LED light source with double-sided flip chips and manufacturing method of LED light source

Abstract

The invention relates to the technical field of semiconductor light-emitting diodes, and discloses a double-sided flip chip LED light source and a manufacturing method thereof, and the manufacturing method comprises the steps: arranging P electrodes and a rigid support array which is higher than the P electrodes on the surface of a P-type layer of an LED epitaxial wafer at intervals; bonding the side of the P-type layer with the substrate, removing the growth substrate through laser lift-off, and generating a strain-induced sunken region on the surface of the N-type layer; a piezoelectric polarization field generated by the supporting array is used for guiding functional molecules to directionally adsorb on the periphery of the strain induction sunken area, and a wettability modification layer is constructed; coating a photosensitive medium layer, jointly constructing a thickness step interface by the wettability modification layer and the strain induction sunken area, and removing the photosensitive medium and preparing the N electrode according to the interface. Through physical morphology and chemical surface energy dual mask boundary control, the self-alignment opening sharpness is improved, the three-dimensional space self-stabilization of the ultrathin epitaxial structure is realized, and the self-alignment performance of the N electrode is improved. And the carrier transport path is shortened.

Description

Technical Field

[0001] This invention relates to a double-sided flip-chip LED light source and its manufacturing method, belonging to the field of semiconductor light-emitting diode technology. Background Technology

[0002] The current manufacturing process involves preparing an epitaxial layer consisting of an N-type layer, an active layer, and a P-type layer on the surface of a growth substrate, bonding the P-type layer to the substrate, and removing the growth substrate using laser lift-off to expose the N-type layer surface and prepare the N-electrode. As the device thickness decreases to below 3 micrometers, the removal of the growth substrate causes instantaneous stress release. During the stress release process, the epitaxial film undergoes nonlinear lateral displacement. This deviation in physical topology leads to spatial positional discrepancies between the subsequently prepared N-electrode and the underlying P-electrode, generating endogenous leakage channels and reducing radiative recombination efficiency.

[0003] The industry typically addresses these issues by increasing electrode spacing or introducing double-sided high-precision alignment equipment. Increasing electrode spacing leads to an increase in the lateral transport distance of charge carriers, resulting in localized voltage drops and heat generation. High-precision alignment equipment is limited by substrate warping and thin film deformation, making it difficult to maintain alignment consistency across the entire wafer area. For example, Chinese invention patent CN1176483C discloses a method for preparing self-supporting gallium nitride substrates using laser lift-off. The epitaxial layer is bonded to a silicon substrate as a temporary support using organic adhesive. After laser lift-off, the substrate is removed using pyrolysis or solvent methods to achieve physical transfer of the epitaxial layer. However, the bonding interface lacks a micro-stress locking mechanism, and the organic adhesive layer is prone to localized rheology under the instantaneous thermal shock of laser irradiation. This cannot suppress the lateral creep or micro-wrinkles generated by the ultrathin epitaxial layer during the lift-off transient. After the epitaxial layer loses the constraint of the growth substrate, the surface topological features drift randomly, failing to provide a stable geometric reference for subsequent electrode fabrication and limiting the electrical injection efficiency of ultrathin structures.

[0004] Therefore, how to avoid interference of the lateral displacement of the thinned epitaxial layer with spatial alignment after stress release, and to achieve the structural self-stabilization and self-alignment of the electrode in three-dimensional space without introducing external measurement alignment, has become the technical problem to be solved by this invention. Summary of the Invention

[0005] To address the problems mentioned in the background art, the technical solution of the present invention is as follows: A method for manufacturing a double-sided flip-chip LED light source, comprising the following steps: Step S101: Provide an LED epitaxial wafer, the LED epitaxial wafer having an N-type layer, an active layer and a P-type layer supported by a growth substrate; Step S102: A number of P electrodes are prepared at intervals on the surface of the P-type layer, and a support array is set in the gap region between adjacent P electrodes. The height of the support array is higher than the height of the P electrodes, and the support array is composed of a rigid insulating medium with a Young's modulus higher than that of the LED epitaxial wafer material. Step S103: Bond the P-type layer to the substrate, remove the growth substrate by laser lift-off to expose the N-type layer surface, and generate a strain-induced recessed area on the N-type layer surface corresponding to the spatial position of the support array. Step S104: Using the polarization electric field gradient generated by the support array in the strain-induced depression region, surface energy modified molecules are directionally adsorbed to the periphery of the strain-induced depression region to construct a wettability modified layer. Step S105: A photosensitive medium layer is coated on the surface of the N-type layer. The surface energy repulsion of the wettability modified layer and the capillary accumulation effect of the strain-induced depression region are used to construct a thickness step interface at the junction of the strain-induced depression region and the non-depression region. Step S106: The exposure dose is set according to the height difference of the thickness step interface. The photosensitive medium in the non-recessed area is removed to expose the area to be etched corresponding to the area above the P electrode. The area to be etched is etched to form the N electrode contact hole, and metal is deposited in the N electrode contact hole to form the N electrode. In this process, the support array acts as a strain locking center during the lateral stress release of the LED epitaxial wafer, and constrains the lateral displacement of the N electrode within the three-dimensional geometric grid determined by the support array through the polarization electric field gradient established on the surface of the N-type layer.

