鎂鋁基合金之電子束銲接研究

Abstract

本研究係以純鎂、AZ31B、AZ61A、AZ91D及純鋁為研究材料，藉熱擠形（前四者）與鑄造製得12mm厚板，銑除表面氧化層後，於高真空環境對接為11mm厚之同、異質電子束銲件，研究內容分為下列三階段： 第一階段係以單因子（加速電壓、射束電流、銲接速率與聚焦位置）趨勢分析獲得AZ系列合金之低壓電子束銲接最佳化參數及銲件強度經驗式，其應力集中銲件最大抗拉強度分別可達原材之78%、83%與82%，非應力集中銲件分別為91%、96%與89%，故知應力集中效應至少達13%、13%與7%，熱影響區與少量氣孔影響至少為9%、4%與11%，且接合效能與製程窗分別為AZ61A＞AZ91D＞AZ31B與AZ91D＞AZ61A＞AZ31B。再者，隨Al含量增加，AZ系列銲道析出物分佈將由稀疏散佈顆粒狀轉為濃密交錯樹枝狀，因而衍生二種破斷模式：不規則斜穿銲道破斷與規則縱向界面破斷，AZ31B與AZ91D分別以前者與後者為主，AZ61A則二者兼具。 第二階段續以田口式分析法探討AZ系列合金之高壓電子束銲接參數特性，並驗證缺值法之可行性，再由變異數分析獲致參數最佳化組合與貢獻度。在此所採用七項控制參數對銲件影響程度依次為擺弧效應＞聚焦位置＞應力釋放＞材料種類＞射束電流≒銲接速率≒加速電壓。各材料之最佳化趨勢均以不執行擺弧、應力釋放與聚焦於工件底面為宜；材料選用則以AZ61A最佳、AZ91D次之、AZ31B最差；加速電壓、射束電流與銲接速率三者則無特定組合可循；這些結果尚可聯結灰關聯分析將參數對銲道缺陷（甚至尺寸）之影響完全量化。此外，多數銲道具銲蝕、銲根凹陷與氣孔等缺陷，前二者由非均勻對稱高斯分佈之電子束能量所造成，氣孔則為原材內存親氧性析出物Mg17Al12重熔分解所生成之氧氣，其存量視銲池析出物多寡與對流排氣優劣而定，此等缺陷均足以形成應力集中並造成嚴重損害。再者，經比較低、高壓電子束銲件可知：銲道剖面積及其形態係由射束功率與銲接速率二者相互搭配而得，因能量轉換效率隨之而異，故輸入能量與銲道剖面積之間未必呈等比例關係。 第三階段則以高壓電子束銲接之最佳化參數進行五種材料之同、異質接合。AZ系列銲件之Al含量以6.0wt.%為界，熱影響區形成機制可劃分為二：晶粒成長與次微米相析出，AZ31B與AZ91D分別以前者與後者為主，AZ61A則二者兼具且形成最窄之熱影響區。再者，當銲道二側Al含量差異過大時，不僅使銲道內部因亂流而產生大量氣孔，且分佈不均之MgAl2、Mg2Al3與Mg17Al12亦因快速冷凝而形成大量裂縫，銲件亦僅於強度較低之銲道界面發生縱向規則破斷。然而，介金屬化合物之生成對銲道並非全然有害，Mg17Al12析出於Mg-Al-Zn銲道即屬此例。

The research materials in this study consist of home-made 12mm-thick pure Mg, AZ31B, AZ61A, and AZ91D extruded plates, and a commercial pure Al ingot. The oxide layer was removed from the plate surfaces, and two plates were welded together under a high vacuum using a butt joint process with similar and dissimilar metals welding (SMW and DMW) without a filler. This process created a 11mm-thick weldment. The follow-up research results can be divided into three stages. Stage 1 indicates that the optimum parameters of low-voltage (LV) electron beam welding (EBW) and the empirical weldment strength formulae for AZ31B, AZ61A, and AZ91D alloys can be obtained by changing one factor (accelerating voltage, beam current, welding speed, and focal position) at a time (COFAAT). The stress and non-stress concentration (SC and NSC) weldments were 78%, 83%, 82% and 91%, 96%, 89% of the three base material strengths, respectively. The harmful SC influence in the weld and grain coarsening in the heat-affected zone (HAZ) reached at least 13%, 13%, 7% and 9%, 4%, 11%, respectively. The joint efficiency and process window of these alloys rank in decreasing order as AZ61A, AZ91D, AZ31B and AZ91D, AZ61A, AZ31B, respectively. Moreover, the distribution of precipitates in the fusion zone (FZ) changes from a relatively small number of scattered particles to a dense population of dendrites as the Al content of the magnesium alloy increases. Alloy weldments break in one of two fracture modes: an irregular FZ fracture, or a regular HAZ fracture. AZ31B usually exhibits the former mode and AZ91D the latter, while AZ61A exhibits both modes equally. Stage 2 illustrates that Taguchi’s method can not only reduce the number of experiments required, but can also produce precise optimum parameters and predicted ultimate tensile strength (UTS) even when some data points are lost. Analysis of variance (ANOVA) results indicate that the strength of a weldment is affected, in order of impact, by beam oscillation, focal position, stress relief, material difference, beam current, welding speed, and accelerating voltage. Optimal parameters generally include a nonoscillating beam, a focus at the bottom, and no stress relief. Joint efficiency ranks in the order of AZ61A, AZ91D, and AZ31B. Beam current, welding speed, and accelerating voltage exhibit no specific change tendencies. The parameters influencing various defects (and their dimensions) can be individually quantified by linking Taguchi’s method with grey relational analysis. Additionally, most welds form undercuts, root concavities, and pores. Beam energy with asymmetrical Gaussian distribution causes the first two; the latter one result from the dissolution of oxidative precipitates to reform molecular oxygen. The number of pores in the weld pool depends on the number of precipitates and the degassing effect. These defects obviously induce SC in the weld, and may seriously decrease its strength. Furthermore, for comparing between LV-EBW and high-voltage (HV) EBW, the area and shape of the weld cross section are related to the input power and welding speed, and they are not proportional to the input energy. Stage 3 analyzes the weld differences that the five different materials are welded with SMW and DMW using the same HV-EBW optimal parameters. The formation mechanism for the HAZ in Mg-Al-Zn weldment at 6.0wt.%Al can be classified into two modes: grain coarsening and submicron-sized crystal precipitation. AZ31B exhibits the former mode, and AZ91D exhibits the latter. AZ61A exhibits both modes and forms the narrowest HAZ. Mixing alloys with different Al contents and fluidities causes turbulence in the weld pool when welding dissimilar metals. The weld pool remains porous after solidification. The streaky distribution of metastable (MgAl2) and intermetallic (Mg2Al3, Mg17Al12) phases increases thermal stress and induces cracks which greatly reduce the weldment strength. These weldments therefore tend to follow the regular HAZ fracture mode. However, the formation of intermetallic compounds is improbably harmful, for example Mg17Al12 precipitating in Mg-Al-Zn weld.