A low-cavity no-clean solder paste welding method for semiconductor packaging

Abstract

This invention belongs to the field of semiconductor packaging and electronic assembly technology, specifically a low-void no-clean solder paste soldering method for semiconductor packaging. The method includes: applying no-clean solder paste to establish a pre-soldering reference; gradient heating to controllable solvent evaporation and complete the reduction reaction; inducing a hydrodynamic imbalance logic during the solder melting period, using the surface tension gradient generated by thermo-chemical synergistic induction to drive Marangoni flow, breaking the bubble interface adhesion balance, and guiding the bubbles to migrate directionally towards the edge and escape in accordance with the flow field vector; finally, performing controlled cooling and crystallization. This application can achieve ultra-low void ratio soldering under normal atmospheric conditions, significantly reducing packaging costs and improving the thermal conductivity reliability of power devices.

Description

Technical Field

[0001] This invention belongs to the field of semiconductor packaging and electronic assembly technology, specifically a method for soldering low-void, no-clean solder paste for semiconductor packaging. Background Technology

[0002] With the development of third-generation semiconductor power devices and advanced packaging technology, wide-bandgap semiconductors such as silicon carbide have become the core of the power electronics field. The thermal management efficiency of semiconductor packaging layer determines the service life and reliability of power modules. Solder joint voids will reduce the effective heat conduction area, increase local thermal resistance, induce hot spot effect, and even chip failure. Achieving ultra-low void ratio soldering connection is the core goal of the industry. Solder voids originate from the thermal decomposition of flux and the deoxidation reaction of the metal surface. Current mainstream no-clean soldering technology improves solder wettability by enhancing flux activity, but its highly active components will generate a large number of gaseous products at high temperatures, exacerbating voids. Furthermore, residual organic matter can easily cause electromigration or leakage current risks, threatening the long-term electrical reliability of devices.

[0003] While vacuum reflow equipment used in high-end packaging can alleviate the void problem, the equipment is expensive and lengthens the production cycle. Furthermore, air bubbles in narrow gap solder joints are locked due to insufficient buoyancy and interfacial tension. Existing solder paste lacks the ability to dynamically control the flow of molten solder, and the reduction in pad size further hinders the removal of air bubbles. Existing technologies are unable to overcome the air bubble locking barrier, and there is an urgent need for new soldering technologies that can actively remove voids.

[0004] Therefore, the present invention provides a method for soldering low-void, no-clean solder paste for semiconductor packaging. Summary of the Invention

[0005] In order to overcome the shortcomings of the prior art, at least one technical problem raised in the background art is solved.

[0006] The technical solution adopted by this invention to solve its technical problem is: a method for soldering low-void, no-clean solder paste for semiconductor packaging, comprising the following process steps: The pre-soldering baseline construction step involves applying a specific formulation of no-clean solder paste to the metallization area of ​​a semiconductor substrate using high-precision printing or dispensing processes. The substrate is typically a copper-clad ceramic substrate or a lead frame with high heat dissipation characteristics. The surface metallization layer has a predetermined metallurgical affinity with the metal alloy powder in the no-clean solder paste. The no-clean solder paste contains spherical solder alloy powder with multi-level particle size distribution and a flux carrier with rheological control characteristics. In this step, the no-clean solder paste establishes an initial physical connection orientation between the chip and the substrate through its inherent viscosity and thixotropy, and reserves the necessary chemical reduction for subsequent thermophysical evolution.

[0007] A gradient heating preheating step is performed, in which the assembly carrying the semiconductor chip is placed in the preheating section of the multi-temperature zone reflow system. During this process, the system temperature rises regularly according to a preset energy gradient, causing the solvent components in the flux carrier to undergo controlled physical volatilization. The gradient heating process ensures that the escape rate of solvent molecules is lower than the critical threshold for bubble nucleus formation by precisely adjusting the heat flux, thereby avoiding the formation of macroscopic physical defects before the solder melts. Meanwhile, after reaching a specific activation energy level, the active reducing component in the flux begins to undergo a heterogeneous deoxidation reaction with the oxides on the surface of the metallization layer and alloy powder, generating a metal soap intermediate with wetting gain effect, and simultaneously adjusting the energy level distribution of the interface, thus setting the interface conditions for the subsequent active desorption of bubbles.

