Diffractive optical element design and manufacturing integration method based on gray laser direct writing photoetching

Abstract

The invention provides a diffractive optical element design and manufacturing integration method based on gray level laser direct writing photoetching, which comprises the following steps that: a manufacturing simulator and a wave optical simulator are established, and the manufacturing simulator represents the spatial frequency response of an optical system through a low-pass filtering model; a mapping relation between exposure dose and a developing result is established by combining nonlinear response of photoresist, and the wave optical simulator predicts optical performance by utilizing output of the manufacturing simulator and combining a wave optical theory; by setting initial manufacturing parameters and performing iterative optimization, the surface appearance of the photoresist is dynamically adjusted to be close to the target optical performance, the optimization process is combined with an automatic differential technology, a gradient optimization algorithm and a block optimization strategy, and the calculation efficiency is improved; and the optimized parameters are used for processing gray laser direct writing equipment to manufacture the diffractive optical element. The problem of manufacturing deviation caused by the optical proximity effect in the prior art can be solved, and meanwhile the overall performance of the diffraction optical element is remarkably improved.

Description

Technical Field

[0001] The present application relates to the technical field of diffractive optical element manufacturing, and in particular to an integrated method for designing and manufacturing diffractive optical elements based on grayscale laser direct writing lithography. Background Art

[0002] Diffractive optical elements (DOEs) have been widely used in high-end optical fields such as LiDAR, AR / VR display, optical computing, communication and imaging due to their lightweight, multifunctional and high-precision characteristics. However, in existing technologies, the design and manufacturing of DOEs are usually separated, resulting in a lack of coordination between design freedom and manufacturing constraints, thus limiting the improvement of overall performance.

[0003] Traditional DOEs design methods are usually based on optical simulation and optimization algorithms, such as the Gerchberg-Saxton algorithm and direct phase optimization. These methods focus on optimizing the target optical performance of DOEs, but lack systematic modeling and compensation for non-ideal effects in the manufacturing process. In the manufacturing stage, grayscale lithography, electron beam lithography and other technologies are used to convert the design pattern into the actual structure. However, due to problems such as the optical proximity effect (OPE), the surface morphology of the actually processed optical components often deviates greatly from the design target, resulting in a decrease in optical performance.

[0004] In the prior art, in order to solve the problem of optical proximity effect, it is usually handled through the following two approaches: one is to optimize the manufacturing process separately, such as using a photoresist exposure dose calibration model (such as the Dill model) to adjust the manufacturing parameters, but this method only compensates in the manufacturing stage and fails to fully utilize the degrees of freedom in the design process; the other is to use a discrete model to optimize the design and manufacturing in steps, such as adjusting the optical performance and manufacturing parameters through multiple iterative optimizations, but this method is inefficient due to the fragmentation of the process, and the optimization results are prone to fall into local optimality.

[0005] In addition, the existing large-scale DOEs design and manufacturing capabilities are relatively limited, and it is difficult to support the processing requirements of ultra-high resolution and complex surface structures. At the same time, due to the high complexity of the model or the large amount of experimental calibration data required, the design and manufacturing costs of traditional methods are high, which is difficult to meet the growing market demand.

[0006] In summary, the design and manufacturing of DOEs in the existing technology have the following major problems: separation of design and manufacturing processes, performance deviation caused by optical proximity effect, low optimization efficiency, insufficient design capabilities for large-scale complex components, and high costs. These problems have severely limited the application of DOEs in the field of high-end optics, and an innovative "end-to-end" integrated design and manufacturing method is urgently needed to solve these bottlenecks and comprehensively improve the performance, efficiency, and applicability of DOEs. Summary of the invention

[0007] The present application aims to solve one of the technical problems in the related art at least to some extent.

[0008] To this end, the first objective of the present application is to propose an integrated method for designing and manufacturing diffractive optical elements based on grayscale laser direct writing lithography.

[0009] The second objective of the present application is to propose an integrated device for designing and manufacturing diffractive optical elements based on grayscale laser direct writing lithography.

[0010] The third objective of the present application is to provide an electronic device.

[0011] A fourth objective of the present application is to provide a computer-readable storage medium.

[0012] A fifth object of the present application is to provide a computer program product.

[0013] To achieve the above-mentioned purpose, the first embodiment of the present application proposes an integrated method for designing and manufacturing a diffractive optical element based on grayscale laser direct writing lithography, comprising:

[0014] Establishing a manufacturing simulator for characterizing the optical proximity effect of a grayscale laser direct writing device and the nonlinear response of a photoresist, the manufacturing simulator comprising a low-pass filter model for simulating the spatial frequency response of an optical system and a nonlinear mapping model between a photoresist exposure dose and a development result;

[0015] Establishing a wave optics simulator, the wave optics simulator is used to calculate optical distribution and optical performance based on the output of the manufacturing simulator in combination with wave optics theory, and predict the diffraction optical properties caused by the surface morphology of the photoresist;

[0016] According to the design goal, an initial exposure dose distribution is set as a manufacturing parameter, the manufacturing simulator is used to predict the initial surface morphology of the photoresist, and the prediction result is input into the wave optics simulator to simulate and obtain the corresponding optical performance distribution;

[0017] defining the difference between the target optical distribution and the simulated optical distribution as a loss function, calculating the gradient of the loss function with respect to the manufacturing parameters by using automatic differentiation technology and a gradient-based optimization method, and optimizing the manufacturing parameters;

[0018] Repeat the optimization steps and gradually iterate the manufacturing parameters to gradually approach the target optical performance until the loss function meets the predetermined performance index or reaches the set maximum number of iterations;

[0019] The optimized manufacturing parameters are used as input parameters and applied to the grayscale laser direct writing equipment for actual processing to manufacture diffractive optical elements.

[0020] Optionally, the method of establishing a manufacturing simulator for characterizing the optical proximity effect of a grayscale laser direct writing device and the nonlinear response of a photoresist comprises:

[0021] Obtaining the modulation transfer function of the laser direct writing device lens, where the modulation transfer function is used to characterize the response characteristics of the optical system to different spatial frequencies and is used as an input of the low-pass filtering model;

[0022] Characterize the nonlinear characteristics of the photochemical reaction and development process of the photoresist material through experiments, and obtain the mapping relationship between the exposure dose and the surface morphology after development;

[0023] A polynomial or a preset fitting function is used to fit the experimental data related to the modulation transfer function and the nonlinear response characteristics, and a differentiable manufacturing simulator is jointly established. The manufacturing simulator can directly predict the optical performance indicators after processing and optimize the process parameters.

[0024] Optionally, obtaining the modulation transfer function of the laser direct writing device lens includes:

[0025] A series of sinusoidal height profiles with different spatial frequencies were produced as basic patterns using grayscale laser direct writing technology;

[0026] By analyzing the spatial frequency and amplitude changes of the above-mentioned sinusoidal pattern, the response characteristics of the optical system to different spatial frequencies are determined, and a modulation transfer function is generated;

[0027] The modulation transfer function is used as a part of a low-pass filter model, and the optical proximity effect in the processing process is modeled in combination with the modulation transfer function.

