A four-step phase shift self-calibration method, device, photolithography system and storage medium of a digital micromirror device maskless photolithography system

Abstract

The application discloses a four-step phase shift self-calibration method, device, photolithography system and storage medium of a digital micromirror device maskless photolithography system, comprising the following steps: S1, camera geometry self-calibration; S2, four-step phase shift projection and phase calculation; S3, phase plane fitting; S4, affine, decoupling, magnification compensation and S5, result storage. The application connects the camera self-calibration based on the displacement table and the DMD geometry calibration based on the four-step phase shift into a one-key full-automatic process, does not require any external standard parts, greatly improves the calibration efficiency and consistency, eliminates the influence of the camera installation error on the calibration accuracy, eliminates the dependence on the full-field absolute phase expansion, significantly improves the robustness to the scenes of limited projection area, low modulation and local pollution, and stores the calibration results through the split objective and the split light source wavelength, effectively compensates the magnification difference caused by the chromatic aberration of different wavelengths, and improves the overlay accuracy.

Description

Technical Field

[0001] This invention relates to the fields of maskless lithography (direct-write lithography) and digital light processing projection imaging technology, and particularly to a four-step phase-shift self-calibration method, apparatus, lithography system, and storage medium for a maskless lithography system of a digital micromirror device. Background Technology

[0002] In maskless lithography systems based on digital micromirror devices (DMDs), the DMD projects the image to be exposed onto the workpiece surface, and after being scaled by the objective lens, the exposure is completed on the photoresist. There is no need to make a physical mask. It has the advantages of flexibility, low cost and short cycle time, and is widely used in the manufacturing of printed circuit boards, microfluidic chips, photonic crystals, and advanced semiconductor packaging.

[0003] To ensure the positional accuracy, dimensional accuracy, and multi-field stitching / overlay accuracy of the exposed images, the system must accurately establish the geometric relationships between several coordinate systems, including: the rotation and magnification relationship between the DMD image plane and the downward (measurement / alignment) camera image plane; the pixel equivalent and rotation relationship between the camera image plane and the physical coordinate system of the motion stage; and the pixel equivalent, rotation, and center offset relationship between the DMD image plane and the upward (alignment) camera image plane. These relationships change with factors such as objective lens magnification switching, light source wavelength (exposure ultraviolet light and alignment / guide light have different magnifications due to chromatic aberration), setup and adjustment conditions, and temperature drift, requiring recalibration during lens changes, light source changes, or periodic maintenance. Therefore, a high-precision, fully automated geometric self-calibration method that eliminates the need for external standard parts is crucial in this field.

[0004] The calibration methods in the prior art mainly include:

[0005] (1) Structured light / projector-camera phase-shift calibration technology. Sinusoidal fringes are projected by a projector (or DMD), and the camera acquires the data. The phase is then determined pixel-by-pixel using a phase-shifting method. The correspondence between camera pixels and projector pixels is established through phase analysis. The geometric relationship between the camera and projector is then determined using Zhang Zhengyou camera calibration or a stereo (pinhole inverse camera) model. This technology primarily serves 3D topography measurement and reconstruction. Typically, a four-step phase shift (phase shift amounts of 0, π / 2, π, and 3π / 2) is employed. Unwrap the phase, where I1, I2, I3, and I4 are the gray values ​​of the striped images acquired when the phase shift is 0, π / 2, π, and 3π / 2, respectively.

[0006] (2) Phase-Shifting Profilometry. Sine fringes are projected and acquired step by step according to phase shift. After unwrapping the phase pixel by pixel, absolute phase unfolding is performed in the entire field of view (multi-frequency heterodyne, Gray code or branch cutting method, etc.). Then, the three-dimensional height of the side is obtained by triangulation. It is a general method for three-dimensional measurement.

[0007] (3) Machine vision displacement stage / hand-eye calibration technology. The image formed by the camera is associated with the coordinate system of the motion displacement stage: the displacement stage is controlled to move a known distance, the pixel displacement of the reference feature point in the image is measured, the pixel equivalent is obtained by "pixel equivalent = known physical displacement / pixel displacement", and the rotation angle of the camera relative to the displacement stage is obtained by the arctangent of the displacement vector. No calibration plate is required.

[0008] (4) Camera-DMD self-calibration based on two-dimensional marker points. Several discrete two-dimensional markers (such as crosses or dot matrix) are projected onto the DMD. The camera detects the centroid of each marker. Then, the affine or homography relationship between the camera and the DMD is fitted by the correspondence of a small number of points. No external standard parts are required. It is a relatively close and simplified self-calibration scheme.

[0009] (5) Geometric calibration of the in-machine camera of the DMD direct writing device. A calibration camera and motion mechanism are set in the direct writing device to acquire images of the patterns projected by each projector and calibrate their position, alignment, focus, intensity, dead pixels, etc., mostly based on the image / reference point (such as erasable photochromic target).

[0010] However, the above solution still has the following technical shortcomings:

[0011] (1) Reliance on external standard parts. Traditional and mainstream fringe projection / camera calibration usually requires standard reticles, standard gratings, checkerboard calibration plates or standard films with known patterns. Manual loading, focusing and image finding are required, which is inefficient and introduces clamping errors. Different magnifications often require different standard parts, which increases system complexity and maintenance costs.

[0012] (2) Absolute phase unfolding is required across the entire field of view. Existing phase shift calibration methods often require absolute phase unfolding across the entire field of view before triangulation or homography. This process is not robust to low-profile regions, incomplete projection, stray light, and other situations, and is prone to phase unfolding errors, leading to calibration failure or decreased accuracy.

[0013] (3) Coupling of camera rotation and DMD rotation. Existing projector-camera calibration combines the installation orientation of the camera and the projector into a single homography or affine matrix, which cannot separate the DMD's own installation rotation error from the camera's own installation rotation error; directly using the stripe tilt angle in the image as the DMD rotation angle will introduce camera installation error and affect the overlay accuracy.

[0014] (4) Oriented towards three-dimensional reconstruction rather than in-plane calibration of lithography. Existing phase shift calibration is mostly used for three-dimensional topography measurement. Its calibration products are three-dimensional reconstruction parameters, and do not directly provide the DMD torsion rotation angle and magnification and wavelength compensation required for maskless lithography. It is not suitable for the in-plane calibration requirements of lithography systems.

[0015] (5) Failure to distinguish the magnification difference of wavelength. The same objective lens has different magnification at the exposure wavelength and the alignment wavelength due to chromatic difference. If compensation is not made separately, systematic errors will be introduced during alignment and overlay, reducing the overlay accuracy.