[0008] During the solder melting triggering stage, the system temperature crosses the liquidus line of the alloy powder, causing the solid solder to transform into a molten liquid with flow characteristics. Due to the thermal decomposition of the organic components inside the flux at instantaneous high temperature, as well as the aggregation of redox reaction byproducts, initial bubble nuclei are inevitably generated inside the weld. During the window period when the solder is in a molten state, a directional migration vector is constructed in the microspace of the weld by inducing hydrodynamic imbalance logic; Specifically, this method alters the local surface tension of the molten solder by establishing a microscopic asymmetric distribution of the thermal field or a chemical potential gradient at the welding interface, thereby generating Marangoni flow driven by the surface tension gradient. This fluid flow behavior has a specific dynamic direction and can generate fluid shear stress sufficient to overcome the adhesion force on the substrate surface.

[0009] The fluid dynamic imbalance logic breaks the bubble's locking effect through the following process: When the above-mentioned directional migration flow field is generated inside the molten solder, the bubbles that were originally attached to the center of the solder pad or the surface of the metallization layer are subjected to continuous momentum impact. Due to the non-uniform velocity distribution of the flow field on the weld cross section, a pressure difference will be generated on both sides of the bubble. When the pressure difference and the accompanying fluid drag exceed the interfacial adhesion force between the bubble and the solid interface, the bubble changes from a locked state to a free migration state. This process realizes an essential transformation from passive diffusion to active driving, so that within the extremely narrow weld gap (usually less than 100 micrometers), the bubble is no longer restricted by weak buoyancy, but migrates directionally towards the edge of the solder in accordance with the preset flow field vector.

[0010] During the directional migration and escape stage of the bubbles, the directional migration flow field continuously guides the desorbed bubbles through the high-viscosity molten solder region. This invention ensures that the solder maintains sufficient fluidity on the bubble migration path by precisely controlling the liquid viscosity temperature change curve of the flux, so as to reduce the viscous resistance of the bubble movement. The flow field effect, through momentum exchange between the solder edge and the central region, enables bubbles to overcome the surface tension barrier when they come into contact with the free interface between the solder and the external environment, thus achieving complete bubble collapse and escape. Due to the adoption of a no-cleaning system, after the flux guides the bubbles to escape, its residual components undergo a further polycondensation reaction at continuous high temperatures, transforming into an inert polymer film with high thermal stability and excellent insulation properties, which firmly wraps around the solder joint. This eliminates the need for subsequent water washing and meets the electrical reliability requirements for long-term operation of power semiconductors.

[0011] The controlled cooling crystallization process involves adjusting the cooling distribution in the cooling section to allow the molten solder, after bubble removal, to solidify sequentially from the edge to the center or from the bottom to the top. This controlled phase transformation process prevents the secondary generation of shrinkage voids caused by excessively rapid cooling and ensures the uniform growth of intermetallic compounds at the interface. The resulting welded joint layer has a dense microstructure and an extremely low void ratio, providing an excellent heat conduction path for semiconductor chips.

[0012] Preferably, the low-void no-clean solder paste soldering method for semiconductor packaging also introduces a physical field coupling enhancement mechanism during the melting window. This mechanism helps to break the pressure balance inside the bubbles by applying ultrasonic micro-vibration or periodic thermal field fluctuations at a preset frequency to the soldering area, causing the micro bubbles to merge into larger bubbles, thereby further reducing the critical energy required for Marangoni flow to drive the bubbles. The physical field coupling enhancement mechanism and the fluid dynamic imbalance logic support each other, ensuring that stable void control can be achieved for semiconductor packaging structures of different sizes and substrates.

[0013] Preferably, the no-clean solder paste contains a specific proportion of surface tension modifier, which can selectively accumulate in the high-temperature zone of reflow soldering. By establishing a preset chemical potential gradient on the solder surface, the intensity of the Marangoni flow is enhanced. This deep coupling between the chemical component and the physical flow field allows the method to achieve the void control level of traditional expensive vacuum reflow soldering under conventional atmospheric reflow conditions, significantly improving production efficiency and reducing fixed asset investment.

[0014] Preferably, in the gradient heating preheating step, the method employs a nonlinear temperature rise curve. By reducing the temperature rise rate during the critical interval of flux activity release, the gaseous byproducts generated by the reduction reaction have sufficient time to be discharged through the solder paste pores before the solder melts, thereby reducing the initial volume of bubbles in the melting stage. This pre-set void management logic and the active driving logic of the melting stage are nested layer by layer to construct a void suppression system throughout the entire life cycle.

