A high-reliability underfill adhesive suitable for large-size chips and its preparation method

Abstract

This invention relates to the field of adhesive technology, specifically to a high-reliability underfill adhesive suitable for large-size chips and its preparation method. The raw materials of the underfill adhesive, by weight, include: 25-35 parts of bisphenol F epoxy resin, 1-3 parts of bisphenol A epoxy resin, 0.5-2 parts of silicone-modified epoxy resin, 0.5-1.5 parts of black paste, 0.8-1.2 parts of coupling agent, 0.1-0.5 parts of flow aid, 50-55 parts of spherical filler, 10-15 parts of curing agent, 0.2-0.8 parts of accelerator premix, and / or 0.5-1 parts of additives with controllable post-curing properties. This invention independently synthesizes the accelerator premix and additives with controllable post-curing properties, resulting in a moderate increase in glass transition temperature and high-temperature modulus, and a moderate decrease in room-temperature modulus, achieving the same level of reliability fatigue as the chip, substrate, and bump substrates, thereby improving the reliability of the chip module.

Description

Technical Field

[0001] This invention relates to the field of adhesive technology, specifically to a high-reliability underfill adhesive suitable for large-size chips and its preparation method. Background Technology

[0002] In chip packaging technology, the bonding layer between the IC chip and the organic substrate contains numerous tiny solder bumps. These bumps have poor deformation adaptability and are extremely sensitive to thermal stress, making chip structural reliability issues more prominent. Using polymer underfill to improve the reliability of packaged chips is a novel method developed in recent years. This method is economical and easy to implement. During chip packaging, the underfill material gradually solidifies in the narrow gap between the IC chip and the organic substrate through thermosetting, protecting the connecting solder bumps. It can also effectively mitigate impact loads, improve the packaged chip's resistance to deformation, moisture, and chemical corrosion, and significantly extend the fatigue life of the packaged chip, thus showing great development potential.

[0003] Because the thermal expansion coefficients of the substrate, chip, and solder joint materials differ, without underfill, the substrate or chip will deform under harsh conditions in real-world applications or reliability testing. This stress will likely cause the bump to break. Applying underfill balances the thermal expansion coefficients of these three components, effectively protecting the bump from breakage. However, when the chip size is large, ordinary underfill is insufficient to resist the warping stress of the chip and substrate. In such cases, the underfill needs to not only balance and counteract warping stress but also adapt to minute changes in the chip, substrate, and bump under specific application conditions or stringent reliability requirements to achieve better reliability performance.

[0004] Based on experience in large-size chip packaging and testing, when a chip module undergoes multiple reflow soldering and high-temperature storage reliability tests, the chip-level underfill adhesive with a pre-cured thickness needs to have appropriate post-curing properties. For example, a moderately increased glass transition temperature, a moderately decreased room temperature modulus, and a moderately increased high-temperature modulus are necessary to achieve the same level of reliability fatigue as the chip, substrate, and bump, thereby ensuring that the chip module successfully passes the high-temperature reliability tests. Summary of the Invention

[0005] To address the aforementioned technical problems, the present invention aims to provide a high-reliability underfill adhesive suitable for large-size chips and its preparation method. Through the self-synthesized accelerators and additives of this invention, combined with optimized formulation design, it is possible to increase the thermal conductivity (Tg), reduce the room temperature / low temperature modulus, and increase the high temperature modulus under reliability testing conditions. This achieves a balance in reliability fatigue levels between the chip, substrate, bump, and underfill material, thereby improving the reliability of the packaged chip and ensuring the chip module successfully passes high-temperature reliability tests.

[0006] To achieve the above objectives, the technical solution adopted is as follows:

[0007] One objective of this invention is to provide a high-reliability underfill adhesive suitable for large-size flip chips, the raw materials of which, by weight, comprise:

[0008] 25-35 parts of bisphenol F epoxy resin, 1-3 parts of bisphenol A epoxy resin, 0.5-2 parts of silicone-modified epoxy resin, 0.5-1.5 parts of black paste, 0.8-1.2 parts of coupling agent, 0.1-0.5 parts of flow aid, 50-55 parts of spherical filler, 10-15 parts of curing agent, 0.2-0.8 parts of accelerator premix and / or 0.5-1 parts of additives for controllable post-curing properties.

