A laser modification processing method and a processing system for eliminating cleavage plane crack deflection

Abstract

The application provides a laser modification processing method and a processing system for eliminating cleavage surface crack deflection. The spatial intersection line of the crystal natural cleavage surface and the macro processing surface is solved by cross multiplication operation, and the intersection line direction is locked as the laser polarization direction. The scanning direction is orthogonal to the polarization direction, and an asymmetric light intensity distribution compressing in the scanning direction and stretching in the polarization direction is constructed. A constraint model of micro-step height and laser focal depth is established, the step height is controlled to be less than the focal depth by controlling the dot spacing, the longitudinal stress is used to bridge the micro-cracks, and large-area smooth peeling is realized. The application does not need mechanical rotating parts, can effectively inhibit crack deflection, eliminate the stepped fracture caused by the inclination of the cleavage surface, and obtain a smooth peeling surface.

Description

Technical Field

[0001] This invention relates to the field of laser processing technology or crystal processing technology, and in particular to a laser-modified processing method and system for eliminating cleavage surface crack deflection. Background Technology

[0002] With the rapid development of the semiconductor, microelectronics, and optoelectronic industries, hard and brittle crystalline materials such as silicon carbide (SiC), silicon wafers, and lithium niobate are increasingly widely used in high-end manufacturing. Laser spalling, as an advanced laser-modified separation technology, induces the formation of a continuous modified layer and a network of directional microcracks by focusing an ultrafast laser at a specific depth within the material. Subsequently, by utilizing internal stress release or the application of slight external force, large-area delamination of the material along a predetermined depth is achieved. This technology has significant advantages such as large processing depth, low material loss, no tool contact damage, and applicability to ultra-hard and brittle materials.

[0003] However, the aforementioned hard and brittle materials often exhibit significant crystal anisotropy and possess specific natural cleavage planes. During laser-induced crack propagation and large-area penetration, microcracks are highly susceptible to deflection along the cleavage planes within the material. Especially during large-area transverse peeling, since cracks need to propagate continuously within a two-dimensional plane, if the laser processing trajectory or polarization direction does not match the cleavage plane, it can easily lead to uncontrolled crack propagation, further causing rough peeling surfaces, localized collapse, step-like fractures, or even overall wafer breakage, severely limiting the yield and surface quality of large-size peeling processes.

[0004] Existing research indicates that the polarization direction of linearly polarized lasers directly affects the asymmetric distribution of plasma during laser-material interaction, thereby influencing the propagation direction of microcracks within the modified region. Aligning the laser polarization direction parallel to the material's cleavage planes or the expected crack propagation direction is a crucial method for achieving directional crack propagation and stable connection.

[0005] Currently, the core problem faced by existing technologies in processing large-area peeling of anisotropic materials is that microcracks are easily induced by lattice cleavage planes, causing directional drift and making it difficult to uniformly control the orientation of the peeling surface, thus failing to form stable and continuous large-area layered fractures. Simultaneously, since the natural cleavage planes of crystals usually have an angle relative to the macroscopic peeling plane, a longitudinal height difference inevitably occurs between adjacent cracks during two-dimensional scanning peeling, resulting in a macroscopic stepped fracture structure. To suppress crack drift caused by material anisotropy and mitigate the surface roughness problem caused by stepped fractures, the laser polarization direction must be parallel to the cleavage plane direction in real time, and the energy distribution of the focal plane must be further directionally controlled.

[0006] However, traditional mechanical rotating waveplate technology typically has a response speed of only milliseconds, which not only easily introduces mechanical vibration errors but also fails to achieve real-time synchronous control of polarization direction during high-frequency dynamic scanning. Against this backdrop, spatial light modulators (SLMs), as programmable pure phase liquid crystal devices, exhibit significant advantages: they require no moving mechanical parts, possess no mechanical inertia, high spatial resolution, and arbitrary wavefront reconstruction capabilities. Compared to traditional mechanical polarization control methods, SLMs can achieve microsecond-level rapid switching of electrically controlled phase and possess the ability to perform multi-dimensional coordinated modulation of the laser wavefront, intensity distribution, and polarization state, providing a new technical approach for large-area, high-quality laser stripping of anisotropic, hard, and brittle materials. Summary of the Invention

[0007] To address the issue that existing laser stripping systems for anisotropic materials are prone to microcracks deviating from the preset processing trajectory due to crystal anisotropy and the orientation of natural cleavage planes, resulting in chaotic crack propagation directions and an inability to guarantee a single, controlled fracture path, this invention provides a laser-based refining process and system that eliminates cleavage plane crack deflection. This method overcomes the limitations of traditional scalar optical field processing by employing high-degree-of-freedom phase modulation, achieving arbitrary shaping of the transverse (perpendicular to the beam propagation direction) spot morphology and control of the local polarization state of the spot within the target area through complete elimination of mechanical rotating components. The core technology utilizes the wavefront reconstruction capability of a spatial light modulator (SLM) to adjust the polarization principal axis direction at the laser focal point in real time, ensuring it remains precisely parallel to the lattice cleavage planes or the tangent of the processing path of the material being processed. This physically forces the asymmetric expansion of plasma, achieving absolute guidance of the microcrack propagation direction and ensuring a high degree of consistency in the fracture surface during the stripping process.

[0008] The present invention achieves the above-mentioned technical objectives through the following technical means.

[0009] A laser-modified processing method for eliminating crack deflection on cleavage surfaces includes the following steps:

[0010] Determine the three-dimensional normal vector of the natural cleavage plane inside the crystal to be processed. And obtain the three-dimensional normal vector of the preset macroscopic processing trajectory surface. ;

[0011] The three-dimensional normal vector of the natural cleavage surface The three-dimensional normal vector of the macroscopic processing trajectory surface Performing the cross product operation yields the intersection direction vector of the natural cleavage surface and the macroscopic processing trajectory surface in three-dimensional space. The intersection direction vector The direction is determined as the target polarization direction;

[0012] Set the scanning direction of the laser focus so that it is located on the macroscopic processing trajectory surface and perpendicular to the target polarization direction;

[0013] A target vector light field is constructed, comprising a target light intensity distribution and a target polarization direction, wherein the energy half-width of the target light intensity distribution along the scanning direction is smaller than the energy half-width along the target polarization direction, and the target polarization direction is the intersection direction vector. The direction; the target vector light field is loaded onto the spatial light modulator and focused inside the crystal to be processed to form a processing spot;

[0014] The laser is controlled to scan and process the crystal to be processed along the scanning direction. During the scanning process, the spacing between adjacent pulses of the laser focus in the scanning direction is such that the height difference of the microcracks induced by the adjacent pulses along the depth direction does not exceed the longitudinal damage range of the laser.

[0015] Furthermore, the intersection direction vector The specific method for determining the direction as the target polarization direction is as follows:

[0016] Intersection direction vector Projecting onto the two-dimensional coordinate system of the focal plane, calculate the direction vector of the intersection line. The azimuth angle of the projection direction relative to the x-axis of the focal plane: θ(x,y)=arctan(L' Y / L' X ),

[0017] Where: L' X L' Y These are the direction vectors of the intersection line. Projection components in the x and y directions of the focal plane;

[0018] The azimuth angle θ(x,y) is used as a constraint parameter for the target polarization direction, so that the principal axis of laser polarization is consistent with the azimuth angle θ(x,y).

