Tunable vortex superlens design method and device, electronic equipment and medium

Abstract

This invention provides a tunable vortex superlens design method, apparatus, electronic device, and medium. The tunable vortex superlens design method includes: generating a first dataset and a second dataset, wherein the first dataset is a crystalline dataset of Sb₂S₃ phase transition material, and the second dataset is an amorphous dataset of Sb₂S₃ phase transition material; training a deep neural network using the first dataset and the second dataset respectively; predicting the positions of particles in the crystalline and amorphous states using a particle swarm optimization algorithm; and determining the target unit structure based on the prediction results and the target phase. The beneficial effect of this invention is that the tunable characteristics of the lens are achieved by changing the phase transition state of Sb₂S₃.

Description

Technical Field

[0001] This invention relates to the field of power and signal measurement and control technology, and in particular to a tunable vortex superlens design method, device, electronic device and medium. Background Technology

[0002] Tunable vortex superlenses are of great significance in high-resolution imaging, optical communication, and sensing technology due to their precise manipulation of electromagnetic waves.

[0003] Metasurfaces, composed of two-dimensional planar arrays of subwavelength nanostructures, have important applications in various optical systems due to their versatility in manipulating wavefronts. Lenses based on metasurfaces are called superlenses. Unlike traditional spatial light modulators that generate vortex beams, superlenses are widely used in vortex beam generation due to their compactness, high integration, and light weight. For example, focusing vortex beams has been achieved using right-hand circularly polarized (RCP) and left-hand circularly polarized (LCP) incident light. Polarization-sensitive multifocal vortex superlenses have been realized using multiple sublenses. However, these vortex superlenses are only focused at a single focal point. Summary of the Invention

[0004] The main objective of this invention is to propose a tunable vortex superlens design method, device, electronic device, and medium, which improves the accuracy, timeliness, and stability of frequency signal measurement of generators and enhances frequency modulation performance.

[0005] One aspect of the present invention provides a method for designing a tunable vortex superlens, comprising: Based on the design request of the tunable vortex superlens, a first dataset and a second dataset are generated. The first dataset is a crystalline dataset of Sb2S3 phase change material, and the second dataset is an amorphous dataset of Sb2S3 phase change material. The crystalline dataset and the amorphous dataset represent the length and width of the unit structure of the Sb2S3 phase change material. A deep neural network is trained on the first dataset and the second dataset respectively to obtain a first model and a second model. The first model is the mapping relationship between the unit structure and the phase response of crystalline state, and the second model is the mapping relationship between the unit structure and the phase response of amorphous state. The particle swarm optimization algorithm is used to set the positions of crystalline particles and amorphous particles. The phase of crystalline particles is predicted by the first model to obtain a first prediction result. The phase of amorphous particles is predicted by the second model to obtain a second prediction result. The target unit structure is determined based on the first prediction result, the second prediction result, and the target phase.

[0006] According to the tunable vortex superlens design method, generating a first dataset and a second dataset based on a tunable vortex superlens design request includes: The length and width of the Sb₂S₃ phase change material in the crystalline state and the Sb₂S₃ phase change material in the amorphous state were scanned using the finite-difference time-domain method. The range of the length and the range of the width were 0.15 μm to 0.45 μm.

[0007] According to the tunable vortex superlens design method, a deep neural network is trained on the first dataset and the second dataset respectively to obtain a first model and a second model, including: The deep neural network includes an input layer, a hidden layer, and an output layer. The input layer receives the first dataset or the second dataset, the hidden layer learns the mapping relationship between the unit structure and the phase response, and the output layer outputs the electromagnetic response of the unit structure at different wavelengths.

[0008] According to the tunable vortex superlens design method, a particle swarm optimization algorithm is used to set the positions of crystalline and amorphous particles. The phase of the crystalline particles is predicted using a first model to obtain a first prediction result, and the phase of the amorphous particles is predicted using a second model to obtain a second prediction result, including: Initialize the particle population, which includes randomly initializing the particle population according to the population size, number of iterations and particle position constraints. The particle position constraints range from 0.15μm to 0.45μm. Each particle includes two independent variables, which are the length and width of the unit structure. Using at least one of the first model and the second model, the phase of the particle is predicted to obtain the first prediction result and the second prediction result, wherein the first prediction result is the crystalline phase and the second prediction result is the amorphous phase.