[0015] Preferably, the welding method is particularly suitable for packaging high-power modules of third-generation semiconductors represented by silicon carbide and gallium nitride. Since the heat dissipation requirements of such devices are extremely high, this method ensures the thermal stability of the device under continuous high pulse current impact by eliminating the interface thermal resistance dead zone, and effectively suppresses the risk of chip cracking caused by thermal stress concentration.

[0016] The beneficial effects of this invention are as follows: 1. The low-void, no-clean solder paste soldering method for semiconductor packaging described in this invention generates a directional migration flow field by inducing hydrodynamic imbalance. This method fundamentally solves the problem of bubbles being locked by interfacial tension in narrow gap solder joints. The generation, desorption, directional migration, and escape of bubbles are all completed under the guidance of a preset dynamic vector, which significantly reduces the void ratio inside the solder joint.

[0017] 2. The low-void no-clean solder paste soldering method for semiconductor packaging described in this invention adopts a no-clean solder paste system and combines it with active degassing logic. This method eliminates the cumbersome cleaning process and overcomes the problem of high void rate in traditional no-clean solder paste. The flux residue is transformed into a high-performance protective film after soldering, ensuring the long-term reliability of power devices under high temperature, high humidity and high pressure environments, and eliminating the hidden dangers of electromigration and leakage current.

[0018] 3. The low-void, no-clean solder paste soldering method for semiconductor packaging described in this invention can achieve an ultra-low void ratio under conventional reflow conditions, completely eliminating the dependence on high-cost vacuum reflow equipment. By simplifying process equipment and shortening the reflow cycle, it significantly improves the output efficiency of mass production lines.

[0019] 4. The low-void, no-clean solder paste soldering method for semiconductor packaging described in this invention achieves a balance between void suppression and wetting performance through precise control of the thermal and chemical fields throughout the entire life cycle. The solder joint microstructure is uniform and the intermetallic compound thickness is moderate, which not only ensures excellent thermal conductivity but also significantly enhances the mechanical properties and fatigue life of the solder joint. Attached Figure Description

[0020] The invention will now be further described with reference to the accompanying drawings.

[0021] Figure 1 This is a flowchart of a method for soldering low-void, no-clean solder paste for semiconductor packaging according to the present invention. Detailed Implementation

[0022] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0023] like Figure 1 As shown, an embodiment of the present invention provides a method for soldering low-void, no-clean solder paste for semiconductor packaging, comprising the following steps: The pre-soldering baseline construction step involves applying a specially formulated no-clean solder paste to the metallized areas of a semiconductor substrate using high-precision printing or dispensing processes. The substrate is typically a copper-clad ceramic substrate or leadframe with high heat dissipation properties. Its surface metallization layer (such as electroless nickel-gold, electroplated silver, or bare copper) has a pre-defined metallurgical affinity with the metal alloy powder in the no-clean solder paste. This affinity ensures that during the subsequent melting process, the solder can rapidly spread on the substrate surface and form a continuous intermetallic compound layer. The no-clean solder paste is designed with the following components... The invention incorporates spherical solder alloy powder with multi-level particle size distribution and flux carrier with rheological control properties. Multi-level particle size distribution refers to achieving the highest packing density by precisely proportioning powder particles of different diameters, thereby reducing the original gap volume inside the solder paste. The flux carrier, through its inherent viscosity and thixotropy, establishes the initial physical connection orientation between the chip and the substrate. This thixotropy ensures that the solder paste has good fluidity when subjected to printing shear force, while maintaining a stable geometry in a static state, thus pre-setting the necessary chemical reduction reserve for subsequent thermophysical evolution.

[0024] A gradient temperature rise preheating step is performed. During this stage, the assembly carrying the semiconductor chip is sent to the preheating section of the multi-temperature zone reflux system. During this process, the system temperature rises regularly according to the preset energy gradient. This energy gradient is not set in a linear fashion, but is optimized nonlinearly based on the saturated vapor pressure curves of solvents with different boiling points in the flux carrier. Specifically, the gradient heating process ensures that the escape rate of solvent molecules is always below the critical threshold for bubble nucleus formation by precisely adjusting the heat flux. If the temperature rise rate is too fast, the violent vaporization of the solvent inside the solder paste will cause bubbles to aggregate and form macroscopic physical defects, or even cause solder bursting. At the same time, after reaching a specific activation energy level, the active reducing components in the flux (such as organic acids, organic amines or synthetic resins) begin to undergo heterogeneous deoxidation reactions with the oxides on the surface of the metallization layer and alloy powder. The essence of this reaction is a chemical reduction process, and the metal soap intermediates generated have a significant wetting gain effect, which can reduce the spreading resistance of the solder after melting. More importantly, this process simultaneously modulates the energy level distribution at the solid-liquid interface, thereby changing the interfacial energy and pre-setting the microscopic physical conditions for the subsequent active desorption of bubbles.