[0006] Preferably, setting the exposure dose based on the height difference of the thickness step interface includes: obtaining the thickness of the photosensitive dielectric layer in the non-recessed area. and the thickness of the photosensitive medium layer in the strain-induced depression region Exposure threshold based on preset unit thickness of photosensitive medium layer Calculate the exposure dose , to increase the exposure dose satisfy: Perform a global scan exposure to expose the photosensitive medium layer in the non-recessed areas and to strain-induced thickness exceeding that in the recessed areas. The photosensitive medium remains in a non-photosensitive state.

[0007] Preferably, the support array is made of silicon nitride or aluminum oxide and is used to constrain the lateral displacement of the LED epitaxial wafer relative to the substrate during the removal of the growth substrate.

[0008] Preferably, before depositing metal to form an N-electrode in step S106, the method further includes: controlling the photosensitive medium layer to form a meniscus with a preset curvature within the strain-induced depression region; performing plasma thinning on the photosensitive medium layer to expose the N-type layer edge region around the strain-induced depression region; using the remaining photosensitive medium layer as a mask, utilizing the residual stress field of the N-type layer in the edge region to enhance the rate of isotropic wet etching, thereby forming a pit array on the surface of the N-type layer around the strain-induced depression region.

[0009] Preferably, the support array includes several transparent support pillars. The transparent support pillars are made of a transparent medium with a radially gradient refractive index, and the sidewalls of the transparent support pillars have a conical structure with a bottom diameter larger than the top diameter, which is used to redirect photons incident on the support array to the N-type layer side through a total internal reflection mechanism.

[0010] Preferably, the support array includes a composite pillar, which consists of a high-modulus core and an elastic buffer layer wrapped around the outside of the high-modulus core. The composite pillar is used to compress the edge of the LED epitaxial wafer by squeezing it with the elastic buffer layer, thereby forming a compressive stress field in the transition region between the P electrode and the N electrode to control the height of the quantum well barrier of the active layer.

[0011] Preferably, the support array is made of aluminum nitride material, which is used to guide the instantaneous thermal energy of the strain-induced depression region to the substrate through the longitudinal heat flow path formed by the support array during laser lift-off, and guide the liquid gallium metal generated by GaN decomposition to rearrange at the edge of the strain-induced depression region to repair lattice damage.

[0012] Preferably, in step S103, during the bonding process between the P-type layer and the substrate, the method further includes: applying an ultrasonic pulse signal to the support array through the substrate; acquiring the reflected echo after passing through the LED epitaxial layer and extracting the harmonic components in the reflected echo that are linearly related to the imprint depth; determining the morphological saturation of the strain-induced depression region based on the amplitude change of the harmonic components, and adjusting the bonding pressure in real time based on the morphological saturation.

[0013] Preferably, after depositing metal to form the N-electrode in step S106, the method further includes: removing the photosensitive dielectric layer and filling the strain-induced recessed region with a reflective dielectric composed of titanium dioxide or aluminum oxide, wherein the total thickness of the LED epitaxial wafer is less than [amount missing]. Furthermore, the surface energy modified molecules have colloidal functional groups and are driven by the Coulomb force of the polarization electric field gradient to adsorb on the edge of the strain-induced depression region to change the wettability of the N-type layer surface.

[0014] A double-sided flip-chip LED light source, comprising: substrate; An LED epitaxial layer is disposed on one side of the substrate. The LED epitaxial layer consists of a P-type layer close to the substrate, an active layer, and an N-type layer away from the substrate. Several P electrodes are disposed between the P-type layer and the substrate, and a support array is disposed in the gap region between adjacent P electrodes. The height of the support array is greater than the height of the P electrodes, and the support array adopts a rigid insulating medium with a Young's modulus higher than that of the LED epitaxial layer material. Several N electrodes are disposed on the surface of the N-type layer; wherein, the surface of the N-type layer has a strain-induced depression region corresponding to the support array in spatial position, and the periphery of the strain-induced depression region has a wettability modified layer formed by the directional adsorption of polar functional molecules containing colloidal functional groups; the N electrodes and P electrodes form a predetermined staggered distribution in three-dimensional space, and the lateral physical displacement of the N electrodes is constrained within a three-dimensional geometric framework determined by the support array.

[0015] Compared with the prior art, the beneficial effects of the present invention are: 1. In double-sided flip-chip LED light sources, the support array establishes a rigid stress lattice inside the epitaxial layer to achieve spatial self-stabilization of the thin film structure. In the current flip-chip process, stress release occurs in the ultra-thin epitaxial layer with a thickness of less than 3 micrometers during the transient removal of the growth substrate, resulting in microscopic wrinkles and lateral drift in the lattice, causing secondary alignment failure. By using a rigid support array with a preset height higher than the P electrode, a load-bearing point is formed at the bonding interface, locking the epitaxial layer within the geometric lattice. After the growth substrate is removed, the support array continues to constrain the lateral displacement of the epitaxial layer, ensuring that it maintains the topological relationship in the initial growth state. This makes the strain-induced depression formed on the surface of the N-type layer physically related to the position of the underlying support, eliminating positional interference during the thinning process. This transforms the spatial alignment accuracy of the P and N electrodes from the accuracy of external equipment to the geometric stability of the structure itself.

