establishes a theoretical model based on the optical characteristic parameters of the residual polymer (such as single / multiphoton absorption coefficient, thermal chemical reaction kinetic parameters, etc.), and then uses numerical algorithms to solve the model to obtain the optimal laser working parameters. The optical characteristic database includes single-photon and multi-photon absorption coefficients, laser-induced thermal chemical reaction kinetic parameters, and photogenerated carrier quantization transport models.

[0043] Single-photon and multi-photon absorption coefficients: The single-photon absorption coefficient describes the efficiency of the residual polymer molecules absorbing a single photon, while the multi-photon absorption coefficient describes the efficiency of simultaneously absorbing multiple photons. These absorption coefficients are closely related to the molecular structure, energy levels, and laser wavelength of the residual polymer, and they determine the strength of the coupling between the laser energy and the residual polymer.

[0044] Laser-induced thermal chemical reaction kinetic parameters: When the residual polymer absorbs laser energy, a series of photochemical and thermal chemical reactions will occur, which lead to processes such as the breaking of molecular bonds and the generation of new species. The kinetic parameters include reaction rate constants, activation energies, association energies, etc., which describe the rate, efficiency, and energy conversion process of the photolysis reaction.

[0045] Photogenerated carrier quantization transport model: Lasers can not only directly cause the breaking of molecular bonds, but also generate quantum carriers such as electron-hole pairs. The transport behavior of these carriers in the residual polymer and carbon nanotubes is very important. The transport model describes the kinetic processes such as the generation, recombination, diffusion, and relaxation of carriers, which affects the photolysis effect of the residual polymer and the energy deposition on the carbon nanotubes.

[0046] The optimal laser working parameters are related to the parameter information of the residual polymer. The parameter information of the residual polymer includes quantity, area ratio, spatial distribution, type, etc. These information will affect the required laser peak power, repetition frequency, scanning speed, etc., and need to be considered in the inverse calculation to obtain the true "optimal" laser working parameters.

[0047] The optimal laser working parameters are also related to the carbon nanotube defect information. For example, the information about carbon nanotube defects (such as the number, size, and position of defects) will also affect the required laser parameters. Because while removing the residual polymer, it is necessary to avoid damaging the carbon nanotubes, so the defect information is also an important input parameter.

[0048] Determine the laser peak power and repetition frequency according to the quantity and area ratio of the residual polymer. Preferably, the optimal peak power and the optimal pulse repetition frequency. For example, the quantity and area ratio of the residual polymer mainly affect the required laser peak power and repetition frequency. For a larger quantity or a larger area of residual polymer, it is necessary to increase the laser peak power to provide sufficient energy to photolyze it per unit time; at the same time, it is also necessary to appropriately increase the repetition frequency to ensure that all residues are covered and removed within a shorter time.

[0049] Determine the spot scanning speed according to the spatial distribution of the residual polymer on the surface of the carbon nanotube. Preferably, the optimal spot scanning speed. For example, the spatial distribution of the residual polymer on the surface of the carbon nanotube mainly affects the selection of the spot scanning speed. If the residues are distributed more dispersedly, it is necessary to reduce the spot scanning speed to ensure that sufficient energy can be obtained for irradiation at each position; if the distribution is more concentrated, the scanning speed can be appropriately increased to improve the overall processing efficiency.

[0050] Determine (or decide) the laser wavelength and laser pulse width according to the type of the residual polymer. Preferably, the optimal wavelength and the optimal pulse width of femtosecond laser or picosecond laser. Due to the differences in molecular structure and optical properties of different types of polymers (such as co-solvents, cross-linking agents, etc.), the corresponding optimal excitation wavelength and the required pulse width will be different. It is necessary to calculate the corresponding optimal wavelength and pulse duration for different types of polymers in combination with the optical property database.

[0051] Based on the parameter information of the residual polymer (such as the type of the residual polymer, etc.) and the defect information of the carbon nanotube, combined with the data in the optical property database, the corresponding optimal laser working parameters can be inversely calculated.

[0052] Therefore, for different types of residual polymers and carbon nanotubes with different defect positions, it is necessary to calculate the corresponding optimal laser parameters such as laser wavelength, pulse width, peak power, repetition frequency and scanning speed by using the optical property database through numerical simulation methods according to their specific parameter information, and then use these optimal parameters for precise fixed-point irradiation, so as to achieve the purpose of efficiently removing the residual polymer and repairing the defects of the carbon nanotube.

[0053] In step S104, based on the optimal laser working parameters, by respectively performing fixed-point irradiation on the residual polymer at the target position and the carbon nanotube at the defect position, the residual polymer is photolyzed and discharged, and the in-situ defects of the carbon nanotube are repaired.

[0054] Specifically, by performing spot irradiation on the residual polymer at the target position to cause photolysis of the residual polymer and discharge it further includes: when irradiating the residual polymer at the target position with femtosecond laser having optimal laser working parameters, the residual polymer molecules absorb multi-energy photons, and the covalent bonds inside the residual polymer molecules are broken, causing the main chain and side chains of the residual polymer to break to promote the decomposition of the residual polymer and effectively discharging the residual polymer molecular clusters remaining on the surface of the carbon nanotubes. Specifically, when the femtosecond laser irradiates the residual polymer at the target position, the residual polymer molecules absorb multi-energy photons; when the multi-energy photons are absorbed simultaneously, the internal electrons of the residual polymer molecules are excited to transition from the ground state to a high-energy excited state or directly ionized; in the case of high-energy excited state or direct ionization, the covalent bonds inside the residual polymer molecules are broken, causing the main chain and side chains of the residual polymer to break, and at the same time, the excited state electrons and hole pairs induce a thermochemical process to promote the decomposition of the residual polymer; and the main chain and side chains of the residual polymer break, effectively discharging the residual polymer molecular clusters remaining on the surface of the carbon nanotubes.

