CN-122026118-A - Tunable terahertz generation method for generating countermeasure network and topological super-constructed surface based on physical driving

CN122026118ACN 122026118 ACN122026118 ACN 122026118ACN-122026118-A

Abstract

The application belongs to the technical field of terahertz, and particularly relates to a tunable terahertz generation method for generating an antagonistic network and a topological super-structured surface based on physical driving, which comprises the steps of constructing a heterogeneous integrated topological super-structured surface unit containing a thin film lithium niobate, an arrayed nano antenna, tungsten disulfide and a graphene layer; generating an countermeasure network by using physical driving, and reversely designing a topological structure sequence meeting wave vector matching conditions by introducing Maxwell's equation set constraint in a loss function; the application realizes the high-efficiency conversion and ultra-fast dynamic tuning of the terahertz full-frequency band by the deep fusion of an AI algorithm and the topological super-structured surface, remarkably improves the output power and the beam quality and ensures the physical feasibility of a design result.

Inventors

Qi Sitong
LEI JIAN
Yang Gaozinan
ZHANG ZIMO

Assignees

商洛学院

Dates

Publication Date: 20260512
Application Date: 20260331

Claims (10)

1. A tunable terahertz generation method for generating a countermeasure network and a topological super-structured surface based on physical driving, comprising the following steps: s1, constructing a heterogeneous integrated topological super-structured surface unit, Arranging microstructures distributed in an array manner on a cutting film lithium niobate substrate with a preset thickness, covering a transition metal chalcogenide thin layer on the surface of the microstructure, integrating a graphene layer in the heterogeneous integrated topological super-structured surface unit, and realizing dynamic control of terahertz wave amplitude and phase by regulating and controlling the charge concentration of the graphene layer; s2, generating a countermeasure network based on physical driving to carry out topological structure reverse design, Acquiring a data set containing geometric parameters, transmission spectrum, phase and nonlinear conversion efficiency of the microstructure by using an electromagnetic simulation algorithm, constructing a generation countermeasure network consisting of a generator and a discriminator, introducing a Maxwell equation set as a physical rule constraint into a loss function of the generation countermeasure network, and searching a topological structure sequence which enables the wave vector mismatch quantity to be towards a preset minimum value in a preset frequency range by using an artificial intelligence algorithm; s3, realizing high-power terahertz radiation output, And using two infrared lasers with wavelengths in a preset infrared band as a pumping source, exciting the heterogeneous integrated topological super-structure surface unit through a difference frequency generation effect, and enhancing the local field strength by utilizing a continuous domain binding state effect generated by the heterogeneous integrated topological super-structure surface unit, thereby generating terahertz waves with average power reaching a preset power threshold and continuously adjustable frequency in a preset frequency range.
2. The method for generating tunable terahertz based on physical driving to combat network and topology super-structured surface according to claim 1, wherein in S1, the thin film lithium niobate substrate is used as a core supporting layer of a nonlinear gain medium, and an interaction space is provided for the difference frequency generation process by using an electro-optical effect and nonlinear optical coefficient of the thin film lithium niobate substrate; The thickness of the thin film lithium niobate substrate is set to be a preset thickness value, a low refractive index buffer layer is arranged below the thin film lithium niobate substrate and is arranged above a high-resistance substrate, so that a high-refractive index difference waveguide system is formed, and local constraint of a pumping light mode is realized.
3. The method for generating tunable terahertz against a network and a topological super-structured surface based on physical driving according to claim 1, wherein in S1, the microstructure adopts arrayed gold nano-antennas, and the geometric parameters of the microstructure include a unit period, a microstructure height, an arm length of the microstructure, an included angle of two arms and an overall rotation angle; The unit period is set within a preset period range, and the microstructure height is set to a preset height value; the geometrical phase is introduced into the space by changing the included angle between the two arms and the integral rotation angle, so that the terahertz wave radiation wavefront is controlled; the microstructure is constructed by defining a pattern on the surface of the thin film lithium niobate substrate and depositing a metal layer.
4. The method for generating tunable terahertz based on a physical driving method for generating a resistive network and a topological super-structured surface according to claim 1, wherein in S1, the transition metal chalcogenide thin layer adopts a single-layer tungsten disulfide, the single-layer tungsten disulfide covers the microstructure surface, and a synergistic enhancement is formed by utilizing the nonlinear polarization rate possessed by the single-layer tungsten disulfide and the bulk nonlinear effect of the thin film lithium niobate substrate, so that nonlinear frequency conversion efficiency under a sub-wavelength scale is improved; The graphene layer is integrated in the inner part or the surface layer of the heterogeneous integrated topological super-structured surface unit, and the Fermi level of the graphene layer is changed by arranging a contact electrode on the microstructure and applying external bias voltage.