[0019] Furthermore, the target light intensity distribution is an asymmetric Gaussian distribution, which can be represented in the local coordinate system (x′, y′) of the focal plane as follows:

[0020] ,

[0021] Where the x′ axis is parallel to the laser scanning direction, and the y′ axis is parallel to the target polarization direction; I0 is the peak light intensity at the focal center; w v w is the half-width at half-maximum energy of the laser spot along the scanning direction. p The half-width of the light spot energy in the target polarization direction is given, and w satisfies the following condition. p ≥2w v .

[0022] Furthermore, the target vector light field is constructed, specifically including the following steps:

[0023] Distribution of target light intensity Coupled with the target polarization direction angle θ(x,y), construct the target vector light field. :

[0024] ,

[0025] In the formula:

[0026] The amplitude component of the target light field;

[0027] The normalized Jones vector is the linearly polarized state of the target, used to describe the polarization direction of the target electric field in two-dimensional space.

[0028] The target vector light field is represented by the Jones matrix. The components are decoupled into two orthogonal scalar components, and the target amplitude constraint of the two orthogonal components on the focal plane is extracted:

[0029] ,

[0030] ,

[0031] In the formula: This represents the initial phase distribution of the target light field.

[0032] Furthermore, the target vector light field is loaded onto the spatial light modulator, specifically including the following steps:

[0033] An initial complex amplitude distribution is set for each of the two orthogonal polarization components on the plane where the spatial light modulator is located. The amplitude of the initial complex amplitude distribution is the incident constant A0, and the phase is a random initial phase. and ;

[0034] Constructing the complex amplitude of the SLM surface in the 0th iteration:

[0035] ,

[0036] ,

[0037] in: This represents the complex amplitude of the x-polarization component in the SLM plane at coordinates (x″, y″) in the SLM plane during the 0th iteration. This represents the complex amplitude of the y-polarization component in the SLM plane at coordinates (x″, y″) in the SLM plane during the 0th iteration.

[0038] In the k-th iteration (k≥1), the focusing effect of the lens is simulated by a two-dimensional fast Fourier transform (FFT), and the light field of the SLM plane is propagated forward to the focal plane to obtain the calculated complex amplitude of the focal plane in the current iteration:

[0039] , ,

[0040] In the formula: This represents the calculated complex amplitude of the x-component on the focal plane in the k-th iteration; Let x'' represent the complex amplitude of the x'' component in the SLM plane during the (k-1)th iteration; This represents the calculated complex amplitude of the y-component on the focal plane in the k-th iteration; Let represent the complex amplitude of the y'' component on the SLM plane in the (k-1)th iteration; Represents the Fourier transform operation; (x′′, y′′) represents the two-dimensional coordinates on the liquid crystal surface of the spatial light modulator (SLM);

[0041] Extracting the phase component of the complex amplitude of the x-component on the focal plane in the k-th iteration. Extract the phase component of the complex amplitude of the y-component on the focal plane in the k-th iteration. ;

[0042] Constrain the target amplitude A x (x,y) and A y The (x,y) components are combined with the phase components to generate a new complex amplitude in the focal plane that incorporates the target constraint. and :

[0043] ,

[0044] ,

[0045] In the formula: This represents the new complex amplitude of the x-component on the lower focal plane after the replacement amplitude; This represents the new complex amplitude of the y-component on the lower focal plane after the amplitude replacement in the k-th iteration;

[0046] The new complex amplitude of the focal plane is backpropagated to the spatial light modulator plane using inverse Fourier transform. The amplitude component obtained from the backpropagation is discarded, the amplitude is reset to the incident constant A0, and the phase component is extracted. and ;

[0047] The forward propagation, focal plane target fusion, and backward propagation steps are repeated until the convergence condition is met, and the final phase distribution is output. and .

[0048] Furthermore, the final phase distribution is applied to the spatial light modulator, specifically including the following steps:

[0049] Two blazed grating carrier phases Λ with different spatial deflection angles are introduced into the liquid crystal surface of the spatial light modulator. x (x'',y'') and Λ y (x'',y''):

[0050] ,

[0051] ,

[0052] Where: λ is the laser wavelength; (α1,β1) is the deflection angle of the blazed grating in the horizontal direction; (α2,β2) is the deflection angle of the blazed grating in the vertical direction; For mold taking operation;

[0053] Final phase distribution and Each with the corresponding carrier phase Λ x (x'',y'') and Λ y (x'', y'') superposition is used to perform complex amplitude synthesis and extract the composite phase hologram. :

[0054] ,

[0055] The composite phase hologram Loaded into the spatial light modulator.

[0056] Furthermore, the two sets of deflection angles (α1,β1) and (α2,β2) satisfy the following constraints:

[0057] Separation crosstalk prevention constraints: ,

[0058] Where K is the crosstalk prevention safety factor, θ div The far-field divergence angle of the system beam;

[0059] Bundle matching constraints: In the formula θ wx and θ wy These are the components of the inherent physical drift angle of the common-path interference conversion component in the horizontal and vertical directions, respectively.

[0060] Furthermore, the constraint condition for the height difference ΔZ of the microcracks induced by adjacent pulses along the depth direction is:

[0061] ,

[0062] Where v is the laser scanning speed; f is the laser repetition frequency; α is the angle between the cleavage surface and the macroscopic processing trajectory surface; η is the safety overlap coefficient; K′ is the morphology constant; n is the material refractive index; λ is the laser wavelength; and NA is the numerical aperture of the objective lens.

[0063] A laser refining system for implementing the laser refining method for eliminating cleavage surface crack deflection, comprising:

[0064] Ultrafast lasers are used to generate pulsed lasers;

[0065] The X-ray diffraction online measurement module is used to measure and calibrate the crystal orientation and three-dimensional cleavage plane normal vector of the sample to be processed;

[0066] Multi-degree-of-freedom rotation and translation stage, used to carry samples and perform displacement scanning;

[0067] The optical field and polarization control module, including a beam expander, polarizer, spatial light modulator and common-path interference conversion component, is used to load a composite phase hologram to achieve asymmetric shaping of laser intensity distribution and locking of polarization direction;

[0068] The objective focusing module is used to focus the shaped light beam into the interior of the material;

[0069] The computer control system has a built-in iterative diffraction algorithm module and a micro-step fracture constraint algorithm module, which are used to dynamically generate composite phase holograms and coordinate with the laser system and multi-degree-of-freedom rotation and translation stage to perform scanning processing.

[0070] The beneficial effects of this invention are as follows:

[0071] 1. The laser-modified processing method for eliminating crack deflection on cleavage surfaces described in this invention determines the direction of the intersection of the two planes in three-dimensional space by performing a cross product operation on the normal vector of the natural cleavage surface and the normal vector of the macroscopic processed surface, and locks this intersection direction as the laser polarization direction. This ensures that the microcracks generated by laser-induced plasma expansion always extend along the common direction of the cleavage surface and the processed surface, thereby effectively suppressing the deflection of cracks to other cleavage surfaces, improving the continuity of the modified layer and the flatness of the peeling surface, and solving the crack drift problem that easily occurs in the peeling of anisotropic materials.