[0009] According to the aforementioned tunable vortex superlens design method, the method further includes: Using at least one of the first model and the second model, the predicted phase value is mapped to the unit circle using trigonometric transformation, and the mapping points are continuous planar coordinate points.

[0010] According to the tunable vortex superlens design method, determining the target unit structure based on the first prediction result, the second prediction result, and the target phase includes: The fitness value is calculated using a fitness function based on the first prediction result, the second prediction result, and the target phase, wherein the target phase includes a crystalline target phase and an amorphous target phase. Find the optimal value for each individual and the optimal value for the group from the fitness values ​​and update the velocity and position of the particles. Determine whether the updated velocity and position of the particles meet the termination condition. If the termination condition is met, save the optimal individual.

[0011] According to the tunable vortex superlens design method, the target phase includes: Based on the tunability of the target vortex superlens, the distribution of the amorphous phase is as follows: ; The distribution of crystalline phases is as follows ; Where (x, y) are the coordinates of any position on the superlens, λ is the wavelength of the incident light, and f is the focal length. The topological charge is in the amorphous state. The topological charge is that of the crystalline state.

[0012] Another aspect of the present invention provides a tunable vortex superlens design apparatus, comprising: The first module is used to generate a first dataset and a second dataset according to the design request of the tunable vortex superlens. The first dataset is a crystalline dataset of Sb2S3 phase change material, and the second dataset is an amorphous dataset of Sb2S3 phase change material. The crystalline dataset and the amorphous dataset represent the length and width of the unit structure of the Sb2S3 phase change material. The second module is used to train a deep neural network using the first dataset and the second dataset respectively to obtain a first model and a second model. The first model is the mapping relationship between the unit structure and the phase response of crystalline state, and the second model is the mapping relationship between the unit structure and the phase response of amorphous state. The third module is used to set the positions of crystalline particles and amorphous particles using a particle swarm optimization algorithm, predict the phase of crystalline particles using the first model to obtain a first prediction result, and predict the phase of amorphous particles using the second model to obtain a second prediction result. The fourth module is used to determine the target unit structure based on the first prediction result, the second prediction result, and the target phase.

[0013] Another aspect of the present invention provides an electronic device, including a processor and a memory; The memory is used to store programs; The processor executes the program to implement the method as described above.

[0014] This invention also discloses a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device can read the computer instructions from the computer-readable storage medium and execute the computer instructions, causing the computer device to perform the methods described above.

[0015] The beneficial effects of this invention are as follows: By combining DNN (Deep Neural Network) with PSO (Particle Swarm Optimization), the efficient and accurate prediction capability of DNN is used to predict the phase value corresponding to the unit structure, and then PSO is used to optimize the structural parameters of the unit structure; by irradiating the low-loss phase change material Sb2S3 in the near-infrared band, the tunable characteristics of the lens (tunable focal length and tunable topological charge) are achieved by changing the phase transition state of Sb2S3.

[0016] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0017] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which: Figure 1 This is a schematic diagram of the tunable vortex superlens according to an embodiment of the present invention.

[0018] Figure 2 This is a schematic diagram of the unit structure and phase change material Sb2S3 according to an embodiment of the present invention.

[0019] Figure 3 This is a flowchart illustrating the design method of a tunable vortex superlens according to an embodiment of the present invention.

[0020] Figure 4 This is a schematic diagram of DNN model construction and phase processing according to an embodiment of the present invention.

[0021] Figure 5 This is a schematic diagram of the prediction process of the particle swarm algorithm according to an embodiment of the present invention.

[0022] Figure 6 This is a comparison chart of training error and test error of the DNN training results in an embodiment of the present invention.

[0023] Figure 7 This is a phase distribution and structural arrangement diagram of a topologically charge-tunable vortex superlens according to an embodiment of the present invention.

[0024] Figure 8 This is a simulation result diagram of a topologically charge-tunable vortex lens according to an embodiment of the present invention.

[0025] Figure 9 This is a phase distribution and structural arrangement diagram of a focal length-tunable vortex superlens according to an embodiment of the present invention.

[0026] Figure 10 This is a simulation result of a focal length-tunable vortex lens according to an embodiment of the present invention.