[0028] Optionally, the experimental characterization of the nonlinear characteristics of the photochemical reaction and development process of the photoresist material and the acquisition of the mapping relationship between the exposure dose and the surface morphology after development includes:

[0029] A series of square patterns with different exposure doses were produced by grayscale laser direct writing technology;

[0030] Measuring the surface depth of the square pattern after development to obtain nonlinear response data between exposure dose and surface topography;

[0031] A nonlinear mapping model of exposure dose and surface structure depth after development is established based on experimental data. The nonlinear mapping model is a component of the manufacturing simulator and is used to further optimize the prediction accuracy of the photoresist surface morphology.

[0032] Optionally, the calculating the gradient of the loss function with respect to the manufacturing parameters by using the automatic differentiation technique and the gradient-based optimization method comprises:

[0033] Calculating the gradient of the loss function with respect to the manufacturing parameters based on the manufacturing simulator and the wave optics simulator using an automatic differentiation technique;

[0034] Adopt a gradient-based optimization algorithm, combined with the Adam optimization algorithm to dynamically adjust the learning rate and momentum factor to improve the optimization convergence efficiency;

[0035] The diffraction efficiency is incorporated into the objective function, and a penalty term is introduced to suppress the zero-order diffraction intensity, and the phase distribution is optimized to redistribute the energy to the desired diffraction order and improve the efficiency of the target diffraction order;

[0036] Batch normalization is performed on parameter updates across multiple optimization iterations to improve the stability and convergence speed of model training.

[0037] Optionally, also include:

[0038] For large-scale or complex designs, a block optimization strategy is used during the optimization process to decompose the large design into multiple sub-problems;

[0039] Use parallel computing technology to process multiple sub-problems simultaneously, and gradually refine the optimization results through multi-scale optimization methods;

[0040] After the optimization is completed, the optimization results of the sub-problems are integrated to form a global design.

[0041] Optionally, the method is applicable to manufacturing a microstructure surface and a continuous free-form surface having a predetermined optical function, comprising:

[0042] Microlens array for beam shaping and focusing;

[0043] Diffraction gratings, used for spectral separation or wavelength selection;

[0044] Phase plate, used for wavefront modulation;

[0045] a multi-focus beam splitter for generating multiple focused beams;

[0046] Computer generated holograms, used for three-dimensional image display or wavefront reconstruction;

[0047] An extended depth of field imaging system for enhancing the depth of field of an imaging system;

[0048] wavefront correctors, used to compensate for aberrations in optical systems;

[0049] Continuous free-form optical surfaces for complex optical functional needs.

[0050] Optionally, the method is applicable to a two-photon laser direct writing device, including the following adjustments:

[0051] Add a two-photon absorption effect model to the manufacturing simulator to characterize the impact of the two-photon excitation process on photoresist processing;

[0052] Considering the formation process of the three-dimensional structure, the optimization strategy is adjusted to adapt to the three-dimensional processing capability of the two-photon direct writing equipment;

[0053] The nonlinear response model is adjusted based on the characteristics of the two-photon laser direct writing device to ensure high resolution and processing accuracy of the structure.

[0054] To achieve the above-mentioned purpose, the second embodiment of the present application proposes an integrated device for designing and manufacturing a diffractive optical element based on grayscale laser direct writing lithography, comprising:

[0055] A first establishment module is used to establish a manufacturing simulator for characterizing the optical proximity effect of the grayscale laser direct writing device and the nonlinear response of the photoresist, wherein the manufacturing simulator includes a low-pass filter model for simulating the spatial frequency response of the optical system and a nonlinear mapping model between the exposure dose of the photoresist and the development result;

[0056] A second establishment module is used to establish a wave optics simulator, wherein the wave optics simulator is used to calculate optical distribution and optical performance based on the output of the manufacturing simulator in combination with wave optics theory, and predict diffraction optical properties caused by the surface morphology of the photoresist;

[0057] A setting module is used to set the initial exposure dose distribution as a manufacturing parameter according to the design goal, use the manufacturing simulator to predict the initial surface morphology of the photoresist, and input the prediction result into the wave optics simulator to simulate and obtain the corresponding optical performance distribution;

[0058] An optimization module, used for defining the difference between the target optical distribution and the simulated optical distribution as a loss function, calculating the gradient of the loss function with respect to the manufacturing parameters by using automatic differentiation technology and a gradient-based optimization method, and optimizing the manufacturing parameters;

[0059] An iteration module is used to repeat the optimization steps and gradually iterate the manufacturing parameters to gradually approach the target optical performance until the loss function meets the predetermined performance index or reaches the set maximum number of iterations;

[0060] The manufacturing module is used to use the optimized manufacturing parameters as input parameters and apply them to the grayscale laser direct writing equipment for actual processing to manufacture diffractive optical elements.

[0061] To achieve the above-mentioned purpose, the third aspect of the present application provides an electronic device, including: a processor, and a memory communicatively connected to the processor;

[0062] The memory stores computer-executable instructions;

[0063] The processor executes the computer-executable instructions stored in the memory to implement the method as described in any one of the first aspects.

[0064] To achieve the above-mentioned purpose, the fourth aspect embodiment of the present application proposes a computer-readable storage medium, in which computer-readable storage medium is stored computer execution instructions, and when the computer execution instructions are executed by a processor, they are used to implement the method as described in any one of the first aspects.

[0065] To achieve the above-mentioned purpose, the fifth aspect of the present application proposes a computer program product, which implements any method in the first aspect when executed by a processor.

[0066] The technical solution provided by the embodiments of the present application brings at least the following beneficial effects:

[0067] First, this application proposes an "end-to-end" integrated design and manufacturing optimization method that seamlessly combines optical design with manufacturing processes, effectively solving the manufacturing deviation problem caused by the optical proximity effect. Compared with the traditional separate design and manufacturing process, this application can dynamically adjust manufacturing parameters, optimize optical performance and manufacturing accuracy as a whole, and significantly improve product performance and consistency.

[0068] Secondly, this application realizes the joint optimization from manufacturing process to optical performance through the integration of physics-based manufacturing model and wave optics simulation model, so that the non-ideal effects in the manufacturing process can be fully modeled and corrected. Through large-scale optimization of automatic differentiation technology, this application significantly improves the efficiency of design and manufacturing, and can support the design and manufacturing needs of large-scale and highly complex optical components.

[0069] Thirdly, the present application can achieve high-precision manufacturing and improve optical performance. For example, in the beam splitter, zero-order diffraction is effectively suppressed, and the uniformity of the diffraction order is improved; in the digital hologram, the imaging quality and structural fidelity are improved; in the extended depth of focus imaging lens, the focal depth range and imaging clarity are significantly improved.

[0070] In addition, this application supports the design and manufacture of ultra-large-scale optical elements (such as diffractive optical elements up to 100 million pixels), and is suitable for multifunctional optical devices such as complex surface profiles, multi-focus imaging, and beam splitting. Its efficient optimization algorithm reduces the cost of experimental calibration and iteration, and significantly shortens the design and manufacturing cycle.