[0016] 2. The polarization electric field and hydrodynamic effects form a physicochemical dual mask, improving the boundary sharpness of self-aligned apertures. To address the issue of smooth transition of adhesive layer thickness at the edge of strain-induced depressions, the piezoelectric polarization field generated by the support array in the epitaxial layer extrusion region induces a polarity difference on the N-type layer surface. The polarization electric field gradient drives functional molecules to be deposited directionally at the periphery of the depression, forming a single-molecule-thick wettable modification layer. The surface energy difference restricts the spreading trajectory of the photosensitive medium. This mechanism, in conjunction with the capillary accumulation effect of the photosensitive medium at the depression, constructs a steep thickness step interface between the depression and non-depression regions, making the boundary of the etchable area clear after global exposure and development. This solves the bottleneck of blurred edges in conventional physical imprinting schemes and ensures the structural stability of the electrode contact area under high-power electric injection conditions.

[0017] 3. The carrier transport path is changed from lateral expansion to vertical counterbalancing, suppressing efficiency decay under high current injection. This invention achieves the shortest distance distribution of P and N electrodes in three-dimensional space through the pre-set staggered distribution logic of the support array. This optimizes the recombination path of holes and electrons from lateral migration across regions to direct vertical counterbalancing. This spatial topology shortens the transport distance of carriers in the low-doped region, significantly reducing the voltage drop and heat loss caused by intrinsic resistance. Furthermore, the pre-set stress field generated by the composite support pillars adjusts the quantum well barrier height of the active layer, compensating for the uneven current distribution that may be caused by electrode staggering. This maintains the spatial stability of the radiative recombination center under ultra-high current density injection, expanding the linear driving boundary of the chip. Attached Figure Description

[0018] Figure 1 This is a flowchart of a self-aligned LED light source manufacturing process according to the present invention; Figure 2 This is a schematic diagram of the architecture of an LED light source manufacturing system according to the present invention. Detailed Implementation

[0019] The present invention will be described below with reference to specific embodiments. It should be understood that the following embodiments are for illustrative purposes only and do not constitute a limitation on the scope of protection of the present invention.

[0020] This invention provides a double-sided flip chip Light source and its manufacturing method, including epitaxial wafer preparation, In stages such as side heterostructure fabrication, interface locking, strain transfer, physical field-induced masking, and electrode self-alignment forming, through... The surface of the mold layer is built higher than A rigid support array of electrodes is used, and the stress field and polarization electric field gradient generated by this array during laser ablation are utilized to achieve the desired effect. The surface of the model layer induces a mask boundary with dual characteristics of physical morphology and chemical potential, thereby achieving... Electrode and Self-aligned construction of electrodes in spatial position; to solve the problem of lateral displacement caused by stress release during the transient removal of the growth substrate in ultrathin epitaxial structures, by performing... Side heterostructure fabrication process, providing a growth substrate made of sapphire material. Epitaxial wafers are sequentially deposited on the surface of the growth substrate. Type layer, active layer and Type layer, and in Several layers of surface spacers were prepared Electrodes, in adjacent A support array is provided in the gap region between the electrodes, and the height of the support array is higher than that of the electrode. The height of the electrodes, wherein the support array is composed of Young's modulus higher than The epitaxial layer material is composed of rigid insulating media such as silicon nitride, aluminum oxide, or aluminum nitride, serving as a strain-locking center; to suppress the deformation of the epitaxial layer during laser lift-off transients through a support array, the center-to-center distance between adjacent support pillars in the support array is... The design follows a physical model of small deflection bending of an elastic thin plate, with the center distance between adjacent support columns in the support array being... The following relationship must be satisfied: ,in, The center-to-center distance of the support columns, in units of , Young's modulus of the epitaxial layer material, in units of , The total thickness of the epitaxial layer, in units of , The maximum allowable deflection of the epitaxial layer is defined in units of... , The pressure per unit area on the bonding interface, in units of By center distance Within the aforementioned threshold range, the rigid sites provided by the support pillars ensure that the lateral wrinkle amplitude of the epitaxial layer after removal of the growth substrate is less than [a certain value]. .

[0021] After completing the above heterogeneous structure construction, a support array will be constructed. Eutectic bonding is performed between the epitaxial layer and the substrate. A support array is used as the main load-bearing sites to lock the geometric position of the epitaxial layer within the three-dimensional space defined by the support pillars. The growth substrate is removed by laser lift-off to expose the epitaxial layer. The surface of the molded layer, due to the support array and There is a height difference between the electrodes The epitaxial layer experiences uneven longitudinal pressure distribution at the bonding interface. After removing the growth substrate, the residual stress inside the epitaxial layer redistributes. Strain-induced indentations are generated on the surface of the substrate layer, corresponding to the spatial position of the support array; by applying ultrasonic pulse signals from the substrate to the support array during the bonding process, the signals passing through the substrate are collected. The reflected echo from the epitaxial layer is used to extract the harmonic components related to the imprint depth. The amplitude variation of these harmonic components determines the morphological saturation of the strain-induced depression region, and the bonding pressure is adjusted to maintain the depth of the strain-induced depression region within a certain range. to Within the range; establish the amplitude of the second harmonic component. With depth The linear mapping relationship is determined by injecting a 50MHz longitudinal wave ultrasonic pulse into the support array through a piezoelectric ceramic transducer within the substrate, extracting the reflected echo signal after passing through the epitaxial layer, and combining it with measured data from a standard sample group with a preset depth gradient to determine the slope coefficient. , to make depth Satisfy the calculation formula ;Amplitude collection When the bonding pressure is below the target threshold, increase the bonding pressure in 50 kPa increments. until amplitude Reaching depth At a saturation state of 150nm, an intensity of not less than [value missing] is generated on the surface of the N-type layer. V / m local piezoelectric field gradient.