[0025] During the solder melting triggering stage, the system temperature field precisely crosses the liquidus line of the alloy powder; At this point, the solid solder lattice collapses and transforms into a molten liquid with flow characteristics. Due to the thermal decomposition of the organic components inside the flux at instantaneous high temperature and the aggregation of byproduct gases produced by the redox reaction, initial bubble nuclei are inevitably generated inside the weld. It should be noted that during the window period when the solder is in a molten state, a directional migration vector is constructed in the extremely narrow weld microspace by inducing hydrodynamic imbalance logic. The essence of this hydrodynamic imbalance is to break the static equilibrium state inside the molten solder. Specifically, this method changes the local surface tension of the molten solder by establishing a microscopic asymmetric distribution of thermal field or chemical potential energy gradient at the welding interface. Since surface tension is a function of temperature and chemical composition, this non-uniform distribution generates a Marangoni flow driven by the surface tension gradient. This fluid flow behavior has a specific dynamic direction, and its velocity vector points from the high surface tension region to the low surface tension region. From a physical mechanism perspective, Marangoni flow can generate fluid shear stress sufficient to overcome the adhesion force on the substrate surface, which is the key driving force for the active removal of voids.

[0026] The fluid dynamics imbalance logic completely breaks the bubble locking effect through the following process: In traditional welding processes, due to the high interfacial adhesion between the bubble and the metallization layer, and the weld gap usually being less than 100 micrometers, the buoyancy force on the bubble is much less than the constraint force caused by surface tension, resulting in the bubble being firmly locked on the surface of the solder pad. This invention induces the aforementioned directional migration flow field, causing the bubbles that were originally in a static adhesive state to be subjected to continuous momentum impact. Due to the typical high non-uniformity of the velocity distribution of the flow field on the weld section, the fluid pressure on the bubbles in the flow field is asymmetrical. According to the Bernoulli effect and the law of fluid dynamic pressure distribution, a significant pressure difference will be generated on both sides of the bubble. When the pressure difference and the accompanying fluid drag force exceed the interfacial adhesion force between the bubble and the solid interface, the mechanical equilibrium of the bubble is broken, and it changes from a locked state to a free migration state. This process realizes the essential transformation from passive diffusion to active driving, so that in the confined microspace, the bubble no longer makes disordered random movements, but migrates directionally towards the edge of the solder in accordance with the preset flow field vector.

[0027] During the directional migration and escape stage of the bubbles, the directional migration flow field continuously guides the desorbed bubbles through the high-viscosity molten solder region. In order to ensure the smoothness of this process, the liquid viscosity temperature change curve of the flux is precisely controlled to ensure that the solder maintains sufficient fluidity on the bubble migration path, thereby reducing the viscous resistance encountered by the bubble movement. During this process, driven by the directional flow field, the migration velocity of the bubble coincides with the flow field vector height; Furthermore, the flow field effect, through momentum exchange between the solder edge and the central region, enables bubbles to have sufficient kinetic energy to overcome the surface tension barrier when they come into contact with the free interface between the solder and the external environment. At the free interface, the internal pressure of the bubble and the external environmental pressure reach a critical equilibrium point, ultimately achieving the complete collapse and escape of the bubble. Due to the adoption of a no-cleaning system, after the flux carrier has completed its task of guiding the bubble to escape, its residual components undergo a further polycondensation reaction under continuous high temperature. This chemical evolution process transforms the residue into an inert polymer film with high thermal stability and excellent insulation properties. This film, like a high-performance encapsulation material, firmly wraps around the solder joint. This inert film not only has extremely high breakdown voltage but also excellent hydrophobicity. It can meet the electrical reliability requirements of power semiconductors for long-term operation in high temperature and high humidity environments without any subsequent water washing or chemical cleaning processes, completely eliminating the risk of electromigration.