[0071] Finally, this application is based on grayscale laser direct writing lithography technology, and realizes the manufacture of quasi-continuous relief structures through a single-step process, without the need for complex multi-step lithography processes, significantly reducing manufacturing costs and resource investment. Overall, this application shows significant advantages in optical performance, manufacturing efficiency, design freedom, and production costs, and can be widely used in high-end optical fields such as laser radar, AR / VR display, optical computing and communications, providing solid support for product development and market competitiveness.

[0072] Additional aspects and advantages of the present application will be given in part in the description below, and in part will become apparent from the description below, or will be learned through the practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0074] Figure 1 A schematic diagram of a flow chart of an integrated method for designing and manufacturing a diffractive optical element based on grayscale laser direct writing lithography provided in an embodiment of the present application;

[0075] Figure 2 A schematic diagram comparing the radiation efficiency and diffraction order uniformity of a one-dimensional beam splitter provided in an embodiment of the present application;

[0076] Figure 3 Schematic diagram of the comparison of surface morphology measurements of the one-dimensional beam splitter provided in the embodiments of the present application.

[0077] Figure 4 A schematic diagram showing the comparison between the holographic projection (Gaussian 1840 oil painting and the emblem of Tsinghua University) provided in the embodiment of the present application;

[0078] Figure 5 A schematic diagram of a super-large angle dot matrix (15×15) structured light projector provided in an embodiment of the present application;

[0079] Figure 6 A schematic diagram of the imaging quality comparison of the long focal depth (0.1-1.2m) near-infrared imaging lenses provided in the embodiments of the present application;

[0080] Figure 7 A schematic diagram of the structure of an integrated device for designing and manufacturing a diffractive optical element based on grayscale laser direct writing lithography provided in an embodiment of the present application. DETAILED DESCRIPTION

[0081] Embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present application, and should not be construed as limiting the present application.

[0082] In view of the problems existing in the prior art, the embodiments of the present application provide a method for designing and manufacturing a diffractive optical element based on grayscale laser direct writing lithography. Figure 1 The following is a flow chart of an integrated method for designing and manufacturing a diffractive optical element based on grayscale laser direct writing lithography provided in an embodiment of the present application. Figure 1 As shown, the method comprises the following steps:

[0083] Step 101, establish a manufacturing simulator that characterizes the optical proximity effect of the grayscale laser direct writing device and the nonlinear response of the photoresist, the manufacturing simulator includes a low-pass filter model that simulates the spatial frequency response of the optical system and a nonlinear mapping model between the exposure dose of the photoresist and the development result.

[0084] In this step, we first need to build a manufacturing simulator to characterize the optical proximity effect of grayscale laser direct writing equipment and the nonlinear response of photoresist. The core of this manufacturing simulator is to accurately model the physical properties of the optical system and photoresist during the grayscale laser direct writing process, thereby providing reliable prediction capabilities for subsequent optimization and manufacturing.

[0085] The Manufacturing Simulator consists of two main modules:

[0086] The first module is a low-pass filter model, which is used to simulate the spatial frequency response characteristics of the optical system, specifically characterizing the resolution loss caused by optical diffraction effects and system limitations during the lithography process. Through this model, the response behavior of the optical system at different spatial frequencies can be effectively captured, providing basic support for the simulation of the manufacturing process.

[0087] The second module is a nonlinear mapping model, which is used to describe the nonlinear response relationship between the exposure dose of photoresist and the development result. By fitting the experimental data, the model can characterize the photochemical reaction characteristics of photoresist at different doses and establish a mapping relationship between it and the final surface morphology depth. Through this nonlinear mapping model, the morphological changes of the photoresist surface in actual processing can be predicted.

[0088] Overall, the manufacturing simulator combines the low-pass filtering model and the nonlinear mapping model, which can fully simulate the physical processes involved in the grayscale laser direct writing process, laying the foundation for optimizing the exposure dose distribution and improving the manufacturing accuracy of optical components.

[0089] In step 101, in order to establish a manufacturing simulator that can fully characterize the optical proximity effect of the grayscale laser direct writing device and the nonlinear response of the photoresist, the following key steps are also included:

[0090] Step 201, obtaining the modulation transfer function of the lens of the laser direct writing device, where the modulation transfer function is used to characterize the response characteristics of the optical system to different spatial frequencies and is used as an input of a low-pass filtering model.

[0091] In the embodiment of the present application, the modulation transfer function of the laser direct writing device lens is first obtained. The modulation transfer function is used to characterize the response characteristics of the optical system to different spatial frequencies, and provides input parameters for the low-pass filter model by analyzing the resolution loss of the optical system at different spatial frequencies. Specifically, the modulation transfer function can effectively simulate the attenuation of high-frequency components caused by optical diffraction and system limitations during the lithography process, and is the key to accurately describing the performance of the optical system.

[0092] As a possible implementation, step 201 specifically obtains the modulation transfer function of the laser direct writing device lens through the following process:

[0093] First, a series of sinusoidal height profiles with different spatial frequencies are produced as basic patterns using grayscale laser direct writing technology. These sinusoidal patterns cover different spatial frequencies and are used to represent the different spatial frequencies of the surface structure of real optical components. The height and frequency of the sinusoidal patterns can be precisely controlled by the laser direct writing device to ensure the reliability of the experimental results.

[0094] Next, the spatial frequency and amplitude changes of the above-mentioned sinusoidal patterns are analyzed to determine the response characteristics of the optical system to different spatial frequencies. Specifically, by measuring the actual amplitude of these sinusoidal waveforms after development and comparing them with the amplitude of the design target, the attenuation of the optical system at a specific spatial frequency can be quantified. These data are used to generate the modulation transfer function, which reflects the transmission efficiency of the optical system for different frequency components and is an important basis for simulating the performance of the optical system.

[0095] Finally, the generated modulation transfer function is used as part of the low-pass filter model to characterize the high-frequency attenuation effect caused by the optical system during the lithography process. At the same time, the optical proximity effect during the processing is modeled in combination with the modulation transfer function to further describe the performance limitations and error sources of the system under different processing conditions.

[0096] Through the above steps, the modulation transfer function not only provides accurate input parameters for the low-pass filter model, but also lays an important physical foundation for the construction of the manufacturing simulator, providing support for optimizing processing accuracy and optical performance.

[0097] Step 202 , characterizing the nonlinear characteristics of the photochemical reaction and development process of the photoresist material through experiments, and obtaining the mapping relationship between the exposure dose and the surface morphology after development.

[0098] In the embodiments of the present application, the nonlinear characteristics of the photochemical reaction and development process of the photoresist material are characterized by experiments. This process involves making a series of large block patterns of different doses through grayscale laser direct writing experiments, and measuring the depth changes of the surface morphology of these patterns after development, so as to obtain nonlinear response data between the exposure dose and the development result. The size of each block must be much larger than the PSF of the laser writing lens to avoid the influence of the optical proximity effect. The contrast curve can be directly measured using equipment such as a confocal microscope, which illustrates the correlation between the depth reached after exposure and development and the exposure dose used. These experimental data are an important basis for constructing a nonlinear mapping model of photoresist, and can describe the dynamic behavior of photoresist under different dose conditions.

