請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/63244
標題: | 無鉛銲料接合之介面反應研究 Analysis of interfacial reaction during the formation of Pb-free solder joint |
作者: | Chee-Key Chung 鄭子企 |
指導教授: | 高振宏(C. Robert Kao) |
關鍵字: | 無鉛銲料,介面反應,動力學,氧化,微添加合金, Pb-free soldering,interfacial reaction,kinetics,oxidation,micro-alloying, |
出版年 : | 2012 |
學位: | 博士 |
摘要: | 本研究著重於分析銲點在銲接過程中之界面反應。本論文一共分成四個部分。第一部分著重於銲點熔化之初期反應。在 217.4°C下使用Sn4Ag0.5Cu焊料在Cu基板上迴銲反應初期主要生成Cu6Sn5和Cu3Sn相。即使溫度在217.4°C這樣相對的低溫下,Cu6Sn5和Cu之間仍會生成薄薄的一層(80~90 nm) Cu3Sn相。而使用高解析穿透式電子顯微鏡(HR-TEM) 發現在Cu6Sn5相和Cu3Sn相的相界上有富銅的區域,因為晶格的錯位,這些區域大概只有3~6 nm的厚度,在Cu6Sn5相的晶界中也可觀察到類似的富銅區域,這些在相界以及晶界上富銅區域的存在可以說明在迴銲反應初期生成Cu6Sn5相和Cu3Sn相的主要機制是靠銅原子在晶界以及相界的擴散來主導的。而在此之前,也就是銲點溶化之前,就能看到有Cu6Sn5相和(Ni,Cu)3Sn4相的生成。我們發現Cu6Sn5相和Cu3Sn相會在銲點與銅基板之間生成,而(Ni,Cu)3Sn4 和 Ni2SnP會在銲點與無電鍍鎳之間生成。在銲點融化之前,焊料和銅基板以及銲點和無電鍍鎳之間的接觸點會作為異質成核的地方。由TEM的觀察結果指出,這些介金屬的成核都是屬於擴散誘導的結晶行為。
第二部分著重於在迴銲過程中,銲點內的介金屬化合物裡形成的奈米孔洞。無電鍍鎳浸金(ENIG)與錫銀銅的反應形成了一層Ni2SnP化合物。這個界面層包含的奈米孔洞會導致多種類型的電子產品中的銲點失效。到目前為止,因為其需要先進的分析技術,對於這些孔洞目前只有少數的研究。所以在此對於無電鍍鎳浸金和Sn3Ag0.5Cu的反應進行了研究。觀察在迴焊以及淬火過程中的微觀結構,以了解這些孔洞的演變。透過不同的迴焊溫度來分析孔洞密度的變化,並分析迴焊次數對於孔洞密度的影響。我們發現界面上的孔洞有兩種類型:一個在Ni3P區而其他在Ni2SnP區。在Ni2SnP中的孔洞會相互連接,形成一個連續孔洞層。孔洞密度會隨著迴焊次數的增加而增加,但與迴焊溫度無關。接下來會討論孔洞的成核與合併的機制。 第三部分著重在對於形成銲點時重要的氧化物。雖然在焊接過程中,氧化物的產生會造成不良影,但由於氧化物的奈米級厚度和不規則的織構,導致很難去測量氧化物的厚度。關鍵的銅氧化層厚度的是由二次離子質譜儀,掃描電子顯微鏡(SEM),能量色散X射線光譜儀(EDX)和TEM來測量。在銅基板塗上有機防腐層(OSP),在150°C下,85% 相對濕度用不同的時間時效處理。當氧化層厚度達18±5 nm時,會有無潤濕性的現象發生。一但氧化層厚度超出這個臨界厚度,無潤濕銲點的比率會指數成長增加。氧化層厚度的成長速率遵循拋物線成長速率的行為,其速率常數為0.6×10-15 cm2 min-1。銅原子和氧原子通過OSP和氧化層的相互擴散導致氧化銅的形成。氧化層的機制將在後面介紹和討論。 第四部分是透過微合金化作用來改善奈米孔洞和氧化物的形成。將不同重量磷(P)以ppm等級的不同濃度加入Sn4Ag0.5Cu的無鉛焊料。對於電解鎳金層和摻雜不同濃度P以及沒有摻雜P的Sn4Ag0.5Cu焊料的界面反應進行研究。焊點迴焊一次後,再經歷超過兩次的迴焊。我們使用SEM和TEM來分析介金屬化合物的厚度以及其成分。使用達格4000儀器來測定焊點的拉伸性能。結果表明,焊料中的磷含量與IMC層的厚度和最大拉伸應變成反比。 藉由TEM觀察出奈米晶層存在於(Cu,Ni)6Sn5和(Cu,Ni)3Sn4之間。這種奈米晶層是上述現象的主要原因。這一層不僅有效抑制IMC的生長,同時也減少了3次迴焊後銲點的拉伸強度。隨著P摻雜焊料中的P含量增加,IMC層從原本厚而且崎嶇的形貌,變成薄而且平坦。 P的摻雜對於銲點的形成的過程中產生了不同的介面反應。 This research analyzed the interface of solder joint during the soldering process. It separated into four parts. Part I focused on early melting of the solder joint. The dominant mechanism for the growth of Cu6Sn5 and Cu3Sn at the early stage was investigated using Sn4.0Ag0.5Cu solder reflowed on Cu. A thin layer of Cu3Sn (80~90 nm) was observed between Cu6Sn5 and Cu even though the temperature was as low as 217.4 °C. High-resolution transmission electron microscopy (TEM) identified Cu-enriched regions at phase boundaries between Cu6Sn5 and Cu3Sn. Such regions were about 3-6nm thick. Similar Cu-enriched regions were also observed at the grain boundary of Cu6Sn5. The existence of these Cu-enriched regions suggested that the grain boundary and phase boundary diffusion of Cu atoms were the dominant mechanisms for the growth of Cu6Sn5 and Cu3Sn at the early stage of soldering reactions. Further on this work, prior to the melting, early formation of Cu6Sn5 and (Ni,Cu)3Sn4 were investigated. At the contacting point, Cu3Sn and Cu6Sn5 were found at the solder-Cu substrate; while (Ni,Cu)3Sn4 and Ni2SnP were identified at the solder-electroless Ni. The contacting points were confirmed as the sites of heterogeneous nucleation. TEM analysis suggested that the nucleation of both intermetallic compounds were attributed to diffusion-induced crystallization. Part II investigates the nano voids inside the intermetallic compound during the formation of solder joint. The reaction of electroless Ni immersion Au with SnAgCu forms a layer of Ni2SnP. This interfacial layer contains nano voids and responsible for many types of