[0009] Preferably, the bisphenol F type epoxy resin comprises at least one of: DIC Japan's 830S and 830CRP, Nippon Kayaku's RE-303S-L, and Dow Chemical's DER354; the bisphenol A type epoxy resin comprises at least one of: DIC Japan's 850-S and 850CRP, Nippon Kayaku's RE-310S, and Dow Chemical's DER383 and DER332; the silicone modified epoxy resin comprises at least one of: Lepmod ES 3201, ES 3201S, and ES 3202 from complexed high-tech materials, HSL-50 from Shanghai Huayi, and YLSE-1500 from Nanjing Yuelai.

[0010] Preferably, the black paste is homemade, because using carbon black alone for coloring easily causes carbon black particles to agglomerate. In this invention, epoxy resin and carbon black are physically blended, and then dispersed and ground using a three-roll mill to form a homogeneous paste. The carbon black used is Cabot M800, and the epoxy resin is a blend of bisphenol A and bisphenol F epoxy resins. The bisphenol A epoxy resin is Dow Chemical's DER332, and the bisphenol F epoxy resin is Dow Chemical's DER354. The resulting black paste is named HG-01.

[0011] Preferably, the coupling agent is at least one of γ-aminopropyltriethoxysilane, γ-mercaptopropyltrimethoxysilane, γ-mercaptopropyltriethoxysilane, γ-glycidoxypropyltrimethoxysilane, and γ-aminopropyltrimethoxysilane. More preferably, it is at least one of KBM-403, KBE-403, KBM-303, KBM-802, KBM-803, KBM-903, and KBE-903 from Shin-Etsu Chemical Co., Ltd. of Japan.

[0012] Preferably, the curing agent is at least one selected from maleic anhydride, benzoic anhydride, phthalic anhydride, succinic anhydride, 4,4'-diamino-3,3'-diethyldiphenylmethane, diethyltoluene diamine, diaminodiphenyl sulfone, m-aminomethylamine, phenylenediamine trimer, and dibenzylamino ether. More preferably, it is at least one selected from Jiangsu Yarui Chemical's MHHPA, Nippon Kayaku's MCD and AA, and Huntsman's T-5000 and 5200.

[0013] Preferably, the flow aid is at least one of the following: AC-1201, AC-1202, and AC-1203 from Jiangsu Haian Petrochemical; triphenylphosphine and triphenylphosphine oxide from Shanghai Aladdin; and BYK-307 and BYK-333 from BYK Chemical.

[0014] Preferably, the spherical packing is at least one of the following from Admatechs Corporation of Japan: SE6050-STE, SO-E2, SO-E2 / 24C, FE920ASQ, 200SE-E4, SC2500-SQ, SE5050-SEJ, and SE2050-SEJ.

[0015] Preferably, the raw materials for preparing the accelerator premix include an accelerator.

[0016] Preferably, the preparation method of the accelerator includes the following steps: adding 3,3',4,4'-benzophenone tetracarboxylic dianhydride, 2-ethyl-4-methylimidazole and ethanol sequentially into a reaction vessel, heating to 50-80°C, stirring for 6-10 hours, then vacuum filtering the reaction product, and finally drying it in an oven at 80-120°C for 2 hours to obtain the accelerator;

[0017] The specific reactions are as follows:

[0018]

[0019] Preferably, the molar ratio of 3,3',4,4'-benzophenone tetracarboxylic dianhydride to 2-ethyl-4-methylimidazole is 1:1-3, and the ratio of the total molar amount of 3,3',4,4'-benzophenone tetracarboxylic dianhydride and 2-ethyl-4-methylimidazole to the molar amount of ethanol is 1:1-3.