[0099] As a possible implementation, step 202 specifically characterizes the nonlinear characteristics of the photochemical reaction and development process of the photoresist material through the following process:

[0100] First, a series of square patterns with different exposure doses are made by grayscale laser direct writing technology. The exposure doses of these square patterns are precisely controlled, covering a range from low doses to high doses, to fully reflect the reaction behavior of photoresists under different exposure conditions. Each pattern is prepared through the same processing and development process to ensure the repeatability of the experiment and the reliability of the data.

[0101] Next, the surface depth of the square pattern after development is measured to obtain the nonlinear response data between the exposure dose and the surface morphology. In this step, the depth change of the pattern surface is measured by high-precision measurement equipment (such as atomic force microscope or white light interferometer), and the post-development depth value corresponding to each exposure dose is recorded. These data reflect the nonlinear behavior of the photoresist during the development process and are the basis for subsequent modeling.

[0102] Finally, a nonlinear mapping model of exposure dose and surface structure depth after development is established based on the above experimental data. By using polynomial fitting or other appropriate fitting methods, the nonlinear relationship between exposure dose and surface depth is expressed as a mathematical model. This nonlinear mapping model is a key component of the manufacturing simulator and can be used to further optimize the prediction accuracy of the photoresist surface morphology. By combining with the modulation transfer function, the nonlinear mapping model can accurately describe the reaction characteristics of the photoresist during the grayscale laser direct writing process, thereby improving the overall accuracy of the manufacturing simulator.

[0103] The above steps provide strong data support for comprehensive modeling of the nonlinear response of photoresists and provide the necessary tools and foundation for end-to-end manufacturing optimization.

[0104] Step 203, use a polynomial or a preset fitting function to fit the experimental data related to the modulation transfer function and the nonlinear response characteristics, and jointly establish a differentiable manufacturing simulator. The manufacturing simulator can directly predict the optical performance indicators after processing and optimize the process parameters.

[0105] In the embodiment of the present application, a polynomial or a preset fitting function is used to fit the experimental data related to the modulation transfer function and the nonlinear response characteristics of the photoresist. Through the fitting process, the modulation transfer function and the nonlinear response data are effectively integrated to form a complete differentiable manufacturing simulator. The manufacturing simulator can not only accurately predict the optical performance indicators after processing, but also optimize the process parameters through automatic differentiation technology, thereby providing real-time feedback in the design and manufacturing stages.

[0106] Through these steps, a manufacturing simulator is fully constructed, which can accurately simulate the optical system characteristics and photoresist behavior involved in the grayscale laser direct writing process, providing a reliable foundation for end-to-end design and manufacturing optimization.

[0107] Step 102, establishing a wave optics simulator, which is used to calculate optical distribution and optical performance based on the output of the manufacturing simulator in combination with wave optics theory, and predict the diffraction optical characteristics caused by the surface morphology of the photoresist.

[0108] The core of this step is to establish a wave optics simulator, which is established to predict the diffraction optical properties caused by the surface morphology of the photoresist. Its core function is to accurately model and calculate the light field propagation process and optical performance.

[0109] Specifically, the wave optics simulator uses the output of the manufacturing simulator as input and the photoresist surface topography data generated by the manufacturing simulator as the basic information for simulating the performance of the diffractive optical element. The surface topography data output by the manufacturing simulator accurately reflects the actual processing results affected by the optical proximity effect and the nonlinear response of the photoresist during the lithography process.

[0110] Furthermore, combined with the wave optics theory, the wave optics simulator can accurately model and calculate the light field propagation process. Through analytical and numerical methods, the wave optics simulator can simulate the diffraction behavior of light in the photoresist surface structure, especially the intensity distribution, phase distribution and energy distribution of high-order diffraction, so as to fully predict the actual optical performance of the diffractive optical element. The main output of the wave optics simulator includes the optical performance of the diffractive optical element at the target wavelength, such as diffraction efficiency, phase distribution consistency and light field distribution quality. These performance indicators can be compared with the target design requirements to provide feedback information for subsequent optimization.

[0111] In the embodiment of the present application, by combining the wave optics simulator with the manufacturing simulator, seamless connection and dynamic update from the manufacturing process to the optical performance can be achieved. The wave optics simulator plays a key role in the entire design and manufacturing optimization process, ensuring that processing deviations can be corrected in a timely manner through optical performance feedback, improving the prediction ability and optimization efficiency of the overall system, and, during the optimization process, all adjustments to manufacturing parameters are dynamically updated between manufacturing simulation and optical simulation. Through this closed-loop update mechanism, the differences between design and manufacturing can be corrected in a timely manner during the optimization process to ensure that the processing results gradually approach the target optical performance.

[0112] It should be noted that this integrated simulation system has significant advantages over the prior art. In the prior art, design and manufacturing are usually separated, and it is difficult to coordinate design freedom and manufacturing constraints, resulting in the inability to efficiently correct manufacturing deviations. This application eliminates this separation by integrating a manufacturing simulator and an optical simulator, so that manufacturing deviations can be dynamically corrected directly through optical performance feedback.

[0113] In addition, this integrated simulation system significantly improves the globality and efficiency of optimization, especially in the design and manufacturing scenarios of large-scale complex DOEs. It can support high-degree-of-freedom design optimization and complex structure processing requirements, providing strong technical support for innovative applications in the field of optics.

[0114] Step 103, according to the design goal, set the initial exposure dose distribution as the manufacturing parameter, use the manufacturing simulator to predict the initial surface morphology of the photoresist, and input the prediction result into the wave optics simulator to simulate and obtain the corresponding optical performance distribution.

[0115] The purpose of this step is to convert the design goals into initial manufacturing parameters, and through the joint simulation of the manufacturing simulator and the wave optics simulator, predict the photoresist surface morphology and the corresponding optical performance distribution under the initial conditions, providing a basis for subsequent optimization.

[0116] First, the initial exposure dose distribution is set according to the design goal. The design goal can include specific optical performance indicators such as diffraction efficiency, phase distribution consistency, and zero-order diffraction suppression rate. The initial exposure dose distribution is a parameter field used to guide processing. Its distribution range and value are determined by the design goal and serve as the core parameter of the input manufacturing simulator.

[0117] Next, the input initial exposure dose distribution is simulated using a manufacturing simulator. The manufacturing simulator calculates the surface morphology that may be formed during the actual processing of the photoresist based on the modulation transfer function and the nonlinear response model of the photoresist. The modulation transfer function characterizes the response of the optical system to different spatial frequencies, while the nonlinear mapping model of the photoresist describes how the exposure dose is converted into the depth of the structure after development. Through the combined effect of these models, the manufacturing simulator generates the initial surface morphology of the photoresist, reflecting the potential errors and non-ideal characteristics in the processing results.