[0028] The controlled cooling and crystallization step involves adjusting the heat distribution within the cooling zone to implement asymmetric heat dissipation control. This allows the molten solder, after bubble removal, to solidify sequentially from the edge to the center or from the bottom to the top. This sequential solidification control logic has profound material science implications. It ensures the smooth movement of the liquid-solid interface and can effectively prevent the secondary generation of shrinkage voids caused by excessive cooling rate. These shrinkage voids are usually caused by the physical vacuum formed by the last cooling of the central area of ​​the solder and the inability to be compensated by liquid solder. The self-compensation of the solder can be achieved through sequential solidification. Meanwhile, the controlled cooling rate ensures uniform growth of intermetallic compounds at the interface, preventing the formation of excessively thick brittle compound layers. The resulting welded connection layer has a dense microstructure and extremely low void ratio (typically less than three percent), providing an excellent heat conduction path for semiconductor chips.

[0029] Furthermore, the low-void no-clean solder paste soldering method for semiconductor packaging introduces a physical field coupling enhancement mechanism during the melting window. This mechanism helps to break the pressure balance inside the bubbles by applying ultrasonic micro-vibration or periodic thermal field fluctuations at a preset frequency to the soldering area. From a sonochemical perspective, the cavitation effect generated by ultrasonic micro-vibration and the microfluidic stream can significantly reduce the apparent viscosity of the fluid and cause the microbubbles to merge under the action of the vibration field, transforming them into larger bubbles. According to Stokes' law, the driving force on larger bubbles in the flow field is more significant, thereby further reducing the critical energy required for the Marangoni flow to drive the bubbles to migrate outward. The physical field coupling enhancement mechanism and the fluid dynamics imbalance logic support each other, ensuring that stable and deterministic void control effects can be achieved for semiconductor packaging structures of different sizes (such as from 1 mm x 1 mm to 20 mm x 20 mm) and different substrates.

[0030] Furthermore, the no-clean solder paste contains a specific proportion of surface tension modifiers, which are typically composed of fluorinated surfactants or silicon oligomers with specific functional groups. These modifiers can selectively accumulate in the high-temperature zone of reflow soldering, i.e., preferentially migrate to the interface between the solder and flux. By establishing a preset chemical potential gradient on the solder surface, the modifier can significantly enhance the intensity of the aforementioned Marangoni flow. This deep coupling between the chemical composition and the physical flow field allows this method to generate sufficient flow field dynamics to drive bubble escape under conventional atmospheric reflow conditions, achieving the void control level of traditional expensive vacuum reflow soldering, significantly improving production efficiency and reducing fixed asset investment costs.

[0031] Furthermore, in the gradient heating preheating step, the method employs a nonlinear temperature rise curve. By reducing the temperature rise rate in the critical range of flux activity release (typically 150 to 180 degrees Celsius), the gaseous byproducts generated by the reduction reaction have sufficient time to be discharged through the micropores between the solder paste powders before the solder melts. This reduces the initial volume of bubbles in the melting stage. This pre-set void management logic and the active driving logic of the melting stage are nested layer by layer to construct a void suppression system throughout the entire life cycle, ensuring the extremely high density of the final bonding layer.

[0032] Furthermore, the welding method is particularly suitable for packaging high-power modules of third-generation semiconductors, such as silicon carbide and gallium nitride. These power devices have extremely high power density and generate extremely high heat flux density during operation. The device is extremely sensitive to voids at the welding interface. Even tiny voids can form thermal resistance centers and induce local hot spots. This method eliminates the dead zone of thermal resistance at the interface, ensuring the thermal stability of the device under continuous high pulse current impact. From the perspective of microscopic failure analysis, this method effectively suppresses the risk of chip cracking caused by thermal stress concentration and ensures the service life of the power module under extreme conditions.

[0033] The present invention provides a low-void, no-clean solder paste soldering method for semiconductor packaging. The process of Marangoni flow driving bubble migration can be described by the balance equation of fluid shear force and resistance. In the confined solder joint space, the shear force generated by the flow field is proportional to the projected area of ​​the bubble. By introducing the chemical potential energy gradient, the derivative of the interfacial tension (i.e., the gradient term) becomes the core driving force for the fluid flow. For example, by adjusting the heating power of different sides of the reflow oven, a tiny temperature difference is created under the chip. This temperature difference directly induces a non-uniform distribution of the solder surface tension. According to physical simulations, the surface tension gradient generated by this temperature difference can drive the molten solder to produce microflows on the order of centimeters per second. This speed is sufficient to expel all bubbles with a diameter greater than ten micrometers to the edge of the solder within a melting window of a few seconds.