5. The method for generating tunable terahertz wave based on physical driving to combat network and topology super-constructed surfaces according to claim 4, The fermi level adjustment process of the graphene layer utilizes the physical characteristic that the conductivity of the graphene layer in a terahertz frequency band changes along with the concentration of carriers to realize dynamic adjustment and control of the intensity and the phase of the output terahertz wave in a preset response time level; The response time is set to be of the order of microseconds so as to adapt to the requirement of quick dynamic regulation; the graphene layer also maintains the thermal stability of the heterogeneous integrated topological super-structured surface unit by absorbing transient heat energy and utilizing the relaxation characteristic of carriers, and inhibits the nonlinear saturation effect under high-power pumping.
6. The tunable terahertz generation method based on physical driving to generate an countermeasure network and a topologically super-structured surface according to claim 1, wherein in S2, the design of the generated countermeasure network adopts a physical enhancement deep learning architecture, and an electromagnetic field physical law is embedded into the optimization process of the generated countermeasure network; The input end of the generator is configured to be a target working frequency and preset conversion efficiency, and the output end of the generator is configured to be a geometric parameter of the microstructure; The generator adopts a depth residual error network structure, and realizes nonlinear mapping from a target performance space to a physical structure space through a jump connection structure comprising a plurality of residual error blocks; The discriminator adopts a multi-layer perceptron structure for evaluating the physical feasibility of the geometric parameters output by the generator.
7. The method for generating tunable terahertz against network and topology super-constructed surfaces based on physical driving according to claim 1, wherein in S2, the physical rule constraint in the loss function is implemented by converting the electric field rotation equation in the maxwell' S equation set into a differential form and introducing a penalty term; The specific components of the physical rule constraint comprise norms of differences between rotation items of electric field intensities and products of angular frequencies and magnetic field intensities; In the iterative optimization process of the generating countermeasure network, geometric parameters which the microstructure should have at each spatial position are calculated through a global optimizing algorithm to form a non-uniform spatial distribution sequence, so that continuous phase compensation in a preset frequency band is realized.
8. The method for generating tunable terahertz based on physical driving to combat network and topology super-structured surfaces according to claim 1, wherein in S2, the acquisition process of the dataset includes creating a simulation model in a three-dimensional cartesian coordinate system, performing full-wave electromagnetic simulation on a plurality of groups of super-structured surfaces with different topology morphologies by a time domain finite difference algorithm; In the simulation process, setting a preset grid size to capture local field enhancement details of the microstructure edge, and extracting electric field distribution, phase mutation and nonlinear polarization intensity in a far field and a near field; The data set covers electromagnetic response characteristics within a predetermined frequency band, providing training samples for the generation of the countermeasure network.
9. The method for generating the tunable terahertz based on the physical driving and the countering network and the topological super-structured surface according to claim 1, wherein in S3, the two infrared lasers with the wavelengths in the preset infrared band are subjected to power boost through an optical fiber amplifier, and the frequency difference of the infrared lasers corresponds to the terahertz frequency of the target output; scanning of the output frequency of the terahertz waves in a preset frequency range is achieved by adjusting the central wavelength of at least one beam of infrared laser in the pumping source; The scanning process comprises step scanning or continuous scanning, wherein the average power of the terahertz waves is controlled by adjusting the input power of the infrared laser, and the output intensity is ensured to follow the square law growth relation of the pumping power.
10. The method for generating tunable terahertz based on physical driving to combat network and topology super-structured surface according to claim 1, wherein the continuous domain confinement state effect is achieved by adjusting the degree of symmetry breaking of the microstructure, introducing topology singular points at specific momentum space points, enabling a radiation mode to enter a confinement state of a high quality factor, and enabling a local field enhancement factor to reach a preset enhancement threshold; the method further comprises the steps of: and sampling the generated terahertz waves by using a real-time stability monitoring module, and adjusting the phase deviation of the pumping source in real time through a feedback control loop to ensure that the power fluctuation is smaller than a preset fluctuation threshold value within a preset working time, wherein the far-field distribution of the terahertz waves presents a highly concentrated Gaussian profile, and the beam quality factor is smaller than a preset quality factor threshold value.

