Ultrasonic bonding process and bonding process reliability evaluation method

Abstract

The invention relates to an ultrasonic bonding process and a bonding process reliability evaluation method, and belongs to the technical field of semiconductor device packaging. The invention relates to an ultrasonic bonding process, which comprises the following steps of: 1, before bonding, carrying out plasma cleaning on a metal ceramic shell to remove surface contamination and an oxide layer; step 2, enabling the aluminum wire to penetrate into a chopper of a bonding machine, enabling the aluminum wire to be in contact with the surface of the gold-plated layer at the bonding pressure of 250cN-280cN by adopting the chopper, and enabling the aluminum wire to be in contact with the surface gold-plated layer of the bonding area in a low-pressure mode; and step 3, the bonding machine applies ultrasonic power of 10.5-12.3 W for 400-500 ms, the aluminum wire begins to rub and destroy the gold-plated layer on the surface of the bonding area under the action of the ultrasonic power, and after the gold-plated layer is destroyed, the aluminum wire penetrates through the gold-plated layer and forms an atomic-scale mutual diffusion interface with the nickel-plated layer. Reliable combination of gold, aluminum and nickel is achieved, meanwhile, a reliability evaluation method is provided, and therefore it is guaranteed that devices, circuits and the like adopting the ultrasonic bonding technology work reliably in the life cycle.

Description

Technical Field

[0001] This application belongs to the field of semiconductor device packaging technology, specifically relating to an ultrasonic bonding process and a method for evaluating the reliability of the bonding process. Background Technology

[0002] Metal or ceramic casings, due to their high thermal conductivity, high insulation, and excellent mechanical strength, are widely used in high-power devices, aerospace electronics, and semiconductor device packaging in high-temperature environments. Wire bonding is frequently used in circuit assembly to achieve electrical interconnections between the substrate and pins, chips and electrodes, and electrodes and electrodes. Wire bonding is a critical process in circuit assembly, and its reliability significantly impacts the yield rate and long-term reliability of the circuit. In integrated circuit assembly, aluminum wires, with their high electrical conductivity, high thermal conductivity, high ductility, and low cost, have become commonly used bonding wires in aerospace devices. Since metal or ceramic casings typically have nickel or gold plating as a protective metal layer to prevent oxidation, gold-aluminum bonding is widespread. The gold-aluminum interface is highly susceptible to the formation of Kirkendall voids, a well-known problem in China. Currently, there are no effective methods, either domestically or internationally, to completely eliminate this problem; the only solutions are to reduce the thickness of the gold layer and decrease the gold content to reduce Kirkendall voids. There is an urgent need for a process to significantly improve the reliability of gold-aluminum bonding and achieve highly reliable bonding of gold, aluminum, and nickel with high strength, low Kirkendall voids, low porosity, and slow degradation. Summary of the Invention

[0003] The technical problem to be solved by this application is to overcome the shortcomings of the prior art and provide an ultrasonic bonding process and a method for evaluating the reliability of the bonding process. This application achieves reliable bonding of three metals: gold, aluminum, and nickel, and at the same time provides a reliability evaluation method, thereby ensuring that devices, circuits, etc. using this ultrasonic bonding process can work reliably throughout their life cycle.

[0004] The technical solution adopted in this application to solve the problems existing in the prior art is:

[0005] An ultrasonic bonding process includes the following steps:

[0006] Step 1: Before bonding, perform plasma cleaning on the metal-ceramic shell to remove surface contaminants and oxide layers;

[0007] Step 2: The aluminum wire is threaded into the bonding machine's cleaver. The cleaver uses a bonding pressure of 250cN-280cN to bring the aluminum wire into contact with the gold-plated surface. The gold-plated surface of the bonding area is contacted using a low-pressure method.

[0008] Step 3: The bonding machine applies an ultrasonic power of 10.5W-12.3W for 400ms-500ms. Under the action of the ultrasonic power, the aluminum wire begins to rub and destroy the gold plating layer on the surface of the bonding area. After the gold plating layer is destroyed, the aluminum wire penetrates the gold plating layer and forms an atomic-level interdiffusion interface with the nickel plating layer.