[0022] To address the potential mask boundary ambiguity issue at the edge of the depression, a polarization-induced electrochemical self-anchoring mechanism is introduced. This mechanism utilizes a support array to anchor the mask within the strain-induced depression region. The longitudinal compression caused by the crystal lattice generates a polarization electric field gradient within the epitaxial layer, due to... The material exhibits piezoelectric effect, generating a local electric field around the strain-induced depression region where stress is concentrated. The epitaxial wafer is immersed in a polar solution containing surface-modified molecules. These surface-modified molecules possess colloidal functional groups and specific polarity. Driven by the Coulomb force of the polarization electric field gradient, the surface-modified molecules are directionally adsorbed onto the periphery of the strain-induced depression region, constructing a wettable modified layer. The surface-modified molecules are selected from surfactants containing long-chain phosphate groups. Under polarization field induction, the monolayer formed at the depression edge reduces the surface free energy of this region to a minimum. The following describes the process of treating the N-type layer surface with oxygen plasma before the directional adsorption of surface-modified molecules. The plasma is set to a power of 100W to 150W and a time of 30s to 60s to remove residual organic matter and expose suspended hydroxyl groups. The epitaxial wafer is then immersed in plasma of a concentration of [missing value]. It is a 0.5% octadecyl phosphoric acid and anhydrous ethanol mixed solution, at a temperature of 25°C. After standing in the environment for 15 minutes, the piezoelectric field generated around the strain-induced depression region was used to drive the directional coulombic adsorption of surface energy-modifying molecules. An auxiliary DC bias electric field of 5V to 15V was applied below the substrate. The auxiliary DC bias electric field and the local piezoelectric field were vector-superimposed to construct a wettability modification layer with a thickness of one molecule layer around the strain-induced depression region. The surface free energy of this region was reduced to below 18.6mN / m.

[0023] To achieve self-aligned apertures, a thickness-step interface is constructed using the synergistic effect of physical morphology and chemical surface energy. The viscosity of the coating on the surface of the mold layer is lower than that of the mold layer. In the photosensitive medium layer, the capillary accumulation effect in the strain-induced recessed region guides the photosensitive medium to converge towards the center, while the surface energy repulsion effect of the wettability-modified layer inhibits the diffusion of the photosensitive medium to the periphery. A thickness step interface with a significant height difference is constructed at the boundary between the strain-induced recessed region and the non-recessed region. The exposure dose is set according to this height difference, so that the exposure dose... The following relationship must be satisfied: ,in, This is the global exposure dose. The exposure threshold per unit thickness of the photosensitive medium layer. The thickness of the photosensitive medium layer in the non-recessed area. The thickness of the photosensitive medium layer in the strain-induced depression region is determined; through global scanning exposure and development, the thinner non-depression photosensitive medium is removed to expose the corresponding area. The area to be etched above the electrode retains a relatively thick photosensitive medium within the recessed region, forming a self-mask structure; the photosensitive medium coating experiences sensitivity drift due to fluctuations in ambient humidity. The humidity parameters are acquired in real time by the environmental sensor within the adaptive lithography system and adjusted according to the correction factor. Adjust global exposure dose Before global scanning exposure, the thickness of the dielectric layer in the non-recessed area is measured online using an in-situ ellipsometer. and the corresponding thickness of the strain-induced depression region Satisfies the calculation formula Determine the global exposure dose ,in Preset the exposure threshold per unit thickness for the photosensitive medium layer, and the dose correction factor. With a value of 0.95, the boundary roughness of the area to be etched after development is maintained below 50nm, and the three-dimensional spatial physical displacement deviation between the central axis of the N electrode and the geometric center of the underlying support array is constrained to within 0.5μm.

[0024] Finally, using the remaining photosensitive dielectric layer as a mask, the area to be etched is formed by dry etching. Polar contact hole, and in Metal is deposited inside the electrode contact hole to form The electrode and support array constrain the lateral displacement of the epitaxial layer relative to the substrate during the lateral stress release process of the epitaxial layer. The physical positions of the electrodes are locked within a three-dimensional geometric lattice defined by a support array; after deposition, the remaining photosensitive medium is removed, and the strain-induced depression region is filled with a reflective medium composed of titanium dioxide or aluminum oxide, wherein the refractive index of the reflective medium filling the strain-induced depression region is... The selection is determined based on the critical angle model of total internal reflection at the main emission wavelength of the active layer; when the wavelength of the blue light emitted by the active layer is... and The refractive index is When choosing a filling material, the refractive index should be... for to The titanium dioxide medium within the range, combined with the reflection from the inclined sidewall of the support column, reflects photons incident on the support site. The light-emitting window on the molded layer side achieves a uniform distribution of optical flow density under high-power injection; the present invention also provides a double-sided flip chip prepared by the above method. Light source, including a substrate and a light source disposed on one side of the substrate. Epitaxial layer, The epitaxial layer consists of layers close to the substrate. Model layer, active layer and layer away from the substrate Layered structure; Several layers are provided between the mold layer and the substrate. Electrodes and adjacent The support array in the electrode gap region has a height greater than [missing information]. Electrode height, The surface of the molding layer is provided with several Electrodes; among which, The surface of the molded layer has strain-induced depressions corresponding to the spatial positions of the support array. The periphery of these strain-induced depressions has a wettable modified layer formed by the directional adsorption of polar functional molecules containing colloidal functional groups. Electrode and The electrodes are arranged in a pre-defined staggered pattern in three-dimensional space.