failure in electronic solder joints, but so far, detailed investigation of these voids has been pursued only in a few studies because of the need for advanced analysis techniques. Thus, the interaction of Sn3Ag0.5Cu with electroless Ni/immersion Au is investigated. The microstructures during reflow are quenched to understand the evolution of these voids. Different peak reflow temperatures are adopted to analyze the change in the void density. The effect of the number of reflow cycles on the void density is then investigated. Two types of interfacial voids were found: one in the Ni3P region and the other in the Ni2SnP region. The voids in Ni2SnP were connected to each other to form a void line. The void density increased with the number of reflow cycles, but not with the peak reflow temperature. The mechanism of the void nucleation and coalescence was then discussed. Part III studies the critical oxide for the formation of solder joint. Though oxide is an undesirable effect in the formation of solder joint, characterizing the critical oxide thickness for soldering is limited due to oxide’s nano-scale thickness and irregular topographic surface. The critical copper oxide thickness was characterized by Time-of-Flight Secondary Ion Mass Spectrometry, scanning electron microscopy (SEM), Energy-Dispersive X-ray spectroscopy (EDX), and TEM. Copper substrates were coated with an Organic-Solderable-Preservative (OSP) layer and baked at 150 °C and 85% Relative Humidity for different amounts of time. The onset of the non-wetting phenomenon occurred when the oxide thickness reached 18 ± 5 nm. As the oxide grew beyond this critical thickness, the percentage of non-wetting solder joint increased exponentially. The growth of the oxide thickness followed a parabolic rate law. The rate constant of oxidation was 0.6 x 10-15 cm2 min-1. Oxidation resulted from interdiffusion of copper and oxygen atoms through the OSP and oxide layers. The oxidation mechanism will be presented and discussed. Part IV is to improve the nano voids and oxide via micro-alloying effect. Different weights of Phosphorous (P) at the level of part per million are introduced to the Sn4Ag0.5Cu Pb-free solder ball. The interfacial reactions of different P-doped and undoped Sn4Ag0.5Cu solders with electrolytic Ni-Au were investigated. The solder joints were reflowed once and then subjected to two more reflow cycles. The thickness of the intermetallic compound (IMC) layer and the IMC composition were analyzed by SEM and TEM. The tensile properties of the solder joint were measured using a Dage 4000 instrument. The results showed that the P content of the solder was inversely proportional to the IMC layer thickness and maximum tensile strain. TEM observations showed that a nanocrystallite layer existed between (Ni,Cu)6Sn5 and (Cu,Ni)3Sn4. This nano-crystallite layer was responsible for the abovementioned relationship. This layer not only suppressed the growth of the IMC layer effectively but also decreased the pull strength of the solder joint after three reflow cycles. As the P content of the P-doped solder was increased, the IMC layer, which originally had a chunky morphology, became thin and flat. The P-doped solder produced a different solder reaction in the formation of solder joint. |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/63244 |
全文授權: | 有償授權 |
顯示於系所單位: | 材料科學與工程學系 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
ntu-101-1.pdf 目前未授權公開取用 | 23.55 MB | Adobe PDF |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。