[0055] Specifically, by performing spot irradiation on the carbon nanotubes at the defect position to repair the in-situ defects of the carbon nanotubes further includes: when irradiating the carbon nanotubes at the defect position with femtosecond laser having optimal laser working parameters, part of the laser energy is directly absorbed by the carbon nanotubes, selectively exciting the local vibration of the defect positions existing in the carbon nanotubes; in the case of local vibration, the carbon atoms are rearranged and reconstructed to repair the defects of the carbon nanotubes in-situ, where the defects of the carbon nanotubes include dislocations, defects or fractures. Specifically, when the femtosecond laser irradiates the carbon nanotubes at the defect position, part of the laser energy is directly absorbed by the carbon nanotubes. The hollow cylindrical structure and quantized electronic states of the carbon nanotubes result in multiple characteristic absorption peaks in the visible light wavelength to near-infrared light wavelength range. Among them, when the laser energy matches the energy levels of multiple characteristic absorption peaks, the π electrons inside the carbon nanotubes are excited to form excitons; by selecting laser parameters according to the defect information of the carbon nanotubes, selectively exciting the local vibration of the defect positions existing in the carbon nanotubes; in the case of local vibration, the carbon atoms are rearranged and reconstructed to repair the defects of the carbon nanotubes in-situ, where the defects of the carbon nanotubes include dislocations, defects or fractures.

[0056] Specifically, when the target position of the residual polymer coincides with the defect position of the carbon nanotube, the target position of the residual polymer is irradiated point by point with femtosecond laser. The laser energy breaks the molecular chains of the residual polymer through multi-energy photon absorption to remove the residual polymer. At the same time, when the femtosecond laser irradiates the defect position of the carbon nanotube, the laser energy is absorbed by the carbon nanotube, exciting the carbon nanotube to perform local atomic vibration, causing the defect area in the carbon nanotube to be reconstructed, and reordering the carbon atoms to in-situ repair the defects of the carbon nanotube. When the target position of the residual polymer is different from the defect position of the carbon nanotube, first, the target position of the residual polymer is irradiated with laser to remove the residual polymer, and then the defect position of the carbon nanotube is separately scanned with laser in another path, and the laser parameters are controlled to excite the local atomic vibration in the defect area of the carbon nanotube to in-situ repair the defects of the carbon nanotube.

[0057] In step S105, the surface topography images of the carbon nanotube wafer after point irradiation are repeatedly collected (i.e., another set of surface topography patterns or the second set of surface topography images), and the remaining amount of the residual polymer and the defect repair status are verified based on the repeatedly collected surface topography images until the residual polymer is completely removed and the defects are repaired. When the second set of surface topography images is usually different from the first set of surface topography images, the carbon nanotube wafer is continuously irradiated point by point to remove the residual polymer and repair the in-situ defects of the carbon nanotube. When the second set of surface topography images is the same as the first set of surface topography images, the continuous point irradiation of the carbon nanotube wafer is stopped.

[0058] Specifically, repeatedly collecting the surface topography images of the carbon nanotube wafer after point irradiation and verifying the remaining amount of the residual polymer and the defect repair status based on the repeatedly collected surface topography images further includes: re-acquiring another surface topography image of the carbon nanotube wafer and comparing it with a set of surface topography images before the process to determine the surface damage and reconstruction degree of the carbon nanotube wafer; measuring the surface roughness parameters and the elemental composition analysis results to verify the removal degree of the residual polymer and the surface reconstruction degree; measuring the capacitance current and carrier mobility of the carbon nanotube wafer to evaluate the electrical performance parameters of the carbon nanotube thin film and indirectly verify the surface quality of the carbon nanotube thin film.

[0059] Reference Figure 9, Another specific embodiment of the present invention discloses an atomic-level laser optimization device for carbon-based integrated circuits, comprising: an image scanning module 901 for scanning the thin film surface of a carbon nanotube wafer to obtain a set of surface topography images; an identification module 902 for using machine vision and image processing algorithms to identify the parameter information of the residual polymer on the surface of the carbon nanotube wafer and the defect information of the carbon nanotubes according to a set of surface topography images, wherein the parameter information of the residual polymer includes the target position of the residual polymer and the defect information of the carbon nanotubes includes the defect position; a laser selection module 903 for determining the optimal laser operating parameters of femtosecond laser based on the parameter information and / or defect information in combination with an optical property database; a wafer processing module 904 for respectively performing spot irradiation on the residual polymer at the target position and the carbon nanotubes at the defect position, causing the residual polymer to undergo photolysis and be discharged, and repairing the in-situ defects of the carbon nanotubes; and a verification module 905 for repeatedly collecting surface topography images of the carbon nanotube wafer after spot irradiation, and verifying the remaining amount of the residual polymer and the defect repair status according to the repeatedly collected surface topography images until the residual polymer is completely removed and the defects are repaired.

[0060] The parameter information of the residual polymer further includes the target position of the residual polymer, the quantity of the residual polymer, the area ratio of the residual polymer, the spatial distribution of the residual polymer on the surface of the carbon nanotubes, and the type of the residual polymer, wherein the type of the residual polymer includes a cosolvent, a surfactant, a crosslinking agent, and / or a sensitizer introduced during the preparation of the carbon nanotube wafer. The defect information of the carbon nanotubes includes the defect position, the number of defects, and the size of the defects.

[0061] Hereinafter, reference is made to Figures 2 to 8 , and a carbon-based integrated circuit atomic-level laser optimization method according to an embodiment of the present invention is described in detail by way of specific examples.