[0027] Figure 11 This is a diagram of the tunable vortex superlens design and analysis device according to an embodiment of the present invention. Detailed Implementation

[0028] The embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings. Throughout the description, the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions. In the following description, suffixes such as "module," "part," or "unit" used to denote elements are used only for the purpose of illustrative purposes and have no specific meaning in themselves. Therefore, "module," "part," or "unit" can be used interchangeably. Terms such as "first," "second," etc., are used only to distinguish technical features and should not be construed as indicating or implying relative importance, or implicitly indicating the number of indicated technical features, or implicitly indicating the sequential relationship of the indicated technical features. In the following description, the consecutive reference numerals for method steps are for ease of review and understanding. Adjusting the implementation order of steps, in conjunction with the overall technical solution of the present invention and the logical relationship between the various steps, will not affect the technical effect achieved by the technical solution of the present invention. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0029] Reference Figure 1 The schematic diagram of the tunable vortex superlens shown illustrates a superlens composed of Sb₂S₃ nanopillars, an ITO layer, and a silicon dioxide (SiO₂) substrate, with right-handed circularly polarized light as the incident light. By adjusting the phase transition state of Sb₂S₃, this embodiment of the invention achieves the tunable characteristics of the superlens. When the unit structure of Sb₂S₃ exhibits the amorphous state (left side), it is in an amorphous state; while when the unit structure exhibits the crystalline yellow state (right side), it is in a crystalline state, thus creating two different focal points. Therefore, by simply controlling the phase transition state of Sb₂S₃, the focusing capability of the vortex superlens can be freely adjusted.

[0030] In some embodiments, reference Figure 2 A schematic diagram of the unit structure and phase change material Sb₂S₃, in which... Figure 2 (a) shows the unit cell structure of Sb2S3 in the amorphous state. Figure 2(b) shows the unit cell structure of Sb₂S₃ in its crystalline state. The unit cell structures for both phase transition states have the same structural parameters. Red (first layer, Sb₂S₃) represents amorphous Sb₂S₃ material, yellow (second layer, C-Sb₂S₃) represents crystalline Sb₂S₃ material, blue (third layer, ITO) represents ITO material, and gray (fourth layer, SiO₂) represents SiO₂ material. ITO material is conductive; therefore, by applying an appropriate current pulse (~0.3V), temperature regulation can be performed rapidly, thereby reversibly regulating the phase transition state of Sb₂S₃. P is the period of the unit cell structure, with a value of 0.5 μm. To ensure the phase of the unit cell structure satisfies... The height H of the Sb₂S₃ unit structure is set to 1 μm. L and W are the length and width of the Sb₂S₃ unit structure. The electromagnetic response can be changed by adjusting L and W of the unit structure to satisfy the phase distribution. Figure 2 Figures (c) and (d) show the refractive index and extinction coefficient curves of Sb₂S₃ in crystalline and amorphous states. It can be seen from the figures that Sb₂S₃ has low loss in the 1 μm-1.5 μm wavelength range (k=10). -5 Furthermore, the significant difference in refractive index between the two states is beneficial for achieving the tunable characteristics of vortex lenses.

[0031] refer to Figure 3 A flowchart illustrating the design method for a tunable vortex superlens, including but not limited to steps S100~S400: S100, based on the design request of the tunable vortex superlens, generates a first dataset and a second dataset. The first dataset is the crystalline dataset of Sb2S3 phase change material, and the second dataset is the amorphous dataset of Sb2S3 phase change material. The crystalline dataset and the amorphous dataset represent the length and width of the unit structure of Sb2S3 phase change material.

[0032] In some embodiments, the length and width of the Sb2S3 phase change material in the crystalline state and the Sb2S3 phase change material in the amorphous state are scanned using the finite-difference time-domain method, with the length and width ranging from 0.15 μm to 0.45 μm.

[0033] S200 uses a deep neural network trained on the first dataset and the second dataset to obtain a first model and a second model. The first model is the mapping relationship between the unit structure and phase response of crystalline state, and the second model is the mapping relationship between the unit structure and phase response of amorphous state.

[0034] In some embodiments, a deep neural network includes an input layer, a hidden layer, and an output layer. The input layer receives a first dataset or a second dataset, the hidden layer learns the mapping relationship between the unit structure and the phase response, and the output layer outputs the electromagnetic response of the unit structure at different wavelengths.

[0035] For example, refer to Figure 4 The DNN model construction and phase processing shown are as follows: Figure 4 (a) Fully connected neural network model Figure 4 (b) Example of phase transition curves for phase change materials in crystalline and amorphous states, with the red circles indicating the phase transition locations. Figure 4 (c) Preprocess the phase to convert it into a projection of xy.