[0025] Example 1: This example illustrates double-sided flip-chip technology. The manufacturing method of the light source, with a total thickness of less than high power In device fabrication scenarios, it demonstrates superior performance under heat-constrained conditions with high current injection density, providing a solution for applications with sapphire growth substrates. Epitaxial wafers, in Forming thickness on the surface of the mold layer for of Electrodes, and in adjacent Thickness is set in the gap area between electrodes for The aluminum nitride support array makes the support array relative to Height difference of electrode surface for The support array is fixed to the ceramic substrate via eutectic bonding. During the removal of the growth substrate using an excimer laser, the rigid physical properties of the support array are utilized to establish a displacement constraint lattice within the epitaxial layer. Under the transient stress redistribution caused by removing the growth substrate, due to the height difference... The resulting interfacial pressure gradient, in The spatial position of the support array corresponding to the surface of the mold layer generates a depth of The strain-induced depression region.

[0026] By utilizing the longitudinal lattice extrusion of the strain-induced depression region using a support array, an intensity of [value missing] is induced within the epitaxial layer. The piezoelectric polarization field places the epitaxial wafer in a polar solution containing a surfactant with long-chain phosphate groups. Driven by the Coulomb force of this piezoelectric polarization field, the surfactant molecules are directionally adsorbed and form a wettable modification layer with a thickness of one molecule, reducing the surface free energy of the strain-induced depression periphery to a certain value. ;exist A photosensitive dielectric layer is spin-coated onto the surface of the model layer. The morphological accumulation effect of the strain-induced depression region and the chemical repulsion effect of the wettability-modified layer work synergistically to increase the thickness of the photosensitive dielectric layer in the non-depression region. for The thickness of the strain-induced depression region for Set the exposure threshold per unit thickness of the photosensitive medium layer. for The energy density is Global exposure dose It satisfies the following relationship: ,in, This is the global exposure dose. The exposure threshold per unit thickness of the photosensitive medium layer. The thickness of the photosensitive medium layer in the non-recessed area. The thickness of the photosensitive medium layer in the strain-induced depression region; removed after development. The photosensitive medium of the area is exposed to the corresponding The area to be etched above the electrode is preserved. The photosensitive medium in the region serves as a self-mask; using the remaining photosensitive medium layer as a mask, it is penetrated by dry etching. Formation of layers Electrode contact holes are formed, and metal is deposited within these holes to create... After electrode deposition, the remaining photosensitive medium is removed, and the strain-induced depression region is filled with a reflective medium composed of titanium dioxide. Utilizing the topological relationship of the support array locked within a three-dimensional lattice, the electrode can be aligned without the use of secondary alignment equipment. Electrodes and the underlying layer The alignment deviation of the electrodes in spatial position is controlled within Within; the light source produced is in Under continuous current density injection, the junction temperature is maintained at a certain level by a longitudinal thermal bridge constructed using an aluminum nitride support array. Below, and the tapered slope of the supporting array sidewall will originally be directed towards Photons from the electrode are redirected to The light-emitting surface on the molded layer side.

[0027] Example 2: In an experimental scenario verifying the three-dimensional spatial self-stabilization accuracy and thermal characteristics of an ultrathin epitaxial structure, a double-sided flip chip was fabricated using the manufacturing method provided in the aforementioned specific embodiments. The light source and control group 1 used a traditional flip-chip process without a support array, while control group 2 used a physical mask method with a support array but without executing the polar molecule adsorption procedure; the experimental platform was equipped with a measurement resolution of [missing information]. The step gauge and temperature measurement accuracy are The transient thermal resistance testing system, the test process is in the presence of The process is performed under vacuum bonding conditions to address the anisotropic warpage of the substrate, and a preset rotation speed is introduced during the spin coating of the photosensitive medium. Speed ​​fluctuation noise; in the parameter setting logic, Electrode thickness The value of is limited by the balance between ohmic contact resistivity and interface reflectivity, and its influencing factors include the metal layer deposition thickness and thermal annealing temperature, as well as the height of the support array. The design considerations lie in the constraint capability of lateral stress release caused by epitaxial thinning and the distribution gradient of bonding pressure; when the epitaxial layer thickness is less than To ensure that the depth of the strain-induced depression region meets the self-mask thickness step requirement, a method is established... The mapping model with strain intensity was calculated to yield... exist to Within the interval, an intensity higher than [the specified value] can be induced. The localized polarization electric field, in a specific operating condition instance, is set... for , for height difference for .

[0028] The test group was conducted in an environment with substrate warping and rotation speed fluctuations. Electrode relative to Electrode alignment deviation is maintained In contrast, the lateral displacement of control group 1 due to stress release exceeded the compensation capacity of the secondary alignment, resulting in an alignment deviation of [missing information]. Regarding intermediate characteristic quantities, the experimental group utilized a piezoelectric polarization field to drive the adsorption of polar molecules, and the strain-induced surface free energy at the periphery of the depression region was increased from... original surface Reduce to This establishes a chemical potential energy barrier based on the physical depression, enabling a thickness step ratio achieve Higher than control group 2 At the final performance output level, the test group in The junction temperature under injected current is The light extraction efficiency was improved compared to the control group 1. This result stems from the fact that the support array acts as a longitudinal thermal bridge, reducing the heat flux density.