[0062] Reference is made to Figure 2, the carbon-based integrated circuit atomic-level laser optimization method according to an embodiment of the present invention (combining the laser energy action at the atomic level and the online detection technology with atomic resolution) includes: (1) preparing a carbon nanotube wafer film sample prepared by a solution method; (2) using atomic-level characterization methods such as scanning tunneling microscopy, high-resolution transmission electron microscopy, or atomic force microscopy to scan the surface of the sample to obtain a surface topography image with molecular-level or even atomic-level resolution; (3) applying machine vision and image processing algorithms to accurately identify the parameter information of different polymer species residues; (4) selecting a picosecond or femtosecond laser with a suitable wavelength and pulse frequency according to the optical property database for fixed-point irradiation of each target position; (5) the precisely controlled picosecond or femtosecond laser couples energy to the molecular chains and atomic bonds in the nano-region through multi-photon absorption, causing the polymer to undergo photolysis and expulsion, and at the same time realizing in-situ defect repair by exciting the local atomic vibration of the carbon nanotubes; (6) collecting and processing the surface image of the processed area again through the above atomic-level characterization method to verify the effect of residual polymer removal and carbon nanotube surface repair. (7) According to the newly obtained sample surface information, intelligently optimize the next laser irradiation parameters and path planning.

[0063] The parameter information of the residual polymer includes: the number of residues; the area ratio of the residues; the spatial distribution of the residues on the surface of the carbon nanotubes; the target position of the residues; the type or species of the residue particles (such as cosolvent residue, surfactant residue, etc.).

[0064] Applied to a picosecond or femtosecond laser system, perform fixed-point precise focusing irradiation on the target position of each residual polymer accurately located by the machine vision algorithm before.

[0065] The optical property database contains the following optical property parameters of various polymer materials: 1. Single-photon and multi-photon absorption coefficients; 2. Laser-induced thermal chemical reaction kinetic parameters; 3. Parameters describing the interaction between light and matter such as the quantum transport model of photo-generated carriers.

[0066] The database is updated in real time. Using online monitoring and machine autonomous learning algorithms, the optical property database is updated in real time to keep it in the latest state.

[0067] Inverse calculation of the optimal laser parameters. Based on the optical property parameters of each residual polymer in the database, using the variational method for numerical simulation, inverse calculate the optimal laser operating parameters that can effectively excite and photolyze the material, including: the optimal wavelength; the optimal pulse width; the optimal peak power; the optimal pulse repetition frequency; the optimal spot scanning speed, etc. Perform precise fixed-point irradiation with the laser parameters calculated above.

[0068] Taking the polymer additive polycarbazole (PCz) as an example, it is a commonly used polymer additive in the growth process of carbon nanotubes, but it is very easy to remain on the surface of carbon nanotubes, affecting the performance of subsequent devices. When precisely controlled femtosecond laser irradiates the residual PCz polymer, the ultrashort pulses (e.g., 10 - 100 femtoseconds) and extremely high peak power density of the femtosecond laser will trigger a non-linear multi-photon absorption process of PCz molecules.

[0069] Multiple low-energy photons are absorbed simultaneously, exciting the internal electrons of PCz molecules from the ground state to high-energy excited states or directly ionizing them.

[0070] In this high-energy excited state, the covalent bonds inside PCz molecules will be broken, and a "photolysis" reaction of photochemical dissociation will occur, that is, the breakage of the polymer main chain and side chains. At the same time, the high-density excited-state electrons and hole pairs will also induce a series of subsequent thermochemical processes, further promoting the decomposition of the polymer.

[0071] Finally, the main chain and side chains of the PCz polymer are cut off, losing the original polymer structure, and the PCz molecular clusters remaining on the surface of the carbon nanotubes are effectively "discharged". This process is efficiently completed within a nanoscale spatial range under precise control.

[0072] Through the non-linear multi-photon absorption effect of femtosecond laser, high energy can be precisely coupled to the molecular chains and atomic bonds in the nanoscale region, inducing the photolysis and discharge of polymer molecules, thereby selectively and non-destructively removing the residual polymers on the surface of carbon nanotubes.

[0073] 1. When precisely controlled femtosecond laser irradiates the surface of carbon nanotubes, in addition to inducing the photolysis and discharge of residual polymers, part of the laser energy will also be directly absorbed by the carbon nanotubes. The unique hollow cylindrical structure and quantized electronic states of carbon nanotubes enable it to have many characteristic absorption peaks in the visible to near-infrared wavelength range. When the laser energy matches the energy levels of these absorption peaks, the π electrons inside the carbon nanotubes will be excited to form excitons (i.e., bound electron-hole pairs).

[0074] In this excited state, the carbon atoms will undergo radial vibrations from the inside out. Due to the high peak power density of the laser, this vibration is highly localized, only concentrated in certain nanoscale regions. By selecting appropriate laser parameters, the local vibrations of those regions with defects or deficiencies in the carbon nanotubes can be selectively excited.

[0075] Driven by this local vibration, the carbon atoms will rearrange and reconstruct, thereby repairing the previously existing dislocation, defect or fracture and other defective structures. This process does not require any external conditions and is spontaneously completed at the intrinsic atomic scale of carbon nanotubes.

[0076] It should be emphasized that the repair process is "in-situ" and does not require destructive treatment of the entire carbon nanotube, effectively maintaining the integrity of its overall structure and electrical properties.

[0077] 2. (1) The case where the positions of the residue and the defect are the same: The positions of the residual polymer and the surface defect of the carbon nanotube coincide. Usually, it is because the residual polymer exists at the defect position of the carbon nanotube, or the residual polymer itself causes the generation of carbon nanotube defects.