[0020] Preferably, the preparation method of the accelerator premix includes the following steps: according to the weight parts, 25-35 parts of bisphenol A epoxy resin and 25-35 parts of alicyclic epoxy resin are mixed and stirred evenly, and then 35-45 parts of accelerator are added, stirred evenly, and then ground to obtain the accelerator premix. The bisphenol A epoxy resin used is Dow Chemical's DER332, and the alicyclic epoxy resin used is Daicel's 2021P.

[0021] Preferably, the preparation method of the additive with controllable post-curing properties includes the following steps: 70-80 parts by weight of bisphenol F type epoxy resin and 20-30 parts by weight of a cycloaliphatic amine curing agent are heated to 40-60°C, mixed and stirred for 1-4 hours, and then 0.5-1 parts by weight of barbituric acid are added, and the mixture is stirred for another 1-2 hours to obtain the additive with controllable post-curing properties. The bisphenol F type epoxy resin used is Dow Chemical's DER354, and the cycloaliphatic amine curing agent is BASF's EC331.

[0022] The second objective of this invention is to provide a method for preparing the high-reliability underfill adhesive suitable for large-size flip chips, comprising the following steps:

[0023] Mix 25-35 parts of bisphenol F epoxy resin, 1-3 parts of bisphenol A epoxy resin, 0.5-2 parts of silicone-modified epoxy resin, 0.5-1.5 parts of black paste, 0.8-1.2 parts of coupling agent, and 0.1-0.5 parts of flow aid for 1-4 hours. Add 50-55 parts of spherical filler in 2-4 portions and continue stirring at 70-90℃ for 6-10 hours. Then cool to room temperature, add 10-15 parts of curing agent, and stir at 25-30℃ for 0.5-2 hours. Add 0.2-0.8 parts of accelerator premix and stir at 25-30℃ for 0.5-2 hours. Finally, add 0.5-1 parts of additive with controllable post-curing properties and stir at 25-30℃ for 0.5-2 hours.

[0024] Preferably, the entire preparation process is carried out under vacuum conditions, with a vacuum degree of not less than -0.08 MPa.

[0025] Compared with existing technologies, the beneficial effects of this invention are as follows: By introducing a self-synthesized accelerator premix and controllable post-curing additives, combined with appropriate formulation design, a high-reliability underfill adhesive suitable for large-size flip chips is finally obtained. This underfill adhesive possesses good toughness, suitable glass transition temperature, modulus, and coefficient of thermal expansion. Most importantly, after multiple reflow soldering, high temperature and humidity, high temperature storage, and thermal cycling reliability processes, the glass transition temperature will moderately increase, the room temperature modulus will moderately decrease, and the high temperature modulus will significantly increase, while improving the interfacial adhesion after reliability testing. This series of special properties can offset the internal stress generated during reliability testing as much as possible, prevent chip warpage, and balance the different changes caused to the chip, substrate, and bump materials during reliability testing. In particular, it can protect the bumps at the corners of large-size chips from being broken by warpage stress, avoiding failure modes such as cracking and delamination in large-size chips. Attached Figure Description

[0026] Figure 1 The diagram shows the failure of the body cracking that occurred after 7 reflow soldering cycles in Comparative Example 1 during reliability testing.

[0027] Figure 2 The diagram shows the corner crack / interface delamination failure that occurred after PCT 1000 in the reliability test for Comparative Example 1.

[0028] Figure 3 The diagram shows the corner crack / interface delamination failure that occurred after HTST1000h in the reliability test for Comparative Example 1. Detailed Implementation

[0029] The principles and features of the present invention are described below. The examples given are only for explaining the present invention and are not intended to limit the scope of the present invention.

[0030] 1. Preparation of accelerators

[0031] 3,3',4,4'-benzophenone tetracarboxylic dianhydride, 2-ethyl-4-methylimidazole, and ethanol were sequentially added to a three-necked flask equipped with a stirrer, thermometer, and condenser. The mixture was heated to 60°C and stirred for 8 hours. The reaction product was then vacuum filtered and dried in an oven at 100°C for 2 hours. The dried powdered solid (named CJ-01) is the accelerator required in this invention. The molar ratio of 3,3',4,4'-benzophenone tetracarboxylic dianhydride to 2-ethyl-4-methylimidazole is 1:2, and the ratio of the total molar amount of 3,3',4,4'-benzophenone tetracarboxylic dianhydride and 2-ethyl-4-methylimidazole to the molar amount of ethanol is 1:2.