[0036] This embodiment uses a fully connected deep neural network to predict the electromagnetic response of the unit cell structure. The schematic diagram of the fully connected neural network model is shown below. Figure 4 As shown in (a), the system consists of one input layer, seven hidden layers, and one output layer. The input layer comprises the L and W units of the unit structure, and the output layer represents the electromagnetic responses at 26 different wavelengths. The structural parameters of the input unit structure can be converted into the corresponding electromagnetic responses. To ensure sufficient learning by the neural network, this embodiment establishes seven hidden layers, each composed of 100, 300, 600, 600, 600, 300, and 100 neurons, respectively.

[0037] S300 uses a particle swarm optimization algorithm to set the positions of crystalline and amorphous particles. It uses a first model to predict the phase of crystalline particles and obtains a first prediction result. It uses a second model to predict the phase of amorphous particles and obtains a second prediction result.

[0038] In some embodiments, referring to the particle swarm optimization algorithm prediction flowchart shown in Figure 5, it includes, but is not limited to, steps S310-S320: S310, Initialize the particle population, including random initialization of the particle population according to the population size, number of iterations and particle position constraints, wherein the particle position constraints range from 0.15μm to 0.45μm, and each particle includes two independent variables, namely the length and width of the unit structure; S320, using at least one of the first model and the second model, the phase of the particle is predicted to obtain a first prediction result and a second prediction result, wherein the first prediction result is a crystalline phase and the second prediction result is an amorphous phase.

[0039] In some embodiments, reference Figure 4(b) illustrates the phase jump problem that occurs when the phase change material is in a crystalline or amorphous state during unit structure simulation. As can be seen from the figure, a phase discontinuity appears in the 1μm-1.5μm working wavelength range. This abrupt phase change severely interferes with the prediction of the DNN, and since the target fitting method of this embodiment only requires phase values, this embodiment solves the phase jump problem by preprocessing the phase values. This embodiment can map each phase value onto the unit circle, such as... Figure 4 As shown in (c), the phase value is transformed into a continuous pair of xy values ​​through trigonometric transformation, thereby achieving stable prediction of the network.

[0040] In some embodiments, reference Figure 6 The training and testing errors of the amorphous and crystalline neural networks in the DNN training results shown are as follows: Figure 6 (a) and Figure 6 (b) represents the training and testing errors of amorphous and crystalline neural networks. Figure 6 (c), Figure 6 (d) shows examples of training results for crystalline DNNs with L=0.4, W=0.365 and L=0.3, W=0.185. Figure 6 (e), Figure 6 (f) shows examples of training results for amorphous DNNs with L=0.16, W=0.225 and L=0.25, W=0.36. Figure 6 (c)- Figure 6 In (f), the dashed line represents the phase curve predicted by the DNN, and the solid line represents the simulation result.

[0041] The training results of DNN are as follows Figure 6 As shown in (a) and (b), after 300 iterations, the mean squared error of the amorphous neural network is 0.056, and the mean squared error of the crystalline neural network is 0.031, both tending towards a stable value. To more clearly demonstrate the prediction performance of the DNN, Figure 6 (c)-(f) show examples of phase curves randomly selected from the test dataset, where the red dashed line and the black solid line represent the phase prediction curve of the DNN and the phase simulation curve of the FDTD, respectively. Figure 6 (c) and (d) show the training results of the crystalline DNN with L=0.4, W=0.365 and L=0.3, W=0.185. Figure 6 Figures (e) and (f) show the training results of amorphous DNN with L=0.16, W=0.225 and L=0.25, W=0.36. As can be seen from the embodiments of the present invention in the above figures, the predicted phase curves of crystalline and amorphous DNNs are in good agreement with the simulated phase curves of FDTD, and the trained DNN can predict the corresponding phase value of each unit structure in just a few milliseconds.

[0042] S400, determine the target unit structure based on the first prediction result, the second prediction result, and the target phase. In some embodiments, the fitness value is calculated using the predicted results and the actual phase required by the tunable superlens. The fitness function f is expressed as f = abs(TrueA - PredictA) + abs(TrueC - PredictC), where TrueA and TrueC represent the target phases required for the amorphous and crystalline states in the tunable vortex lens, respectively, and PredictA and PredictC represent the amorphous and crystalline phases predicted by the DNN, respectively. Individual and swarm optimal values ​​are found, and the particle velocity and position are updated. Finally, it is determined whether the termination condition has been met. If the termination condition is met, the optimal individual is saved. Here, "individual" and "swarm" refer to a single particle and a swarm of particles, respectively.