[0078] During femtosecond laser irradiation, first perform fixed-point irradiation on the position of the residual polymer. The laser energy breaks the polymer molecular chain through multi-photon absorption, achieving the removal and discharge of the residue.

[0079] At the same time, the laser energy irradiated to these positions will also be absorbed by the remaining carbon nanotube structure, exciting the local atomic vibration of the carbon nanotube. Driven by this vibration, the defect area in the carbon nanotube will be reconstructed, and the atoms will be rearranged in an orderly manner, thus repairing the defect.

[0080] Therefore, in this case of position coincidence, only one-step precise laser irradiation is required to simultaneously remove the residual polymer and repair the surface defect of the carbon nanotube.

[0081] (2) The case where the positions of the residue and the defect are different: If the positions of the residual polymer and the carbon nanotube defect are separated, two-step treatment is required:

[0082] In the first step, first perform precise laser irradiation on the position of the residual polymer to remove the residue.

[0083] In the second step, separately perform laser scanning on the surface defect position of the carbon nanotube along another path, precisely control the parameters to excite the local atomic vibration in the defect area, and achieve in-situ defect repair.

[0084] In the second step, due to the absence of the influence of the residual polymer, the laser parameters can be adjusted more freely, focusing on optimizing the vibration mode and reconstruction path of the carbon atoms in the defect area, so as to obtain a higher-quality repair effect.

[0085] It should be noted that in the case where the positions of the residue and the defect are separated, the process route will be relatively more complex. It is necessary to accurately distinguish the positions of the two, and set different laser scanning paths and parameters respectively, which will make the operation process more refined.

[0086] Generally speaking, whether the positions of the residue and the defect are the same or different, the method of the present invention can efficiently remove the residual polymer and repair the surface defect of the carbon nanotube through precise laser irradiation, thereby comprehensively improving the quality of the carbon nanotube material.

[0087] For the newly obtained surface information of the sample, after each round of laser irradiation treatment, atomic-level characterization techniques such as atomic force microscopy are reused to obtain the latest surface topography image of the treated area. By comparing with the original image, the following information can be detected: the removal of residual polymer (whether there is still residue); newly generated defects or damages on the surface of carbon nanotubes; the change of surface roughness; the change of surface element composition.

[0088] Intelligently optimize the laser irradiation parameters, use machine vision algorithms to intelligently analyze the newly obtained surface images, and automatically extract key parameters such as residue coverage and defect quantity. Then input these parameters into the previously trained artificial intelligence model, and combine with the previously established optical property database to intelligently optimize and calculate the optimal laser parameters for the next step, including laser wavelength, pulse width, peak power, pulse repetition frequency, spot scanning speed, etc.

[0089] Optimize the laser scanning path planning. Based on the new surface topography image, use image processing and pattern recognition algorithms to automatically identify and extract the precise position coordinates of the residual polymer and newly generated defects. Input this position coordinate information into the path planning module, and combine with the previously optimized laser parameters to plan the scanning path in three-dimensional space. This path will cover all target positions that need further treatment, and the order and scanning direction of the path will also be automatically optimized to obtain the highest processing efficiency.

[0090] Closed-loop control and feedback update. Conduct the next round of precise laser irradiation treatment according to the optimized laser parameters and scanning path. Obtain the surface information after treatment again, intelligently analyze and evaluate the treatment effect. If the requirements are not met, enter the next round of parameter and path optimization, and continuously iterate until the quality requirements are met. At the same time, the new data obtained during the optimization process will also be used to continuously update the optical property database and artificial intelligence model, forming a closed-loop automatic optimization control.

[0091] The beneficial effects of the above technical solutions are as follows: By integrating atomic-level precise imaging, picosecond laser precise treatment with specific wavelengths, and intelligent dynamic control and optimization technologies, this method constructs a technical platform for automatically detecting and controlling the micro-nano structure and quality of the carbon nanotube surface in-situ. It realizes the non-destructive identification and efficient removal of various residual polymers, and at the same time repairs the carbon nanotube defects at the atomic level. This series of precise on-line characterization, quality evaluation and laser optimization technologies can realize the intelligent analysis and accurate improvement of the surface quality of carbon-based chip materials, making it applicable to automated large-scale processes without obstacles, simplifying operations, reducing costs, increasing production and yield; and is expected to accelerate the wide application of two-dimensional materials such as carbon nanotubes in new-generation micro-nano electronic devices and other fields.

[0092] The residual polymers of carbon nanotubes prepared by the solution method mainly include cosolvents, surfactants, crosslinking agents, sensitizers, etc. introduced during the preparation process.

[0093] The beneficial effects are as follows: the types of various residual polymers that may be introduced during the preparation of carbon nanotubes by the solution method, including cosolvents, surfactants, crosslinking agents, sensitizers, etc. This provides a basic reference range and classification standard for the detection and identification of surface residues of carbon nanotubes, and also provides a basis for subsequent selection of matching laser irradiation parameters. In addition, clarifying the composition of the possible residues also helps to further analyze their influence mechanisms on the electronic structure and electrical transport properties of carbon nanotubes, and guides the establishment of a quantitative characterization and evaluation system.

[0094] Machine vision and image processing algorithms for the intelligent analysis method of carbon nanotube surface topography based on deep learning include: constructing and pre-training a convolutional neural network model: based on a large amount of atomic-level image data of carbon nanotube surface topography collected and annotating the residual polymer regions, using this data set to train the convolutional neural network model to obtain the recognition and classification model specific to the method of the present invention; intelligent analysis of carbon nanotube surface images: for the atomic-level surface images of actual samples, directly input them into the pre-trained convolutional neural network model, automatically detect and classify different types of residual polymers, and output characteristic parameters such as their quantity, area ratio, and spatial distribution; precise positioning of residues: combining computer vision techniques such as region growing algorithm and template matching to achieve precise automatic positioning of individual residue particles, and giving their center coordinate information, with a positioning accuracy reaching the nanometer level.

