A red copper IGBT heat dissipation substrate cold forging forming nickel plating process

Abstract

The present application relates to the technical field of high-power semiconductor device manufacturing, in particular to a red copper IGBT heat dissipation substrate cold forging forming nickel plating process, which comprises blanking and material detection, dynamic step annealing, self-adaptive composite lubrication, cold forging forming, edge cutting and needle cutting and machining, stress relief annealing, cleaning and ultrasonic assisted pickling, variable parameter non-contact sand blasting, predictive pre-bending forming and real-time regulation and control chemical nickel plating; the present application can realize accurate compensation of welding deformation by real-time capturing of welding temperature gradient and combining professional model to predict welding deformation amount and dynamically correcting pre-bending angle, can make the red copper substrate still maintain excellent flatness after welding by using the correlation between the thermal expansion characteristics of red copper and the welding temperature fluctuation, avoid the problem of excessive deformation after welding caused by insufficient pre-bending angle adaptability in the traditional process, meet the strict requirements of high-power IGBT module on the flatness of heat dissipation substrate packaging, and improve the structural stability of the substrate under long-term thermal cycle working conditions.

Description

Technical Field

[0001] This invention relates to the field of high-power semiconductor device manufacturing technology, specifically a cold forging and nickel plating process for a copper IGBT heat dissipation substrate. Background Technology

[0002] Copper, due to its high thermal conductivity and good plasticity, has become the mainstream material for IGBT heat sink substrates. Currently, the cold forging and nickel plating process for copper heat sink substrates mainly follows these technical paths: Traditional processes use fixed annealing parameters and single graphite lubrication, removing excess material through machining, resulting in a material utilization rate of only 60%. While the forging-extrusion composite process for irregularly shaped copper terminals improves material utilization to 80%, it still relies on empirical parameter adjustments, leading to a cold forging crack rate as high as 5%. The industry commonly uses empirical pre-bending angles and rigid fixtures for fixing, failing to quantify the impact of copper's thermal expansion coefficient, resulting in post-weld flatness control only within 0.05mm, which cannot meet the stringent requirement of <0.05mm for IGBT module packaging. Existing chemical nickel plating often uses sodium citrate as a single complexing agent and a fixed pH value, resulting in a plating porosity >1 / cm². 2 The bonding strength is <30MPa, and red rust appears after 192 hours of salt spray testing. Existing technologies for cold forging and nickel plating of copper IGBT heat sink substrates suffer from poor material performance consistency, insufficient precision in welding deformation control, and defects in plating quality and reliability. Fixed annealing parameters lead to grain size fluctuations, residual stress retention rate >10%, high cold forging cracking rate, uncontrollable lubrication layer thickness during cold forging, lack of DLC coating on the mold surface, and susceptibility to copper adhesion and stress concentration. Empirical pre-bending angles do not consider the nonlinear relationship between welding temperature gradient and deformation, resulting in actual deformation compensation errors >0.1mm. The high CTE characteristics of copper are not quantified and incorporated into the compensation formula, leading to deterioration of flatness after long-term thermal cycling. Traditional sandblasting processes easily embed copper shavings, and activation treatment relies on sulfuric acid pickling, resulting in insufficient interfacial bonding. Summary of the Invention

[0003] The purpose of this invention is to provide a cold forging and nickel plating process for copper IGBT heat dissipation substrates to solve the aforementioned technical defects.

[0004] To achieve the above effects, the technical solution adopted by the present invention is as follows: a cold forging and nickel plating process for a copper IGBT heat dissipation substrate, comprising the following steps:

[0005] Step 1, Blanking and Material Inspection: Select T2 copper strip and use laser cutting to obtain copper blanks;

[0006] Step 2, Dynamic Step Annealing: The step annealing parameters output in the above steps are transmitted to the temperature controller of the annealing furnace as the process instructions for the first stage of dynamic step annealing, guiding the temperature and time control of the annealing process; during the annealing process, the temperature controller continuously collects the real-time temperature in the annealing furnace through the temperature sensor, and makes annealing temperature correction according to the current target temperature in each stage.