[0118] The photoresist surface topography generated by the manufacturing simulator is then passed as input data to the wave optics simulator. The wave optics simulator simulates the propagation and diffraction behavior of the light field based on wave optics theory. The simulation results include the distribution of optical performance, such as the diffraction efficiency distribution at the target wavelength, the deviation of the phase distribution, and the distribution of energy between different diffraction orders. The wave optics simulator can be used to further evaluate the impact of manufacturing results on optical performance.

[0119] This process closely links the design goals and the processing, so that the direct impact of manufacturing parameters on optical performance can be preliminarily evaluated through simulation. At the same time, this joint simulation reveals the difference between manufacturing reality and design goals, providing an important starting point and data support for the subsequent adjustment of manufacturing parameters based on optimization methods. This comprehensive analysis can effectively reduce the separation between design and manufacturing and improve the overall performance prediction ability of the system.

[0120] Step 104, defining the difference between the target optical distribution and the simulated optical distribution as a loss function, using automatic differentiation technology and a gradient-based optimization method to calculate the gradient of the loss function with respect to the manufacturing parameters, and optimizing the manufacturing parameters.

[0121] The purpose of this step is to optimize the surface morphology and optical properties of the photoresist by defining a loss function and using automatic differentiation technology and gradient-based optimization methods to dynamically adjust the manufacturing parameters, so as to gradually approach the design target.

[0122] First, the difference between the target optical distribution and the simulated optical distribution is defined as the loss function. The target optical distribution is the ideal optical performance generated according to the design goal, such as the diffraction efficiency distribution at a specific wavelength, the phase distribution consistency and other indicators; the simulated optical distribution is the optical performance corresponding to the processing result calculated by the wave optics simulator. The loss function reflects the difference between the actual processing and the ideal design, and its specific form can include the following:

[0123] (1) Deviation of diffraction efficiency, such as the square of the difference between the target diffraction order efficiency and the simulated diffraction order efficiency;

[0124] (2) The consistency of phase distribution, such as the error between the target phase distribution and the simulated phase distribution;

[0125] (3) Other constraints, such as the suppression strength of the zero-order diffraction or the energy distribution of higher-order diffraction.

[0126] Next, the automatic differentiation technique is used to calculate the gradient of the loss function with respect to the manufacturing parameters. Automatic differentiation technology efficiently calculates the partial derivatives of the loss function with respect to the manufacturing parameters through analytical methods, avoiding the problem of large amount of calculation in numerical difference methods. Combined with the differentiable characteristics of the manufacturing simulator and the wave optics simulator, automatic differentiation technology can dynamically track the changes of the loss function during the optimization process and generate accurate gradient information.

[0127] Subsequently, a gradient-based optimization algorithm is used to adjust the parameters. Combined with the Adam optimization algorithm, the optimization process can converge to the global optimal solution faster by dynamically adjusting the learning rate and momentum factor. The Adam algorithm not only speeds up the optimization speed but also reduces the risk of falling into the local optimal solution by balancing the momentum and update amplitude of the gradient. The goal of the optimization is to gradually reduce the loss function value so that the simulated optical distribution gradually approaches the target optical distribution.

[0128] In addition, during multiple optimization iterations, batch normalization is performed on the parameter update results to improve the training stability and convergence speed of the model. Batch normalization can balance the gradient scales of different parameters, avoid excessive oscillation or convergence stagnation during the optimization process, and thus improve the overall optimization efficiency and model robustness.

[0129] This process achieves closed-loop optimization from manufacturing parameters to optical performance, and ensures that the processing results are closer to the design goals by dynamically adjusting the manufacturing parameters, while fully considering the actual constraints and non-ideal effects in the processing process. This gradient-based optimization method significantly improves the optimization efficiency and provides technical support for achieving high-precision design and manufacturing integration.

[0130] In addition, for large-scale or complex designs, in order to improve optimization efficiency and processing complexity, the embodiments of the present application adopt a block optimization strategy, a multi-scale optimization method, and a parallel computing technology, which specifically include the following contents:

[0131] First, for large designs, this application decomposes the entire design area into multiple manageable sub-problems through a block optimization strategy. Each sub-problem corresponds to a part of the design area, and its complexity is significantly reduced, making the optimization task easier to handle. The block strategy can be divided according to the spatial distribution, complexity, or target functional requirements of the design to ensure that each sub-problem is independent and efficient during the calculation and optimization process.

[0132] Secondly, this application uses parallel computing technology to process multiple sub-problems at the same time. During the optimization process, by assigning sub-problems to different computing units, parallel computing greatly improves the overall optimization speed. At the same time, the optimization results of each sub-problem are independent of each other, avoiding computing bottlenecks and resource conflicts that may occur during the global optimization process. The high efficiency of parallel processing enables this method to support the design and manufacture of ultra-large-scale and high-resolution optical components.

[0133] In addition, this application further improves the optimization efficiency through a multi-scale optimization method. The optimization process starts with a rough structure, first preliminarily optimizing the overall performance of the global design to quickly approach the target optical performance. After the global performance meets the preliminary requirements, the design is gradually refined to accurately optimize the high-resolution characteristics and local details of each sub-area. The layer-by-layer advancement mechanism of the multi-scale optimization method can make full use of computing resources while achieving high-precision optimization on the basis of ensuring global performance.

[0134] Finally, after the optimization is completed, the embodiment of the present application integrates the optimization results of all sub-problems to form an overall design. During the integration process, the boundary conditions between the sub-areas and the design parameters of the adjacent areas are coordinated to ensure the continuity of the overall design and the consistency of the optical performance.

[0135] This combination of block optimization and multi-scale optimization is particularly suitable for the design and manufacture of complex optical components, such as wavefront controllers with high degrees of freedom, ultra-high resolution microlens arrays, and computational holograms with complex phase distributions. Through decomposition, parallel computing, and step-by-step refinement, this method significantly improves the optimization efficiency while ensuring the accuracy and consistency of global design performance, providing strong technical support for the development of large-scale complex optical components.

[0136] Step 105, repeating the optimization steps, gradually iterating the manufacturing parameters to gradually approach the target optical performance, until the loss function meets the predetermined performance index or reaches the set maximum number of iterations.

[0137] The core of this step is to iteratively adjust the manufacturing parameters multiple times by repeating the optimization steps, so that the processing results gradually approach the design goals (such as surface relief profile), thereby optimizing the optical performance and achieving the final manufacturing requirements.

[0138] Specifically, in this process, the aforementioned optimization steps 103-104 are first repeatedly performed, including predicting the photoresist surface morphology through a manufacturing simulator, simulating the optical performance distribution using a wave optics simulator, calculating the loss function, and adjusting the parameters based on the gradient. The goal of each iterative optimization is to reduce the loss function value so that the optical performance distribution of the actual processing gradually approaches the target optical performance distribution.

[0139] It should be noted that, through the introduction of automatic differentiation technology, the method of the present application can accurately calculate the gradient of the loss function to the manufacturing parameters in each iteration, providing a clear optimization direction for the update of the manufacturing parameters. In each iteration, the manufacturing parameters are dynamically adjusted in combination with a gradient-based optimization algorithm (such as the Adam optimization algorithm). The optimization process uses the learning rate and momentum factor for adaptive adjustment to accelerate the convergence speed and avoid falling into the local optimal solution.