[0032] 2. Preparation of accelerator premix

[0033] 30 parts of bisphenol A type epoxy resin (Dow Chemical's DER332) and 30 parts of alicyclic epoxy resin (Dairu's 2021P) were mixed and stirred evenly. Then, 40 parts of CJ-01 were added and stirred evenly. The mixture was then ground three times with a three-roll mill to obtain the accelerator premix (named CP-01).

[0034] 3. Preparation of additives with controllable post-curing properties

[0035] 79.5 parts of bisphenol F type epoxy resin (Dow Chemical's DER354) and 29.5 parts of alicyclic amine curing agent (BASF's EC331) were heated to 50°C and mixed and stirred for 2 hours. Then, 1 part of barbituric acid was added, and the mixture was stirred and stirred for another hour. The resulting additive (named YP-01) had controllable post-curing properties.

[0036] 4. Preparation of black paste

[0037] 40 parts of bisphenol A epoxy resin and 40 parts of bisphenol F epoxy resin were mixed and stirred for 0.5 hours. 20 parts of carbon black were added in two batches, and the mixture was stirred for 1 hour after each addition. The mixture was then ground three times with a three-roll mill and stirred for 0.5 hours to obtain the final black paste.

[0038] 5. Preparation of high-reliability underfill adhesive suitable for large-size flip chips

[0039] 30 parts of bisphenol F epoxy resin, 2 parts of bisphenol A epoxy resin, 1 part of silicone-modified epoxy resin, 0.5 parts of black paste, 1 part of coupling agent, and 0.1 parts of flow aid were mixed and stirred for 1 hour. 52 parts of spherical filler were added in three portions, and stirring continued at 80℃ for 8 hours. The mixture was then cooled to room temperature, and 12 parts of curing agent were added, stirring at 25℃ for 0.5 hours. Next, 0.2 parts of accelerator premix were added, stirring at 25℃ for 0.5 hours. Finally, 1 part of an additive with controllable post-curing properties was added, and stirring was continued at 25℃ for 0.5 hours. The final preparation was completed. The entire preparation process was carried out under vacuum conditions, preferably maintaining a vacuum level of no less than -0.08 MPa throughout the process.

[0040] The formulation of the high-reliability underfill adhesive suitable for large-size flip chips is shown in Table 1.

[0041] Table 1. Preparation formulations for examples and comparative examples (unit: parts by weight)

[0042]

[0043]

[0044] Test Results and Analysis

[0045] Table 2. Reliability Results of Typical Formulation Large-Size Chips (Pass / Fail)

[0046]

[0047] Table 3. Comparison of key performance characteristics between the examples and comparative examples

[0048]

[0049]

[0050]

[0051] As shown in Table 2, it is clear that Comparative Example 1, lacking the self-made accelerator premix and controllable post-curing additives, exhibited the following issues during the reliability test: Figure 1 The failure and damage observed in / 2 / 3 did not present similar issues in the reliability test of Example 5, demonstrating a reliability level comparable to that of competing products. (Reliability tests for Example 5, Comparative Example 1, and Competitor X were all conducted on large-size flip chips with a diameter of 20*21mm). The reason for this problem is primarily that the high-temperature modulus change of the bottom filler in Comparative Example 1 before and after the reliability test was relatively low, failing to synchronize with the microscopic changes in the silicon wafer, substrate, and bump material after the reliability test. Especially for large-size chips, where reliability failure modes mainly occur at high temperatures, the synchronization of these microscopic changes is crucial.

[0052] As shown in Table 2, Comparative Example 1 is the base formulation. Due to reliability issues, after comparing the performance of competing products and conducting numerous DOE experiments, a self-made accelerator premix and a controllable post-curing additive were obtained to prepare the formulation and achieve the performance of competing products.