[0007] Step 3, Adaptive composite lubrication: The bottom layer of the annealed copper billet is coated with a borate coating, and the outer layer is coated with a nano-graphite emulsion;

[0008] Step 4, Cold forging: Using a DLC-coated mold, the copper billet is cold forged into a copper substrate;

[0009] Step 5, edge trimming and machining: Use a CNC lathe to machine positioning holes on the copper substrate and remove the burrs from the copper substrate;

[0010] Step 6, Stress-relief annealing: Perform stress-relief annealing treatment on the copper substrate;

[0011] Step 7, Cleaning and Ultrasonic-Assisted Acid Pickling: Perform ultrasonic acid pickling on the copper substrate.

[0012] Step 8, Variable parameter non-contact sandblasting: Use spherical ceramic pellets to perform non-contact sandblasting on the copper substrate;

[0013] Step 9, Predictive Pre-bending: Set the initial pre-bending angle to 1.8°, output the corrected pre-bending angle based on the matching relationship between the pre-bending angle and the predicted deformation, convert the corrected pre-bending angle into a control command, and transmit it to the servo control system of the pre-bending machine. After receiving the command, the pre-bending machine adjusts the mold angle from the initial value, and then pre-bends the copper substrate to form a pre-bending shape that matches the predicted welding deformation before welding.

[0014] Step 10, Real-time Control of Electroless Nickel Plating: The copper substrate, after being processed in the previous step, is first immersed in sodium pyrophosphate activation solution for pre-immersion activation; then the copper substrate is subjected to ultrasonic-assisted acid washing; finally, the copper substrate is transferred to an electroless nickel plating bath for nickel plating.

[0015] Preferably, in step 1, the oxygen content of the blank is detected by a spectrometer, the thickness of the blank is detected by a laser thickness gauge, the oxygen content and thickness of the blank are input into the material property parameter mapping module, and the step annealing parameters are output by the material property parameter mapping module.

[0016] Preferably, the material property parameter mapping module outputs the stepped annealing parameters based on the oxygen content and thickness of the blank as follows: the oxygen content of the copper blank is denoted as O, and the thickness of the copper blank is d; a pre-trained random forest model is used to construct a nonlinear mapping between material properties and annealing parameters; wherein the random forest model contains two sub-modules: a temperature prediction sub-module and a time prediction sub-module.

[0017] Let the input feature vector be denoted as X = [O, d], then the temperature prediction sub-model is T1 = f RF1 (X), the time prediction sub-model is t1 = f RF2 (X); T1 is the temperature of the first stage of stepped annealing, t1 is the time of the first stage of stepped annealing, f RF1 f RF2 The random forest function is trained on historical data. The model training data covers the mapping relationship between different combinations of oxygen content and thickness and the corresponding optimal annealing parameters.

[0018] Preferably, in step 2, the target annealing temperature for the first stage is set to T. 1-目标 The target annealing time for the first stage is t. 1-目标 The target annealing temperature for the second stage is T. 2-目标 The target annealing time for the second stage is t. 2-目标 The real-time furnace temperature collected by the temperature sensor is T(t), and the allowable fluctuation range of the furnace temperature is ΔT = ±2℃. The heating power output by the temperature controller is P(t); the furnace temperature deviation ΔT(t) = T(t) - T 目标 ;T 目标 The target temperature for the current stage is T. (The first stage is set to T.) 1-目标 The second stage takes T 2-目标 The first stage target annealing temperature T 1-目标 =T1, the first stage target annealing time t 1-目标 =t1, the target annealing temperature T in the second stage 2-目标 =250℃, second-stage target annealing time t 2-目标 =2h; The heating power P(t) is dynamically adjusted according to the real-time furnace temperature deviation ΔT(t) through the built-in closed-loop control algorithm of the temperature controller to ensure that the furnace temperature is stable within the target range. Inside.

[0019] Preferably, in step 3, the thickness of the borate coating is 3.5 μm, and the initial thickness of the nano-graphite emulsion is 5 μm. Preferably, in step 4, a piezoelectric stress sensor is used to monitor the cold forging stress σ in real time. When σ > 320 MPa, the lubrication system is triggered to add 1 μm of nano-graphite emulsion, adjusting the total thickness of the lubrication layer to 9 μm-10 μm, and the friction coefficient μ < 0.03. Preferably, in step 8, the surface of the sandblasted copper substrate is tested with a roughness tester and Ra = 1.2 μm before proceeding to the next process.