[0043] In some embodiments, the present invention provides two tunable vortex superlenses, namely, tunable focal length and tunable topological charge.

[0044] To achieve the tunability of the vortex superlens, the lens needs to satisfy the following phase distribution, with the phase distribution corresponding to the amorphous state being: (1) The phase distribution corresponding to the crystalline state is as follows: (2) Where (x, y) are the coordinates of any position on the superlens, λ is the incident light wavelength, set to 1.064 μm, f is the focal length, and the focal point coordinates are (0, 0). For a topologically tunable superlens, in this embodiment of the invention, the topological charge of the vortex lens in crystalline and amorphous states is set to... =1 and =2, and the focal length is set to 30μm. According to formulas (1) and (2), the embodiments of the present invention design tunable vortex superlenses with different topological charges. For example, when Sb2S3 is amorphous, it generates focused vortex light with a topological charge of 1; when Sb2S3 is crystalline, it generates focused vortex light with a topological charge of 2.

[0045] In some embodiments, reference Figure 7 The diagram shows the phase distribution and structural arrangement of a topologically charge-tunable vortex superlens, in which... Figure 7 (a) and Figure 7 (b) shows the phase distribution of the lens along the x-axis when Sb2S3 is in an amorphous state and a crystalline state, respectively. Figure 7 (c) shows the structural arrangement of the superlens. Red (top) represents the lens structure arrangement in the amorphous state, and yellow (bottom) represents the lens structure arrangement in the crystalline state. Figure 7 (a) and (b) describe the phase distribution of the lens along the x-axis when Sb2S3 is in the amorphous and crystalline states, where the solid line represents the actual target phase and the pentagram represents the DNN predicted phase. It can be seen that the predicted phase and the actual phase fit well. Figure 7 (c) describes the geometric arrangement of the required phase of the superlens at each position. Red represents the structural arrangement of the lens when Sb₂S₃ is amorphous, and yellow represents the structural arrangement of the lens when Sb₂S₃ is crystalline. As can be seen from the figure, the crystalline and amorphous states have the same structural distribution. Here, in this embodiment of the invention, the tunability of the topological charge can be achieved simply by controlling the phase transition state of Sb₂S₃.

[0046] In some embodiments, reference Figure 8 The simulation results of a topologically charged tunable vortex lens are shown in the figure. Figure 8 (a) and Figure 8 (d)xy are the focused vortex light field diagrams of topological charges 1 and 2 in the plane, respectively. Figure 8 (b) and Figure 8 (e)xz are focused vortex light field diagrams for topological charges 1 and 2 in the plane, respectively. Figure 8 (c) and Figure 8 (f) Enlarged view of the corresponding Ex phase distribution of the focused optical vortex in the xy plane. Figure 8 (a) and (d) are the light field diagrams on the xy plane when the lens is in the amorphous and crystalline states, respectively. It can be seen from the figures that the vortex beam has a good focusing effect on the xy plane, forming a circular annular aperture. It can be clearly observed that as the topological charge increases, a larger optical vortex ring is generated, which is consistent with the theory. Figure 8 Figures (b) and (e) show the light field diagrams of the superlens in the longitudinal plane (y=0) and the corresponding scan lines along the x-axis when Sb₂S₃ is in the amorphous and crystalline states. As can be seen from the figures, the superlens can be focused onto a specific focal plane with a focal length of 30 μm, which is the same as the initial setting value in the embodiment of the present invention. Some side lobes appear near the focal point, which may be due to the mutual influence of the unit structures. Figure 8 (c) and (f) show magnified views of the Ex phase distribution of the optical vortex in the xy plane in the amorphous and crystalline states of the lens. It can be seen from the figures that within the vortex optical field, the phase of Ex varies along the circumferential direction for a fixed period, and the number of this period corresponds to the topological charge.