[0053] Comparison of data from Comparative Example 2 and Comparative Example 1 shows that if only EC331, the main raw material in the controllable post-curing additive, is used to prepare the formulation, although the high temperature modulus before and after reflow soldering approaches the trend of the competing products and the DSC curing peak onset is significantly reduced, its storage stability deteriorates and the open time is very short, which is not conducive to the actual application process.

[0054] A comparison of the data from Comparative Example 3 and Comparative Example 1 shows that if only the self-synthesized accelerator CJ-01 is used to prepare the formulation, apart from a slight decrease in the DSC curing onset and an increase in the high-temperature modulus before and after reflow soldering, the other properties are similar to Comparative Example 1 and fail to reach the performance level of competitor X. Furthermore, because the self-synthesized product is a powder solid, its particle size is relatively large and it is prone to agglomeration; direct use may affect the flow and filling performance.

[0055] A comparison of the data from Comparative Example 4 and Comparative Example 1 shows that if the formulation is prepared using the self-synthesized accelerator CJ-01 and the main raw material EC331 in the controllable post-curing additive, the DSC curing onset is significantly reduced, and the high temperature modulus before and after reflow soldering is close to that of competitor X. However, like Comparative Example 2, the storage stability deteriorates and the open time is very short, which is not conducive to the actual application process. This indicates that EC331 will have high activity when used directly in this formulation, which is not conducive to stability and operating time.

[0056] Examples 1 and 2 primarily aim to verify the appropriate dosage of the self-prepared accelerator premix CP-01. Compared with Comparative Example 3, the performance differs slightly. However, as mentioned earlier, preparing the accelerator premix allows the powdered accelerator to be dispersed as much as possible, resulting in more stable use and no adverse effects on flow filling. Compared with the basic formulation of Comparative Example 1, the DSC curing peak onset is reduced, and the high-temperature modulus before and after reflow soldering is significantly increased, approaching the level of competitor X. Since the accelerator CJ-01 in CP-01 is a product of the reaction between acid anhydride and imidazole, its basic activity is low, especially in amine systems. In addition, the premix incorporates alicyclic epoxy resin, ultimately enabling the continued release of some activity at high temperatures, thereby improving the high-temperature modulus.

[0057] Examples 3 and 4 primarily aimed to verify the appropriate dosage of the self-prepared additive YP-01, which allows for controllable post-curing performance. It was observed that a higher dosage of this additive resulted in higher high-temperature modulus before and after reflow soldering, with a greater rate of increase. Compared to Comparative Example 2, it was clearly shown that when EC331 was used as a raw material to prepare YP-01 before being incorporated into the formulation, the DSC curing onset was reduced. More importantly, storage stability was significantly improved, and the open time was greatly extended, which is beneficial for actual application processes. The main reason for this phenomenon is that EC331 undergoes a pre-crosslinking reaction with the epoxy resin, significantly reducing its reactivity. Furthermore, barbituric acid acts as an inhibitor, preventing excessive reaction. However, some of the EC331's activity and its alicyclic structure are still retained, allowing for further partial reaction at high temperatures, thus increasing the high-temperature modulus.

[0058] Example 5 innovatively combines the accelerator premix CP-01 with the additive YP-01, which allows for controllable post-curing performance, following Examples 1-4. Comparing the test results with Examples 1-4 and Comparative Examples 1-4, it can be seen that after optimizing the dosage of CP-01 and YP-01, their synergistic effect in the formulation resulted in a balanced performance across many aspects. Compared to competitor X, the high-temperature modulus and growth rate after reflow soldering reached or even slightly exceeded the level of competitor X; the room-temperature storage stability and operational open time were also better; and the high-temperature pudding mold adhesion before and after UHast was also higher. This is because both CP-01 and YP-01 release some activity at high temperatures, and due to their higher activity compared to other resins and curing agents in the formulation, and the certain reactivity between them, their combined effect resulted in the performance demonstrated in the tests, ultimately achieving the goal of passing the reliability test.