[0016] (6) Inherent drawbacks of the two-dimensional marker point method. Sparse sampling, with only a few point pairs constraining affine parameters, results in a large fitting variance; the centroid positioning accuracy is affected by image noise, defocus, uneven illumination, and marker distortion, and a single point anomaly can significantly skew the fitting result; the magnification sensitivity is limited by the spacing between marker points, making it difficult to distinguish the sub-0.1% magnification compensation required for lithographic overlay; there are problems with marker recognition and matching; local dust spots or contamination can cause the corresponding markers to fail, resulting in poor robustness. Summary of the Invention

[0017] The purpose of this invention is to provide a four-step phase-shift self-calibration method, apparatus, lithography system, and storage medium for a maskless lithography system for digital micromirror devices, addressing the shortcomings of existing technologies, so as to achieve fully automatic and highly robust geometric self-calibration without the need for external standard parts.

[0018] To solve the above-mentioned technical problems, the present invention adopts the following technical solution:

[0019] A four-step phase-shift self-calibration method for a maskless lithography system for digital micromirror devices, characterized by comprising the following steps:

[0020] S1. Camera geometry self-calibration:

[0021] S1.1 Control the digital micromirror device to project the marking pattern, acquire the first image at the reference position of the motion displacement stage through the image acquisition device and extract the reference feature points, and record the reference coordinates of the motion displacement stage;

[0022] S1.2 Control the motion displacement stage to move a known distance in multiple directions and return, acquire images after each movement and extract reference feature points;

[0023] S1.3 Calculate the pixel equivalent of the image acquisition device based on the known distance and pixel displacement in each direction, and combine the pixel equivalents in each direction to obtain the pixel equivalent of the image acquisition device;

[0024] S1.4 Calculate the rotation angle of the image acquisition device relative to the motion stage based on the displacement direction in each direction and the pixel displacement direction. ;

[0025] S2. Four-step phase-shifting projection and phase calculation:

[0026] S2.1 controls the digital micromirror device to project sinusoidal fringe patterns with a period of T and phase shifts of 0, π / 2, π, and 3π / 2 respectively along the first and second directions, and acquires the corresponding fringe images through the image acquisition device;

[0027] S2.2 Calculates the wrapping phase pixel by pixel using the four-step phase shift formula. Harmony and reconciliation system B;

[0028] S3, Phase plane fitting:

[0029] S3.1 Select effective pixels according to modulation index B, expand the neighborhood phase of the effective pixels using the phase plane model, then perform least squares plane fitting and iterative weighting until convergence, and output the phase gradient in the first direction and the phase gradient in the second direction.

[0030] S3.2 does not perform full-field absolute phase expansion throughout the entire process;

[0031] S4, Affine, Decoupling and Magnification Compensation:

[0032] S4.1 Calculate the affine transformation parameters from the image plane of the image acquisition device to the image plane of the digital micromirror device based on the phase gradient and period T, including the comprehensive rotation angle of the image acquisition device relative to the digital micromirror device. And scaling in both directions , ;

[0033] S4.2 According to Calculate the torsional rotation angle of the digital micromirror device ;

[0034] S4.3 Based on the pixel equivalent p of the image acquisition device cam0 The magnification compensation coefficient k in both directions is calculated based on the scaling factor and the theoretical magnification M of the current objective lens. X k Y ;

[0035] S5. Result Storage:

[0036] Correcting the rotation angle of the digital micromirror device and magnification compensation coefficient k X k Y Stored in association with the current objective magnification and the current light source wavelength.

[0037] This self-calibration method integrates stage-based camera self-calibration with four-step phase-shift DMD geometric calibration into a one-click fully automated process, requiring no external standard parts. This significantly improves calibration efficiency and consistency. This method decouples camera rotation from DMD rotation, eliminating the impact of camera installation errors on calibration accuracy. Furthermore, it directly calculates affine transformations using the full-field phase gradient, eliminating the dependence on full-field absolute phase expansion. It can also achieve dense sub-pixel measurements on hundreds of thousands of effective pixels across the entire field, significantly improving robustness to scenarios with limited projection areas, low-profile conditions, and local contamination. By storing calibration results separately for objective lenses and light source wavelengths, it effectively compensates for magnification differences caused by chromatic aberration at different wavelengths, thereby improving overlay accuracy.

[0038] A four-step phase-shift self-calibration device for a maskless lithography system of a digital micromirror device, used to implement the four-step phase-shift self-calibration method for a maskless lithography system of a digital micromirror device as described above, characterized in that it includes:

[0039] The image acquisition and feature extraction module is used to control the image acquisition device to acquire images and extract the positions of reference feature points;

[0040] The displacement stage motion control module is used to control the displacement stage to move a known distance in multiple directions and return.

[0041] The pattern projection module is used to control the projection of marking patterns and four-step phase-shifting sinusoidal fringe patterns onto the digital micromirror device;

[0042] The phase calculation module is used to calculate the wrapping phase and modulation degree pixel by pixel based on the acquired phase shift fringe pattern;

[0043] The phase plane fitting module is used to select effective pixels based on the modulation degree, expand the neighborhood phase of the effective pixels using the phase plane model, and then perform least squares plane fitting and iterative weighting until convergence, outputting the phase gradient.

[0044] The parameter calculation module is used to calculate the affine transformation parameters based on the phase gradient. Calculate the torsional rotation angle of the digital micromirror device and the magnification compensation coefficient;

[0045] The result storage module is used to associate and store the calibration results with the current objective lens magnification and the current light source wavelength.

[0046] The device achieves full automation of the above method in a modular manner. The data flow between modules is clear and the coupling is low, which makes it easy to integrate and transfer values ​​in different types of maskless lithography systems. At the same time, the device does not require manual intervention or external standard parts, and can achieve one-click calibration, which significantly reduces the maintenance cost and calibration time of the equipment.

[0047] A maskless photolithography system for digital micromirror devices, characterized in that it comprises:

[0048] Digital micromirror devices, as spatial light modulators, are used to project marked patterns and four-step phase-shifted sinusoidal fringe patterns;

[0049] Image acquisition device, used to acquire images of the projected pattern of a digital micromirror device;

[0050] A motion displacement stage is used to support the workpiece and serve as the physical length reference of the system;

[0051] Objective lens and objective lens switching device, providing multiple switchable magnifications;

[0052] The light source device includes an exposure ultraviolet light source and an alignment / guiding light source;

[0053] A control and calibration device for performing a four-step phase-shift self-calibration method for a maskless lithography system of a digital micromirror device as described above.