[0009] After the ultrasonic time is exhausted, the aluminum wire completes its deformation, and a strong Al / Au / Ni composite interface is formed in the bonding region.

[0010] Preferably, the plasma cleaning uses a mixture of Ar and H2 gas, and the oxygen content on the nickel plating layer surface is reduced to below 5 at% after cleaning.

[0011] Preferably, the plasma ionization power is 280-310W and the processing time is 7-9min.

[0012] Preferably, the diameter of the aluminum wire is 8-10 mil, and its plastic deformation is controlled within the range of 1 to 3 times.

[0013] Preferably, the bonding machine is the F&K5350 semi-automatic coarse aluminum wire bonding machine.

[0014] A method for evaluating the reliability of a bonding process, used to evaluate bonded samples prepared by the above-described ultrasonic bonding process, includes the following steps:

[0015] S01. Prepare bonded samples using the ultrasonic bonding process described above;

[0016] S02. Perform temperature cycling tests on the bonded samples for more than 900 cycles to accelerate the diffusion of the metal interface to form the IMC layer.

[0017] S03. Perform destructive bond strength testing on the samples that have undergone temperature cycling tests and record the data;

[0018] S04. Perform bond point shear force testing on the samples that have completed the bond strength test, and record the data;

[0019] S05. Perform longitudinal cutting on the bonding interface, including the horizontal half of the wedge bonding point. Select at least two cross sections and analyze the interface using scanning electron microscopy (SEM) to evaluate reliability.

[0020] The criteria for SEM analysis include: the metal compound layer is continuous without cracks and has no obvious Kirkendall voids.

[0021] Preferably, in step S02, the temperature range of the temperature cycling test is -55℃ to 150℃.

[0022] Preferably, in step S02, the temperature cycling test has a heating / cooling rate of 12-14℃ / min and a holding time of 12-16min for both high and low temperatures.

[0023] Preferably, in step S03, the tensile rate of the destructive bond strength test is 680-720 μm / s, until the bond wire breaks. 10. A method for evaluating the reliability of a bonding process according to claim 6, characterized in that:

[0024] In step S04, the shear rate of the bonding point shear force test is 95-103 μm / s until the bonding point is pushed away.

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

[0026] (1) Improved interface reliability: The aluminum wire plastic deformation fully penetrates the gold layer and the bottom nickel layer to achieve mutual diffusion bonding. After 1000 temperature cycles (-55-150℃), the length of the bonding interface crack is ≤5μm (measured by SEM longitudinal profile).

[0027] (2) Mechanical property optimization: The initial bonding strength of φ250μm aluminum wire is 350g-500g with a standard deviation of <5%; the bonding strength retention rate after 1000 temperature cycle tests is ≥95%.

[0028] (3) Process compatibility: Applicable to existing equipment (F&K 5350 semi-automatic coarse aluminum wire bonding machine), with a yield rate of 99.7% (compared to 95% for conventional processes). Attached Figure Description

[0029] The present application will be further described below with reference to the accompanying drawings and embodiments.

[0030] Figure 1 This application describes an ultrasonic bonding process.

[0031] Figure 2 This is a flowchart of the bonding process reliability evaluation method in this application.

[0032] Figure 3 This is a schematic diagram of the longitudinal cross-section of the bonding interface under conventional bonding processes.

[0033] Figure 4 This is a schematic diagram of the longitudinal cross-section of the bonding interface under the low-pressure long-term ultrasonic bonding process of this application. Detailed Implementation

[0034] The ultrasonic bonding process and the method for evaluating the reliability of the bonding process described in this application are further described in detail with reference to the accompanying drawings, but this is not intended to limit the scope of this application.

[0035] Depend on Figure 1As shown, an ultrasonic bonding process includes the following steps:

[0036] Step 1: The nickel plating layer of the metal-ceramic shell is wrapped with a gold plating layer. The gold plating layer acts as a buffer and protects the nickel plating layer from oxidation.

[0037] Before bonding, the metal-ceramic shell is plasma cleaned to remove surface contaminants and oxide layers while improving surface activity.

[0038] The plasma cleaning uses a mixture of Ar and H2 gas, with a plasma ionization power of 300W and a processing time of 8 minutes. After cleaning, the oxygen content on the nickel plating layer surface is reduced to below 5 at%.