[0095] The intelligent analysis method for the surface topography image of carbon nanotubes based on deep learning has innovative improvements compared with the existing technologies. The main improvements are reflected in the following aspects: 1. Introduction of deep learning technology: This algorithm utilizes the convolutional neural network model in deep learning to perform intelligent analysis on the atomic-level topography image of carbon nanotubes. Compared with traditional machine vision methods, deep learning can automatically learn the feature representation of images and has stronger recognition ability for complex nano-scale image patterns. 2. The algorithm is specifically for carbon nanotube images: Most traditional image recognition algorithms are general, but this algorithm is specifically trained based on a large amount of labeled data of the surface topography images of carbon nanotubes. This makes the algorithm have higher accuracy and applicability when identifying features such as residual polymers on carbon nanotubes. 3. Automatic detection and classification: This algorithm can automatically detect and classify different types of residual polymers on the surface of carbon nanotubes and output multi-dimensional feature parameters such as their quantity, area ratio, and spatial distribution. Traditional algorithms often require manual marking and counting, with low efficiency. 4. Sub-nanometer level precise positioning: By combining technologies such as region growing and template matching, the algorithm can achieve sub-nanometer level (in the order of 1 nanometer) precise positioning of a single residual polymer particle and give its precise center coordinates. This accuracy far exceeds that of ordinary algorithms and meets the atomic-level operation requirements of this invention. 5. End-to-end integration of data: This algorithm realizes end-to-end integration from data acquisition, annotation, model training to final application deployment and can be continuously optimized and expanded according to requirements. Existing algorithms are often scattered modular solutions and are not tightly integrated enough.

[0096] The beneficial effects are as follows: A convolutional neural network deep learning model specifically for the surface topography image of carbon nanotubes is constructed, improving the accuracy of detection and classification of residual polymers and laying a foundation for subsequent fixed-point laser treatment. Automatic and efficient surface scanning and quality evaluation of carbon nanotube samples are realized, and parameters such as quantity, area, and distribution can be detected, replacing the traditional cumbersome manual analysis. By combining a more precise positioning algorithm, a positioning accuracy of 1 nanometer order of magnitude for a single residue particle can be achieved, meeting the requirements of atomic-level treatment and repair. Overall, intelligent, automatic detection and diagnosis of the surface of carbon nanotubes are realized, and the speed of characterization and quality control is improved. Combining the whole method provides a key technical means for in-situ and non-destructive evaluation and improvement of the quality of carbon-based chip materials.

[0097] The optical property database includes optical property parameters of photo-induced laser-matter interaction such as single-photon and multi-photon absorption coefficients of various polymer materials, kinetic parameters of laser-induced thermochemical reactions, and a quantitative transport model of photo-generated carriers. The online monitoring and machine autonomous learning algorithm are used to update the optical property database in real time. Based on the optical property parameters of each residual polymer in the database, the variational method numerical simulation is used to inversely calculate the key parameters such as the optimal laser working wavelength, pulse width, peak power, pulse frequency, and spot scanning speed for exciting and effectively photodegrading the material. The above calculation results are converted into a precise focused laser scanning path in three-dimensional space to perform atomic-level cleaning of residual polymer removal and form a closed-loop control for in-situ repair of carbon nanotubes.

[0098] The beneficial effects are as follows: A database containing optical property parameters of various polymer materials, such as single-photon and multi-photon absorption coefficients, kinetic parameters of photo-induced thermochemical reactions, etc., is constructed, which can accurately describe the interaction process between light and materials. The online monitoring and machine autonomous learning algorithm are used to update the database in real time to keep it up-to-date. Based on this database, the optimal laser working parameters required for removing various residual polymers can be accurately calculated by variational method numerical simulation. The calculation results are converted into a precise laser scanning path in three-dimensional space to achieve atomic-level cleaning and removal of residual polymers. A closed-loop control for in-situ repair of carbon nanotubes is formed, that is, the entire closed-loop process of detecting, calculating, removing residues, and repairing carbon nanotubes.

[0099] The verification of the removal of residual polymers and the repair of the carbon nanotube surface specifically includes: (a) Re-acquiring the atomic-level topographic image of the surface of the processed area and comparing it with the image before the process to observe the surface particle damage situation and the degree of reconstruction; (b) Measuring the surface roughness parameters and the elemental composition analysis results to quantitatively verify the removal degree of polymer elements and the degree of surface reconstruction; (c) Conducting methods such as capacitance current testing and carrier mobility measurement to evaluate the electrical performance parameters of the carbon nanotube film and indirectly verify the effect of the improvement of its surface quality; (d) Preparing carbon nanotube film-based devices and testing their key performance indicators to verify the effectiveness of the method in improving the subsequent device integration and application effects.

[0100] The beneficial effects are as follows: Re-acquire the atomic-level surface topographic image of the processed area, compare it with the image before the process, and directly observe the surface particle damage and reconstruction situation. Measure the surface roughness and elemental composition analysis results to quantitatively analyze the removal degree of residual polymer elements and the degree of surface reconstruction. Conduct electrical tests such as capacitance, current, and carrier mobility to indirectly evaluate the improvement effect of the carbon nanotube film surface quality. Prepare carbon nanotube-based devices and test their key performance indicators to verify the effectiveness of the method for subsequent device integration and application.