[0140] It is understandable that during multiple iterations, the optimization will gradually approach the target performance while continuously evaluating the current loss function value. If the loss function meets the predetermined performance indicators (e.g., the diffraction efficiency meets specific requirements, the phase distribution error is within the allowable range, etc.), the optimization is stopped and the final manufacturing parameters are output. If the loss function fails to meet the performance indicators, the optimization process is limited according to the set maximum number of iterations to avoid unnecessary computational overhead.

[0141] In the embodiments of the present application, this iterative optimization process continuously narrows the gap between the actual processing results and the design goals, ensuring that the final manufacturing parameters can meet the optical performance requirements while fully considering the constraints and non-ideal effects in the processing process. The final iterative optimization results provide a reliable parameter basis for subsequent actual processing, and at the same time lay a solid technical guarantee for the high-precision manufacturing of complex optical components.

[0142] Step 106 , using the optimized manufacturing parameters as input parameters and applying them to a grayscale laser direct writing device for actual processing to manufacture a diffractive optical element.

[0143] The goal of this step is to actually apply the optimized manufacturing parameters to the grayscale laser direct writing equipment to complete the manufacturing of the diffractive optical element.

[0144] In the embodiment of the present application, the manufacturing parameters obtained by iterative optimization in the above steps are used as the final input parameters. These parameters generally include the design requirements of the photoresist surface morphology and its corresponding exposure dose distribution. After multiple optimizations, the optical proximity effect, nonlinear response characteristics and target optical performance requirements have been fully considered.

[0145] Next, the optimized parameters are imported into the grayscale laser direct writing device. The device accurately maps the optimized design onto the photoresist surface by controlling the laser power, beam scanning path and dose distribution. The high resolution and continuous control capability of grayscale laser direct writing technology can achieve quasi-continuous relief surface processing and meet the high-precision manufacturing requirements of complex diffractive optical elements.

[0146] After that, the photoresist structure after direct writing enters the development process. Through development, the three-dimensional morphology of the photoresist surface is revealed, forming a diffraction structure with optical functions. This structure is highly consistent with the design requirements and can minimize processing errors and performance deviations.

[0147] Finally, the manufactured diffractive optical element can be subsequently tested and verified for optical performance. By measuring key indicators such as diffraction efficiency, phase distribution, and zero-order suppression rate, the consistency between the actual processing results and the design goals can be further verified. If the test results meet expectations, it indicates that the manufacturing process is successful; if there are deviations, the manufacturing parameters can be further optimized based on the feedback information.

[0148] It should be noted that this step converts the optimized manufacturing parameters into actual processing results, completing the end-to-end process from design to manufacturing. This processing method based on optimized parameters significantly improves the manufacturing accuracy and performance of diffractive optical elements, enabling them to be widely used in high-end optical fields such as laser radar, AR / VR display, and optical imaging.

[0149] In addition, the present application has wide applicability and can manufacture microstructure surfaces and continuous free-form surfaces with a variety of predetermined optical functions, which is suitable for a variety of high-end optical application scenarios.

[0150] Specifically, the method of this application supports the design and manufacture of microlens arrays, can achieve high-precision beam shaping and focusing functions, and is suitable for the fields of optical imaging and display. For diffraction gratings, the method of this application can effectively improve the efficiency of spectral separation or wavelength selection by optimizing manufacturing parameters, meeting the needs of optical communications and spectroscopic devices.

[0151] The manufacturing of phase plates benefits from the high-precision phase distribution optimization capability of the method of the present application, and can be used for precise wavefront modulation and is widely used in phase correction of optical imaging systems. Multi-focal beam splitters can achieve precise control of multiple focal points through the method of the present application, and are used for beam splitting and the design of high-power optical devices.

[0152] In the application of computer-generated holograms, the method of this application can optimize the structural fidelity and imaging quality of digital holograms, realize three-dimensional image display or wavefront reconstruction functions, and improve the performance of optical display devices. At the same time, the extended depth of field imaging system can optimize the optical design through the method of this application, enhance the depth of field range and clarity of the imaging system, and is suitable for high-resolution optical microscopes and shooting equipment.

[0153] The design of the wavefront corrector compensates for aberrations in the optical system through the method of this application, effectively improving the resolution and quality of optical imaging. In addition, for continuous free-form optical surfaces, the method of this application can accurately control the structural morphology of complex optical surfaces, realize multifunctional optical design requirements, and is suitable for high-end optical application scenarios such as lidar, AR / VR display and optical computing.

[0154] In summary, the method of the present application has a wide range of applications and can support the manufacture of optical components ranging from microstructures to large-scale complex optical components, providing comprehensive technical support for innovative design and high-precision manufacturing in the field of optics.

[0155] In addition, the method proposed in this application is also applicable to two-photon laser direct writing equipment, and necessary adjustments can be made to its characteristics to support high-precision three-dimensional structure manufacturing.

[0156] Specifically, this application can characterize the processing behavior of photoresist during two-photon excitation by adding a two-photon absorption effect model to the manufacturing simulator. The two-photon absorption effect is the core mechanism of two-photon laser direct writing. By modeling this process, the nonlinear response characteristics of photoresist under two-photon excitation can be accurately described, providing a reliable physical basis for the manufacturing simulator.

[0157] Considering the formation process of the three-dimensional structure, the present application can also adjust the optimization strategy to adapt to the three-dimensional processing capabilities of the two-photon laser direct writing device. Two-photon laser direct writing technology has the advantage of ultra-high resolution and can achieve precise processing in three-dimensional space. By adjusting the optimization strategy, this method can make full use of the three-dimensional degrees of freedom of the equipment, take into account both local characteristics and overall design requirements during the optimization process, and achieve optimized manufacturing of complex three-dimensional structures.

[0158] Based on the characteristics of two-photon laser direct writing equipment, this application also supports the adjustment of nonlinear response models to ensure high resolution and processing accuracy of the structure. Combined with the characteristics of two-photon direct writing technology, by adjusting model parameters, the processing path and exposure dose can be further optimized, and the structural details can be accurately controlled to meet the needs of high-precision manufacturing.

[0159] Through these adjustments, the method of the present application can not only adapt to the unique requirements of two-photon laser direct writing technology, but also significantly improve the manufacturing capabilities of three-dimensional optical components, including complex free-form surfaces, micro-nanostructure arrays, etc., further expanding the scope of application of the method.

[0160] Some examples are shown below to demonstrate the effect of the method of the present application.

[0161] Figure 2 Schematic diagram comparing the diffraction efficiency and diffraction order uniformity of the one-dimensional beam splitter provided in an embodiment of the present application at wavelengths of 650nm and 1550nm and different beam splitting modes (1×3, 1×5, ±15°). Figure 2 The performance comparison of three design methods is shown in the figure: Naive (original design), Conventional (traditional design method) and E2E (end-to-end optimization method of this application).