[0039] Step 2: The aluminum wire is inserted into the bonding machine's wedge. The wedge is used to bring the aluminum wire into contact with the gold-plated layer surface at a bonding pressure of 250cN-280cN. The bonding pressure is 60%-70% of that of conventional bonding processes. The gold-plated layer on the bonding area is contacted using a low-pressure method.

[0040] The aluminum wire has a diameter of 10 mil, and its plastic deformation is controlled within the range of 1 to 3 times. The F&K 5350 semi-automatic coarse aluminum wire bonding machine is used.

[0041] Step 3: Apply ultrasonic power of 10.5W-12.3W for 400ms-500ms. Under the action of ultrasonic power, the aluminum wire begins to abrade and break the gold plating layer on the surface of the bonding region. After the gold plating layer is broken, the aluminum wire penetrates the gold plating layer and forms an atomic-level interdiffusion interface with the nickel plating layer. After the ultrasonic time is exhausted, the aluminum wire completes the corresponding deformation, and a strong Al / Au / Ni composite interface is formed in the bonding region.

[0042] The ultrasonic power is 60%-70% of that of conventional bonding processes, and the bonding time is 200%-250% of that of conventional bonding processes.

[0043] The reduction in bonding pressure and ultrasonic power decreases the deformation rate of the aluminum wire, thus prolonging the effective bonding time. The extended ultrasonic time ensures sufficient friction between the aluminum wire and the gold plating layer, allowing the wire to penetrate the gold plating layer and ultimately achieve a good bond with the nickel layer.

[0044] Figure 4 This is a longitudinal cross-sectional view of the bonding interface between the aluminum wire and the nickel plating layer, achieved under low-pressure, long-term bonding conditions using the aforementioned ultrasonic bonding process. Figure 3This is a longitudinal cross-sectional view of the bonding interface under conventional bonding processes. A comparison of the two clearly shows that the bonding region formed by the ultrasonic bonding process exhibits a dense Al / Au / Ni composite interface, significantly improving interfacial bonding strength. After ultrasonic bonding, the interface displays a uniform intermetallic diffusion layer with no unbonded areas, demonstrating excellent peel resistance. SEM analysis shows that the Al / Au / Ni interface forms a continuous, defect-free transition layer, with the Au-Al IMC thickness controlled within the ideal range of 0.1-0.2 μm.

[0045] Depend on Figure 2 As shown, a method for evaluating the reliability of a bonding process includes the following two evaluation methods:

[0046] Evaluation Method 1 includes the following steps:

[0047] S01. Prepare bonded samples using the ultrasonic bonding process described above;

[0048] S02. Perform a temperature cycling test (-55℃-150℃) on the bonded sample 1000 times to accelerate the diffusion of the metal interface to form the IMC layer. The temperature cycling test has a heating / cooling rate of 13.3℃ / min, and the high and low temperature holding times are 15min each.

[0049] S03. Perform a destructive bond strength test on the sample that has undergone temperature cycling test. The tensile rate of the destructive bond strength test is 700 μm / s until the bond wire breaks, and record the data.

[0050] S04. Perform a bond point shear force test on the sample that has completed the bond strength test. The shear rate of the bond point shear force test is 100 μm / s until the bond point is pushed away and removed. Record the data.

[0051] S05. Perform a longitudinal cut on the bonding interface, including the midpoint of the wedge-shaped bonding point in the horizontal direction. Select at least two cross-sections and analyze the interface using scanning electron microscopy (SEM) to assess reliability.

[0052] In the longitudinal cutting analysis of the bonding interface, the cutting position is selected at the horizontal halfway point of the wedge-shaped bonding point, mainly based on a comprehensive consideration of the following mechanical properties, microstructural characteristics, and standardized testing requirements:

[0053] 1. Representativeness of stress distribution

[0054] The central region (at the 1 / 2 mark) of the wedge-shaped bond point is the core area for stress transfer, where the stress state is closest to the average stress level of the bond interface. While the edge regions may suffer from data distortion due to stress concentration or uneven deformation, the central region effectively reflects the overall shear strength and tensile capacity of the bond point. Experiments show that bond point failure often begins with interfacial peeling or root fracture in the central region. Cutting this area allows for effective observation of typical failure mechanisms (such as oxide layer detachment or plastic deformation of the metal).