[0034] The polymer substrate selected in the flux carrier has a specific thermosetting logic. In the final stage after the bubbles escape, as the temperature rises further to the peak temperature, the polymer molecular chains undergo a rapid condensation reaction. This process transforms the flux from a liquid state into a hard solid film. This transformation must occur before the solder begins to cool and solidify to ensure that the film can play an auxiliary encapsulation pressure role during the solder shrinkage process. This precise coordination of phase transition time not only improves the electrical insulation performance of the solder joint, but also significantly enhances the solder joint's resistance to moisture absorption and chemical corrosion.

[0035] When implementing the pre-soldering baseline construction step, for copper-clad ceramic substrates, the alloy powder selected by the system is a tin-silver-copper alloy with a silver content of 3.0%. This alloy undergoes a eutectic reaction at 217 degrees Celsius and has good mechanical properties. By controlling the pressure and time of the dispensing needle, the high consistency error of the solder paste is ensured to be controlled within ±5 micrometers. Maintaining this precision is crucial for the subsequent establishment of a symmetrical hydrodynamic flow field. If the solder paste thickness distribution is uneven, random local pressure gradients will be generated, interfering with the preset Marangoni flow vector, thereby reducing the degassing efficiency.

[0036] During the gradient heating preheating step, the system collects the temperature data of the assembly in real time through thermocouples distributed in the reflux furnace. The temperature control algorithm uses a combination of proportional-integral-derivative (PID) logic and fuzzy control to ensure that the actual temperature rise curve is highly coincident with the preset energy gradient curve. In this stage, solvent components such as isopropanol and diethylene glycol ether in the flux evaporate in order of boiling point. Experiments have shown that maintaining this stage for 60 to 120 seconds can reduce the oxygen content inside the solder to an extremely low level, thereby eliminating the chemical barrier for the next step of wetting and flow.

[0037] During the directional migration of bubbles, the drag force generated by the flow field must overcome the adhesion work between the bubbles and the solder pad metal layer. The magnitude of the adhesion work is closely related to the roughness of the metal surface and the surface tension of the solder. By introducing a surface tension modifier, this invention reduces the static surface tension of the solder by 15% to 20%, which directly reduces the critical starting pressure required for bubble desorption. Meanwhile, the ultrasonic frequency applied by the physical field coupling enhancement mechanism is set between 20 kHz and 40 kHz. This frequency range can generate significant microfluidic effects without mechanically damaging the fragile semiconductor chip structure.

[0038] In the controlled cooling crystallization step, the cooling rate is strictly controlled between two and five degrees Celsius per second. This rate is chosen based on the equilibrium crystallization theory and actual production efficiency. Too slow cooling will cause the intermetallic compounds (such as copper-tin compounds) to grow excessively, forming coarse and brittle needle-like structures, reducing the resistance of the solder joint to vibration fatigue. Too fast cooling will lock in residual stress and even generate microcracks inside the solder. The asymmetric heat sink guides heat to preferentially dissipate from the bottom of the substrate, thereby driving the liquid-solid interface to move upward from the bottom, and completely pushing any remaining tiny air bubbles to the free surface.

[0039] Example 1: The solder paste used is a tin-silver-copper no-clean solder paste containing 3% silver and 0.5% copper. The substrate is a copper-clad ceramic substrate, and the chip is a silicon carbide (SiC) power chip with a size of 5 mm x 5 mm. The soldering process parameters are as follows: The temperature rise rate of the preheating section is 1.5 degrees Celsius per second; The peak melting temperature is 248 degrees Celsius; The melting time is 60 seconds. During the melting period, a 5-degree Celsius thermal gradient is established through a lateral infrared heat source to induce Marangoni flow. At the same time, 30 kHz ultrasonic micro-vibration is applied. The cooling section adopts sequential solidification control with a cooling rate of 3 degrees Celsius per second.

[0040] Comparative Example 1 (Traditional Reflow Soldering): Using the same no-clean solder paste, substrate and chip, a traditional linear temperature rise process is performed with a temperature rise rate of 1.2 degrees Celsius per second and a peak temperature of 245 degrees Celsius. No thermal gradient flow field is established, no physical field enhancement is applied, and uniform air cooling is used.