[0101] On the other hand, refer toFigure 4 , an embodiment of the present invention provides an atomic-level laser optimization device for carbon-based integrated circuits, including: a sample loading unit, an atomic-level imaging unit, a laser irradiation unit, an image processing and laser control and path planning unit, and a quality control unit.

[0102] The beneficial effects are as follows: It realizes the atomic-level laser optimization processing of carbon-based integrated circuits and improves the processing accuracy; is equipped with a complete laser irradiation unit, which can achieve precise laser control and path planning; integrates units such as atomic-level imaging and image processing, and can perform atomic-level characterization and analysis on samples; sets up a quality control unit, which can comprehensively evaluate the effect of laser processing; the modular design has a high system integration degree and is easy to debug and maintain.

[0103] The laser irradiation unit uses a picosecond or femtosecond pulsed light source with adjustable wavelength and is equipped with a precision optical system. The optical system includes a focusing lens, a scanning platform, and a process monitoring module. The focusing lens can focus the laser to the sub-micron range on the sample surface; the scanning platform can perform precise two-dimensional scanning on the sample surface according to the pre-planned path; the process monitoring module can detect the changes in surface topography and laser parameters in real time, ensuring the accuracy and stability of laser irradiation.

[0104] The beneficial effects are as follows: It uses a picosecond or femtosecond pulsed light source with adjustable wavelength, has a short pulse width, high peak power, and concentrated energy, which is conducive to realizing refined surface processing. It is equipped with a precision optical system, realizing precise focusing and scanning of the laser, and has high irradiation accuracy. The focusing lens enables the laser to be focused to the sub-micron range, which is beneficial to atomic-level surface processing. The scanning platform can perform two-dimensional scanning according to the pre-planned path, with a large scanning range and strong accessibility. The process monitoring module can monitor the surface topography and laser parameters in real time, timely feedback the operating state of the device, and ensure the accuracy and stability of processing. Each module works together to achieve an accurate, controllable, and stable laser processing process as a whole.

[0105] The quality control unit realizes the automatic analysis of the surface topography image of the sample by calling machine vision algorithms, and can give quality evaluation results such as the surface roughness parameter, the number of defects, and the residue coverage of the sample surface; and combines the quality evaluation results to send updated laser parameters or scanning path data to the laser control and path planning unit, thereby realizing the full-automatic, closed-loop control and process optimization of the device.

[0106] The beneficial effects are as follows: By invoking machine vision algorithms to automatically analyze the surface topography of samples, the evaluation of key surface quality parameters is realized, such as roughness, defect quantity, residue coverage, etc. The quality evaluation results can be used for closed-loop control and process optimization, and the automatic operation of the device is optimized by feeding back and updating laser parameters or scanning path data. The full-automatic and closed-loop control of the entire device is achieved without manual intervention, making the processing more intelligent and precise. The continuously optimized processing flow can gradually improve the processing quality and meet higher quality requirements. The risk of misoperation caused by human factors is reduced, and the stability and consistency of processing are improved. The control strategy can be flexibly adjusted according to different samples and processing requirements to meet diverse processing needs.

[0107] The residual polymers introduced during the preparation of carbon nanotube array films by the solution method mainly include the following categories: (1) Co-solvents: To enhance the solubility of carbon nanotubes, organic co-solvents with good deferring properties need to be added, such as xylene, toluene, acetonitrile, N-methylpyrrolidone, etc.; (2) Surfactants: Carbazole-based homopolymers, cationic surfactants such as [C11H23(CH3)3N]Br (undecyltrimethylammonium bromide) are added to reduce the bundling state between carbon nanotubes; (3) Cross-linking agents: Hydrazine-based compounds, alkyd compounds, and isocyanate compounds are added to perform end-group functionalization on the sidewalls and ends of carbon nanotubes to form a three-dimensional cross-linked support network; (4) Other additives: Other functional oligomers or small-molecule substances such as plasticizers, stabilizers, and dispersants can also be introduced. By precisely controlling the types and dosages of the above additives, it is helpful to obtain carbon nanotube thin film materials with more excellent properties.

[0108] The atomic-level characterization techniques are scanning tunneling microscopy, atomic force microscopy, or scanning transmission microscopy based on transmission electron microscopy. Specifically, scanning tunneling microscopy obtains the electron cloud distribution images of individual atoms and molecules by detecting the local quantum tunneling current on the sample surface; atomic force microscopy uses a non-contact probe with atomic-level resolution to map the precise three-dimensional distribution of atoms on the sample surface; the scanning transmission microscopy based on transmission electron microscopy can visually reproduce the atomic structure arrangement of thin film materials by detecting quantized backscattered electrons and transmitted electron signals. All atomic-level microscopy techniques can be used in this method to perform real-time, on-line, and in-situ detection of the atomic-level state of the carbon-based chip surface, providing theoretical guidance for subsequent precise laser optimization.

[0109] Take a 4-inch Si / SiO2 wafer, and then grow a densely arranged single-walled carbon nanotube grid film on its surface by the solution method, as Figure 5As shown. The average length of the carbon nanotubes is about 50 micrometers, and the average diameter is about 1.5 nanometers. During the growth process of the carbon nanotubes, a typical carbazole-based homopolymer (poly[9-(1-octylonoyl)-9H-carbazole-2,7-diyl] (PCz)) was used as a polymer additive. However, after growth, the PCz polymer was not removed and still adhered to the surface of some carbon nanotubes, which can be observed through Figure 7 the Raman test results. For example, referring to Figure 7 , the Raman shift on the abscissa reflects the molecular vibration frequency and is related to the molecular structure, while the Raman intensity on the ordinate reflects the Raman activity of the molecule and is related to the change in molecular polarizability. Therefore, the sample exhibits certain problems of resistivity and current leakage.