[0054] This lithography system integrates a self-calibration function, which realizes fully automatic calibration of geometric relationships within the system without the need for external calibration instruments or disassembly and assembly of standard parts. The calibration operation can be completed during equipment operation intervals without affecting normal production cycle. Furthermore, the calibration results can be automatically loaded with objective lens and light source switching, ensuring the long-term stability of the equipment under different working conditions.

[0055] A computer-readable storage medium storing computer instructions, characterized in that: when the computer program is executed by a processor, it implements the four-step phase-shift self-calibration method of a maskless lithography system for a digital micromirror device as described above.

[0056] The present invention, by adopting the above-described technical solution, has the following beneficial effects:

[0057] 1. This invention requires no external standard parts. It uses the displacement stage as the length reference and the DMD self-projection stripes as the angle and magnification reference, eliminating the need for standard parts loading and focusing, reducing costs, and improving efficiency and consistency.

[0058] 2. This invention integrates camera geometry self-calibration and DMD four-step phase-shift calibration into a single process, eliminating the need for manual point-by-point picking or manual stage movement, achieving one-click fully automatic operation.

[0059] 3. This invention is achieved through... Eliminating the camera's own installation and rotation errors improves the accuracy of DMD torsion angle calibration, thus decoupling the DMD rotation angle from the camera rotation angle.

[0060] 4. This invention can avoid full-field absolute phase expansion. The phase plane fitting adopts the median adaptive modulation threshold, neighborhood phase expansion and iterative weighted least squares, and only performs plane fitting on the local phase gradient, which is robust to scenarios with limited projection area and low-frequency modulation.

[0061] 5. This invention achieves full-field dense sub-pixel measurement, continuously encodes the phase to the entire field of view with four-step phase shift, increases the sample size from a few marker points to hundreds of thousands of effective pixels across the entire field, achieves overfitting, is insensitive to single-point errors, dust spots, contamination, and local defocus, and can distinguish sub-0.1% magnification differences.

[0062] 6. This invention directly produces two in-plane corrections for photolithography, and compensates for magnification and chromatic aberration according to objective lens and wavelength, thereby improving alignment and overlay accuracy.

[0063] 7. This invention directly images the DMD under a conjugate optical path and reuses the bottom results, while simultaneously calibrating the pixel equivalent, rotation angle, and center offset of the aligned camera. Attached Figure Description

[0064] The present invention will be further described below with reference to the accompanying drawings:

[0065] Figure 1 This is a flowchart of a four-step phase-shift self-calibration method, apparatus, lithography system, and self-calibration method in a storage medium for a maskless lithography system of a digital micromirror device according to the present invention.

[0066] Figure 2 This is a schematic diagram of the photolithography system in this invention;

[0067] Figure 3 This is a flowchart of the four-step phase-shift self-calibration process for the bottom camera in this invention;

[0068] Figure 4 This is a schematic diagram showing the reverse displacement of feature points within the camera image plane as the displacement stage moves a known distance d along ±X / ±Y in this invention.

[0069] Figure 5 This is a schematic diagram of four phase-shifted fringes in each of the X and Y directions in this invention;

[0070] Figure 6 This is a flowchart of step S3, phase plane fitting, in this invention;

[0071] Figure 7 This is a schematic diagram of the camera image plane to the DMD image plane in this invention. The included angle has been enlarged, and the DMD image plane is reduced in size by the objective lens, occupying a small proportion in the camera's field of view.

[0072] Figure 8 This is a flowchart of the self-calibration of the conjugate optical path of the upward-looking camera in this invention.

[0073] Wherein: 1-DMD; 2-Light source; 2a-Exposure ultraviolet light source; 2b-Alignment / guiding light source; 3-Objective lens; 4-Downward camera; 5-Upward camera; 6-Motion displacement stage; 7-Workpiece; 8-Control and calibration device. Detailed Implementation

[0074] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0075] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0076] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion.

[0077] like Figure 1 As shown, this invention provides a four-step phase-shift self-calibration method for a maskless lithography system for digital micromirror devices, comprising the following steps:

[0078] S1. Camera geometry self-calibration:

[0079] S1.1 Control the digital micromirror device to project the marking pattern, acquire the first image at the reference position of the motion displacement stage through the image acquisition device and extract the reference feature points, and record the reference coordinates of the motion displacement stage;

[0080] The marking pattern is a cross shape. The positions of the reference feature points are extracted using the image moment centroid method, and the centroid coordinates are calculated using the following formula (c x ,c y ):

[0081] ;

[0082] Among them, M 00 M 10 M 01 The image moments are used to mark the pattern. The cross-shaped pattern has good anisotropic symmetry, and when combined with the image moment centroid method, it can achieve sub-pixel-level feature localization accuracy. It is also simple to calculate and has strong noise resistance, providing a reliable foundation for the accurate calculation of subsequent pixel equivalents and rotation angles.

[0083] S1.2 Control the motion displacement stage to move a known distance in multiple directions and return, acquire images after each movement and extract reference feature points;

[0084] Multiple directions include four directions: +X, -X, +Y, and -Y. The motion displacement stage moves a known distance d along +X, -X, +Y, and -Y in sequence and returns to the reference position.

[0085] The pixel equivalent in each direction is calculated using the following formula:

[0086] ;

[0087] Among them, L i Let be the pixel displacement magnitude in the i-th direction;

[0088] The pixel equivalent p of the image acquisition device cam0 Take the average value of the pixel equivalents in the four directions;

[0089] Rotation angle of the image acquisition device relative to the motion displacement stage The displacement directions in the four directions are uniformly converted into the angle α of the +X axis of the motion displacement stage in the image plane of the image acquisition device. i According to the circumference average formula The method employs symmetrical movement in four directions and averages the results, which effectively counteracts the effects of anisotropic errors in the motion displacement stage and straightness errors in the guide rail. The circumferential averaging formula improves the robustness of the rotation angle calibration and eliminates the need for additional angle measuring instruments.