[0055] 2. Microstructural integrity assessment

[0056] At the 1 / 2 position, the degree of plastic deformation of the bonded wire (such as neck width and bonding zone indentation depth) best represents the rationality of the process parameters. For example, when the width-to-diameter ratio (bond width / lead diameter) is 1.2, the neck structure is optimized, with uniform deformation and no defects; if the cut deviates from the center, it may be misjudged as an excessively thin neck or excessive deformation (actually an edge effect). The central region is the focus of ultrasonic energy, allowing clear observation of the thickness of the intermetallic diffusion layer, micropores, and contaminant distribution (such as the effect of residual organic matter on the bonding force).

[0057] 3. Standardized testing and comparability

[0058] The standard requires the wedge bonding width to be between 1.2 and 3 times the lead diameter, and the inspection must cover the core area of ​​the bonding point. Cutting off halfway is an industry-standard practice to ensure lateral data comparability. Micro-cracks or material buildup may exist at the bonding point edges due to wedge pressing; center cutting avoids this type of process noise.

[0059] 4. Avoid the vulnerable areas of the neck.

[0060] The bonding neck (near the lead transition region) is a stress concentration and high-risk area for fracture, but deformation in this region is significantly affected by parameters (e.g., neck widening leads to strength reduction when the width-to-diameter ratio is >1.5). Cutting off half of the bonding neck allows for independent analysis of the bonding strength, avoiding interference from neck failure with the core interface evaluation.

[0061] The criteria for SEM analysis include: the metal compound layer is continuous without cracks and has no obvious Kirkendall voids.

[0062] SEM analysis is a materials science method based on the phase transformation behavior of shape memory alloys. In bonding processes, it significantly improves connection reliability and lifespan by optimizing interfacial contact mechanics and residual stress distribution. SEM refers to the ability of a material to recover its initial shape after deformation through heating or unloading; its essence is the reversible process of thermoelastic martensitic phase transformation.

[0063] Evaluation Method Two includes the following steps:

[0064] S01. Prepare bonded samples using the ultrasonic bonding process described above;

[0065] S02. Perform a temperature cycling test (-55℃-150℃) on the bonded sample 1000 times to accelerate the diffusion of the metal interface to form the IMC layer. The temperature cycling test has a heating / cooling rate of 13.3℃ / min, and the high and low temperature holding times are 15min each.

[0066] During the temperature cycling test, the humidity of the test environment was maintained at 70%-85%RH to accelerate the assessment of the risk of intermetallic compound oxidation and interface corrosion.

[0067] Simultaneously, a constant acceleration of 20,000g or a vibration load (50-2000Hz) was applied to expose the problem of fatigue crack initiation.

[0068] S03. Perform a destructive bond strength test on the sample that has undergone temperature cycling test. The tensile rate of the destructive bond strength test is 700 μm / s until the bond wire breaks, and record the data.

[0069] S04. Perform a bond point shear force test on the sample that has completed the bond strength test. The shear rate of the bond point shear force test is 100 μm / s until the bond point is pushed away and removed. Record the data.

[0070] S05. Perform a longitudinal cut on the bonding interface, including the midpoint of the wedge-shaped bonding point in the horizontal direction. Select at least two cross-sections and analyze the interface using scanning electron microscopy (SEM) to assess reliability.

[0071] Existing SEM cross-sectional analysis has been expanded to include 3D reconstruction and dynamic monitoring. This involves adding cutting sites in the neck transition zone (bonding point edge) and the bonding root, comparing the thickness difference of the intermetallic compound (IMC) between the center and the edge. Normally, this should be <2 μm. If the edge IMC is excessively thick (>3 μm), it may be due to uneven bonding pressure.

[0072] Increase in-situ high-temperature SEM observations, observe IMC growth kinetics under in-situ heating at 150℃, and establish a predictive model of temperature-time-IMC thickness (such as the Arrhenius equation).