[0072] 2. The laser-modified processing method for eliminating crack deflection on cleavage surfaces described in this invention constructs an orthogonal processing mode in which the scanning direction and polarization direction are perpendicular to each other, and combines it with asymmetric light intensity distribution. That is, the half-width of the light spot energy is compressed in the scanning direction to cut off the thermal overlap of the preceding and following pulses, and the half-width of the light spot energy is stretched in the polarization direction to enhance the lateral guidance capability of the crack. This achieves synergistic optimization of heat accumulation suppression and crack directional guidance, thereby avoiding material thermal damage and ensuring the stable propagation of microcracks along the predetermined direction during large-area high-frequency scanning processing.

[0073] 3. The laser-modified processing method for eliminating cleavage surface crack deflection described in this invention establishes a quantitative constraint model between the micro-step height and the laser focal depth. By setting the dot spacing, the longitudinal step height caused by the tilt of the cleavage surface is made smaller than the longitudinal damage range of the laser. The focal depth stress of the laser itself is used to perform secondary tearing and stitching of the misaligned micro-crack gaps. At the micro level, an ordered step structure is preserved, and at the macro level, a smooth and continuous peeling surface is achieved. Thus, the influence of the step-like fracture caused by the tilt of the cleavage surface on the surface quality is eliminated in principle.

[0074] 4. The laser refining process for eliminating cleavage surface crack deflection described in this invention uses a spatial light modulator (SLM) to achieve purely electronically controlled coordinated modulation of polarization direction and light intensity distribution, completely eliminating the mechanical rotating parts in traditional schemes. The response speed reaches the microsecond level, enabling real-time tracking of changes in the processing trajectory and synchronous adjustment of the polarization direction during large-area high-speed scanning. At the same time, the precise reconstruction of the vector light field is completed through a single-device phase multiplexing algorithm, simplifying the system structure and improving processing efficiency and long-term operational stability. Attached Figure Description

[0075] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. The drawings described below are some embodiments of the present invention. For those skilled in the art, it is obvious that other drawings can be obtained from these drawings without creative effort.

[0076] Figure 1 This is a flowchart of the laser-modified processing method for eliminating cleavage surface crack deflection as described in this invention.

[0077] Figure 2 This is a schematic diagram of the laser-modified processing system for eliminating cleavage surface crack deflection as described in this invention.

[0078] Figure 3 This is a geometrical diagram illustrating the orthogonal relationship between the polarization direction locking line and the processing direction being parallel or perpendicular to the polarization in the method of this invention.

[0079] Figure 4 This is a schematic diagram illustrating how the stepped cracks are reduced to a negligible level by constraining the spacing between points in this invention.

[0080] Figure 5 This is a schematic diagram showing the through-hole effect of the modified layer in the deep lateral peeling process inside the silicon carbide (SiC) ingot in Embodiment 1 of the present invention.

[0081] Figure 6 The morphology of the modified crack is continuous, parallel, and uniform when the polarization direction is parallel to the spatial intersection line;

[0082] Figure 7 This refers to the crack deflection, local collapse, and rough fracture morphology formed when the polarization direction is not parallel to the spatial intersection line.

[0083] In the picture:

[0084] 1-Laser system; 2-Beam expander; 3-Polarizer; 4-Reflector; 5-Reflector; 6-Convex lens group; 7-Beam splitter; 8-Spatial light modulator; 9-CCD; 10-Dichroic mirror; 11-Processed sample; 12-Translation stage; 13-Computer control system; 14-Online X-ray diffraction measurement module. Detailed Implementation

[0085] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0086] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "axial," "radial," "vertical," "horizontal," "inner," and "outer," etc., indicating orientation or positional relationships based on the orientation or positional relationships shown in the accompanying drawings, are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined with "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0087] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0088] like Figure 1 As shown, the laser refining processing method based on polarization and light intensity spatial shaping described in this invention is used to solve the problems of crack deflection and roughness of the peeling surface induced by natural cleavage planes in the laser refining and peeling process of anisotropic hard and brittle materials (such as silicon carbide), and includes the following steps:

[0089] Step S01: Crystal orientation detection and spatial intersection polarization direction calculation, with the following characteristics:

[0090] The internal crystal orientation of the crystal to be processed is determined by X-ray diffraction (XRD), and the normal vector of its natural cleavage plane (such as the 4° or 8° offset plane of SiC) in three-dimensional space is obtained. , where: n cx Let n be the component of the cleavage plane normal vector along the X-axis. cy Let n be the component of the cleavage plane normal vector along the Y-axis. cz The component of the cleavage surface normal vector along the Z-axis;

[0091] Based on the processing requirements, a macroscopic processing trajectory surface (i.e., the desired peeling plane) is preset in the computer. The normal vector of this plane can be expressed as: n mx To define the components of the normal vector of the pre-defined macroscopic machining trajectory surface along the X-axis, n my To define the component of the normal vector of the pre-defined macroscopic machining trajectory surface in the Y-axis direction, n mz The component of the normal vector of the preset macroscopic machining trajectory surface in the Z-axis direction;

[0092] By performing a cross product operation on the normal vector of the pre-defined macroscopic machining trajectory surface and the normal vector of the natural cleavage surface, the intersection direction vector of the natural cleavage surface and the macroscopic machining surface in three-dimensional space is obtained:

[0093] ,

[0094] The specific expressions for the three components are as follows:

[0095] The component L of the intersection direction vector in the X-axis direction X =n cy n mz -ncz n my ;

[0096] The component L of the intersection direction vector in the Y-axis direction Y =n cz n mx -n cx n mz ;

[0097] The component L of the intersection direction vector along the Z-axis Z =n cx n my -n cy n mx ;

[0098] Since the spatial intersection direction resides simultaneously within both the natural cleavage surface and the macroscopically processed surface, this invention defines the intersection direction as the target polarization direction. To achieve the subsequent construction of the focal plane vector light field, the spatial intersection direction is projected onto the two-dimensional coordinate system of the focal plane, and the azimuth angle of the spatial intersection line relative to the x-axis of the focal plane is calculated: θ(x,y)=arctan(L' Y / L' X ),

[0099] Where: θ(x,y) is the azimuth angle of the spatial intersection line relative to the x-axis of the focal plane, i.e., the target polarization direction angle; L' X L' Y Let L' and Y be the projection components of the spatial intersection vector onto the focal plane in the x and y directions, respectively. Since the x and y axes of the focal plane are parallel to the x and y axes of the laboratory coordinate system, respectively, the projection components of the spatial intersection vector onto the focal plane are its x and y components, i.e., L'. x =L x L' Y =L Y It should be noted that the focal plane (x, y) here is a fixed global coordinate system used to describe the absolute position of any point on the focal plane.

[0100] The obtained target polarization direction angle θ(x,y) will be used as a constraint parameter for the subsequent construction of the target vector light field and the phase encoding of the spatial light modulator (SLM), thereby ensuring that the laser polarization principal axis is always parallel to the direction of the crystal's natural cleavage plane, and realizing the directional and stable propagation of microcracks along the cleavage plane.