[0020] Preferably, the process of outputting the corrected pre-bending angle based on the matching relationship between the pre-bending angle and the predicted deformation is as follows:

[0021] An infrared temperature sensor captures the temperature gradient ΔTD during the welding process in real time. ΔTD is then input into a pre-trained backpropagation (BP) neural network model, based on the theoretical thermal expansion formula ΔL. 理论 =α·L0·ΔTD is the deviation from the actual length of the copper substrate, and the output deformation ΔL is the coefficient of thermal expansion of copper, α = 16.5 × 10⁻⁶. -6 / K, L0 is the original length of the copper substrate;

[0022] Based on the matching relationship between the pre-bending angle and the predicted deformation, a modified formula is adopted: The corrected pre-bending angle θ is calculated, where The radian value corresponding to the deformation, multiplied by Convert to angle.

[0023] Preferably, in step 10, during the electroless nickel plating process, the pH sensor detects the pH value of the plating solution in real time at a sampling interval of ≤1s. When the pH value of the plating solution is detected to be lower than the lower limit of the preset target range, the plating solution control unit is triggered to add sodium citrate to the plating solution at an addition amount of 2g / L. During the electroless nickel plating process, the temperature sensor monitors the temperature of the plating solution in real time at a sampling interval of ≤1s. When the temperature of the plating solution is detected to be lower than the lower limit of the preset target range, the plating solution control unit is triggered to increase the concentration of sodium hypophosphite in the plating solution by an increase amount of 1g / L.

[0024] Preferably, in step 10, after electroless nickel plating, the coating thickness is measured using an eddy current thickness gauge, and the coating porosity is measured using a porosity meter; wherein the coating thickness is controlled within the range of 12-15 μm, and the coating porosity is controlled within 0.5 porosities / cm. 2 Within.

[0025] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0026] 1. Through the collaborative design of blanking detection and dynamic step annealing, the problem that the fixed parameters in the traditional process cannot adapt to the characteristics of different batches of copper materials is solved. It can accurately adjust the annealing conditions according to the actual oxygen content and thickness of the copper blanks, effectively control the grain growth state and fully eliminate the rolling stress, ensuring that the mechanical properties and processing properties of different batches of substrates are consistent. At the same time, the combined use of adaptive composite lubrication and DLC-coated molds can dynamically optimize the lubrication effect according to the actual stress changes during the cold forging process, avoid stress concentration or copper sticking phenomena that are prone to occur under traditional lubrication methods, and greatly reduce the crack risk in the cold forging forming stage, providing a copper substrate base with stable quality and uniform performance for subsequent processes, and ensuring the processing reliability of the overall process from the source.

[0027] 2. By real-time capturing the welding temperature gradient and combining with a professional model to predict the welding deformation amount, dynamically correcting the pre-bending angle to achieve precise compensation for welding deformation, and utilizing the correlation between the thermal expansion characteristics of copper and the welding temperature fluctuation, the copper substrate can still maintain excellent flatness after welding, avoiding the problem of excessive welding deformation due to insufficient adaptability of the pre-bending angle in the traditional process, meeting the stringent requirements of high-power IGBT modules for the packaging flatness of the heat dissipation substrate, and at the same time improving the structural stability of the substrate under long-term thermal cycling conditions and extending the service life of the IGBT module.

[0028] 3. By dynamically monitoring and precisely adjusting the plating solution parameters, the problem of unstable coating quality caused by parameter fluctuations in the traditional nickel plating process is effectively solved. It can form a nickel plating layer with higher density and stronger bonding force, significantly improving the corrosion resistance of the copper substrate and meeting the usage requirements under complex working conditions. In addition, the combination of variable parameter non-contact sandblasting and ultrasonic-assisted pickling can ensure that the surface roughness of the substrate meets the requirements of nickel plating while avoiding overprocessing or impurity residues, further ensuring the coating quality. The collaborative design of the full-process technology and the detection link reduces the manual intervention and rework frequency in the traditional process, improves the production efficiency and product qualification rate, and can adapt to the dual requirements of industrialized large-scale production for process stability and efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention, and therefore should not be regarded as limiting the scope. For those of ordinary skill in the art, other related drawings can be obtained based on these drawings without creative efforts.

[0030] Figure 1 is a flowchart of a cold forging forming nickel plating process for a copper IGBT heat dissipation substrate according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0031] The present invention will be further explained below with reference to the accompanying drawings and specific embodiments.