[0047] In some embodiments, reference Figure 9 Phase distribution and structural arrangement diagram of a focal length-tunable vortex superlens. Figure 9 (a) and Figure 9(b)Sb2S3 represents the phase distribution of the lens along the x-axis in the amorphous and crystalline states, respectively. Figure 9 (c) The structural arrangement of the superlens, with red (top) representing the lens structure arrangement in the amorphous state and yellow (bottom) representing the lens structure arrangement in the crystalline state. According to formulas (1) and (2), this embodiment of the invention designs a focal length-adjustable vortex superlens with a topological charge of 1. That is, when Sb₂S₃ is amorphous, the vortex light is focused at 20 μm; when Sb₂S₃ is crystalline, the vortex light is focused at 25 μm. Figure 9 Figures (a) and (b) describe the phase distribution of the lens along the x-axis when Sb2S3 is in amorphous and crystalline states. The solid blue line represents the actual target phase, and the pentagram represents the DNN predicted phase. As can be seen from the figures, the predicted phase and the actual phase fit well. Figure 9 (c) describes the geometric arrangement of the phase required by the superlens at each position. Red represents the structural arrangement of the lens when Sb2S3 is amorphous, and yellow represents the structural arrangement of the lens when Sb2S3 is crystalline. As can be seen from the figure, the crystalline and amorphous states have the same structural distribution. In this embodiment of the invention, the focal length can be adjusted simply by adjusting the phase transition state of the phase change material without changing the structure.

[0048] In some embodiments, reference Figure 10 Simulation results of a focal length-tunable vortex lens. Figure 10 (a) and Figure 10 (d)xy is the focused vortex light field diagram with focal lengths of 20μm and 25μm in the plane. Figure 10 (b) and Figure 10 (e) shows the focused vortex light field diagrams with focal lengths of 20 μm and 25 μm in the xz plane. Figure 10 (c) and Figure 10 (f) Normalized intensity distribution along the x-axis in the focal planes with focal lengths of 20 μm and 25 μm. Figure 10 (a) and Figure 10 (b) The light field diagrams in the xy plane are shown for the superlens in both amorphous and crystalline states. The results show that the vortex beam has a good focusing effect in the xy plane, forming a ring-shaped aperture in both cases. Figure 10 (b) and Figure 10 (e) shows the light field diagrams of the superlens in the longitudinal plane (y=0) and the corresponding scan lines along the x-axis when Sb2S3 is in an amorphous and crystalline state, respectively. As can be seen from the figure, when the superlens is in a crystalline state, the focal length obtained by simulation is 25 μm; when the superlens is in an amorphous state, the focal length obtained is 20 μm, which is the same as the initial setting value of the embodiment of the present invention. Figure 10 (c) and Figure 10(f) shows the normalized intensity distribution along the x-axis in the focal plane when Sb2S3 is in the amorphous and crystalline states, respectively. The simulation results show that the intensity distribution of the generated vortex light is not uniform. This is mainly due to the mutual influence of the unit structures and the different light transmittance of unit structures of different sizes. This will be solved in the future with the continuous development of phase change materials.

[0049] Figure 11 This is a diagram of a tunable vortex superlens design and analysis device according to an embodiment of the present invention. The device includes a first module 1110, a second module 1120, a third module 1130, and a fourth module 1140.

[0050] The system comprises the following modules: A first module generates a first dataset and a second dataset based on the design request for a tunable vortex superlens. The first dataset represents the crystalline state dataset of Sb₂S₃ phase change material, and the second dataset represents the amorphous state dataset. The crystalline and amorphous datasets represent the length and width of the unit cell structure of the Sb₂S₃ phase change material, respectively. A second module trains a deep neural network using the first and second datasets to obtain a first model and a second model. The first model represents the mapping relationship between the crystalline unit cell structure and the phase response, and the second model represents the mapping relationship between the amorphous unit cell structure and the phase response. A third module uses a particle swarm optimization algorithm to set the positions of crystalline and amorphous particles. The first model predicts the phase of the crystalline particles to obtain a first prediction result, and the second model predicts the phase of the amorphous particles to obtain a second prediction result. A fourth module determines the target unit cell structure based on the first prediction result, the second prediction result, and the target phase.