[0090] S1.3 Calculate the pixel equivalent of the image acquisition device based on the known distance and pixel displacement in each direction, and combine the pixel equivalents in each direction to obtain the pixel equivalent of the image acquisition device;

[0091] S1.4 Calculate the rotation angle of the image acquisition device relative to the motion stage based on the displacement direction in each direction and the pixel displacement direction. ;

[0092] S2. Four-step phase-shifting projection and phase calculation:

[0093] S2.1 controls the digital micromirror device to project sinusoidal fringe patterns with a period of T and phase shifts of 0, π / 2, π, and 3π / 2 respectively along the first and second directions, and acquires the corresponding fringe images through the image acquisition device;

[0094] The first direction is the X direction, and the sinusoidal fringes in the X direction are vertical fringes. The second direction is the Y direction, and the sinusoidal fringes in the Y direction are horizontal fringes. By using orthogonal vertical and horizontal fringes, the phase gradients in the two orthogonal directions X and Y can be independently calibrated, thereby completely resolving all parameters of the two-dimensional affine transformation (including rotation, scaling, and shearing), with complete information and mutual decoupling.

[0095] S2.2 Calculates the wrapping phase pixel by pixel using the four-step phase shift formula. Harmony and reconciliation system B;

[0096] Wrap phase Calculate pixel-by-pixel using the following formula:

[0097] ;

[0098] Modulation B is calculated pixel-by-pixel using the following formula:

[0099] ;

[0100] Wherein, I1, I2, I3, and I4 are the grayscale values ​​of the striped images acquired when the phase shift is 0, π / 2, π, and 3π / 2, respectively. The four-step phase-shifting algorithm has the advantage of being insensitive to changes in background light and contrast, and is computationally efficient. The modulation index, as a quantitative indicator of reliability, provides a basis for the adaptive selection of subsequent effective pixels.

[0101] S3, Phase plane fitting:

[0102] S3.1 Select effective pixels according to modulation index B, expand the neighborhood phase of the effective pixels using the phase plane model, then perform least squares plane fitting and iterative weighting until convergence, and output the phase gradient in the first direction and the phase gradient in the second direction.

[0103] Effective pixels are selected using an adaptive modulation threshold, B. thr According to the median B of the adjustment system med Sure:

[0104] ;

[0105] Where ρ is a preset proportionality coefficient, and B floor As a preset lower limit threshold for modulation, only B≥B is retained. thr The pixels are selected as valid pixels. An adaptive threshold based on the median modulation index is adopted, which can dynamically select criteria according to the actual modulation index distribution of the image. It has good adaptability to different lighting conditions and different objective magnifications. The dual threshold design (proportional coefficient and absolute lower limit) takes into account both relative screening and absolute quality control, which can effectively remove low signal-to-noise ratio pixels and background pixels, improve the accuracy and reliability of subsequent plane fitting, and does not rely on the absolute consistency of the modulation index of the entire field.

[0106] The phase plane model is as follows:

[0107] ;

[0108] Where (u0, v0) are the coordinates of the image center, a and b are the phase gradients along the u and v directions, c is a constant term, and (u, v) are the coordinates of the current pixel, where u is the x-coordinate (column coordinate, unit in pixels) of the pixel in the camera image plane coordinate system, and v is the y-coordinate (row coordinate, unit in pixels) of the pixel in the camera image plane coordinate system. Establishing a phase plane model with the image center as the origin can reduce the correlation between plane fitting parameters and improve vertical stability. By combining neighborhood phase expansion with least squares plane fitting, there is no need to perform full-field absolute phase expansion, which greatly reduces the algorithm complexity and avoids the problem of expansion failure caused by error accumulation or low-profile regions in large-scale phase expansion.

[0109] S3.2 does not perform full-field absolute phase expansion throughout the entire process;

[0110] S4, Affine, Decoupling and Magnification Compensation:

[0111] S4.1 Calculate the affine transformation parameters from the image plane of the image acquisition device to the image plane of the digital micromirror device based on the phase gradient and period T, including the comprehensive rotation angle of the image acquisition device relative to the digital micromirror device. And scaling in both directions , ;

[0112] S4.2 According to Calculate the torsional rotation angle of the digital micromirror device ;

[0113] S4.3 Based on the pixel equivalent p of the image acquisition device cam0 The magnification compensation coefficient k in both directions is calculated based on the scaling factor and the theoretical magnification M of the current objective lens. X k Y ;

[0114] In the affine, decoupling, and magnification compensation steps:

[0115] make , , , ;where a X The gradient coefficient of the phase along the u direction, b, is obtained by fitting the sinusoidal fringes in the first direction (X direction) to the phase plane. X The gradient coefficient of the phase along the v direction, a, is obtained by fitting the sinusoidal fringes in the first direction (X direction) to the phase plane. Y The gradient coefficients of the phase along the u direction, b, are obtained by fitting the sinusoidal fringes in the second direction (Y direction) to the phase plane. Y The gradient coefficients of the phase along the v direction are obtained by fitting the phase plane to the sinusoidal fringes in the second direction (Y direction); phase plane fitting is performed on the two sets of fringes in the first and second directions respectively, yielding a pair of gradient coefficients (a) for each.X ,b X ) and (a Y ,b Y Together, they form the affine mapping matrix from the camera image plane to the DMD image plane. .

[0116] Overall rotation angle of the image acquisition device relative to the digital micromirror device ;

[0117] The scaling factor in both directions is: , ;

[0118] Theoretical pixel size of digital micromirror devices , where p base This is the reference pixel size for digital micromirror devices;

[0119] Actual pixel size of digital micromirror devices , ;

[0120] The magnification compensation coefficient is: , Using an affine mapping model, the complete affine transformation parameters (including rotation and independent scaling in both directions) can be directly derived from the phase gradients in two orthogonal directions. The physical meaning is clear, the calculation is concise, and... By decoupling the DMD rotation angle from the camera rotation angle, the true installation rotation angle of the DMD itself is obtained, eliminating the influence of camera rotation error on the DMD torsion angle calibration in traditional methods. At the same time, since the phase gradient is obtained by fitting the plane of hundreds of thousands of pixels across the entire field, it is not sensitive to single-point errors, dust spots, and local defocus, and can distinguish magnification differences of sub-0.1%, meeting the high precision requirements of photolithography overlay.

[0121] S5. Result Storage:

[0122] Correcting the rotation angle of the digital micromirror device and magnification compensation coefficient k X k Y Stored in association with the current objective magnification and the current light source wavelength.

[0123] During associated storage, the corresponding magnification compensation coefficients are stored separately for different light source wavelengths under the same objective magnification. The light source wavelengths include the exposure ultraviolet light wavelength and the alignment / guide light wavelength. By storing the magnification compensation coefficients for different wavelengths separately, the magnification difference caused by color difference between the exposure ultraviolet light and the alignment / guide light for the same objective is eliminated, ensuring the consistency of the exposure pattern and the alignment pattern in terms of spatial scale, and significantly improving the alignment and overlay accuracy.