[0162] The results show that the end-to-end optimization method significantly improves the diffraction efficiency compared to the traditional method, especially the energy distribution on the target diffraction order is more concentrated, and the diffraction order uniformity is higher, reducing the possible performance deviation in the optical system. This figure verifies the effect of end-to-end optimization on uniformity improvement and significant improvement in diffraction efficiency.

[0163] Figure 3 A schematic diagram of the surface morphology measurement comparison of the one-dimensional beam splitter (650nm, 1550nm) provided in the embodiment of the present application. The figure shows the surface morphology changes after processing using the Naive, Conventional and E2E methods, as well as the fit comparison between these design methods and the target morphology.

[0164] The results show that the end-to-end optimization design makes the processed shape closer to the ideal shape, significantly improves the fit with the target design, and reduces the processing error, especially in the restoration of complex surface contours. This experiment shows that the advantages of end-to-end optimization in manufacturing accuracy make the actual processing results more in line with the design requirements.

[0165] Figure 4 This is a schematic diagram of holographic projection provided in the embodiment of the present application, which includes the projection comparison results of Gauss 1840 oil painting (continuous optical image) and Tsinghua University logo (binary image). This figure compares the effects of no proximity effect correction, traditional correction method and E2E optimization method.

[0166] The results show that end-to-end optimization significantly improves image quality, especially in terms of detail restoration and edge clarity, and significantly reduces the intensity of zero-order diffraction, improving the contrast and layering of the projected image. This experiment demonstrates the application potential of end-to-end optimization in high-quality holographic projection and provides technical support for the accurate reproduction of complex images.

[0167] Figure 5 This is a schematic diagram of a super-large angle (~90°×90°) dot matrix (15×15) structured light projector provided in an embodiment of the present application. This figure shows the differences in dot matrix uniformity and zero-order diffraction suppression effects of different design methods (Naive, Conventional, E2E).

[0168] The results show that the end-to-end optimization method achieves a larger available projection angle while ensuring the uniformity of the dot-matrix spot distribution, and the zero-order diffraction intensity is significantly reduced, reducing unnecessary interference in the light field and improving system performance. This experiment shows that end-to-end optimization can meet the design requirements of optical devices with large angles and high uniformity.

[0169] Figure 6 This is a schematic diagram for comparing the imaging quality of a near-infrared imaging lens with a long focal depth (0.1-1.2m) provided in an embodiment of the present application. The figure compares the performance of a traditional correction method and an E2E optimization method in terms of imaging clarity and focal depth range.

[0170] The results show that end-to-end optimization significantly improves the imaging quality in a longer focal depth range (0.05m to 1.2m), and effectively compensates for optical aberrations in the edge and center areas, achieving higher resolution and better depth perception. The experimental results show that end-to-end optimization not only enhances the imaging capability of the lens, but also significantly expands the focal depth range.

[0171] In order to implement the above-mentioned embodiments, the present application also proposes an integrated device for designing and manufacturing diffractive optical elements based on grayscale laser direct writing lithography. Figure 7 A schematic diagram of the structure of an integrated device for designing and manufacturing a diffractive optical element based on grayscale laser direct writing lithography provided in an embodiment of the present application. Figure 2 As shown, the device comprises:

[0172] The first establishment module 100 is used to establish a manufacturing simulator for characterizing the optical proximity effect of the grayscale laser direct writing device and the nonlinear response of the photoresist, wherein the manufacturing simulator includes a low-pass filter model for simulating the spatial frequency response of the optical system and a nonlinear mapping model between the exposure dose of the photoresist and the development result;

[0173] A second establishment module 200 is used to establish a wave optics simulator, which is used to calculate optical distribution and optical performance based on the output of the manufacturing simulator in combination with wave optics theory, and predict the diffraction optical characteristics caused by the surface morphology of the photoresist;

[0174] A setting module 300 is used to set an initial exposure dose distribution as a manufacturing parameter according to a design goal, predict the initial surface morphology of the photoresist using a manufacturing simulator, and input the prediction result into a wave optics simulator to simulate and obtain a corresponding optical performance distribution;

[0175] An optimization module 400 is used to define the difference between the target optical distribution and the simulated optical distribution as a loss function, calculate the gradient of the loss function with respect to the manufacturing parameters using an automatic differentiation technique and a gradient-based optimization method, and optimize the manufacturing parameters;

[0176] Iteration module 500, used to repeat the optimization steps, gradually iterate the manufacturing parameters to gradually approach the target optical performance, until the loss function meets the predetermined performance index or reaches the set maximum number of iterations;

[0177] The manufacturing module 600 is used to use the optimized manufacturing parameters as input parameters and apply them to the grayscale laser direct writing equipment to perform actual processing to manufacture the diffractive optical element.

[0178] In order to implement the above embodiments, the present application also proposes an electronic device, comprising: a processor, and a memory communicatively connected to the processor; the memory stores computer-executable instructions; the processor executes the computer-executable instructions stored in the memory to implement the method provided by the above embodiments.

[0179] In order to implement the above embodiments, the present application also proposes a computer-readable storage medium, in which computer-executable instructions are stored. When the computer-executable instructions are executed by a processor, they are used to implement the methods provided by the above embodiments.

[0180] In order to implement the above embodiments, the present application also proposes a computer program product, including a computer program, which implements the methods provided by the above embodiments when executed by a processor.

[0181] The collection, storage, use, processing, transmission, provision and disclosure of user personal information involved in this application are in compliance with relevant laws and regulations and do not violate public order and good morals.

[0182] It should be noted that personal information from users should be collected for legitimate and reasonable purposes and should not be shared or sold outside of these legitimate uses. In addition, such collection / sharing should be carried out after receiving the user's informed consent, including but not limited to notifying the user to read the user agreement / user notice and sign the agreement / authorization including authorization of relevant user information before the user uses the function. In addition, any necessary steps should be taken to protect and safeguard access to such personal information data and ensure that others who have access to personal information data comply with its privacy policy and procedures.

[0183] The present application is expected to provide an implementation scheme for users to selectively block the use or access of personal information data. That is, the present disclosure is expected to provide hardware and / or software to prevent or block access to such personal information data. Once the personal information data is no longer needed, the risk can be minimized by limiting data collection and deleting the data. In addition, when applicable, such personal information is de-identified to protect the privacy of the user.

[0184] The storage medium mentioned above may be a read-only memory, a magnetic disk or an optical disk, etc. Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limiting the present application. A person of ordinary skill in the art may change, modify, replace and modify the above embodiments within the scope of the present application.

[0185] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this application can be executed in parallel, sequentially or in different orders, as long as the expected results of the technical solution of this application can be achieved, and this document is not limited here.

[0186] The above specific implementations do not constitute a limitation on the protection scope of this application. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of this application should be included in the protection scope of this application.