[0101] In actual processing, when the macroscopic processing trajectory needs to be changed (e.g., switching from X-axis scanning to Y-axis scanning), the relative geometric relationship between the processing trajectory plane and the natural cleavage plane of the crystal will be broken. The system will recalculate the required change in spatial angle according to the above formula and simultaneously control the multi-axis moving platform (such as the Theta rotary stage) to dynamically rotate the sample to be processed, rather than rotating the spatial light modulator (SLM) or polarization adjustment mechanism. Through the rotation of the physical platform, the natural cleavage plane of the sample is passively aligned to the new processing trajectory, thereby ensuring that the laser polarization direction, processing direction and lattice cleavage plane always maintain the optimal matching relationship, so as to accurately induce microcracks with the most efficient stress release.

[0102] Step S02: Set the relationship between the processing direction and the lateral propagation of the micro-oblique crack, as follows:

[0103] After determining the target polarization direction (i.e., the intersecting line vector), the moving scanning direction of the laser focus (i.e., the laser direct writing direction) is set so that it is always located on the macroscopic processing surface and remains perpendicular (orthogonal) to the target polarization direction.

[0104] This design is based on the fact that the target polarization direction is locked within the cleavage plane. Laser-induced plasma expansion drives the microcracks to extend laterally strictly along the polarization direction within the inclined cleavage plane. Since there is a fixed angle (e.g., 8°) between the cleavage plane and the macroscopic machining surface, this lateral extension inevitably results in the cracks exhibiting an inclined state in space. When the laser focus travels in a direction perpendicular to the polarization, a natural, periodic height difference—the so-called micro-step—is generated between these inclined cracks induced by adjacent pulses in the depth direction (Z-axis). To suppress the macroscopic stepped cracks generated when these inclined microcracks connect in space, high-frequency spotting in the machining direction is needed in subsequent processes to significantly reduce the spot spacing between adjacent pulses, thereby minimizing the physical damage of the stepped cracks in the Z-axis (depth direction).

[0105] Step S03: Decoupling and amplitude constraint extraction of the focal plane target vector light field based on the Jones matrix, as detailed below:

[0106] A target light intensity distribution with an asymmetric elliptical shape is constructed to suppress high-frequency heat accumulation along the processing direction and fully promote the lateral propagation of microcracks within the inclined cleavage plane, as detailed below:

[0107] A local coordinate system (x′, y′) is introduced within the focal plane, rotating with the laser spot direction. The x′ axis is parallel to the laser scanning direction, and the y′ axis is parallel to the target polarization direction. This local coordinate system and the fixed global focal plane coordinate system (x, y) lie in the same physical plane and are linked by a rotational transformation. That is, as the scanning and polarization directions change, the local coordinate system (x′, y′) rotates relative to the global focal plane coordinate system (x, y), thus conveniently describing the change in the overall direction of the elliptical laser spot. The target light intensity distribution in this local coordinate system is expressed as:

[0108] ,

[0109] In the formula:

[0110] For any coordinate point at the focal plane The target light intensity distribution is used to describe the energy distribution state of the focal region after being shaped by the spatial light modulator (SLM); where the focal plane x′ axis is parallel to the laser scanning direction and the focal plane y′ axis is parallel to the target polarization direction.

[0111] I0 is the peak light intensity at the focal center;

[0112] w v The half-width of the spot energy in the processing direction is minimized by extremely compressing it to reduce thermal overlap during high-frequency scanning and avoid thermal damage.

[0113] w p Let w be the energy half-width of the light spot in the polarization direction. Stretching this energy expands the lateral guidance range of the microcrack within the inclined cleavage plane, where the constraint condition is: w p ≥2 w v .

[0114] The coefficient "2" in the exponent term is the standard energy attenuation coefficient in the definition of a Gaussian beam, used to characterize the decrease in light intensity to 1 / e of the peak light intensity. 2 The corresponding half-width position of the light spot.

[0115] It should be noted that since the global focal plane coordinate system (x, y) does not change with the beam direction, it is difficult to conveniently describe the major and minor axes of the rotated elliptical beam using xy directly. Therefore, a local coordinate system (x′y′) is introduced, where the x′ axis is parallel to the laser scanning direction and the y′ axis is parallel to the target polarization direction. x′ and y′ will rotate relative to the global focal plane coordinate system as the scanning direction and polarization direction change, thus representing the rotation of the overall direction of the elliptical beam. The intensity distribution of the target elliptical beam is essentially defined along the x′ and y′ directions, but since the entire focal plane still needs to be described using global coordinates (x, y), x′ and y′ can be expressed as functions of x and y through coordinate transformation.

[0116] The target light intensity distribution is coupled with the target polarization direction angle θ(x,y) determined in step S01 to construct the target vector light field on the focal plane:

[0117] ,

[0118] In the formula:

[0119] For the amplitude components of the target light field, the global coordinates (x,y) must first be transformed into local coordinates (x′,y′) through rotation transformation before being input into the light intensity expression during numerical calculation;

[0120] The normalized Jones vector is the linearly polarized state of the target, used to describe the polarization direction of the target electric field in two-dimensional space. and It can be considered as the weight of the electric field components in the two orthogonal directions of the target polarization direction, x and y;

[0121] This represents the initial phase distribution of the target light field.

[0122] The target vector light field couples the target light intensity distribution, polarization direction, and wavefront phase into the same vector light field, achieving synchronous constraint of the energy distribution, polarization direction, and wavefront structure of the focal plane.

[0123] By using Jones matrix decoupling and spatial light modulator (SLM) phase encoding, the focal plane vector light field is accurately reconstructed, so that the laser polarization direction is always parallel to the natural cleavage plane of the crystal, and the asymmetric light intensity distribution is used to enhance the directional extension capability of microcracks along the cleavage plane.

[0124] Based on this, the target vector light field is decoupled into two orthogonal scalar components using the Jones matrix. The target amplitude constraints of the two orthogonal scalar components on the focal plane are extracted respectively, serving as boundary conditions for subsequent iterative algorithms.

[0125] ,

[0126] ,

[0127] The two target amplitude constraints simultaneously contain the spatial distribution shape information of the asymmetric light spot and the polarization state component weight information, which will be used as the input of the iterative algorithm in step S04 to drive the phase encoding of the spatial light modulator (SLM). Finally, the vector processing light field with polarization direction locked to the spatial intersection line of the cleavage surface and the processing surface and the light intensity orthogonally asymmetrically distributed will be accurately reconstructed in the focal plane.

[0128] Step S04: Focal plane energy control and composite phase hologram generation based on vector decoupling, as detailed below:

[0129] To address the processing lag caused by the reliance on mechanical half-wave plate rotation in existing technologies, this invention completely eliminates mechanical polarization adjustment devices. Instead, it uses a phase multiplexing algorithm with a single spatial light modulator (SLM) to collaboratively reconstruct the polarization state and light intensity distribution through purely electronic control. In the iterative algorithm of this step, (x′′, y′′) represents the two-dimensional coordinates on the SLM liquid crystal surface, and (x, y) represents the global two-dimensional coordinates on the focal plane. The SLM surface and the focal plane are interconnected through a Fourier transform relationship. The forward propagation of the iterative algorithm maps the SLM surface coordinates to the focal plane coordinates, and the backward propagation maps the focal plane coordinates back to the SLM surface coordinates. The coordinate axes of the two coordinate systems are aligned, with the origin located at the center of the optical axis. The specific steps are as follows:

[0130] S4.1: On the input plane of the spatial light modulator (SLM), set the initial optical field for the two orthogonal polarization basis vectors (x'' direction and y'' direction). Set the incident laser as a uniform plane wave with a constant amplitude A0; simultaneously assign random initial phase distributions to the x'' and y'' polarization components. and Thus, the complex amplitude of the SLM surface in the 0th iteration is constructed:

[0131] ,

[0132] ,

[0133] in: This represents the complex amplitude of the x-polarization component in the SLM plane at coordinates (x″, y″) in the SLM plane during the 0th iteration. This represents the complex amplitude of the y-polarization component in the SLM plane at coordinates (x″, y″) in the SLM plane during the 0th iteration.