[0073] TEM-EDS line scan analysis was used to analyze the diffusion of interface elements (such as Kirkendall voids in Au / Al bonds) to supplement conventional SEM morphology observation.

[0074] To avoid missing local defects (such as micropores) that may be missed by a single cross-section, multidimensional analysis is used to reveal the spatiotemporal evolution of interface failure.

[0075] In steps S03 and S04 above, high-speed video recording at speeds greater than 1000 fps is added to capture the necking deformation or interfacial delamination behavior at the moment the bonded wire breaks. If the fracture occurs at the bonding point interface, it indicates insufficient bonding of the intermetallic compound; if the fracture occurs at the necking point, it reflects that the ultrasonic energy is too high, causing material damage.

[0076] The embodiments of this application have been described in detail above with reference to the accompanying drawings. However, this application is not limited to the above embodiments. Within the scope of knowledge possessed by those skilled in the art, various changes can be made without departing from the spirit of this application.

Claims

1. An ultrasonic bonding process characterized by, The method comprises the following steps: Step 1: Before bonding, the metal ceramic shell is subjected to plasma cleaning to remove surface contamination and oxide layer; Step 2: The aluminum wire is inserted into the wedge of the bonding machine, and the wedge is used to contact the gold-plated layer surface at a bonding pressure of 250 cN-280 cN, and the low-pressure mode is used to contact the surface gold-plated layer of the bonding area; Step 3: The bonding machine applies ultrasonic power of 10.5 W-12.3 W for 400 ms-500 ms, and the aluminum wire starts to rub and destroy the surface gold-plated layer of the bonding area under the action of ultrasonic power. After the gold-plated layer is destroyed, the aluminum wire penetrates the gold-plated layer and forms an atomic-level interdiffusion interface with the nickel-plated layer; After the ultrasonic time is consumed, the aluminum wire completes the deformation, and the bonding area forms a firm Al / Au / Ni composite interface.

2. The ultrasonic bonding process according to claim 1, wherein: In step S01, the plasma cleaning uses Ar and H2 mixed gas, and the oxygen content on the surface of the nickel-plated layer after cleaning is reduced to below 5at%.

3. The ultrasonic bonding process according to claim 2, wherein: In step S01, the plasma ionization power is 280-310 W, and the processing time is 7-9 min.

4. The ultrasonic bonding process according to claim 1, wherein: In step S02, the diameter of the aluminum wire is 8-10 mil, and the plastic deformation amount is controlled within 1-3 times.

5. The ultrasonic bonding process according to any one of claims 1-4, wherein: The bonding machine uses F&K5350 type semi-automatic coarse aluminum wire bonding machine.

6. A method for evaluating the reliability of a bonded sample prepared by the ultrasonic bonding process according to claim 5, characterized by, The method comprises the following steps: S01, using the above ultrasonic bonding process to prepare a bonding sample; S02, subjecting the bonding sample to a temperature cycle test of more than 900 times to accelerate the diffusion of the metal interface to form an IMC layer; S03, destructive bonding strength test is carried out on the sample after temperature cycle test, and the data is recorded; S04, bonding point shear force test is carried out on the sample after bonding strength test, and the data is recorded; S05, the bonding interface is longitudinally cut, the longitudinal cutting position includes 1 / 2 of the horizontal direction of the wedge-shaped bonding point, at least two cross sections are selected, the interface is analyzed by using a scanning electron microscope (SEM), and the reliability is evaluated; The judgment standard of the SEM analysis includes: the metal compound layer is continuous and has no cracks, and there is no obvious Kirkendall cavity.

7. The bonding process reliability evaluation method according to claim 6, wherein: In step S02, the temperature range of the temperature cycle test is -55℃-150℃.

8. The bonding process reliability evaluation method according to claim 6, wherein: In step S02, the temperature cycle test has a temperature rising and falling rate of 12-14℃ / min, and the high and low temperature holding time is 12-16 min.

9. The bonding process reliability evaluation method according to claim 6, wherein: In step S03, the tensile rate of the destructive bonding strength test is 680-720 μm / s until the bonding wire breaks.

10. The method of claim 6, wherein the shear speed rate of the shear test of the bonding point is 95-103 μm / s until the bonding point is removed by the flat push. ​

Images