Claims

1. A method for designing and manufacturing a diffractive optical element based on grayscale laser direct writing lithography, characterized in that: The following steps are involved: Establishing a manufacturing simulator for characterizing the optical proximity effect of a grayscale laser direct writing device and the nonlinear response of a photoresist, the manufacturing simulator comprising a low-pass filter model for simulating the spatial frequency response of an optical system and a nonlinear mapping model between a photoresist exposure dose and a development result; Establishing a wave optics simulator, the wave optics simulator is used to calculate optical distribution and optical performance based on the output of the manufacturing simulator in combination with wave optics theory, and predict the diffraction optical properties caused by the surface morphology of the photoresist; According to the design goal, an initial exposure dose distribution is set as a manufacturing parameter, the manufacturing simulator is used to predict the initial surface morphology of the photoresist, and the prediction result is input into the wave optics simulator to simulate and obtain the corresponding optical performance distribution; defining the difference between the target optical distribution and the simulated optical distribution as a loss function, calculating the gradient of the loss function with respect to the manufacturing parameters by using automatic differentiation technology and a gradient-based optimization method, and optimizing the manufacturing parameters; Repeat the optimization steps and gradually iterate the manufacturing parameters to gradually approach the target optical performance until the loss function meets the predetermined performance index or reaches the set maximum number of iterations; The optimized manufacturing parameters are used as input parameters and applied to the grayscale laser direct writing equipment for actual processing to manufacture diffractive optical elements.

2. The method according to claim 1, characterized in that The manufacturing simulator for establishing the optical proximity effect and nonlinear response of the photoresist of the grayscale laser direct writing device comprises: Obtaining the modulation transfer function of the laser direct writing device lens, where the modulation transfer function is used to characterize the response characteristics of the optical system to different spatial frequencies and is used as an input of the low-pass filtering model; Characterize the nonlinear characteristics of the photochemical reaction and development process of the photoresist material through experiments, and obtain the mapping relationship between the exposure dose and the surface morphology after development; A polynomial or a preset fitting function is used to fit the experimental data related to the modulation transfer function and the nonlinear response characteristics, and a differentiable manufacturing simulator is jointly established. The manufacturing simulator can directly predict the optical performance indicators after processing and optimize the process parameters.

3. The method according to claim 2, characterized in that The step of obtaining the modulation transfer function of the lens of the laser direct writing device includes: A series of sinusoidal height profiles with different spatial frequencies were produced as basic patterns using grayscale laser direct writing technology; By analyzing the spatial frequency and amplitude changes of the above-mentioned sinusoidal pattern, the response characteristics of the optical system to different spatial frequencies are determined, and a modulation transfer function is generated; The modulation transfer function is used as a part of a low-pass filter model, and the optical proximity effect in the processing process is modeled in combination with the modulation transfer function.

4. The method according to claim 3, characterized in that: The method of experimentally characterizing the nonlinear characteristics of the photochemical reaction and development process of the photoresist material and obtaining the mapping relationship between the exposure dose and the surface morphology after development includes: A series of square patterns with different exposure doses were produced by grayscale laser direct writing technology; Measuring the surface depth of the square pattern after development to obtain nonlinear response data between exposure dose and surface topography; A nonlinear mapping model of exposure dose and surface structure depth after development is established based on experimental data. The nonlinear mapping model is a component of the manufacturing simulator and is used to further optimize the prediction accuracy of the photoresist surface morphology.

5. The method according to claim 4, characterized in that The method of calculating the gradient of the loss function with respect to the manufacturing parameters by using the automatic differentiation technique and the gradient-based optimization method comprises: Calculating the gradient of the loss function with respect to the manufacturing parameters based on the manufacturing simulator and the wave optics simulator using an automatic differentiation technique; Adopt a gradient-based optimization algorithm, combined with the Adam optimization algorithm to dynamically adjust the learning rate and momentum factor to improve the optimization convergence efficiency; The diffraction efficiency is incorporated into the objective function, and a penalty term is introduced to suppress the zero-order diffraction intensity, and the phase distribution is optimized to redistribute the energy to the desired diffraction order and improve the efficiency of the target diffraction order; Batch normalization is performed on parameter updates across multiple optimization iterations to improve the stability and convergence speed of model training.

6. The method according to claim 5, characterized in that Also includes: For large-scale or complex designs, a block optimization strategy is used during the optimization process to decompose the large design into multiple sub-problems; Use parallel computing technology to process multiple sub-problems simultaneously, and gradually refine the optimization results through multi-scale optimization methods; After the optimization is completed, the optimization results of the sub-problems are integrated to form a global design.

7. The method according to claim 6, characterized in that The method is applicable to manufacturing a microstructure surface and a continuous free-form surface with a predetermined optical function, and comprises: Microlens array for beam shaping and focusing; Diffraction gratings, used for spectral separation or wavelength selection; Phase plate, used for wavefront modulation; a multi-focus beam splitter for generating multiple focused beams; Computer generated holograms, used for three-dimensional image display or wavefront reconstruction; An extended depth of field imaging system for enhancing the depth of field of an imaging system; wavefront correctors, used to compensate for aberrations in optical systems; Continuous free-form optical surfaces for complex optical functional needs.

8. The method according to claim 7, characterized in that The method is applicable to two-photon laser direct writing equipment, including the following adjustments: Add a two-photon absorption effect model to the manufacturing simulator to characterize the impact of the two-photon excitation process on photoresist processing; Considering the formation process of the three-dimensional structure, the optimization strategy is adjusted to adapt to the three-dimensional processing capability of the two-photon direct writing equipment; The nonlinear response model is adjusted based on the characteristics of the two-photon laser direct writing device to ensure high resolution and processing accuracy of the structure.

9. An integrated device for designing and manufacturing diffractive optical elements based on grayscale laser direct writing lithography, characterized in that: include: A first establishment module is used to establish a manufacturing simulator for characterizing the optical proximity effect of the grayscale laser direct writing device and the nonlinear response of the photoresist, wherein the manufacturing simulator includes a low-pass filter model for simulating the spatial frequency response of the optical system and a nonlinear mapping model between the exposure dose of the photoresist and the development result; A second establishment module is used to establish a wave optics simulator, wherein the wave optics simulator is used to calculate optical distribution and optical performance based on the output of the manufacturing simulator in combination with wave optics theory, and predict diffraction optical properties caused by the surface morphology of the photoresist; A setting module is used to set the initial exposure dose distribution as a manufacturing parameter according to the design goal, use the manufacturing simulator to predict the initial surface morphology of the photoresist, and input the prediction result into the wave optics simulator to simulate and obtain the corresponding optical performance distribution; An optimization module, used for defining the difference between the target optical distribution and the simulated optical distribution as a loss function, calculating the gradient of the loss function with respect to the manufacturing parameters by using automatic differentiation technology and a gradient-based optimization method, and optimizing the manufacturing parameters; An iteration module is used to repeat the optimization steps and gradually iterate the manufacturing parameters to gradually approach the target optical performance until the loss function meets the predetermined performance index or reaches the set maximum number of iterations; The manufacturing module is used to use the optimized manufacturing parameters as input parameters and apply them to the grayscale laser direct writing equipment for actual processing to manufacture diffractive optical elements.

10. An electronic device, characterized in that: include: A processor, and a memory communicatively connected to the processor; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory to implement the method according to any one of claims 1 to 8.

Images