[0134] S4.2: In the k-th iteration (k≥1), the focusing effect of the lens is simulated by using a two-dimensional fast Fourier transform (FFT) to propagate the light field from the SLM plane to the focal plane in the forward direction, thus obtaining the calculated complex amplitude of the focal plane in the current iteration:

[0135] , ,

[0136] In the formula: This represents the calculated complex amplitude of the x-component on the focal plane in the k-th iteration; This represents the complex amplitude of the x-polarization component in the SLM plane at coordinates (x″, y″) in the (k-1)th iteration. This represents the calculated complex amplitude of the y-component on the focal plane in the k-th iteration; Let x represent the complex amplitude of the y-polarization component in the SLM plane at coordinates (x″, y″) in the (k-1)th iteration. This represents the Fourier transform operation.

[0137] Extracting the phase component of the complex amplitude of the x-component on the focal plane in the k-th iteration. Extract the phase component of the complex amplitude of the y-component on the focal plane in the k-th iteration. ;

[0138] S4.3: Apply a target constraint at the focal plane, wherein the target constraint is the target amplitude constraint A extracted in step S03. x (x,y) and A y (x,y); To achieve this target constraint, the amplitude component calculated at the focal plane after forward propagation is discarded, while the phase component calculated at the focal plane after forward propagation is retained. The target amplitude constraint and the retained phase component are then recombine to generate a new complex amplitude at the focal plane that incorporates the target constraint. and :

[0139] ,

[0140] ,

[0141] In the formula: This represents the new complex amplitude of the x-component on the lower focal plane after the replacement amplitude; This represents the new complex amplitude of the y-component on the lower focal plane after the amplitude replacement in the k-th iteration;

[0142] While preserving the phase information, the target amplitude is used instead of the calculated amplitude, thereby introducing the target light intensity distribution and polarization state component weight information into the iterative process.

[0143] S4.4: The new complex amplitude of the focal plane incorporating target constraints. and A two-dimensional inverse Fourier transform (IFFT) is performed to propagate the complex amplitude distribution backward from the focal plane to the input plane where the spatial light modulator (SLM) is located, resulting in the updated SLM surface complex amplitude. and :

[0144] , ,

[0145] Since spatial light modulators (SLMs) can only modulate the phase of the incident beam and cannot change the uniform amplitude of the incident beam, the method obtained from the backpropagation described above is discarded. and The amplitude component is obtained by resetting the amplitude to the incident constant A0 and extracting only its phase component. and :

[0146] , ,

[0147] in and It is represented as the SLM surface phase in the k-th iteration and will be used as the input phase for the forward propagation step in the (k+1)-th iteration.

[0148] S4.5: Repeat steps S4.2 to S4.4 until any one of the following convergence conditions is met:

[0149] Condition 1: In the k-th iteration, the actual normalized amplitude of the focal plane light field obtained in the k-th iteration ( and ) and the target amplitude constraint A extracted in step S3 x (x,y) and A y The relative root mean square error (RMSE) between (x, y) is less than a preset error threshold ϵ (e.g., 10). -3 (or 1%)

[0150] Condition 2: The iterative loop reaches the maximum number of iterations N set by the system. max .

[0151] The preferred convergence condition is determined using relative root mean square error, which normalizes the root mean square deviation between the actual amplitude and the target amplitude of the focal plane to the root mean square value of the target amplitude, making the error criterion a dimensionless quantity. When any of the above conditions is met, the iteration terminates, and the (k+1)th iteration is not entered again. The final output is the phase distribution extracted in the kth iteration. and The phase distribution is then used as the independent phase distribution ultimately applied to the spatial light modulator, denoted as... and .at this time, and The weighting information of the asymmetric light intensity distribution and the specified polarization direction has been encoded at the mathematical level.

[0152] S4.6: Spatial carrier multiplexing generates the final hologram, as detailed below:

[0153] To enable a single spatial light modulator (SLM) to simultaneously and independently modulate the two orthogonal phases output in step S4.5 and These two phase components need to be physically separated in the spatial frequency domain. To achieve this, two blazed grating carrier phases Λ with different spatial deflection angles are introduced. x (x'',y'') and Λ y (x'',y''):

[0154] ,

[0155] ,

[0156] Where: λ is the laser wavelength; (α1,β1) is the deflection angle of a set of blazed gratings in the horizontal direction, used to carry the phase distribution of the polarization component in the x'' direction. (α2,β2) represents the deflection angle of another set of blazed gratings in the vertical direction, used to carry the phase distribution of the polarization component in the y'' direction. ; For the modulus extraction operation, the continuous linear phase is wrapped in the interval of 0~2π, thus mathematically constructing a blazed grating with a periodic sawtooth structure, which perfectly matches the hardware modulation depth range of 0~2π of pure phase SLM devices, ensuring that the diffraction energy is highly concentrated on the +1st diffraction order.

[0157] The values ​​of the deflection angles (α1, β1) and (α2, β2) mentioned above are not arbitrary. The following two mathematical constraint models need to be constructed before the algorithm is generated:

[0158] The first constraint model is the separation degree anti-crosstalk constraint, which means that the spatial angle difference between the two sets of angles must be greater than the far-field divergence angle θ of the system beam. div To avoid energy crosstalk on the intermediate Fourier transform plane:

[0159] ,

[0160] Where K is the system's anti-crosstalk safety factor, which is usually set to K≥2.

[0161] The second constraint model is a beam-matching constraint, meaning the relative deflection angle vector difference between the two carrier phases must precisely match the inherent physical walk-off angle of the common-path interferometry conversion component in the subsequent physical optical path. The common-path interferometry conversion component can be a birefringent crystal or a polarization beam splitter, and its walk-off angle is determined by the component's own physical characteristics. Let the components of the inherent physical walk-off angle of the common-path interferometry conversion component in the horizontal and vertical directions of the physical optical path be θ. wx and θ wy The horizontal and vertical directions are aligned with the x and y directions in the global coordinate system of the focal plane, respectively; the mathematical expression for this constraint is:

[0162] ;

[0163] After setting the carrier phase that satisfies the above two constraints, the final phase distribution is... and Each with the corresponding carrier phase Λ x (x'',y'') and Λ y (x'', y'') are superimposed to perform complex amplitude synthesis, and finally, the composite phase hologram to be loaded onto the liquid crystal panel of the spatial light modulator is extracted. :

[0164] ,

[0165] The composite phase hologram After being loaded into the spatial light modulator, the carrier phase is physically reversed and canceled out by the subsequent optical common-path interference conversion module. The two orthogonal polarization components are vector superimposed on the focal plane, thereby accurately reconstructing the vector processing light field with both the target asymmetric focal plane energy distribution and the target polarization direction in the form of pure mathematical phase encoding. The vector processing light field is described by the global coordinates (x,y) of the focal plane.