[0110] To solve this problem, first, atomic force microscopy technology was used to scan the surface of a 1-square-micrometer area of the carbon nanotube wafer at 28 °C and 20% relative humidity to obtain a surface topography image with a resolution of 0.48 nanometers. The atomic-level image was input into a convolutional neural network model pre-trained using a dataset annotation, and its algorithm flow is as Figure 3 shown, to automatically detect and identify the areas where the PCz polymer exists in the image. Combining the region segmentation and particle matching algorithms for growth, the central coordinates of each PCz residue particle were output, and the positioning accuracy reached the order of 1 nanometer.

[0111] According to the database of the inherent optical property parameters of the material, a single-frequency third-harmonic generation (THG) femtosecond pulsed laser with a laser wavelength of 405 ± 5 nanometers and pulse widths of 10, 20, 30, 40, 50, 60, 70, 80, 90, 100 femtoseconds was selected as the optimal irradiation source. Based on the key parameters such as the optimal laser power density, pulse repetition frequency, exposure time, and scanning speed calculated by the variational method for the PCz polymer material, a precisely focused spot scanning laser irradiation was carried out, and each central position of the PCz polymer residue was irradiated in turn. The laser couples energy to the PCz polymer molecular chain through non-linear multi-photon absorption, causing it to undergo a photochemical reaction and desorb from the surface of the carbon nanotubes. At the same time, local atomic vibrations are excited, realizing the reconstruction of the atomic clusters on the CNT tube surface. The optimized carbon nanotube thin film image is as Figure 6 shown.

[0112] The atomic resolution topography image of the surface of the carbon nanotubes within a 1-square-micrometer range was obtained again. By comparing with the original image, the results show that all the PCz residual polymers have been removed, and no other new defects or damages appear on the surface. Through Figure 8 the Raman test, it was observed that the 3 polymer peaks disappeared. For example, referring to Figure 8, the abscissa Raman shift reflects the molecular vibration frequency and is related to the molecular structure. The ordinate Raman intensity reflects the Raman activity of the molecule and is related to the change in molecular polarizability. At the same time, by comparing the results of roughness, surface element composition analysis, transmission electron microscopy, etc., it is further confirmed that the surface reconstruction of carbon nanotubes is very ideal, reaching atomic-level flatness. To accurately evaluate the treatment effect, the resistivity parameter of the carbon nanotube thin film sample was obtained by four-probe testing: the resistivity after laser optimization decreased by about 60% compared with the original sample. The results of this example fully verify that the method can achieve atomic-level precise residue cleaning, significantly improving the electrical performance and stability of carbon nanotube devices.

[0113] Those skilled in the art can understand that all or part of the processes of implementing the methods of the above embodiments can be completed by instructing relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium. Among them, the computer-readable storage medium is a disk, an optical disc, a read-only memory, or a random access memory, etc.

[0114] The above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed by the present invention should be covered by the protection scope of the present invention.

Claims

1. An atomic-level laser optimization method for carbon-based integrated circuits, characterized in that Including: Scanning the thin film surface of a carbon nanotube wafer to obtain a set of surface topography images; Using machine vision and image processing algorithms to identify parameter information of residual polymers on the surface of the carbon nanotube wafer and defect information of the carbon nanotubes based on the set of surface topography images, wherein the parameter information of the residual polymers includes the target positions of the residual polymers and the defect information of the carbon nanotubes includes the defect positions; Determining the optimal laser operating parameters of femtosecond laser based on the parameter information and / or the defect information in combination with an optical property database; Based on the optimal laser operating parameters, respectively performing spot irradiation on the residual polymers at the target positions and the carbon nanotubes at the defect positions, causing the residual polymers to undergo photolysis and be discharged, and repairing the in-situ defects of the carbon nanotubes; and Repeatedly collecting surface topography images of the carbon nanotube wafer after spot irradiation, and verifying the remaining amount of the residual polymers and the defect repair status based on the repeatedly collected surface topography images until the residual polymers are completely removed and the defects are repaired; 2. The method for optimizing a carbon-based integrated circuit at the atomic level according to claim 1, wherein The parameter information of the residual polymers further includes the quantity of the residual polymers, the area ratio of the residual polymers, the spatial distribution of the residual polymers on the surface of the carbon nanotubes, and the types of the residual polymers, wherein the types of the residual polymers include cosolvents, surfactants, crosslinking agents, and / or sensitizers introduced during the preparation process of the carbon nanotube wafer; and The defect information of the carbon nanotubes further includes the number of defects and the size of the defects of the carbon nanotubes; 3. The method for optimizing a carbon-based integrated circuit at the atomic level by laser according to claim 2, wherein Using machine vision and image processing algorithms to identify parameter information of residual polymers on the surface of the carbon nanotube wafer and identify parameter information of the carbon nanotubes based on the set of surface topography images further includes: Constructing a convolutional neural network model, and training the convolutional neural network model using historical surface topography images of the carbon nanotube wafer to generate a classification model, wherein the historical surface topography images of the carbon nanotube wafer are labeled with the parameter information of the residual polymers and / or the defect information of the carbon nanotubes; Directly inputting a set of surface topography images of the carbon nanotube wafer to be processed into the classification model to utilize the classification model to automatically detect the residual polymers and determine the types of the residual polymers, and automatically detect the defect information, so as to output the parameter information of the residual polymers and the defect information of the carbon nanotubes through the classification model; and Combining region growing algorithm and template matching method to automatically locate individual residual polymer particles to obtain the center coordinate data of the residual polymers and automatically locate the defect positions of the carbon nanotubes to obtain the center coordinate data of the carbon nanotubes; 4. The method for optimizing a carbon-based integrated circuit at the atomic level by laser according to claim 1, wherein Determining the optimal laser operating parameters of femtosecond laser based on the parameter information and / or the defect information in combination with an optical property database includes: Based on the parameter information of the residual polymer and / or the defect information of the carbon nanotubes, combined with the optical property parameters in the optical property database, the optimal laser operating parameters for photodecomposing and discharging the residual polymer and / or repairing the in-situ defects of the carbon nanotubes are inversely calculated, where the optimal laser operating parameters include: optimal wavelength, optimal pulse width, optimal peak power, optimal pulse repetition frequency, optimal spot scanning speed; the optical property database includes single-photon and multi-photon absorption coefficients, laser-induced thermochemical reaction kinetic parameters, and photo-generated carrier quantization transport models.