[0124] It also includes the upward camera calibration procedure:

[0125] Under the conjugate optical path, the digital micromirror device is directly imaged by the upward-looking camera;

[0126] The digital micromirror device projects local rectangular and cross patterns in the central region. Images are acquired by an upward-looking camera, and the centroids of the two patterns are calculated. The average of the two centroids is compared with the center of the pattern to obtain the center offset of the upward-looking camera. , ;

[0127] The digital micromirror device is controlled to project four-step phase-shifting fringes in the central region, and the phase gradients a and b are obtained by phase plane fitting.

[0128] Read the calibrated magnification compensation coefficient under the current objective lens guide light. and the torsional rotation angle of digital micromirror devices ;

[0129] Based on magnification compensation coefficient Calculate the actual pixel size of the digital micromirror device under the current objective lens. Then calculate the pixel equivalent of the upward-looking camera. Based on phase gradients a and b and torsional rotation angle Calculate the rotation angle of the upward-looking camera ;

[0130] The pixel equivalent of the upward-looking camera Rotation angle and center offset , Stored in association with the current objective lens.

[0131] Pixel equivalent of upward-looking camera Calculate as follows:

[0132] ;

[0133] ;

[0134] Where, p base Let M be the reference pixel size of the digital micromirror device (DMD), T be the theoretical magnification of the current objective lens, T be the period of the four-step phase-shifting fringes, and a be the phase gradient output by the phase plane fitting. The theoretical pixel size of the DMD is corrected to the actual pixel size by reusing the magnification compensation coefficient calibrated at the bottom. Then, the pixel equivalent of the upward-looking camera is calculated by combining this with the phase gradient. No additional length reference or standard is required; the calculation method can be directly derived from the phase gradient of the phase-shifting fringes, avoiding the limitations of traditional methods that rely on measurements from known spacing calibration points.

[0135] The calibration of the upward-looking camera under the conjugate optical path reuses the DMD torsion rotation angle and magnification compensation coefficient that have been calibrated at the bottom, without the need for additional standard parts. By combining local rectangles and cross patterns to obtain the centroid mean, the influence of single pattern centroid extraction error on center offset is effectively reduced. This method makes the pixel equivalent, rotation angle and center offset of the upward-looking camera all calibrable, providing complete geometric parameters for high-precision upward-looking alignment.

[0136] This self-calibration method integrates stage-based camera self-calibration with four-step phase-shift DMD geometric calibration into a one-click fully automated process, requiring no external standard parts. This significantly improves calibration efficiency and consistency. This method decouples camera rotation from DMD rotation, eliminating the impact of camera installation errors on calibration accuracy. Furthermore, it directly calculates affine transformations using the full-field phase gradient, eliminating the dependence on full-field absolute phase expansion. It can also achieve dense sub-pixel measurements on hundreds of thousands of effective pixels across the entire field, significantly improving robustness to scenarios with limited projection areas, low-profile conditions, and local contamination. By storing calibration results separately for objective lenses and light source wavelengths, it effectively compensates for magnification differences caused by chromatic aberration at different wavelengths, thereby improving overlay accuracy.

[0137] A four-step phase-shift self-calibration device for a maskless lithography system of a digital micromirror device, used to implement the four-step phase-shift self-calibration method of a maskless lithography system of a digital micromirror device as described above, comprising:

[0138] The image acquisition and feature extraction module is used to control the image acquisition device to acquire images and extract the positions of reference feature points;

[0139] The displacement stage motion control module is used to control the displacement stage to move a known distance in multiple directions and return.

[0140] The pattern projection module is used to control the projection of marking patterns and four-step phase-shifting sinusoidal fringe patterns onto the digital micromirror device;

[0141] The phase calculation module is used to calculate the wrapping phase and modulation degree pixel by pixel based on the acquired phase shift fringe pattern;

[0142] The phase plane fitting module is used to select effective pixels based on the modulation degree, expand the neighborhood phase of the effective pixels using the phase plane model, and then perform least squares plane fitting and iterative weighting until convergence, outputting the phase gradient.

[0143] The parameter calculation module is used to calculate the affine transformation parameters based on the phase gradient. Calculate the torsional rotation angle of the digital micromirror device and the magnification compensation coefficient;

[0144] The result storage module is used to associate and store the calibration results with the current objective lens magnification and the current light source wavelength.

[0145] The device achieves full automation of the above method in a modular manner. The data flow between modules is clear and the coupling is low, which makes it easy to integrate and transfer values ​​in different types of maskless lithography systems. At the same time, the device does not require manual intervention or external standard parts, and can achieve one-click calibration, which significantly reduces the maintenance cost and calibration time of the equipment.

[0146] A maskless lithography system for digital micromirror devices includes:

[0147] Digital micromirror devices, as spatial light modulators, are used to project marked patterns and four-step phase-shifted sinusoidal fringe patterns;

[0148] An image acquisition device is used to acquire images of the projected pattern of the digital micromirror device (DMD). The device includes a downward-looking camera and a top-looking camera. The downward-looking camera acquires images of reference feature points on the moving stage and the fringe image projected by the DMD. The top-looking camera directly images the DMD under a conjugate optical path. Through the division of labor between the downward-looking and top-looking cameras, comprehensive calibration of the bottom measuring camera and the top alignment camera can be completed in the same calibration process. The downward-looking camera calibrates the geometric parameters of the DMD using the moving stage as a length reference, while the top-looking camera reuses the calibration results to complete its own calibration. Their collaborative work ensures the uniformity and accuracy of all coordinate systems in the system. The DMD is a binary digital micromirror device, which presents 8-bit grayscale sinusoidal fringes through static projection by modulating the time duty cycle or dithering of each micromirror. This fully utilizes the high-speed flipping advantage of the DMD, and the static projection method avoids the impact of motion blur during dynamic scanning on calibration accuracy, ensuring the clarity and phase accuracy of the fringe image.

[0149] A motion displacement stage is used to support the workpiece and serve as the physical length reference of the system;

[0150] Objective lens and objective lens switching device, providing multiple switchable magnifications;

[0151] The light source device includes an exposure ultraviolet light source and an alignment / guiding light source;

[0152] A control and calibration device for performing a four-step phase-shift self-calibration method for a maskless lithography system of a digital micromirror device as described above.