[0032] Example 1

[0033] Please see Figure 1 As shown, this embodiment discloses a cold forging and nickel plating process for a copper IGBT heat dissipation substrate, including the following steps:

[0034] Step 1, Blanking and Material Inspection: Select T2 copper strip with a thickness of 6-12mm and an oxygen content ≤0.001%. Use laser cutting to obtain copper blanks. Detect the oxygen content of the blanks using a spectrometer and the thickness of the blanks using a laser thickness gauge. Input the oxygen content and thickness of the blanks into the material property parameter mapping module, and use the material property parameter mapping module to output the step annealing parameters.

[0035] Specifically, the material property parameter mapping module outputs the stepped annealing parameters based on the oxygen content and thickness of the blank as follows: the oxygen content of the copper blank is denoted as O, and the thickness of the copper blank is d; a pre-trained random forest model is used to construct a nonlinear mapping between material properties and annealing parameters; the random forest model contains two sub-modules: a temperature prediction sub-module and a time prediction sub-module.

[0036] Let the input feature vector be denoted as X = [O, d], then the temperature prediction sub-model is T1 = f RF1 (X), the time prediction sub-model is t1 = f RF2 (X); T1 is the temperature of the first stage of stepped annealing, t1 is the time of the first stage of stepped annealing, f RF1 f RF2 The model is a random forest function trained on historical data. The training data covers the mapping relationship between different oxygen content, thickness combinations and corresponding optimal annealing parameters. The prediction accuracy after training exceeds 98%.

[0037] It should be noted that the random forest sub-model first constructs a historical dataset covering the oxygen content, thickness, and corresponding optimal annealing parameters of copper and divides it into training, validation, and test sets; then, through hyperparameter optimization, it trains multiple decision trees using Bootstrap sampling and integrates them to obtain random forest sub-models for temperature and time prediction respectively; finally, the model is deployed to achieve real-time mapping of material properties to annealing parameters.

[0038] In summary, by dynamically adapting the stepped annealing parameters, the following advantages are achieved: First, it addresses the process compatibility issues arising from differences in oxygen content and thickness among different batches of copper, ensuring stable billet grain size at ASTM 8-9 level and effective rolling stress relief, thereby improving material performance consistency. Second, it provides high-quality substrates for subsequent cold forging and nickel plating processes, reducing defects such as cold forging cracks and insufficient coating adhesion, thus improving product qualification rate and process stability. Third, it avoids over- / under-annealing, reducing rework rate and energy consumption, improving production efficiency, saving costs, and supporting the overall process upgrade towards intelligence and precision.

[0039] Step 2, Dynamic Step Annealing: The step annealing parameters output in the above steps are transmitted to the temperature controller of the annealing furnace as the process instructions for the first stage of dynamic step annealing, guiding the temperature and time control of the annealing process; during the annealing process, the temperature controller continuously collects the real-time temperature in the annealing furnace through the temperature sensor, and makes annealing temperature correction according to the current target temperature in each stage.

[0040] Specifically, the target annealing temperature for the first stage is set to T. 1-目标 The target annealing time for the first stage is t. 1-目标 The target annealing temperature for the second stage is T. 2-目标 The target annealing time for the second stage is t. 2-目标 The real-time furnace temperature collected by the temperature sensor is T(t), the allowable fluctuation range of the furnace temperature is ΔT=±2℃, and the heating power output by the temperature controller is P(t).

[0041] Therefore, the furnace temperature deviation ΔT(t) = T(t) - T 目标 ;T 目标 The target temperature for the current stage is T. (The first stage is set to T.) 1-目标 The second stage takes T 2-目标 The first stage target annealing temperature T 1-目标 =T1, the first stage target annealing time t 1-目标 =t1, the target annealing temperature T in the second stage 2-目标 =250℃, second-stage target annealing time t 2-目标 =2h; where the specific values ​​of the second-stage target annealing temperature and the second-stage target annealing time are the optimal fixed parameters determined after verification through a large number of process experiments based on the recrystallization behavior and grain size control requirements of copper.

[0042] The heating power P(t) is dynamically adjusted based on the real-time furnace temperature deviation ΔT(t) using a built-in closed-loop control algorithm in the temperature controller, ensuring that the furnace temperature remains stable within the target range. T目标 -ΔT,T 目标 Within +ΔT]; taking the first stage as an example, the control logic is:

[0043] When T(t) > T1-目标 At +2, the temperature controller reduces the output heating power P(t), reduces heat input, and causes the furnace temperature to drop;

[0044] When T(t) < T 1-目标 At -2 degrees Celsius, the temperature controller increases the output heating power P(t) to increase heat input and cause the furnace temperature to drop.