5. The method for optimizing a carbon-based integrated circuit at the atomic level according to claim 4, wherein The step of photodecomposing and discharging the residual polymer at the target position by performing fixed-point irradiation on the residual polymer further includes: When irradiating the residual polymer at the target position with femtosecond laser having the optimal laser operating parameters, the residual polymer molecules absorb multi-energy photons, and the covalent bonds inside the residual polymer molecules are broken, causing the main chain and side chains of the residual polymer to break to promote the decomposition of the residual polymer and effectively discharging the residual polymer molecular clusters remaining on the surface of the carbon nanotubes.

6. The method for optimizing a carbon-based integrated circuit at the atomic level according to claim 5, wherein The step of repairing the in-situ defects of the carbon nanotubes by performing fixed-point irradiation on the carbon nanotubes at the defect position further includes: When irradiating the carbon nanotubes at the defect position with femtosecond laser having the optimal laser operating parameters, part of the laser energy is directly absorbed by the carbon nanotubes, selectively exciting the local vibrations at the defect positions existing in the carbon nanotubes; In the case of the local vibrations, the carbon atoms are rearranged and reconstructed to repair the in-situ defects of the carbon nanotubes, where the defects of the carbon nanotubes include dislocations, defects or fractures.

7. The method for optimizing a carbon-based integrated circuit at the atomic level by laser according to claim 6, wherein When the target position of the residual polymer and the defect position of the carbon nanotubes are the same, by performing fixed-point irradiation on the target position of the residual polymer with femtosecond laser, the laser energy breaks the molecular chains of the residual polymer present through multi-energy photon absorption to remove the residual polymer; at the same time, when the femtosecond laser irradiates the defect position of the carbon nanotubes, the laser energy is absorbed by the carbon nanotubes, exciting the carbon nanotubes to perform local atomic vibrations, causing the defect regions in the carbon nanotubes to be reconstructed, and reordering the carbon atoms to repair the in-situ defects of the carbon nanotubes; When the target position of the residual polymer and the defect position of the carbon nanotubes are different, first perform laser irradiation on the target position of the residual polymer to remove the residual polymer, and then perform laser scanning on the defect position of the carbon nanotubes alone in another path, controlling the laser parameters to excite the local atomic vibrations in the defect regions of the carbon nanotubes to repair the in-situ defects of the carbon nanotubes.

8. The method for optimizing a carbon-based integrated circuit at the atomic level according to claim 6, wherein The step of repeatedly collecting surface topography images of the carbon nanotube wafer after fixed-point irradiation and verifying the remaining amount of the residual polymer and the defect repair status based on the repeatedly collected surface topography images further includes: Re-acquire another surface topography image of the carbon nanotube wafer and compare it with a set of surface topography images before the process to determine the surface damage and reconstruction degree of the carbon nanotube wafer; Measure the surface roughness parameters and elemental composition analysis results to verify the degree of removal of the residual polymer and the degree of surface reconstruction; Measure the capacitance current and carrier mobility of the carbon nanotube wafer to evaluate the electrical performance parameters of the carbon nanotube film and indirectly verify the surface quality of the carbon nanotube film.

9. An atomic-level laser optimization device for a carbon-based integrated circuit, characterized in that Including: An image scanning module for scanning the film surface of the carbon nanotube wafer to obtain a set of surface topography images; An identification module for using machine vision and image processing algorithms to identify the parameter information of the residual polymer on the surface of the carbon nanotube wafer and the defect information of the carbon nanotubes according to the set of surface topography images, wherein the parameter information of the residual polymer includes the target position of the residual polymer and the defect information of the carbon nanotubes includes the defect position; A laser determination module for determining the optimal laser operating parameters of the femtosecond laser based on the parameter information and / or the defect information in combination with the optical property database; A wafer processing module for respectively performing spot irradiation on the residual polymer at the target position and the carbon nanotubes at the defect position, causing the residual polymer to undergo photolysis and be discharged, and repairing the in-situ defects of the carbon nanotubes; and A verification module for repeatedly collecting another set of surface topography images of the carbon nanotube wafer after spot irradiation and verifying the remaining amount of the residual polymer and the defect repair status according to the another set of surface topography images until the residual polymer is completely removed and the defects are repaired.

10. The carbon-based integrated circuit atomic-level laser optimization device according to claim 9, wherein, The parameter information of the residual polymer further includes the quantity of the residual polymer, the area ratio of the residual polymer, the spatial distribution of the residual polymer on the surface of the carbon nanotubes, and the type of the residual polymer, wherein the type of the residual polymer includes the co-solvent, surfactant, cross-linking agent, and / or sensitizer introduced during the preparation of the carbon nanotube wafer; and The defect information of the carbon nanotubes includes the number and size of the defects of the carbon nanotubes.

Images