Description

Tunable terahertz generation method for generating countermeasure network and topological super-constructed surface based on physical driving Technical Field The application belongs to the technical field of terahertz, and particularly relates to a tunable terahertz generation method for generating an antagonistic network and a topological super-constructed surface based on physical driving. Background Terahertz (THz) technology is used as a key field for connecting microwaves and infrared light in electromagnetic spectrum, and plays an increasingly important role in high-speed wireless communication, precise spectrum detection, nondestructive imaging and other front-end technologies. The high-performance terahertz source is used as a core component of a terahertz system, and the performance of the terahertz source directly determines the detection distance, imaging resolution and data transmission rate of the whole system. Along with the fusion development of metamaterial and micro-nano optoelectronics, the realization of wideband, high-power and tunable terahertz radiation by utilizing a nonlinear optical effect has become a research hotspot and an important development direction in the current photoelectric technology field. The difference frequency generation technology based on the nonlinear super-structured surface provides a new path for generating terahertz radiation by manipulating the interaction of an optical field and a substance on a micro-nano scale. The technology generally utilizes a nano antenna array or a heterogeneous integrated material system with a specific geometric shape to generate terahertz waves through a second-order nonlinear polarization effect under the excitation of pump light. By adjusting the structural parameters of the super-structured surface or introducing external electrical regulation, the output terahertz wave frequency, phase and amplitude can be flexibly controlled. However, conventional terahertz generation schemes still face serious challenges in terms of both high power output and wide-band continuous tuning. The existing nonlinear crystal is limited by the intrinsic absorption loss and phase velocity mismatch of the material, and is difficult to maintain high-efficiency conversion within an extremely wide range of 0.1 to 10THz, so that the output power is generally difficult to reach the order of hundred milliwatts, and the requirement of a high-performance application scene cannot be met. Meanwhile, the design of the traditional super-structured surface is highly dependent on exhaustive numerical simulation, and the lack of deep fusion of a physical mechanism and an artificial intelligence algorithm leads to the failure of realizing efficient reverse design and global optimization aiming at a complex topological structure. In addition, the prior art lacks dynamic compensation capability for broadband alignment phase matching when processing multi-source heterogeneous material integration, and is susceptible to nonlinear saturation effect under high-power pumping, resulting in low conversion efficiency and degradation of beam quality. Accordingly, a tunable terahertz generation method based on physical driving to generate a countering network and a topologically super-structured surface is desired. Disclosure of Invention The invention aims to provide a tunable terahertz generation method for generating an antagonistic network and a topological super-structure surface based on physical driving, which can effectively solve the problems in the background technology. In order to achieve the above purpose, the technical scheme adopted by the invention is as follows: A tunable terahertz generation method for generating an antagonistic network and a topological super-constructed surface based on physical driving comprises the following specific steps: S1, constructing a heterogeneous integrated topological super-structured surface unit, namely arranging an arrayed V-shaped gold nano antenna or silicon nano column on an X-axis cutting film lithium niobate substrate with a preset thickness, covering a single-layer tungsten disulfide on the surface of a microstructure, and realizing dynamic control of amplitude and phase of terahertz waves through charge concentration regulation and control of a graphene layer; s2, generating a countermeasure network based on physical driving to carry out topological structure reverse design, namely, utilizing a time domain finite difference algorithm to simulate and acquire a data set containing geometric parameters, a transmission spectrum, a phase and nonlinear conversion efficiency, constructing the generated countermeasure network consisting of a generator and a discriminator, introducing Maxwell equation set constraint into a loss function, and automatically searching a topological structure sequence which meets the condition that the wave vector mismatch quantity approaches to a preset minimum value in a preset frequency range th