[0045] When T 1-目标 -2≤T(t)≤T 1-目标 At +2, maintain the current output heating power P(t) to ensure stable furnace temperature.

[0046] During the second stage of annealing, the furnace temperature is similarly controlled at T. 2-目标 Within ±2℃ range.

[0047] In summary, by monitoring and automatically correcting the furnace temperature in real time, the following advantages are achieved: First, it ensures consistent annealing results across different batches of copper billets, stabilizing the billet grain size at ASTM 8-9 level and achieving a residual stress relief rate of ≥90%, thus improving the consistency of material properties and providing a high-quality, uniform substrate for subsequent cold forging and nickel plating processes. Second, it reduces defects such as cold forging cracks and insufficient coating adhesion caused by fluctuations in annealing quality, lowering rework and scrap rates and improving production efficiency. Third, precise temperature control optimizes annealing energy consumption, contributing to overall process cost control and intelligent upgrades.

[0048] Step 3, Adaptive composite lubrication: The bottom layer of the annealed copper billet is coated with a borate coating, and the outer layer is coated with a nano-graphite emulsion; wherein, the thickness of the borate coating is 3.5μm, and the initial thickness of the nano-graphite emulsion is 5μm.

[0049] Step 4, Cold Forging: Using a DLC-coated mold with a hardness of HV4000, and setting the cold forging rate to 5mm / s, the copper billet is cold forged into a copper substrate. A piezoelectric stress sensor is used to monitor the cold forging stress σ in real time. When σ > 320MPa, the lubrication system is triggered to add 1μm nano-graphite emulsion, and the total thickness of the lubrication layer is adjusted to 9μm-10μm, with a friction coefficient μ < 0.03.

[0050] Step 5, edge trimming and machining: Use a CNC lathe to machine positioning holes on the copper substrate and remove the burrs from the copper substrate.

[0051] Step 6, Stress-relief annealing: The copper substrate is subjected to stress-relief annealing at a fixed temperature of 300℃ for 1.5h. The hardness of the copper substrate is controlled at HV75±3. The furnace temperature is monitored by a temperature sensor, and an alarm command is issued when the temperature exceeds the tolerance.

[0052] Step 7, Cleaning and Ultrasonic Acid Pickling: Set the frequency to 40kHz, and use a 10% sulfuric acid solution combined with a 2% hydrogen peroxide solution to perform ultrasonic acid pickling on the copper substrate at a temperature of 50℃ for 10 minutes.

[0053] Step 8, Variable parameter non-contact sandblasting: Use spherical ceramic pellets with a specification of Φ0.2mm, set the sandblasting pressure to 0.3MPa, and the sandblasting time on the copper substrate to 30s. After sandblasting, the surface of the copper substrate is tested by a roughness tester and Ra=1.2μm before proceeding to the next process.

[0054] Step 9, Predictive Pre-bending: Set the initial pre-bending angle to 1.8°. Based on the matching relationship between the pre-bending angle and the predicted deformation, output the corrected pre-bending angle. Convert the corrected pre-bending angle into a control command and transmit it to the servo control system of the pre-bending machine. After receiving the command, the pre-bending machine adjusts the mold angle from the initial value and then pre-bends the copper substrate to form a pre-bending shape that matches the predicted welding deformation before welding.

[0055] Specifically, the process of outputting the corrected pre-bending angle based on the matching relationship between the pre-bending angle and the predicted deformation is as follows:

[0056] An infrared temperature sensor captures the temperature gradient ΔTD during the welding process in real time. ΔTD is then input into a pre-trained backpropagation (BP) neural network model, based on the theoretical thermal expansion formula ΔL. 理论 =α·L0·ΔTD is the deviation from the actual length of the copper substrate, and the output deformation ΔL is the coefficient of thermal expansion of copper, α = 16.5 × 10⁻⁶. -6 / K, L0 is the original length of the copper substrate;

[0057] Based on the matching relationship between the pre-bending angle and the predicted deformation, a modified formula is adopted: The corrected pre-bending angle θ is calculated, where The radian value corresponding to the deformation, multiplied by Convert to angle.