[0153] This lithography system integrates a self-calibration function, which realizes fully automatic calibration of geometric relationships within the system without the need for external calibration instruments or disassembly and assembly of standard parts. The calibration operation can be completed during equipment operation intervals without affecting normal production cycle. Furthermore, the calibration results can be automatically loaded with objective lens and light source switching, ensuring the long-term stability of the equipment under different working conditions.

[0154] like Figure 2As shown, the lithography system specifically includes: a DMD1, which serves as a spatial light modulator; one implementation is a binary DMD, which presents 8-bit grayscale sinusoidal fringes through static projection by modulating the time duty cycle or dithering of each micromirror; a light source 2, which includes an exposure ultraviolet light source 2a and an alignment / guiding light source 2b, the two having different wavelengths; an objective lens and an objective lens wheel 3, providing multiple switchable magnifications M; a downward-looking camera 4 and an upward-looking camera 5, which can directly project images onto the DMD through conjugate optical paths; a motion displacement stage 6, including X / Y axes, serving as the physical length reference of the system; a workpiece 7; and a control and calibration device 8, which executes a self-calibration process and persistently saves the calibration results according to the objective lens and light source categories.

[0155] like Figure 4 The diagram shows the self-calibration of the four-directional geometric displacement of the displacement stage in step S1 of this invention. A DMD projects a crosshair, and a downward-looking camera captures an image at a reference position. After binarization, the centroid of the bright area is determined from the image moments as the reference feature point, and the reference coordinates of the displacement stage are recorded. The displacement stage moves a known distance d along +X, -X, +Y, and -Y sequentially and returns. The pixel displacement magnitude L in direction i is... i Pixel equivalent The camera pixel equivalent is taken as the average of four directions; the four displacement directions are uniformly converted into the angle α of the displacement stage + X-axis in the image plane. i The average rotation angle of the camera is obtained by circumference. This is to avoid the angle jumping around ±π.

[0156] like Figure 5 The diagram shows four phase-shifted fringes in each of the X and Y directions of this invention. For both the X and Y directions, the DMD projects four sinusoidal fringes (vertical fringes in the X direction and horizontal fringes in the Y direction) with a period of T and phase shifts of 0, π / 2, π, and 3π / 2 respectively, using static grayscale. The camera captures four images accordingly. The wrapping phase is calculated pixel-by-pixel using a four-step phase shift formula. With adjustment system .

[0157] like Figure 6 The diagram shows a flowchart of step S3, phase plane fitting, of the present invention. The median B of the modulation index is taken. med Adaptive threshold (For example, ρ=0.3, B) floor =5), only keep B≥B thr The number of valid pixels is considered; if the number of valid pixels is too low, the test fails. (Using a phase plane model) After performing neighborhood phase expansion on the effective pixels, least squares plane fitting is performed and iteratively weighted until convergence is obtained to obtain the phase gradients a and b in that direction; no full-field absolute phase expansion is performed throughout the process.

[0158] like Figure 7The diagram shown illustrates the affine mapping and rotational decoupling from the camera image plane to the DMD image plane according to the present invention. Let... , , , The scaling factor in both directions is: , Camera rotation angle relative to DMD DMD torsional rotation angle Theoretical DMD pixel size , where p base Reference pixel size for digital micromirror devices; actual DMD pixel size , The magnification compensation coefficient is: , Similarly, they can be stored according to the wavelength of the objective lens and the light source. With (k) X ,k Y ).

[0159] like Figure 8 The diagram shows a flowchart of the self-calibration process of the conjugate optical path of the upward-looking camera according to the present invention. The camera directly images the DMD using the conjugate optical path; a local rectangle and a cross are projected onto the center of the DMD, and the mean of the two centroids is calculated and compared with the image center to obtain the center offset. , Four-step phase-shifting fringes are projected locally at the center of the DMD, and gradients a and b are obtained by phase plane fitting; the bottom-calibrated fringes are reused. Calculate the pixel equivalent of the aligned camera based on the wavelength magnification compensation. With rotation angle Associated with the current objective lens storage.

[0160] A computer-readable storage medium storing computer instructions, wherein the computer program, when executed by a processor, implements the four-step phase-shift self-calibration method of a maskless lithography system for a digital micromirror device as described above.

[0161] Example 1 (Bottom Measurement Camera Self-Calibration)

[0162] This embodiment is as follows: Figure 3 The system shown executes a four-step phase-shift self-calibration process for the bottom measurement camera. The system is as follows: Figure 2 The DMD is a binary device that statically presents 8-bit grayscale sinusoidal stripes using time duty cycle / grayscale dithering. The nominal physical pixel size of the downward-looking camera is p. cam0 =1.85µm, adaptive threshold set to ρ=0.3, B floor =5, the specific steps are as follows:

[0163] Step 1: Preparation and Verification: Select the objective lens (magnification M) and calibration light source (guide light or ultraviolet light), verify that the camera, DMD, and X / Y axis are ready, and turn on and stabilize the light source.

[0164] Step 2: Camera Geometric Self-Calibration: The DMD projects a crosshair, the downward-looking camera acquires an image at the reference position, and binarizes it using the Otsu's method. The image moments are then analyzed. 00 M 10 M 01 Calculate the centroid Using the reference feature point, record the absolute coordinates (X0, Y0) of the displacement stage reference; move it sequentially along +X, -X, +Y, -Y by a known distance d (e.g., d = 0.5 mm, which can be adjusted according to the field of view), and after each stabilization, capture the image, extract the centroid, and move it back d to return to the reference; the pixel displacement in each direction is used to determine the reference point. , Length L i Calculate pixel equivalent Take the average of the four directions as p cam If the pixel displacement in a certain direction is too small, an error message will be displayed indicating that d should be increased; the angle α should be calculated according to the direction. i Average of the back circumference .

[0165] Step 3, precise repositioning: Use the absolute motion command to move the displacement stage back to (X0,Y0), reset and clear the DMD.

[0166] Step 4, Four-step phase-shift fringe projection and acquisition: Let the period be T (DMD pixels, e.g., T=32), and perform phase shift analysis in the X and Y directions. Generate grayscale stripes, X direction , Y direction DMD static projection, camera captures frame by frame and converts to grayscale, four frames each for X and Y.

[0167] Step 5, Phase and Modulation Calculation: Pixel-by-pixel , The phase diagrams and modulation diagrams in the X and Y directions are obtained.