[0059] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A high reliability underfill adhesive suitable for large size chips, characterized in that, The raw materials include, by weight fraction: 25-35 parts of bisphenol F type epoxy resin, 1-3 parts of bisphenol A type epoxy resin, 0.5-2 parts of silicon modified epoxy resin, 0.5-1.5 parts of black paste, 0.8-1.2 parts of coupling agent, 0.1-0.5 parts of flow aid, 50-55 parts of spherical filler, 10-15 parts of curing agent, 0.2-0.8 parts of accelerator premix, and 0.5-1 part of additive with controllable post-curing performance; The preparation method of the additive with controllable post-curing performance comprises the following steps: According to weight parts, 70-80 parts of bisphenol F type epoxy resin and 20-30 parts of alicyclic amine curing agent are heated to 40-60℃, mixed and stirred for 1-4h, then 0.5-1 part of barbituric acid is added, and the mixing and stirring is continued for 1-2h to obtain the additive with controllable post-curing performance; The preparation method of the accelerator premix comprises the following steps: According to weight parts, 25-35 parts of bisphenol A type epoxy resin and 25-35 parts of alicyclic epoxy resin are mixed and stirred uniformly, then 35-45 parts of accelerator is added, stirred uniformly, and ground to obtain the accelerator premix; The accelerator has the following structural formula: 。 2. The high-reliability underfill adhesive for large-sized chips according to claim 1, wherein The preparation method of the accelerator comprises the following steps: 3,3',4,4'-benzophenonetetracarboxylic dianhydride, 2-ethyl-4-methylimidazole, and ethanol are sequentially added into a reaction container, heated to 50-80℃, stirred for 6-10h, then the reaction product is vacuum filtered, and finally dried in an oven at 80-120℃ for 2h to obtain the accelerator.

3. The high-reliability underfill adhesive suitable for large-size chips according to claim 2, wherein, The molar ratio of 3,3',4,4'-benzophenonetetracarboxylic dianhydride to 2-ethyl-4-methylimidazole is 1:2, and the ratio of the total moles of 3,3',4,4'-benzophenonetetracarboxylic dianhydride and 2-ethyl-4-methylimidazole to the moles of ethanol is 1:1-3.

4. The high-reliability underfill adhesive for large-sized chips according to claim 1, wherein The coupling agent is at least one of γ-aminopropyltriethoxysilane, γ-mercaptopropyltrimethoxysilane, γ-mercaptopropyltriethoxysilane, γ-glycidyloxypropyltrimethoxysilane, and γ-aminopropyltrimethoxysilane.

5. The high-reliability underfill adhesive for large-sized chips according to claim 1, wherein The curing agent is at least one of maleic anhydride, benzoic anhydride, phthalic anhydride, succinic anhydride, 4,4'-diamino-3,3'-diethyl diphenyl methane, diethyl toluene diamine, diaminodiphenyl sulfone, m-aminomethylamine, benzene diamine trimer, and dibenzyl amino ether.

6. A method for preparing the high-reliability underfill adhesive for large-sized chips according to any one of claims 1 to 5, characterized by, The method comprises the following steps: Mix and stir bisphenol F type epoxy resin 25-35 parts, bisphenol A type epoxy resin 1-3 parts, silicon modified epoxy resin 0.5-2 parts, black paste 0.5-1.5 parts, coupling agent 0.8-1.2 parts, flow aid 0.1-0.5 parts for 1-4 hours; add spherical filler 50-55 parts in 2-4 times, continue to stir at 70-90℃ for 6-10 hours; then reduce to room temperature, first add curing agent 10-15 parts, control temperature 25-30℃ and stir for 0.5-2 hours; then add accelerator premix 0.2-0.8 parts, control temperature 25-30℃ and stir for 0.5-2 hours; finally add controllable post-curing additive 0.5-1 parts, control temperature 25-30℃ and stir for 0.5-2 hours.

7. The method for preparing the high-reliability underfill adhesive suitable for large-size chips according to claim 6, characterized in that, The whole preparation process is carried out under vacuum condition, and the vacuum degree is not less than-0.08Mpa.

Images