[0058] In summary, by capturing the welding temperature gradient in real time using an infrared temperature sensor and combining it with a BP neural network to accurately predict welding deformation and dynamically correct the pre-bending angle, the limitations of the traditional fixed pre-bending angle are broken. This allows the pre-bending angle to adapt to real-time welding temperature fluctuations, precisely controlling the post-weld deformation to ≤0.03mm, far superior to the ≤0.05mm of the traditional process. This significantly improves the flatness accuracy of the IGBT heat sink substrate, meeting the stringent requirements of high-power semiconductor devices for package flatness. It reduces rework and scrap caused by flatness deviations, improves product qualification rate, enhances process stability, and better adapts to the needs of large-scale production. It also enables the process to have dynamic adaptability, coping with temperature fluctuations under different operating conditions and enhancing process robustness. At the same time, it provides a flatter substrate for subsequent processes such as nickel plating, indirectly ensuring the uniformity of plating quality and helping the overall process develop towards intelligence and high precision.

[0059] Step 10, Real-time Control of Electroless Nickel Plating: The copper substrate treated in the previous process is first immersed in sodium pyrophosphate activation solution for pre-immersion activation: the temperature of the activation solution is controlled at 50℃ and the activation time is 5min, in order to remove the potential passivation film on the copper surface and enhance the adhesion of the subsequent nickel plating layer; then the copper substrate is subjected to ultrasonic-assisted pickling. The pickling system is a 10% sulfuric acid solution combined with a 2% hydrogen peroxide solution. The copper substrate is subjected to ultrasonic pickling treatment at a temperature of 50℃ for 10min to remove residual impurities on the surface. After that, the copper substrate is transferred to the electroless nickel plating bath for nickel plating treatment.

[0060] It should be noted that the plating solution formula is as follows: nickel sulfate: 25g / L; sodium hypophosphite: 30g / L; composite complexing agent (sodium citrate + glycine): 45g / L; stabilizer (thiourea): 1ppm.

[0061] Specifically, during the electroless nickel plating process, a pH sensor monitors the pH value of the plating solution in real time at sampling intervals of ≤1 second. When the pH value of the plating solution is detected to be lower than the lower limit of the preset target range, the plating solution control unit is triggered to add sodium citrate to the plating solution at a concentration of 2 g / L. Sodium citrate acts as a complexing agent, and its concentration changes alter the complexation equilibrium of the plating solution: increasing the concentration of sodium citrate decreases the concentration of free nickel ions in the plating solution, thus reducing the hydrolysis reaction. Shift the balance to the left, H + As the concentration decreased, the pH value increased. After 5 minutes of reaction and mixing with the plating solution, the pH value stabilized at 4.6, which is within the target control range of 4.5 ± 0.1.

[0062] During the electroless nickel plating process, a temperature sensor monitors the plating bath temperature in real time at sampling intervals of ≤1 second. When the temperature of the plating bath is detected to be lower than the lower limit of the preset target range, the plating bath control unit is triggered to increase the concentration of sodium hypophosphite in the plating bath by 1 g / L. Sodium hypophosphite is a reducing agent in electroless nickel plating, and it participates in the reduction reaction. Ni 2+ +2e - →Ni - As the reaction is exothermic, increasing the sodium hypophosphite concentration accelerates the reduction reaction rate, increases the heat released by the reaction, and raises the temperature of the plating solution. After 3 minutes of reaction and heat transfer, the temperature of the plating solution stabilizes at 86℃, which is within the target control range of 87℃±2℃.

[0063] After electroless nickel plating, the coating thickness is measured using an eddy current thickness gauge, and the coating porosity is measured using a porosity meter. The coating thickness must meet the design requirement of 12-15 μm, and the coating porosity must be below the control requirement of 0.5 porosities / cm². 2 The surface coating has good density and can effectively block the penetration of corrosive media.

[0064] In summary, this invention solves the problem of traditional processes where fixed parameters cannot adapt to the characteristics of different batches of copper materials by combining material feeding detection and dynamic stepped annealing. It allows for precise adjustment of annealing conditions based on the actual oxygen content and thickness of the copper billet, effectively controlling grain growth and fully eliminating rolling stress, ensuring consistent mechanical and processing properties across different batches of substrate. Simultaneously, the combined use of adaptive composite lubrication and DLC-coated molds dynamically optimizes lubrication based on actual stress changes during cold forging, avoiding stress concentration or copper adhesion issues common with traditional lubrication methods. This significantly reduces the risk of cracking during the cold forging stage, providing a stable and uniform copper substrate for subsequent processes, thus ensuring overall process reliability from the outset.