[0051] Exemplarily, with the cooperation of the first, second, third, and fourth modules in the device, the embodiment device can implement any of the aforementioned tunable vortex superlens design methods. Specifically, the first module is used to generate a first dataset and a second dataset based on the tunable vortex superlens design request. The first dataset is a crystalline dataset of Sb₂S₃ phase change material, and the second dataset is an amorphous dataset of Sb₂S₃ phase change material. The crystalline and amorphous datasets represent the length and width of the unit structure of the Sb₂S₃ phase change material, respectively. The second module is used to employ a deep neural network to generate the first dataset... The system trains on the second dataset to obtain a first model and a second model. The first model represents the mapping relationship between the crystalline unit structure and the phase response, and the second model represents the mapping relationship between the amorphous unit structure and the phase response. The third module is used to set the positions of crystalline and amorphous particles using a particle swarm optimization algorithm. The first model is used to predict the phase of crystalline particles to obtain a first prediction result, and the second model is used to predict the phase of amorphous particles to obtain a second prediction result. The fourth module is used to determine the target unit structure based on the first prediction result, the second prediction result, and the target phase.

[0052] The beneficial effects of this invention are as follows: By combining DNN (Deep Neural Network) with PSO (Particle Swarm Optimization), the efficient and accurate prediction capability of DNN is used to predict the phase value corresponding to the unit structure, and then PSO is used to optimize the structural parameters of the unit structure; by irradiating the low-loss phase change material Sb2S3 in the near-infrared band, the tunable characteristics of the lens (tunable focal length and tunable topological charge) are achieved by changing the phase transition state of Sb2S3.

[0053] This invention also provides an electronic device, which includes a processor and a memory; The memory stores the program; The processor executes a program to perform the aforementioned tunable vortex superlens design method; the electronic device has the function of carrying and running a software system for the tunable vortex superlens design provided in the embodiments of the present invention, such as a computer, a speed controller, and a speed control device.

[0054] This invention also provides a computer-readable storage medium storing a program that is executed by a processor to implement the tunable vortex superlens design method described above.

[0055] In some alternative embodiments, the functions / operations mentioned in the block diagrams may not occur in the order shown in the operation diagrams. For example, depending on the functions / operations involved, two consecutively shown blocks may actually be executed substantially simultaneously, or the blocks may sometimes be executed in reverse order. Furthermore, the embodiments presented and described in the flowcharts of this invention are provided by way of example to provide a more comprehensive understanding of the technology. The disclosed methods are not limited to the operations and logic flows presented herein. Alternative embodiments are contemplated in which the order of various operations is altered and sub-operations described as part of a larger operation are executed independently.

[0056] This invention also discloses a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device can read the computer instructions from the computer-readable storage medium and execute the computer instructions, causing the computer device to perform the aforementioned tunable vortex superlens design method.

[0057] Furthermore, although the invention has been described in the context of functional modules, it should be understood that, unless otherwise stated, one or more of the described functions and / or features may be integrated into a single physical device and / or software module, or one or more functions and / or features may be implemented in a separate physical device or software module. It is also understood that a detailed discussion of the actual implementation of each module is unnecessary for understanding the invention. Rather, given the properties, functions, and internal relationships of the various functional modules in the apparatus disclosed herein, the actual implementation of the module will be understood within the scope of conventional skill of an engineer. Therefore, those skilled in the art can implement the invention as set forth in the claims using ordinary techniques without excessive experimentation. It is also understood that the specific concepts disclosed are merely illustrative and not intended to limit the scope of the invention, which is determined by the full scope of the appended claims and their equivalents.

[0058] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0059] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-including system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0060] More specific examples of computer-readable media (a non-exhaustive list) include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0061] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0062] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0063] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

[0064] The above is a detailed description of the preferred embodiments of the present invention, but the present invention is not limited to the embodiments described. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of the present invention, and these equivalent modifications or substitutions are all included within the scope defined by the claims of this application.

Claims

1. A method of designing a tunable vortex superlens, characterized in that, include: Based on the design request of the tunable vortex superlens, a first dataset and a second dataset are generated, including: scanning the length and width of Sb2S3 phase change material in a crystalline state and an amorphous state using the finite-difference time-domain method, wherein the length and width range from 0.15 μm to 0.45 μm. The first dataset is a crystalline dataset of Sb2S3 phase change material, and the second dataset is an amorphous dataset of Sb2S3 phase change material. The crystalline dataset and the amorphous dataset represent the length and width of the unit cell structure of the Sb2S3 phase change material. A deep neural network is trained on the first dataset and the second dataset respectively to obtain a first model and a second model. The first model is the mapping relationship between the unit structure and the phase response of crystalline state, and the second model is the mapping relationship between the unit structure and the phase response of amorphous state. The particle swarm optimization algorithm is used to set the positions of crystalline particles and amorphous particles. The phase of crystalline particles is predicted by the first model to obtain a first prediction result. The phase of amorphous particles is predicted by the second model to obtain a second prediction result. The target unit structure is determined based on the first prediction result, the second prediction result, and the target phase, wherein the target phase includes: Based on the tunability of the target vortex superlens, the distribution of the amorphous phase is as follows: ； The distribution of crystalline phases is as follows ； Where (x, y) are the coordinates of any position on the superlens, λ is the wavelength of the incident light, and f is the focal length. The topological charge is in the amorphous state. The topological charge is that of the crystalline state.