[0166] Step S05: High-frequency point spacing constraint and stress bridging processing for micro-step fracture, as detailed below:

[0167] The composite phase hologram generated in step S04 After being loaded into the spatial light modulator, the target vector processing light field is reconstructed at the focal plane inside the material. The polarization direction of the vector processing light field is parallel to the spatial intersection line between the cleavage surface and the macroscopic processing surface determined in step S01, and the light intensity distribution exhibits the orthogonal asymmetric shape described in step S03. The laser focus is controlled to dynamically scan along the processing direction, which is located within the macroscopic processing surface and is perpendicular to the target polarization direction.

[0168] Because the cleavage planes of the material have an inherent tilt angle relative to the macroscopic processing surface in three-dimensional space (e.g., the common 4° or 8° oblique cutting angle of SiC), although the target polarization direction has constrained the initiation and propagation of microcracks within the cleavage plane, when traveling in the processing direction perpendicular to the polarization direction, parallel microcracks induced by adjacent pulses will inevitably experience physical misalignment in the depth direction (Z-axis direction), forming micro-step-like fractures. To overcome this problem, a high-frequency point-to-point constraint algorithm for micro-step fracture is established, utilizing the longitudinal focal depth stress of the laser to bridge the micro-steps, specifically:

[0169] Let the laser scanning speed be v and the repetition frequency be f, then the center-to-center distance between adjacent pulse dots is: d p =v / f;

[0170] Let α be the angle between the material cleavage plane and the macro-processed surface, then the physical height difference (i.e., the micro-step height) of adjacent lateral oblique cracks along the depth direction. for: ;

[0171] The longitudinal damage length L of the laser focal depth z Approximately expressed as: Where K' is the morphology constant, ranging from 1.0 to 2.0; n is the refractive index of the material, dimensionless; λ is the laser wavelength; and NA is the numerical aperture of the objective lens, dimensionless.

[0172] To achieve secondary tearing and stitching (i.e., stress bridging) of misaligned lateral diagonal cracks using laser longitudinal focal depth stress, the following constraints must be met:

[0173] ,

[0174] The constraint can be understood as the microstep height ΔZ not exceeding the product of the focal depth Lz and the safety overlap coefficient η. Here, η is the safety overlap coefficient, with a value ranging from 0.1 ≤ η ≤ 0.5.

[0175] The system uses this to infer the distance to the critical point. In actual processing, the system sets the dot spacing d. p Much smaller than d p,max High-frequency dense dotting is performed. Since the half-width w of the spot energy in the processing direction was already set in step S03... v Compression and high-frequency scanning do not induce thermal melting of the material; instead, the longitudinal focal depth stress tears and bridges the misaligned microcracks, forming an ordered micro-step fracture structure at the microscopic level. Because the height ΔZ of the microstep is less than the longitudinal damage range L of the laser... z Macroscopically, the negative impact of the stepped cracks is eliminated, and a smooth peeling surface is ultimately presented.

[0176] Example 1: Large-area lateral spalling under crystal orientation homing

[0177] The embodiment aims to plan a two-dimensional horizontal large-area grating scanning spalling located 350 μm deep inside the crystal. For example... Figure 3 As shown, during surface scanning, the laser processing direction is perpendicular to the polarization direction (orthogonal mode). On the one hand, the characteristic that the polarization direction is parallel to the cleavage plane needs to be used to strongly guide the crack; on the other hand, due to the influence of the 8° cleavage plane deflection angle, traveling perpendicular to the polarization direction will inevitably lead to step-like misalignment of adjacent cracks.

[0178] Step S01: Obtain the normal vector of the 8° cleavage plane inside the SiC wafer to be processed in three-dimensional space using the XRD module. Simultaneously, obtain the normal vector of the preset horizontal macroscopic processing peel surface. .

[0179] By performing a cross product operation on the normal vector of the pre-defined macroscopic machining trajectory surface and the normal vector of the natural cleavage surface, the intersection direction vector of the natural cleavage surface and the macroscopic machining surface in three-dimensional space is obtained: ;

[0180] Calculate the azimuth angle of the spatial intersection line relative to the x-axis of the focal plane: θ(x,y)=arctan(-n cx / n cy Angle θ is locked by the system as the absolute polarization principal axis direction for the peeling process in this embodiment.

[0181] Step S02: After determining the target polarization direction (i.e., the intersecting line vector), set the moving scanning direction of the laser focus (i.e., the laser direct writing direction) so that it is always located on the macroscopic processing surface and remains perpendicular (orthogonal) to the target polarization direction.

[0182] To verify the effect of polarization direction locking spatial intersection line on crack directional propagation, the morphology of the modified layer with polarization direction parallel to the spatial intersection line and with polarization direction not parallel to the spatial intersection line were compared under the same laser processing parameters, the same scanning path and the same material crystal orientation.

[0183] like Figure 6 As shown, when the laser polarization direction is parallel to the spatial intersection of the natural cleavage plane and the macroscopically processed surface... At this time, laser-induced microcracks can extend laterally in a directional, continuous, and parallel manner along the target polarization direction. The modified layer as a whole exhibits a regular arrangement, with consistent crack orientation and significantly reduced local deflection and disordered bifurcation, which is beneficial for the subsequent formation of a continuous and smooth peeling surface.

[0184] like Figure 7 As shown, when the laser polarization direction is not parallel to the spatial intersection line... When laser-induced cracks are no longer effectively constrained by the target intersection direction, they are easily induced by the anisotropic cleavage planes of the crystal, causing directional drift and resulting in bifurcation, interlacing, collapse, and uncontrolled penetration in local areas. This morphology manifests as a decrease in the continuity of the modified layer, disordered crack orientation, and obvious rough fracture structure in local areas.

[0185] Depend on Figure 6 and Figure 7The comparison shows that laser internal modification alone cannot guarantee stable crack propagation along the preset peeling direction; only by locking the polarization direction to the spatial intersection of the natural cleavage plane and the macroscopic processed surface can crack deflection be effectively suppressed, and the continuity of the modified layer and the smoothness of the peeled surface be improved. This result further proves the necessity of solving the spatial intersection polarization direction in step S01 and the orthogonality between the processing direction and the target polarization direction in step S02.

[0186] Step S03: Set any coordinate point at the focal plane Target light field The processing direction is extremely compressed, half width w v =1.5μm; polarization-guided stretching, half-width w p =6.0μm, satisfying w p ≥2w v Constraints.

[0187] Using the obtained target polarization direction angle θ, the target vector light field is decoupled into two orthogonal independent amplitude constraint boundary conditions through the Jones matrix:

[0188] X-direction amplitude constraint: ;

[0189] Y-direction amplitude constraint: ;

[0190] Step S04: Extract A from the above steps x (x,y) and A y (x, y) is used as the focal plane constraint and substituted into the GS iterative algorithm. In the k-th iteration, after performing a forward Fourier transform (FFT) to reach the focal plane, the calculated amplitude is forcibly discarded and replaced with the target constraint A. x (x,y) and A y (x,y); then perform an inverse Fourier transform (IFFT) to return to the SLM plane. When the iteration satisfies the root mean square error (RMSE) being less than the system's preset threshold ϵ, the output is a convergent orthogonal pure phase distribution. and .

