| 摘要: | 此研究中，我們開發了一種助銲性環氧樹脂混合Sn58Bi銲料粉體的漿料，並於回銲過程中，依靠熔融銲料液滴的自聚集行為，形成銲點後，圍繞銲點周圍的環氧樹脂固化後形成底膠填充保護銲點，於製程中一步驟完成覆晶接合與底膠充填，我們將此新材料稱之為自聚集錫膠。
助銲性環氧樹脂對回銲過程中熔融銲料液滴的自聚集行為有極大的影響，因此選擇二羧酸化合物和環氧基的固化系統以確保銲料在回銲過程中有效去除表面氧化物並潤濕銲墊以形成銲點，並使用熱差式掃描量熱分析儀（DSC）分析各種二羧酸化合物-環氧單體系統的固化反應進程。此外也探討使用潛伏性催化劑調整環氧樹脂的固化過程，以降低環氧樹脂固化進程對自聚集銲點生成之影響。
從DSC動態掃描模式下的分析，發現銲料粉體與助銲劑所生成的羧酸錫鹽可影響環氧樹脂的固化進程且其與具路易斯鹼性的催化劑共同存在時將加速環氧樹脂固化反應進程。以化學分析電子光譜儀(ESCA),傅立葉轉換紅外光譜儀(FTIR)證明羧酸錫鹽的生成，以及從羧酸錫鹽結構推導其化學性質並建立其催化環氧樹脂固化進程的機制。
通過理解以上這些機制，我們成功配製了固化後機械性質良好且具備助銲性的環氧樹脂，並製備成自聚集錫膠進行初步的回銲測試。
通過應用分枝理論的計算，對自聚集錫膠形成自聚集銲點的過程進行最佳化。理論推導表明，該環氧樹脂固化系統的凝膠點轉化率約為0.78-0.8。推導出的凝膠點轉化率可以幫助確定DSC熱流曲線上固化反應放熱峰與銲料熔融吸熱峰的最佳相對位置，從而使環氧樹脂的固化進程對熔融銲料液滴自聚集的行為之影響降至最低。
為使自聚集錫膠應用於密集接點的覆晶接合，使用超細銲料粉體有其必要性，但在回流的早期階段，超細銲料粉末（2-8 μm）在熔化前往往會沉積和聚集，形成緻密結構，阻礙助銲性環氧樹脂完全潤濕銲料粉體表面和去除表面氧化物。因此通過在環氧樹脂初始固化過程中使含有雙胺基的單體與環氧單體反應生成微凝膠，於未熔融銲料粉體的堆積結構間創建空間，阻止粉體間彼此接近。通過應用分枝理論計算出凝膠點以確定在不形成大網絡結構的情況下可添加的最大雙胺單體量，以及利用DSC與流變儀證明微凝膠的生成。最終回銲測試表明，含有雙胺單體的樣品有效破壞了未熔化銲料顆粒的緻密堆積結構，使助銲性環氧樹脂能潤濕至銲料粉體的表面清除氧化物，使熔融銲料完全自聚集到銲墊上，而不含二胺單體的樣品則無法做到這一點。
在最佳化自聚集錫膠後，進行了回銲和可靠性測試。回銲測試顯示所有的LED樣品均正常工作，Sn58Bi銲料完全自聚集到銲墊上，沒有斷路或橋接。EDX分析確認界面形成了Au-Sn金屬間化合物（IMC），Ni擴散極少，環氧樹脂完全包覆銲點。推力測試表明平均強度為127 g，較使用傳統錫膏增強2.54倍。可靠性測試顯示剪切強度下降40-50%，但仍高於使用傳統錫膏回銲之樣品。經高溫高濕測試後樣品剪切強度下降歸因於環氧樹脂和銲點界面處分層以及環氧樹脂硬度下降。溫度循環測試後剪切強度下降歸因於界面接著處出現裂紋，這是由於熱膨脹係數差異導致。
;In this study, we developed a slurry contained Sn58Bi solder powder and fluxing epoxy resin. During the reflow process, the molten solder droplets self-assemble to form solder joints, then epoxy resin surrounding formed solder joints cures to serve as underfill. This process achieves flip-chip bonding and underfilling in single step, and we term this new material self-assembly solder resin.
The fluxing epoxy resin significantly impacts the self-assembly behavior of molten solder droplets during reflow. Therefore, we selected a curing system based on dicarboxylic acid monomers and diepoxide monomer to ensure the effective removal of surface oxides of solder powder and wettability of molten solder on pads during reflow. We used differential scanning calorimetry (DSC) to analyze the curing reaction process of various dicarboxylic acid-epoxy monomer systems. Additionally, we explored using latent catalysts to adjust the curing process of the epoxy resin to minimize its’ impact on the formation of self-assembling solder joints.
Dynamic DSC analysis revealed that the tin carboxylate salts formed between the solder powder and the flux significantly affect the epoxy resin′s curing process. The coexisting of Lewis base catalyst and tin carboxylate salts will further accelerate the curing reaction. We used X-ray photoelectron spectroscopy (XPS) and Fourier transform infrared spectroscopy (FTIR) to confirm the formation of tin carboxylate salts and deduced their chemical properties to establish the mechanism by which they catalyze the epoxy resin curing process. By understanding these mechanisms, we successfully formulated a fluxing epoxy resin with good mechanical properties after curing and creating a self-assembly solder resin for initial reflow tests.
Using branching theory calculations, we optimized the self-assembly process of the solder joints formed by self-assembly solder resin. The theoretical derivation indicated that the gelation conversion rate of this epoxy curing system is approximately 0.78-0.8. This gelation point conversion rate helps determine the optimal relative positions of the exothermic peak of curing reaction and the endothermic peak of the solder melting on the DSC thermogram, minimizing the impact of the epoxy resin curing process on the self-assembly behavior of molten solder droplets.
For the self-assembly solder resin to be applicable in fine pitch flip-chip bonding, the use of ultra-fine solder powder (2-8 μm) is necessary. However, during the early stages of reflow, ultra-fine solder powder tends to sediment and agglomerate before melting, forming dense structures that hinder the fluxing epoxy resin from fully wetting the powder surface and removing surface oxides. To address this, we reacted a diamine monomer with the epoxy monomer during the initial curing process to form microgels, creating space within the packed structure of the unmelt solder powder, preventing solder powder particles from approaching each other. Branching theory calculations determined the maximum amount of diamine monomer that can be added without forming a large network structure, and DSC and rheometer tests confirmed the formation of microgels. Final reflow tests showed that samples containing diamine monomers effectively disrupted the dense packing structure of the unmelt solder particles, allowing the fluxing epoxy resin to wet the solder powder surface and remove oxides. In contrast, samples without diamine monomers failed to achieve this.
After optimizing self-assembly solder resin, we conducted reflow and reliability tests. The reflow tests showed that all LED samples functioned normally, with Sn58Bi solder fully self-assembling onto the solder pads without open circuits or bridges. Energy-dispersive X-ray (EDX) analysis confirmed the formation of Au-Sn intermetallic compounds (IMC) at the interface, with minimal Ni diffusion, and the epoxy resin completely encapsulated the solder joints. The shear test showed an average strength of 127 g, 2.54 times higher than sample reflowed with traditional solder paste. Reliability tests indicated a 40-50% reduction in shear strength, still higher than that of samples reflowed with traditional solder paste. The decrease in shear strength after high-temperature and high-humidity testing was attributed to delamination at the epoxy resin-solder joint interface and a decrease in epoxy resin hardness. After temperature cycling tests, the reduction in shear strength was due to cracks at the interface, caused by differences in the coefficients of thermal expansion. |