[0029] Table 1: Comparison of performance parameters of different groups under simulated disturbance conditions

[0030] Gradient verification is performed on the boundaries of key parameters when the height difference... When below the lower limit of the range, such as in test group 1 Due to the weakening of the longitudinal pressure gradient, the induced piezoelectric field strength is insufficient to support dense monolayer adsorption, and the thickness step ratio decreases to This results in jagged residue at the edges of the exposed mask; and when When the upper limit is exceeded, as in experimental group 3 Although the thickness step ratio remains at However, due to excessive shear stress at the interface between the support pillar and the epitaxial layer, microcracks were induced during the peeling transient, causing the junction temperature to rise back to [a certain value]. .

[0031] Example 3: This example combines Figures 1 to 2 This document describes a double-sided flip-chip LED light source and its manufacturing method, such as... Figure 1 As shown, step S101 provides an LED epitaxial wafer consisting of an N-type layer, an active layer, and a P-type layer supported by a growth substrate. Next, step S102 involves fabricating P-electrodes at intervals on the surface of the P-type layer and setting a support array, composed of a high-modulus rigid insulating dielectric and with a height higher than the P-electrodes, in the gap between adjacent P-electrodes. Step S103 involves bonding the P-type layer to the substrate and removing the growth substrate by laser lift-off to expose the surface of the N-type layer. A strain-induced depression region is generated on the surface of the N-type layer corresponding to the spatial position of the support array. Then, step S104 utilizes the support array... The polarization electric field gradient generated by the support array directs the surface energy modified molecules to adsorb onto the periphery of the strain-induced depression region to construct a wettability modified layer. Then, step S105 is performed to coat a photosensitive medium layer on the surface of the N-type layer. The thickness step interface is constructed at the junction of the strain-induced depression region and the non-depression region by utilizing the surface energy repulsion and capillary accumulation effect. Finally, step S106 is completed to set the exposure dose according to the height difference of the thickness step interface, remove the photosensitive medium in the non-depression region to expose the area to be etched corresponding to the region above the P electrode, etch the area to be etched to form an N electrode contact hole, and deposit metal in the hole to form an N electrode.

[0032] like Figure 2As shown, the manufacturing system starts with a micro-nano fabrication workstation, which integrates a P-electrode deposition module, a rigid support array construction module, and an insulating dielectric curing unit. The LED epitaxial wafers with the support array are transported to the intelligent bonding system via epitaxial wafer transfer. The system is equipped with a vacuum eutectic bonding stage, an ultrasonic pulse generator, and a reflected echo impedance analyzer. The resulting substrate bond enters a laser lift-off system containing an excimer laser, a stress release control module, and a strain-induced depression monitoring module. The resulting wafers are then transferred to a surface modification processing station. This station contains a polar molecule adsorption tank, a polarized electric field guiding environment, and a wettability modification layer construction unit. It is connected to an adaptive lithography system equipped with a photosensitive dielectric coating machine and thickness step recognition logic. Finally, the device fabrication is completed by a metallization and etching station containing a contact hole etching unit, an N-electrode metal deposition unit, and a spatial grid displacement constraint mechanism.

[0033] Example 4: This example is used to specifically illustrate double-sided flip-chip technology. The precise calibration of physical field coupling parameters and the closed-loop process specifications in the manufacturing method of the light source, specifically for an epitaxial layer thickness of... And the alignment deviation is required to be lower than In high-reliability application scenarios, to address the strain-induced indentation depth deviation caused by non-uniform pressure at the bonding interface, an initial state definition procedure is executed, providing a Young's modulus of... The aluminum nitride support array will The epitaxial wafer is placed in a temperature control system with an accuracy of And possess A pressure-resolution vacuum bonding stage injects a center frequency of [value missing] into the support array through a piezoelectric ceramic transducer within the substrate. Longitudinal wave ultrasonic pulses, real-time acquisition of second harmonic amplitudes after passing through the epitaxial layer The parameter setting logic lies in establishing a depth by utilizing the monotonic linear relationship between interface contact stiffness and epitaxial layer micro-deformation. With harmonic amplitude The mapping model determines the target harmonic threshold through a preset depth-harmonic correspondence curve. When the acquired data... When below the target threshold, Increase bonding pressure for step units until Reach the corresponding depth for The saturation point produces a localized polarization electric field intensity at... A potential gradient sufficient to drive molecular adsorption is formed on the surface of the lamella.

[0034] To eliminate the diffusion effect at the chemical mask boundary, a standardized operating procedure for preparing the wettability modification layer was followed. An adsorption solution was prepared using anhydrous ethanol as the solvent and octadecyl phosphoric acid as the solute, wherein the mass concentration of octadecyl phosphoric acid was [not specified]. Set as The epitaxial wafer is immersed in the solution at a temperature of Let it stand in an environment Minutes, using strain-induced depression periphery The polarized electric field drives polar molecules to undergo directional Coulomb adsorption, forming a dense, colloidal monolayer at the edge of the depression; a spin-coating method is used to cover the surface with a viscosity of [missing value]. The photosensitive medium layer, driven by the capillary force of the strain-induced depression region and the repulsive force of the peripheral wettable modified layer, has a thickness in the strain-induced depression region. achieve The thickness in the non-depression area for .