[0041] Comparative Example 2 (Vacuum Reflow Soldering): Using the same materials, a vacuum reflow process was performed, reducing the chamber pressure to ten millibars during solder melting, and the cycle was repeated three times.

[0042] X-ray void detection and shear strength testing were performed on the weld joints completed in the above embodiments and comparative examples, and the production efficiency index (UPH) was recorded. The comparison data is shown in Table 1.

[0043] Table 1: Comparison of data between the method of the present invention and traditional processes Based on the above comparative data, the following technical conclusions can be drawn: the method of the present invention (Example 1) performs extremely well in terms of void ratio control, with an average void ratio of only 0.8%, which is far lower than the conventional reflow process of 8.5%, and even better than the expensive vacuum reflow process (1.2%). This fully demonstrates the great technical advantages of the induced fluid dynamics imbalance logic and the bubble directional migration flow field in actively expelling bubbles. The weld shear strength of Example 1 reached 88 MPa, the highest among the three, which is attributed to the extremely low void ratio and the dense and uniform microstructure resulting from the controlled cooling and crystallization process.

[0044] From the perspective of economic efficiency and production efficiency, the method of this invention can be implemented in a conventional atmospheric / protective atmosphere reflow furnace, with a UPH (output per hour) of 1,200 pieces, which is comparable to conventional reflow but far exceeds vacuum reflow (350 pieces per hour). Vacuum reflow requires frequent vacuuming and venting cycles, which greatly prolongs the process cycle and results in extremely high equipment maintenance costs. This invention achieves a generational improvement in void control capability without increasing equipment complexity through deep coupling of physical and chemical fields.

[0045] Further analysis of the X-ray detection images of Example 1 reveals that not only has the total number of voids been significantly reduced, but their spatial distribution also shows a clear tendency to migrate towards the edge. In the critical heat dissipation area of ​​the chip at the center of the weld, a near-zero void connection has been achieved, which is of decisive significance for the reliability of high-power semiconductor devices. The inert polymer film formed by flux residues exhibited excellent chemical stability in the test. After a continuous 1,000-hour double 85 (temperature 85 degrees Celsius, humidity 85%) aging test, no corrosion was observed around the solder joint, and the surface insulation resistance remained above 10 to the power of 12 ohms.

[0046] From a deeper mechanical perspective, this method, by inducing a directional migration flow field, not only removes air bubbles but also microscopically stirs the alloy components inside the molten solder. This stirring effect refines the grain size after solidification, making the growth of the intermetallic compound (IMC) at the interface more continuous and smooth. Under a microscope, the thickness of the IMC layer in Example 1 is uniformly controlled between three and five micrometers. This thickness ensures reliable bonding while avoiding the risk of brittle fracture.

[0047] For third-generation semiconductor applications, the thermal expansion coefficient of silicon carbide chips differs from that of ceramic substrates. During thermal cycling, the solder layer will be subjected to huge shear stress. Because this invention achieves an ultra-low void ratio and there are no stress concentration bubble singularities, the thermal resistance of the packaged module increases by less than 5% after 3,000 thermal shock cycles.

[0048] Furthermore, the surface tension modifier contained in the no-clean solder paste of the present invention has gaseous small molecules as its thermal decomposition products at reflow temperature. These molecules can escape with the bubbles and will not remain inside the polymer film to form impurity points. This high level of chemical purity further ensures the electrochemical reliability of the solder joints.

[0049] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for soldering low-void, no-clean solder paste for semiconductor packaging, characterized in that, The process includes the following steps: Perform the pre-soldering reference construction step, apply no-clean solder paste to the metallization area to be soldered on the semiconductor substrate, and establish the initial physical connection orientation between the semiconductor chip and the semiconductor substrate; A gradient heating preheating step is performed, in which the solvent components in the no-clean solder paste are volatilized by adjusting the heat flux, and the active reducing components in the no-clean solder paste are used to remove the oxides on the surface of the metallized area to be soldered. The solder melting triggering step is executed, and the system temperature crosses the liquidus line of the solder alloy powder, causing the solder to transform into a molten liquid state; During the period when the solder is in a molten liquid state, a directional migration flow field is induced inside the molten liquid solder by establishing a microscopic asymmetric distribution of thermal field or chemical potential gradient at the welding interface. The fluid shear force generated by the directional migration flow field breaks the adhesion balance of bubbles at the interface of the metallization area to be welded, driving the bubbles to migrate along a preset dynamic vector toward the edge of the molten solder and escape.