[0065] By capturing the welding temperature gradient in real time and combining it with a professional model to predict the amount of welding deformation, the pre-bending angle is dynamically corrected to achieve precise compensation for welding deformation. By utilizing the correlation between the thermal expansion characteristics of copper and welding temperature fluctuations, the copper substrate can maintain excellent flatness after welding, avoiding the problem of excessive deformation after welding caused by insufficient adaptability of the pre-bending angle in traditional processes. This meets the stringent requirements of high-power IGBT modules for the flatness of the heat dissipation substrate packaging, while improving the structural stability of the substrate under long-term thermal cycling conditions and extending the service life of the IGBT module.

[0066] By dynamically monitoring and precisely adjusting the plating solution parameters, the problem of unstable coating quality caused by parameter fluctuations in the traditional nickel plating process is effectively solved. A nickel plating layer with higher density and stronger bonding force can be formed, significantly improving the corrosion resistance of the copper substrate and meeting the usage requirements under complex working conditions. In addition, the combination of variable parameter non-contact sandblasting and ultrasonic-assisted pickling can ensure that the surface roughness of the substrate meets the requirements of nickel plating while avoiding overprocessing or impurity residue, further ensuring the coating quality. The collaborative design of the full-process technology and the detection link reduces the manual intervention and rework frequency in the traditional process, improves the production efficiency and product qualification rate, and can adapt to the dual requirements of industrial scale production for process stability and efficiency. At the same time, the content not described in detail in this specification belongs to the prior art well-known to those skilled in the art.

[0067] The present invention is not limited to the above optional embodiments, and anyone can obtain other various forms of products under the inspiration of the present invention. The above specific embodiments should not be construed as limiting the protection scope of the present invention, and the protection scope of the present invention should be defined by the claims, and the specification can be used to interpret the claims.

Claims

1. A cold forging and nickel plating process for a copper IGBT heat dissipation substrate, characterized in that, Includes the following steps: Step 1, Blanking and Material Inspection: Select T2 copper strip and use laser cutting to obtain copper blank; use a spectrometer to detect the oxygen content of the blank and a laser thickness gauge to detect the thickness of the blank; input the oxygen content and thickness of the blank into the material property parameter mapping module, and use the material property parameter mapping module to output the step annealing parameters. Step 2, Dynamic Step Annealing: The step annealing parameters output in the above steps are transmitted to the temperature controller of the annealing furnace as the process instructions for the first stage of dynamic step annealing, guiding the temperature and time control of the annealing process. During the annealing process, the temperature controller continuously collects the real-time temperature inside the annealing furnace through the temperature sensor, and adjusts the annealing temperature according to the current target temperature in stages. Step 3, Adaptive composite lubrication: The bottom layer of the annealed copper billet is coated with a borate coating, and the outer layer is coated with a nano-graphite emulsion; Step 4, Cold forging: Using a DLC-coated mold, the copper billet is cold forged into a copper substrate; Step 5, edge trimming and machining: Use a CNC lathe to machine positioning holes on the copper substrate and remove the burrs from the copper substrate; Step 6, Stress-relief annealing: Perform stress-relief annealing treatment on the copper substrate; Step 7, Cleaning and Ultrasonic-Assisted Acid Pickling: Perform ultrasonic acid pickling on the copper substrate. Step 8, Variable parameter non-contact sandblasting: Use spherical ceramic pellets to perform non-contact sandblasting on the copper substrate; Step 9, Predictive Pre-bending Forming: Set the initial pre-bending angle to 1.8°, output the corrected pre-bending angle based on the matching relationship between the pre-bending angle and the predicted deformation, convert the corrected pre-bending angle into a control command, and transmit it to the servo control system of the pre-bending machine. After receiving the command, the pre-bending machine adjusts the mold angle from the initial pre-bending angle to the corrected pre-bending angle, and then pre-bends the copper substrate to form a pre-bending shape that matches the predicted welding deformation before welding. The process of outputting the corrected pre-bending angle based on the matching relationship between the pre-bending angle and the predicted deformation is as follows: Infrared temperature sensors capture the temperature gradient during the welding process in real time. ,Will Input a pre-trained BP neural network model, based on the theoretical thermal expansion formula Deviation from the actual length of the copper substrate, output deformation amount Among them, the coefficient of thermal expansion of copper is... , This represents the original length of the copper substrate. Based on the matching relationship between the pre-bending angle and the predicted deformation, a modified formula is adopted: The corrected pre-bending angle was calculated. ,in The radian value corresponding to the deformation, multiplied by Convert to angle; Step 10, Real-time Control of Electroless Nickel Plating: The copper substrate, after being processed in the previous step, is first immersed in sodium pyrophosphate activation solution for pre-immersion activation; then the copper substrate is subjected to ultrasonic-assisted acid washing; finally, the copper substrate is transferred to an electroless nickel plating bath for nickel plating.