[0191] To achieve single-device control, a space frequency carrier Λ is introduced. x (x'',y'') and Λ y (x'',y''), the final composite phase hologram sent to the SLM is generated using the following multiplexing and synthesis formula:

[0192] ;

[0193] After the SLM loads the hologram, it perfectly reconstructs a processing light field at the focal plane with polarization parallel to the 8° cleavage plane and light intensity distributed in an extremely asymmetric manner of 1.5μm×6.0μm.

[0194] Step S05: The selected laser wavelength is λ = 1064 nm, the objective lens numerical aperture NA = 0.65, the refractive index of SiC is n ≈ 2.58, and the morphology constant is K′ = 1.5. Calculate the longitudinal damage length of the laser focal depth: ;

[0195] Although transverse cracks are generated, due to the 8° inclination of the cleavage surface, each drilling point in the machining direction will create a stepped cliff on the Z-axis compared to the previous drilling point. Let the drilling distance be d. p The vertical drop of the steps is ΔZ=d p ⋅tan(8°). To ensure that the longitudinal faults of these stepped cracks can be overcome by the longitudinal focal depth stress L of the laser itself. z Completely cover and perform secondary tear sutures, establishing the constraint condition: ΔZ≤η⋅L z , that is, d p ⋅tan(8°)≤η⋅L z Set the safety overlap coefficient η=0.3, and solve for the critical point distance d. p,max ≈20.7μm.

[0196] The system control unit works in conjunction with the laser to set the actual high-frequency dense dot spacing d. p =4μm (e.g., scan speed v=400mm / s, repetition rate f=100kHz). At this point, the physical height difference of the stepped crack induced by adjacent pulses is only:

[0197] ΔZ=d p ⋅tan(8°)=4μm×tan(8°)≈0.56μm;

[0198] like Figure 4 As shown, this tiny longitudinal misalignment ΔZ = 0.56 μm is far smaller than the damage range of the laser focal depth Lz = 9.7 μm. Therefore, with high-frequency scanning, the gaps between these stepped cracks are instantly and completely overcome by the intense longitudinal focal depth stress, achieving seamless connection of the stepped cracks.

[0199] like Figure 5 As shown, thanks to the minimal heat accumulation ensured by asymmetric shaping, this embodiment, without ablating the material, successfully achieved large-area smooth and continuous penetration of numerous inclined microcracks at the same depth using ultra-high frequency dotting combined with algorithm constraints. The surface roughness Ra of the peeled surface was reduced to below 0.5 μm, perfectly eliminating the macroscopic negative impact of stepped cracks. Figure 5The left side shows the macroscopic morphology of the sample after laser processing, and the area indicated by the red box is the selected local observation area. Figure 5 The right side shows the magnified microstructure of this region. It can be seen that, under the condition that the polarization direction is parallel to the spatial intersection of the natural cleavage plane and the macroscopic processed surface, and the processing direction is orthogonal to the polarization direction, the modified cracks corresponding to each scanning path continuously extend along the predetermined direction and interpenetrate with each other, ultimately forming a complete and continuous macroscopic peeling surface. No large-scale disordered deflection or blocky fractures appear in local areas. This result demonstrates that this embodiment achieves silicon carbide peeling without ablation and verifies the effectiveness of the polarization direction control and algorithm constraints in achieving directional crack penetration and stable peeling.

[0200] like Figure 2 As shown, the laser refining processing system based on dynamic polarization direction control of the present invention includes a laser system 1, a spatial light modulator 8, a translation stage 12, and an online X-ray diffraction measurement module 14.

[0201] The laser system 1 generates the pulsed laser required for processing. A beam expander 2 and a polarizer 3 are sequentially arranged along the laser propagation direction. The beam expander 2 enlarges the beam diameter to match the aperture of subsequent devices, and the polarizer 3 converts the laser into linearly polarized light. The linearly polarized light is deflected by a first reflecting mirror 4 and a second reflecting mirror 5, and then enters a 4f relay optical system composed of a convex lens group 6. The convex lens group 6 ensures the flatness of the beam wavefront during transmission. A beam splitter 7 is positioned after the convex lens group 6 to guide the incident beam to a spatial light modulator 8 and simultaneously guide the returning beam to a CCD 9.

[0202] The spatial light modulator 8 is a pure phase liquid crystal device loaded with a composite phase hologram, used to modulate the wavefront of the incident beam, simultaneously achieving asymmetric shaping of the light intensity distribution and locking of the polarization direction. The beam modulated by the spatial light modulator 8 passes through the beam splitter 7 and is reflected by the dichroic mirror 10, and then focused by the objective lens (not separately labeled in the figure, but integrated above the processed sample) into the processed sample 11.

[0203] The processed sample 11 is placed on the translation stage 12, which carries the processed sample and performs multi-degree-of-freedom displacement scanning. The X-ray diffraction online measurement module 14 is positioned above and to the side of the processed sample 11 to measure and calibrate the crystal orientation and three-dimensional cleavage plane normal vectors of the processed sample 11. The CCD 9 receives the returned beam reflected by the beam splitter 7 and performs real-time imaging monitoring of the processing area. The computer control system 13 is electrically connected to the laser system 1, the spatial light modulator 8, the CCD 9, the XYZ translation stage 12, and the X-ray diffraction online measurement module 14. It incorporates an iterative diffraction algorithm module and a micro-step fracture constraint algorithm module to dynamically generate composite phase holograms and coordinate with the laser system 1 and the translation stage 12 to perform synchronous scanning processing. Under the unified scheduling of the computer control system 13, the entire system sequentially completes the processing flow of crystal orientation detection, phase hologram generation, beam modulation and focusing, high-frequency synchronous scanning, and closed-loop monitoring.

[0204] It should be understood that although this specification is described according to various embodiments, not every embodiment contains only one independent technical solution. This way of describing the specification is only for clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment can also be appropriately combined to form other implementation methods that can be understood by those skilled in the art.

[0205] The detailed descriptions listed above are merely specific illustrations of feasible embodiments of the present invention and are not intended to limit the scope of protection of the present invention. All equivalent embodiments or modifications made without departing from the spirit of the present invention should be included within the scope of protection of the present invention.