[0168] Step 6, Phase plane fitting: Take B med With threshold B thr , retain valid pixels; take the initial value of the gradient from the median difference between adjacent valid pixels and take the initial value of the phase mean as a constant; perform neighborhood phase expansion and least squares fitting with a planar model and iterate weighted until convergence, output a, b and the number of valid pixels; do not perform full-field absolute phase expansion.

[0169] Step 7, Affine and Magnification Compensation: by a X b X a Y b Y Calculate A, B', C, D, and s with T. Xs Y , ,have to ;Depend on and Calculate the magnification compensation k X k Y .

[0170] Step 8, Result Storage: [The sentence is incomplete and requires more context to be translated accurately.] Compared to wavelength-distinguished (k X ,k Y It associates the current objective lens name with the storage (guidance light stores the guidance light field, ultraviolet light stores the ultraviolet light field) to compensate for color differences of different wavelengths; it can generate a phase pseudo-color preview for manual verification.

[0171] Example 2 (Self-calibration of upward-looking camera, conjugate optical path)

[0172] DMD centers on a local rectangle and crosshair, aligned with the camera to find the mean of the two centroids (d1, d2) and the image center (I). cx ,I cy By comparison, we can obtain , A four-step phase-shifting fringe is projected locally at the center of the DMD (the area outside the region is completely black; if necessary, the projected coordinates are flipped 180° to match the optical path). a and b are obtained by fitting the four-step phase shifting fringe to the phase plane. The calibrated values ​​under the current objective guide light are then read. and ,but Targeting the camera pixel equivalent Rotation angle ;Will , , , Associate with the current objective lens storage.

[0173] Other variations: the marking pattern is not limited to crosses (it can be circular spots, squares, etc.); feature extraction is not limited to the image moment centroid (it can be template matching, sub-pixel corner points); the phase shift step can be extended to N steps (N≥3); the fringe direction is not limited to two orthogonal directions; plane fitting can be robustly performed using Huber weights or RANSAC; the length reference can be combined with a grating ruler in addition to the displacement stage; the calibration results can be persistently saved according to the two-dimensional index of the objective lens and the light source.

[0174] The above are merely specific embodiments of the present invention, but the technical features of the present invention are not limited thereto. Any simple changes, equivalent substitutions, or modifications made based on the present invention to achieve substantially the same technical effect are all covered within the protection scope of the present invention.

Claims

1. A four-step phase-shift self-calibration method for a maskless lithography system for digital micromirror devices, characterized in that... Includes the following steps: S1. Camera geometry self-calibration: S1.1 Control the digital micromirror device to project the marking pattern, acquire the first image at the reference position of the motion displacement stage through the image acquisition device and extract the reference feature points, and record the reference coordinates of the motion displacement stage; S1.2 Control the motion displacement stage to move a known distance in multiple directions and return, acquire images after each movement and extract the reference feature points; S1.3 Calculate the pixel equivalent of the image acquisition device based on the known distance and pixel displacement in each direction, and combine the pixel equivalents in each direction to obtain the pixel equivalent of the image acquisition device; S1.4 Calculate the rotation angle of the image acquisition device relative to the motion displacement stage based on the displacement direction in each direction and the pixel displacement direction. ; S2. Four-step phase-shifting projection and phase calculation: S2.1 Control the digital micromirror device to project sinusoidal fringe patterns with a period of T and phase shifts of 0, π / 2, π, and 3π / 2 respectively along the first and second directions, and acquire the corresponding fringe images through the image acquisition device; S2.2 Calculates the wrapping phase pixel by pixel using the four-step phase shift formula. Harmony and reconciliation system B; S3, Phase plane fitting: S3.1 Select effective pixels according to the modulation index B, expand the neighborhood phase of the effective pixels using the phase plane model, fit the least squares plane and iterate until convergence, and output the phase gradient in the first direction and the phase gradient in the second direction. S3.2 does not perform full-field absolute phase expansion throughout the entire process; S4, Affine, Decoupling and Magnification Compensation: S4.1 Calculate the affine transformation parameters from the image plane of the image acquisition device to the image plane of the digital micromirror device based on the phase gradient and the period T, including the combined rotation angle of the image acquisition device relative to the digital micromirror device. And scaling in both directions , ; S4.2 According to Calculate the torsional rotation angle of the digital micromirror device. ; S4.3 According to the pixel equivalent p of the image acquisition device cam0 The scaling factor and the theoretical magnification M of the current objective lens are used to calculate the magnification compensation coefficient k in both directions. X k Y ; S5. Result Storage: The torsion rotation angle of the digital micromirror device and the magnification compensation coefficient k X k Y Stored in association with the current objective magnification and the current light source wavelength.

2. The four-step phase-shift self-calibration method for a maskless lithography system for digital micromirror devices according to claim 1, characterized in that: The marking pattern in step S1.1 is a cross-shaped pattern. The position of the reference feature point is extracted using the image moment centroid method, and the centroid coordinates are calculated according to the following formula (c x ,c y ): ； Among them, M 00 M 10 M 01 Image moments for marking patterns.

3. The four-step phase-shift self-calibration method for a maskless lithography system for digital micromirror devices according to claim 1, characterized in that: The multiple directions mentioned in step S1.2 include four directions: +X, -X, +Y, and -Y. The motion displacement stage moves a known distance d along +X, -X, +Y, and -Y in sequence and returns to the reference position. The pixel equivalent in each direction is calculated using the following formula: ； Among them, L i Let be the pixel displacement magnitude in the i-th direction; The pixel equivalent p of the image acquisition device cam0 Take the average value of the pixel equivalents in the four directions; The rotation angle of the image acquisition device relative to the motion displacement stage The displacement directions in the four directions are uniformly converted into the angle α of the +X axis of the motion displacement stage in the image plane of the image acquisition device. i According to the circumference average formula get.

4. The four-step phase-shift self-calibration method for a maskless lithography system for digital micromirror devices according to claim 1, characterized in that: In step S2.1, the first direction is the X direction, and the sinusoidal stripes in the X direction are vertical stripes. The second direction is the Y direction, and the sinusoidal stripes in the Y direction are horizontal stripes.

5. The four-step phase-shift self-calibration method for a maskless lithography system for digital micromirror devices according to claim 1, characterized in that: The wrapping phase in step S2.2 Calculate pixel-by-pixel using the following formula: ； The modulation index B is calculated pixel-by-pixel according to the following formula: ； Wherein, I1, I2, I3, and I4 are the gray values ​​of the striped images acquired when the phase shift is 0, π / 2, π, and 3π / 2, respectively.