[0035] To determine the global exposure dose Gradient exposure experiments were performed on blank silicon wafers to measure the exposure threshold per unit thickness of the photosensitive medium under preset development conditions. for The global exposure dose is calculated and executed based on the height difference of the thickness step interface. To increase the global exposure dose The following relationship must be satisfied: ,in, This is the global exposure dose. Exposure threshold per unit thickness The thickness of the photosensitive medium layer in the non-recessed area. The thickness of the photosensitive medium layer in the strain-induced depression region, This is the dose correction factor, used in this embodiment. Values The edge roughness of the self-mask window formed after development is lower than that of... In subsequent dry etching procedures, The central axis of the electrode coincides with the geometric center of the underlying support array, achieving three-dimensional spatial self-stabilization under complex thermomechanical load conditions.

[0036] Example 5: In a scenario involving the bonding and deployment of multiple batches of gallium nitride epitaxial wafers, how to establish the second harmonic amplitude? With strain-induced depression depth A quantitative mapping benchmark was established between the two, and calibration and data filling procedures were performed. Several standard sample groups with preset depth gradients were provided. The depth of each standard sample group was measured and calibrated by focused ion beam cutting and scanning electron microscopy. Data were collected from each sample group at a center frequency of [missing value] using a vacuum bonding stage. The reflected echo under pulse excitation is used to extract harmonic components and record normalized amplitudes, thus establishing an array of data points. The slope coefficients are obtained by linear fitting of the constructed lookup table using the least squares method. This enables real-time data collection during the on-site deployment phase. Signal based on relational formula The deformation of the epitaxial layer was calculated.

[0037] When the system faces scenarios where fluctuations in ambient humidity interfere with the sensitivity of the photosensitive medium layer, a pre-calibration procedure is performed to correct the exposure threshold per unit thickness. When the viscosity of spin coating is After applying the photosensitive medium, the thickness of the non-recessed area was measured using an in-situ ellipsometer. and the thickness of the strain-induced depression region The initial value is obtained by acquiring real-time environmental parameters through a temperature and humidity sensor on the substrate side and comparing them with the calibration curve to calculate the sensitivity drift. Based on the correction factor Adjust exposure dose The value, while maintaining the developing temperature at And the development time is Under these conditions, the shaping of the self-mask boundary in space is controlled by global exposure, so that... The axial deviation of the electrode contact hole is maintained at Within the dynamic tolerance range.

[0038] Example 6: For epitaxial layer thickness less than In the fabrication of ultra-thin display chips, stress cracks during the peeling process are suppressed by optimizing the geometric features of the support array. A structural geometry calibration procedure is executed, setting the cross-section of each support pillar in the support array to be circular with tapered slopes that gradually widen towards the substrate sidewall. The angle between the sidewall and the horizontal plane of the substrate is... The value of is determined based on the balance between the critical angle of total internal reflection and the stress concentration factor. The von Mises stress distribution at the base of the support column under different angles is simulated using finite element analysis to determine the appropriate value when the included angle is... In to Within the specified range, the peak value of the interfacial shear force during the peeling transient is lower than the fracture toughness threshold of the epitaxial layer; in application examples, the height of the support array is set. for The diameter of the top of the support column is included angle for At this time, the stress buffer width generated by the sidewall of the support array The following relationship must be satisfied: in, This is the stress buffer width, in units of , The height of the supporting column, in units of , The angle between the sidewall and the horizontal plane of the substrate, in units of . ; calculation yielded for This size controls the topographic transition gradient of the strain-induced depression region within the capture range of the physical layer self-mask.

[0039] When the system encounters a situation where fluctuations in the interface adhesion of different batches of epitaxial wafers interfere with the laser lift-off yield, an online calibration procedure for the energy density of epitaxial layer interface decomposition is executed, using a wavelength of [wavelength value missing]. The excimer laser was used to perform test pulse scanning, and the mechanical vibration amplitude during the transient separation of the growth substrate and epitaxial layer was monitored by an infrared displacement sensor. Identify Energy density critical value for thermal decomposition of the interface layer Set the working energy density The following relationship must be satisfied: ,in, Operating energy density, in units of , This is the critical energy density value, in units of... , Energy gain coefficient; Energy gain coefficient The value of is determined based on the trade-off between thermal diffusion damage depth and exfoliation properties, and the energy gain coefficient is determined through a linear evolution model of material damage depth as a function of energy. The value is to In actual deployment examples, the critical energy density value was measured. for Set the energy gain coefficient for Execution working energy density for Global scanning stripping, by controlling the overlap rate of the laser spot. While avoiding overheating and melting damage to the epitaxial layer, the generated piezoelectric polarization field strength reaches .

[0040] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0041] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for manufacturing a double-sided flip-chip LED light source, characterized in that, Includes the following steps: Step S101: Provide an LED epitaxial wafer, the LED epitaxial wafer having an N-type layer, an active layer and a P-type layer supported by a growth substrate; Step S102: A number of P electrodes are prepared at intervals on the surface of the P-type layer, and a support array is set in the gap region between adjacent P electrodes. The height of the support array is higher than the height of the P electrodes, and the support array is composed of a rigid insulating medium with a Young's modulus higher than that of the LED epitaxial wafer material. Step S103: Bond the P-type layer to the substrate, remove the growth substrate by laser lift-off to expose the N-type layer surface, and generate a strain-induced recessed area on the N-type layer surface corresponding to the spatial position of the support array. Step S104: Using the polarization electric field gradient generated by the support array in the strain-induced depression region, surface energy modified molecules are directionally adsorbed to the periphery of the strain-induced depression region to construct a wettability modified layer. Step S105: A photosensitive medium layer is coated on the surface of the N-type layer. The surface energy repulsion of the wettability modified layer and the capillary accumulation effect of the strain-induced depression region are used to construct a thickness step interface at the junction of the strain-induced depression region and the non-depression region. Step S106: The exposure dose is set according to the height difference of the thickness step interface. The photosensitive medium in the non-recessed area is removed to expose the area to be etched corresponding to the area above the P electrode. The area to be etched is etched to form the N electrode contact hole, and metal is deposited in the N electrode contact hole to form the N electrode. In this process, the support array acts as a strain locking center during the lateral stress release of the LED epitaxial wafer, and constrains the lateral displacement of the N electrode within the three-dimensional geometric grid determined by the support array through the polarization electric field gradient established on the surface of the N-type layer.