2. The cold forging and nickel plating process for a copper IGBT heat dissipation substrate according to claim 1, characterized in that, The material property parameter mapping module outputs the stepped annealing parameters based on the oxygen content and thickness of the blank in the following way: The oxygen content of the copper blank is denoted as... The thickness of the copper blank is A pre-trained random forest model is used to construct a nonlinear mapping between material properties and annealing parameters; the random forest model contains two sub-modules: a temperature prediction sub-model and a time prediction sub-module. Let the input feature vector be denoted as The temperature prediction sub-model is then... The time prediction sub-model is ; This refers to the temperature of the first stage of stepped annealing. This refers to the first stage of stepped annealing. , The random forest function is trained on historical data. The model training data covers the mapping relationship between different combinations of oxygen content and thickness and the corresponding optimal annealing parameters.

3. The cold forging and nickel plating process for a copper IGBT heat dissipation substrate according to claim 2, characterized in that, In step 2, the target annealing temperature for the first stage is set as follows: The target annealing time for the first stage is... The target annealing temperature for the second stage is The target annealing time for the second stage is... The real-time furnace temperature collected by the temperature sensor is The allowable fluctuation range of furnace temperature is The thermostat controls the output heating power to be Furnace temperature deviation ; The target temperature for the current stage is [temperature value], and the first stage is [temperature value]. The second phase ; Target annealing temperature for the first stage The first stage target annealing time The second stage target annealing temperature The second stage target annealing time The temperature controller uses a built-in closed-loop control algorithm to adjust the real-time furnace temperature deviation. Dynamically adjust heating power To ensure that the furnace temperature remains stable within the target range Inside.

4. The cold forging and nickel plating process for a copper IGBT heat dissipation substrate according to claim 1, characterized in that, In step 3, the thickness of the borate coating is 3.5 μm, and the initial thickness of the nano-graphite emulsion is 5 μm.

5. The cold forging and nickel plating process for a copper IGBT heat dissipation substrate according to claim 1, characterized in that, In step 4, a piezoelectric stress sensor is used to monitor the cold forging stress σ in real time. When σ > 320 MPa, the lubrication system is triggered to add 1 μm nano-graphite emulsion, and the total thickness of the lubrication layer is adjusted to 9 μm-10 μm, with a friction coefficient μ < 0.

03.

6. The cold forging and nickel plating process for a copper IGBT heat dissipation substrate according to claim 1, characterized in that, In step 8, the surface of the sandblasted copper substrate is tested with a roughness tester and found to be Ra=1.2μm before proceeding to the next process.

7. The cold forging and nickel plating process for a copper IGBT heat dissipation substrate according to claim 1, characterized in that, In step 10, during the electroless nickel plating process, the pH sensor detects the pH value of the plating solution in real time at a sampling interval of ≤1s. When the pH value of the plating solution is detected to be lower than the lower limit of the preset target range, the plating solution control unit is triggered to add sodium citrate to the plating solution at a rate of 2g / L. During the electroless nickel plating process, the temperature sensor monitors the temperature of the plating solution in real time at a sampling interval of ≤1s. When the temperature of the plating solution is detected to be lower than the lower limit of the preset target range, the plating solution control unit is triggered to increase the concentration of sodium hypophosphite in the plating solution by 1g / L.

8. The cold forging and nickel plating process for a copper IGBT heat dissipation substrate according to claim 1, characterized in that, In step 10, after electroless nickel plating, the coating thickness is measured using an eddy current thickness gauge, and the coating porosity is measured using a porosity meter; the coating thickness is controlled within the range of 12-15 μm, and the coating porosity is controlled within 0.5 porosities / cm. 2 Within.

Images