2. The method of designing a tunable vortex superlens according to claim 1, wherein, The step of training a deep neural network using the first dataset and the second dataset respectively to obtain a first model and a second model includes: The deep neural network includes an input layer, a hidden layer, and an output layer. The input layer receives the first dataset or the second dataset, the hidden layer learns the mapping relationship between the unit structure and the phase response, and the output layer outputs the electromagnetic response of the unit structure at different wavelengths.

3. The method of claim 2, wherein, The process involves using a particle swarm optimization algorithm to set the positions of crystalline and amorphous particles, predicting the phase of crystalline particles using a first model to obtain a first prediction result, and predicting the phase of amorphous particles using a second model to obtain a second prediction result, including: Initialize the particle population, which includes randomly initializing the particle population according to the population size, number of iterations and particle position constraints. The particle position constraints range from 0.15μm to 0.45μm. Each particle includes two independent variables, which are the length and width of the unit structure. Using at least one of the first model and the second model, the phase of the particle is predicted to obtain the first prediction result and the second prediction result, wherein the first prediction result is the crystalline phase and the second prediction result is the amorphous phase.

4. The method of designing a tunable vortex superlens according to claim 3, wherein, The method further includes: Using at least one of the first model and the second model, the predicted phase value is mapped to the unit circle using trigonometric transformation, and the mapping points are continuous planar coordinate points.

5. The method of claim 2, wherein, The step of determining the target unit structure based on the first prediction result, the second prediction result, and the target phase includes: The fitness value is calculated using a fitness function based on the first prediction result, the second prediction result, and the target phase, wherein the target phase includes a crystalline target phase and an amorphous target phase. Find the optimal value for each individual and the optimal value for the group from the fitness values ​​and update the velocity and position of the particles. Determine whether the updated velocity and position of the particles meet the termination condition. If the termination condition is met, save the optimal individual.

6. A tunable vortex superlens design apparatus, characterized by, include: The first module is used to generate a first dataset and a second dataset according to the design request of the tunable vortex superlens. This includes scanning the length and width of the Sb₂S₃ phase change material in a crystalline state and an amorphous state using the finite-difference time-domain method. The length and width range from 0.15 μm to 0.45 μm. The first dataset is a crystalline dataset of the Sb₂S₃ phase change material, and the second dataset is an amorphous dataset of the Sb₂S₃ phase change material. The crystalline dataset and the amorphous dataset represent the length and width of the unit cell structure of the Sb₂S₃ phase change material. The second module is used to train a deep neural network with the first dataset and the second dataset respectively to obtain a first model and a second model. The first model is the mapping relationship between the unit structure and the phase response of crystalline state, and the second model is the mapping relationship between the unit structure and the phase response of amorphous state. The third module is used to set the positions of crystalline particles and amorphous particles using a particle swarm optimization algorithm, predict the phase of crystalline particles using the first model to obtain a first prediction result, and predict the phase of amorphous particles using the second model to obtain a second prediction result. The fourth module is used to determine the target unit structure based on the first prediction result, the second prediction result, and the target phase, wherein the target phase includes: Based on the tunability of the target vortex superlens, the distribution of the amorphous phase is as follows: ； The distribution of crystalline phases is as follows ； Where (x, y) are the coordinates of any position on the superlens, λ is the wavelength of the incident light, and f is the focal length. The topological charge is in the amorphous state. The topological charge is that of the crystalline state.

7. An electronic device, comprising: Including the processor and memory; The memory is used to store programs; The processor executes the program to implement the tunable vortex superlens design method as described in any one of claims 1-5.

8. A computer-readable storage medium, characterized in that, The storage medium stores a program, which is executed by a processor to implement the tunable vortex superlens design method as described in any one of claims 1-5.

Images