Claims

1. A laser-modified processing method for eliminating cleavage surface crack deflection, characterized in that, Includes the following steps: Determine the three-dimensional normal vector of the natural cleavage plane inside the crystal to be processed. And obtain the three-dimensional normal vector of the preset macroscopic processing trajectory surface. ; The three-dimensional normal vector of the natural cleavage surface The three-dimensional normal vector of the macroscopic machining trajectory surface Performing the cross product operation yields the intersection direction vector of the natural cleavage surface and the macroscopic processing trajectory surface in three-dimensional space. The intersection direction vector The direction is determined as the target polarization direction; Set the scanning direction of the laser focus so that it is located on the macroscopic processing trajectory surface and perpendicular to the target polarization direction; A target vector light field is constructed, comprising a target light intensity distribution and a target polarization direction, wherein the energy half-width of the target light intensity distribution along the scanning direction is smaller than the energy half-width along the target polarization direction, and the target polarization direction is the intersection direction vector. The direction; the target vector light field is loaded onto the spatial light modulator and focused inside the crystal to be processed to form a processing spot; The laser is controlled to scan and process the crystal to be processed along the scanning direction. During the scanning process, the spacing between adjacent pulses of the laser focus in the scanning direction is such that the height difference of the microcracks induced by the adjacent pulses along the depth direction does not exceed the longitudinal damage range of the laser. The construction of the target vector light field specifically includes the following steps: Construct a local coordinate system (x′y′) that rotates with the direction of the laser spot, where the x′ axis is parallel to the laser scanning direction and the y′ axis is parallel to the target polarization direction; Distribution of target light intensity Coupled with the target polarization direction angle θ(x,y), construct the target vector light field. : ， In the formula: The amplitude component of the target light field; The normalized Jones vector is the linearly polarized state of the target, used to describe the polarization direction of the target electric field in two-dimensional space. θ(x,y) is the direction vector of the intersection line. The azimuth angle of the projection direction relative to the x-axis of the focal plane; The target vector light field is represented by the Jones matrix. Decoupled into two orthogonal scalar components, and extract the target amplitude constraint of the two orthogonal components on the focal plane: ， ， In the formula: The initial phase distribution of the target light field; The target vector light field is loaded onto the spatial light modulator, specifically including the following steps: An initial complex amplitude distribution is set for each of the two orthogonal polarization components on the plane where the spatial light modulator is located. The amplitude of the initial complex amplitude distribution is the incident constant A0, and the phase is a random initial phase. and ; Constructing the complex amplitude of the SLM surface in the 0th iteration: ， ， in: This represents the complex amplitude of the x-polarization component in the SLM plane at coordinates (x″, y″) in the SLM plane during the 0th iteration. This represents the complex amplitude of the y-polarization component in the SLM plane at coordinates (x″, y″) in the SLM plane during the 0th iteration. In the k-th iteration, where k≥1, the focusing effect of the lens is simulated using a two-dimensional fast Fourier transform to propagate the light field from the SLM plane to the focal plane in the forward direction, thus obtaining the calculated complex amplitude of the focal plane in the current iteration: ， ， In the formula: This represents the calculated complex amplitude of the x-component on the focal plane in the k-th iteration; This represents the complex amplitude of the x-polarization component in the SLM plane at coordinates (x″, y″) in the (k-1)th iteration. This represents the calculated complex amplitude of the y-component on the focal plane in the k-th iteration; Let x represent the complex amplitude of the y-polarization component in the SLM plane at coordinates (x″, y″) in the (k-1)th iteration. Represents the Fourier transform operation; (x′′, y′′) represents the two-dimensional coordinates on the SLM plane; Extracting the phase component of the complex amplitude of the x-component on the focal plane in the k-th iteration. Extract the phase component of the complex amplitude of the y-component on the focal plane in the k-th iteration. ; Constrain the target amplitude A x (x,y) and A y The (x,y) components are combined with the phase components to generate a new complex amplitude in the focal plane that incorporates the target constraint. and : ， ， In the formula: This represents the new complex amplitude of the x-component on the lower focal plane after the replacement amplitude; This represents the new complex amplitude of the y-component on the lower focal plane after the amplitude replacement in the k-th iteration; The new complex amplitude of the focal plane is backpropagated to the SLM plane using inverse Fourier transform. The amplitude component obtained from the backpropagation is discarded, the amplitude is reset to the incident constant A0, and the phase component is extracted. and ; The forward propagation, focal plane target fusion, and backward propagation steps are repeated until the convergence condition is met, and the final phase distribution is output. and .

2. The laser-modified processing method for eliminating cleavage surface crack deflection according to claim 1, characterized in that, The intersection direction vector The specific method for determining the direction as the target polarization direction is as follows: Intersection direction vector Projecting onto the two-dimensional coordinate system of the focal plane, calculate the direction vector of the intersection line. The azimuth angle of the projection direction relative to the x-axis of the focal plane: θ(x,y)=arctan(L' Y / L' X ), Where: L' X L' Y These are the direction vectors of the intersection line. Projection components in the x and y directions of the focal plane; The azimuth angle θ(x,y) is used as a constraint parameter for the target polarization direction, so that the principal axis of laser polarization is consistent with the azimuth angle θ(x,y).

3. The laser-modified processing method for eliminating cleavage surface crack deflection according to claim 1, characterized in that, The target light intensity distribution is an asymmetric Gaussian distribution, which is represented in the local coordinate system (x′, y′) of the focal plane as follows: ， Where the x′ axis is parallel to the laser scanning direction, and the y′ axis is parallel to the target polarization direction; I0 is the peak light intensity at the focal center; w v w is the half-width at half-maximum energy of the laser spot along the scanning direction. p The half-width of the light spot energy in the target polarization direction is given, and w satisfies the following condition. p ≥2w v .

4. The laser-modified processing method for eliminating cleavage surface crack deflection according to claim 1, characterized in that, The final phase distribution is then applied to the spatial light modulator, specifically including the following steps: Two blazed grating carrier phases Λ with different spatial deflection angles are introduced into the liquid crystal surface of the spatial light modulator. x (x'',y'') and Λ y (x'',y''): ， ， Where: λ is the laser wavelength; (α1,β1) is the deflection angle of the blazed grating in the horizontal direction; (α2,β2) is the deflection angle of the blazed grating in the vertical direction; For mold taking operation; Final phase distribution and Each with the corresponding carrier phase Λ x (x'',y'') and Λ y (x'', y'') superposition is used to perform complex amplitude synthesis and extract the composite phase hologram. : ， The composite phase hologram Loaded into the spatial light modulator.

5. The laser-modified processing method for eliminating cleavage surface crack deflection according to claim 4, characterized in that, The two sets of deflection angles (α1, β1) and (α2, β2) satisfy the following constraints: Separation crosstalk prevention constraints: , Where K is the crosstalk prevention safety factor, θ div The far-field divergence angle of the system beam; Bundle matching constraints: In the formula θ wx and θ wy These are the components of the inherent physical drift angle of the common-path interference conversion component in the horizontal and vertical directions, respectively.

6. The laser-modified processing method for eliminating cleavage surface crack deflection according to claim 1, characterized in that, The constraint condition for the height difference ΔZ of the microcracks induced by adjacent pulses along the depth direction is: ， Where v is the laser scanning speed; f is the laser repetition frequency; α is the angle between the cleavage surface and the macroscopic processing trajectory surface; η is the safety overlap coefficient; K′ is the morphology constant; n is the material refractive index; λ is the laser wavelength; and NA is the numerical aperture of the objective lens.

7. A laser refining system for implementing the laser refining method for eliminating cleavage surface crack deflection as described in any one of claims 1-6, characterized in that, include: Ultrafast lasers are used to generate pulsed lasers; The X-ray diffraction online measurement module is used to measure and calibrate the crystal orientation and three-dimensional cleavage plane normal vector of the sample to be processed; A multi-degree-of-freedom rotation and translation stage is used to carry samples and perform displacement scanning; The optical field and polarization control module, including a beam expander, polarizer, spatial light modulator and common-path interference conversion component, is used to load a composite phase hologram to achieve asymmetric shaping of laser intensity distribution and locking of polarization direction; The objective focusing module is used to focus the shaped light beam into the interior of the material; The computer control system has a built-in iterative diffraction algorithm module and a micro-step fracture constraint algorithm module, which are used to dynamically generate composite phase holograms and coordinate with the laser system and multi-degree-of-freedom rotation and translation stage to perform scanning processing.

Images