2. The method for manufacturing a double-sided flip-chip LED light source according to claim 1, characterized in that, Setting the exposure dose based on the height difference of the thickness step interface includes: obtaining the thickness of the photosensitive medium layer in the non-recessed area. and the thickness of the photosensitive medium layer in the strain-induced depression region Exposure threshold based on preset unit thickness of photosensitive medium layer Calculate the exposure dose , to increase the exposure dose satisfy: Perform a global scan exposure to expose the photosensitive medium layer in the non-recessed areas and to strain-induced thickness exceeding that in the recessed areas. The photosensitive medium remains in a non-photosensitive state.

3. The method for manufacturing a double-sided flip-chip LED light source according to claim 1, characterized in that, The support array, made of silicon nitride or aluminum oxide, is used to constrain the lateral displacement of the LED epitaxial wafer relative to the substrate during the removal of the growth substrate.

4. The method for manufacturing a double-sided flip-chip LED light source according to claim 1, characterized in that, Before depositing metal to form an N-electrode in step S106, the method further includes: controlling the photosensitive medium layer to form a meniscus with a preset curvature within the strain-induced depression region; performing plasma thinning on the photosensitive medium layer to expose the N-type layer edge region around the strain-induced depression region; using the remaining photosensitive medium layer as a mask, utilizing the residual stress field of the N-type layer in the edge region to enhance the rate of isotropic wet etching, thereby forming a pit array on the surface of the N-type layer around the strain-induced depression region.

5. A method for manufacturing a double-sided flip-chip LED light source according to claim 1, characterized in that, The support array includes several transparent support pillars, which are made of a transparent medium with a radially gradient refractive index. The sidewalls of the transparent support pillars are tapered structures with a bottom diameter larger than the top diameter, which are used to redirect photons incident on the support array to the N-type layer side through a total internal reflection mechanism.

6. The method for manufacturing a double-sided flip-chip LED light source according to claim 1, characterized in that, The support array includes a composite pillar, which consists of a high-modulus core and an elastic buffer layer wrapped around the outside of the high-modulus core. The composite pillar is used to compress the edge of the LED epitaxial wafer by squeezing it with the elastic buffer layer, thereby forming a compressive stress field in the transition region between the P electrode and the N electrode to control the height of the quantum well barrier of the active layer.

7. A method for manufacturing a double-sided flip-chip LED light source according to claim 1, characterized in that, The support array is made of aluminum nitride and is used to guide the instantaneous thermal energy of the strain-induced depression region to the substrate through the longitudinal heat flow path formed by the support array during laser lift-off. It also guides the liquid gallium metal generated by GaN decomposition to rearrange at the edge of the strain-induced depression region to repair lattice damage.

8. The method for manufacturing a double-sided flip-chip LED light source according to claim 1, characterized in that, In step S103, during the bonding process between the P-type layer and the substrate, the method further includes: applying an ultrasonic pulse signal to the support array through the substrate; acquiring the reflected echo after passing through the LED epitaxial layer and extracting the harmonic components in the reflected echo that are linearly related to the imprint depth; determining the morphological saturation of the strain-induced depression region based on the amplitude change of the harmonic components, and adjusting the bonding pressure in real time based on the morphological saturation.

9. A method for manufacturing a double-sided flip-chip LED light source according to claim 1, characterized in that, After depositing metal to form the N-electrode in step S106, the method further includes: removing the photosensitive dielectric layer and filling the strain-induced recessed region with a reflective dielectric composed of titanium dioxide or aluminum oxide, wherein the total thickness of the LED epitaxial wafer is less than Furthermore, the surface energy modified molecules have colloidal functional groups and are driven by the Coulomb force of the polarization electric field gradient to adsorb on the edge of the strain-induced depression region to change the wettability of the N-type layer surface.

10. A double-sided flip-chip LED light source, characterized in that, include: substrate; An LED epitaxial layer is disposed on one side of the substrate. The LED epitaxial layer consists of a P-type layer close to the substrate, an active layer, and an N-type layer away from the substrate. Several P electrodes are disposed between the P-type layer and the substrate, and a support array is disposed in the gap region between adjacent P electrodes. The height of the support array is greater than the height of the P electrodes, and the support array adopts a rigid insulating medium with a Young's modulus higher than that of the LED epitaxial layer material. Several N electrodes are disposed on the surface of the N-type layer; wherein, the surface of the N-type layer has a strain-induced depression region corresponding to the support array in spatial position, and the periphery of the strain-induced depression region has a wettability modified layer formed by the directional adsorption of polar functional molecules containing colloidal functional groups; the N electrodes and P electrodes form a predetermined staggered distribution in three-dimensional space, and the lateral physical displacement of the N electrodes is constrained within a three-dimensional geometric framework determined by the support array.

Images