2. The method for soldering low-void, no-clean solder paste for semiconductor packaging according to claim 1, characterized in that: In the gradient heating preheating step, the heat flux is adjusted to control the molecular escape rate of the solvent component to be lower than the critical threshold for bubble nucleus formation. After reaching the preset activation energy level, the active reducing component undergoes a heterogeneous deoxidation reaction with the oxides on the surface of the metallization region to be welded and the solder alloy powder, generating a metal soap intermediate with wetting gain effect, and changing the energy level distribution of the interface of the metallization region to be welded, thus establishing interface conditions for the subsequent active desorption of bubbles.

3. The method for soldering low-void, no-clean solder paste for semiconductor packaging according to claim 1, characterized in that: The directional migration flow field is composed of Marangoni flow driven by surface tension gradient; In the solder melt triggering step, by constructing the directional migration flow field within the weld microspace, a dynamic flow direction from high surface tension region to low surface tension region is generated inside the molten solder, thereby generating the fluid shear stress that overcomes the adhesion force on the substrate surface.

4. The method for soldering low-void, no-clean solder paste for semiconductor packaging according to claim 3, characterized in that: By utilizing the directional migration flow field to break the bubble's locking effect, when the bubble adhering to the surface of the metallization area to be welded is subjected to momentum impact, the non-uniform velocity distribution of the directional migration flow field on the weld cross section generates a pressure difference on both sides of the bubble. When the sum of the pressure difference and the accompanying fluid drag exceeds the interfacial adhesion force between the bubble and the solid interface, the bubble is driven to change from a locked state to a free migration state, causing the bubble to migrate directionally towards the solder edge in accordance with the flow field vector of the directional migration flow field.

5. A method for soldering low-void, no-clean solder paste for semiconductor packaging according to claim 4, characterized in that: During the directional migration and escape stage of bubbles, the viscous resistance of bubble movement is reduced by adjusting the liquid viscosity temperature change curve of the flux carrier. The directional migration flow field drives the bubbles to overcome the free interface tension barrier between the molten solder and the external environment by exchanging momentum between the edge and center regions of the solder, thereby enabling the bubbles to burst and escape.

6. The method for soldering low-void, no-clean solder paste for semiconductor packaging according to claim 5, characterized in that: During the high-temperature sustained stage after the bubbles escape, the flux carrier component in the no-clean solder paste undergoes a polycondensation reaction and transforms into an inert polymer film with preset thermal stability and insulating properties. The inert polymer film wraps around the solder joint to provide electrical reliability protection.

7. The method for soldering low-void, no-clean solder paste for semiconductor packaging according to claim 1, characterized in that: After the bubbles escape, a controlled cooling and crystallization step is performed. By adjusting the cold distribution in the cooling section, the molten solder solidifies sequentially from the edge to the center or from the bottom to the top. The sequential solidification process prevents the formation of shrinkage voids caused by excessively rapid cooling and guides the intermetallic compounds to form a uniformly growing bonding layer at the interface.

8. The method for soldering low-void, no-clean solder paste for semiconductor packaging according to claim 1, characterized in that: During the period when the solder is in a molten liquid state, a physical field coupling enhancement mechanism is also introduced to apply ultrasonic micro-vibration or periodic thermal field fluctuations of a preset frequency to the welding area. The sonochemical cavitation effect generated by the ultrasonic micro-vibration helps to break the pressure balance inside the bubble, causing bubbles of different volumes to merge, thereby reducing the critical energy required for the directional migration flow field to drive the bubble.

9. A method for soldering low-void, no-clean solder paste for semiconductor packaging according to claim 3, characterized in that: The no-clean solder paste contains a surface tension modifier, which selectively accumulates in the high-temperature zone of reflow soldering, establishing a preset chemical potential gradient on the surface of the molten solder to enhance the strength of the Marangoni flow.

10. A method for soldering low-void, no-clean solder paste for semiconductor packaging according to claim 1, characterized in that: In the gradient heating preheating step, a nonlinear temperature rise curve is used; The temperature rise rate is reduced in the active release range of the active reducing component, so that the gaseous byproducts generated by the reduction reaction are discharged through the interparticle pores of the no-clean solder paste before the solder melting triggering step, thereby reducing the initial volume of bubbles in the melting stage.

Images