6. The four-step phase-shift self-calibration method for a maskless lithography system for digital micromirror devices according to claim 1, characterized in that: The effective pixels in step S3.1 are selected by an adaptive modulation threshold, wherein the adaptive modulation threshold B thr According to the median B of the modulation index med Sure: ； Where ρ is a preset proportionality coefficient, and B floor As a preset lower limit threshold for modulation, only B≥B is retained. thr The pixels that are selected as valid pixels are those that are not selected.

7. The four-step phase-shift self-calibration method for a maskless lithography system for digital micromirror devices according to claim 3, characterized in that: The phase plane model in step S3.1 is as follows: ； Where (u0,v0) are the coordinates of the image center, a and b are the gradients of the phase along the u and v directions, c is a constant term, and (u,v) are the coordinates of the current pixel.

8. The four-step phase-shift self-calibration method for a maskless lithography system for digital micromirror devices according to claim 7, characterized in that: In step S4, the affine, decoupling, and magnification compensation steps are as follows: make , , , , where a X The gradient coefficients of the phase along the u direction, obtained by fitting the sinusoidal fringes in the first direction to the phase plane, and b X The gradient coefficients of the phase along the v direction obtained by fitting the sinusoidal fringes in the first direction to the phase plane, a Y The gradient coefficients of the phase along the u direction, b, are obtained by fitting the sinusoidal fringes in the second direction to the phase plane. Y The gradient coefficient of the phase along the v direction is obtained by fitting the sinusoidal fringes in the second direction to the phase plane; The combined rotation angle of the image acquisition device relative to the digital micromirror device ; The scaling rates in the two directions are: , ; The theoretical pixel size of the digital micromirror device , where p base This refers to the reference pixel size of the digital micromirror device; The actual pixel size of the digital micromirror device , ; The magnification compensation coefficient is: , .

9. The four-step phase-shift self-calibration method for a maskless lithography system for digital micromirror devices according to claim 1, characterized in that: During the associated storage in step S5, the corresponding magnification compensation coefficients are stored for different light source wavelengths under the same objective lens magnification. The light source wavelengths include the exposure ultraviolet light wavelength and the alignment / guide light wavelength.

10. The four-step phase-shift self-calibration method for a maskless lithography system for digital micromirror devices according to claim 1, characterized in that: It also includes the upward camera calibration procedure: Under the conjugate optical path, the digital micromirror device is directly imaged by the upward-looking camera; The digital micromirror device projects a local rectangular pattern and a cross pattern in the central region. Images are acquired by the upward-looking camera, and the centroids of the two patterns are calculated. The average of the two centroids is compared with the center of the pattern to obtain the center offset of the upward-looking camera. , ; The digital micromirror device is controlled to project four-step phase-shifting fringes in the central region, and the phase gradients a and b are obtained by phase plane fitting. Read the calibrated magnification compensation coefficient under the current objective lens guide light. and the torsional rotation angle of the digital micromirror device ; According to the magnification compensation coefficient Calculate the actual pixel size of the digital micromirror device under the current objective lens. Then, the pixel equivalent of the upward-looking camera is calculated. ; Based on the phase gradients a and b and the torsional rotation angle Calculate the rotation angle of the upward-looking camera. ; The pixel equivalent of the upward-looking camera Rotation angle and center offset , Stored in association with the current objective lens.

11. The four-step phase-shift self-calibration method for a maskless lithography system for digital micromirror devices according to claim 10, characterized in that: The pixel equivalent of the upward-looking camera Calculate as follows: ； ； Where, p base denoted as the reference pixel size of the digital micromirror device, M as the theoretical magnification of the current objective lens, T as the period of the four-step phase shift fringe, and a as the phase gradient output by the phase plane fitting.

12. A four-step phase-shift self-calibration device for a maskless lithography system of a digital micromirror device, used to implement the four-step phase-shift self-calibration method for a maskless lithography system of a digital micromirror device as described in any one of claims 1 to 11, characterized in that, include: The image acquisition and feature extraction module is used to control the image acquisition device to acquire images and extract the positions of reference feature points; The displacement stage motion control module is used to control the displacement stage to move a known distance in multiple directions and return. The pattern projection module is used to control the projection of marking patterns and four-step phase-shifting sinusoidal fringe patterns onto the digital micromirror device; The phase calculation module is used to calculate the wrapping phase and modulation degree pixel by pixel based on the acquired phase shift fringe pattern; The phase plane fitting module is used to select effective pixels according to the modulation degree, expand the neighborhood phase of the effective pixels using the phase plane model, perform least squares plane fitting and iterative weighting until convergence, and output the phase gradient. The parameter calculation module is used to calculate the affine transformation parameters based on the phase gradient. Calculate the torsional rotation angle of the digital micromirror device and the magnification compensation coefficient; The result storage module is used to associate and store the calibration results with the current objective lens magnification and the current light source wavelength.

13. A maskless photolithography system for digital micromirror devices, characterized in that, include: Digital micromirror devices, as spatial light modulators, are used to project marked patterns and four-step phase-shifted sinusoidal fringe patterns; An image acquisition device is used to acquire images of the projected pattern of the digital micromirror device; A motion displacement stage is used to support the workpiece and serve as the physical length reference of the system; Objective lens and objective lens switching device, providing multiple switchable magnifications; The light source device includes an exposure ultraviolet light source and an alignment / guiding light source; A control and calibration device for performing a four-step phase-shift self-calibration method for a maskless lithography system for a digital micromirror device as described in any one of claims 1 to 11.

14. The maskless lithography system for a digital micromirror device according to claim 13, characterized in that: The image acquisition device includes a downward-looking camera and an upward-looking camera. The downward-looking camera is used to acquire images of reference feature points on the motion displacement stage and stripe images projected by the digital micromirror device. The upward-looking camera is used to directly image the digital micromirror device under a conjugate optical path.

15. A maskless lithography system for a digital micromirror device according to claim 13, characterized in that: The digital micromirror device is a binary digital micromirror device, which presents 8-bit grayscale sinusoidal stripes by modulating the time duty cycle or dithering of each micromirror in a static projection manner.

16. A computer-readable storage medium storing computer instructions thereon, characterized in that: When the computer program is executed by the processor, it implements a four-step phase-shift self-calibration method for a maskless lithography system for digital micromirror devices as described in any one of claims 1 to 11.

Images