鑽石線切割多晶矽晶片之蒸氣製絨研究

Abstract

太陽光電產業快速發展，仍以矽晶電池為主流，佔比超過92%，其中多晶矽佔比近70%。由於晶片製造仍需長晶與切片，使得晶片成本很難進一步下降。特別是切片成本目前已經高於長晶成本一倍之多，一般多在$0.1/Wp以上。目前砂線切割(SiC slurry slicing, SC)的切割損失高達40%，而鑽石線切割是降低切割成本的方法，鑽石線切割有下列幾個優勢: (1) 水性切割液的污染低;(2) 製程簡單; (3) 切割速度快 ; (4)切割損傷淺; (5) 切割損失（kerf loss）少等優點。當然未來在切割矽損失的回收可能性高，製程成本下降的優勢可以預期。但由於鑽石線切割晶片表面過於光滑，切割造成的缺陷少且分布於線痕上，使得一般SC晶片混酸製絨難以應用在DW晶片。因此DW晶片可以其他方式製絨，例如:反應離子蝕刻法(reactive ion etching, RIE)、金屬催化蝕刻(metal catalystic texturing, MCT)、噴砂等。然而目前這些製成仍無法穩定生產且成本高以及後續汙染物處理等問題。單晶矽因DW切割導入，成本大幅下降，多晶矽若無DW切割，成本無法有效下降，將失去與單晶矽的競爭優勢。 因此，本論文提出操作簡單、低排放、製成穩定的製絨技術，蒸氣蝕刻(vapor texture etching, VTE)，透過小型蒸氣蝕刻實驗，我們得知低反射率之表面結構與蝕刻機制的關係，比較蒸氣蝕刻與液相蝕刻，液相蝕刻是利用晶片切割損傷層製絨，而蒸氣蝕刻與晶片起始形貌無關。從小型蒸氣蝕刻，DW晶片製絨後平均反射率約19%(400-1100nm)，因此我們將將實驗規模放大，並針對蒸氣濃度、溫度、流速、晶片間距、添加物等變數比較晶片形貌，再由蒸氣與空氣配比與氣體流量，我們將可以將控制蝕刻厚度與晶片的黑度、均勻性。

In recent years, the low-cost diamond-wire (DW) slicing has been widely used for single-crystalline silicon (sc-Si) wafers, and this makes the sc-Si solar cells very attracted to the market. On the other hand, the adoption of DW slicing for multi-crystalline (mc-Si) wafers in slow due to the poor texturing quality in the existing production lines using acid solutions. The DW sliced wafers are too smooth to be textured well with the acid solutions and this leads the higher reflectivity and the lower efficiency for the mc-Si solar cells . Several effective methods, such as reactive ion etching (RIE), metal catalystic texturing (MCT)], and sand blasting pre-texturing , have been proposed. However, they are either too costly or too troublesome. Therefore, this paper proposes simple and low-emission, stable texturing technology, vapor texture etching (VTE). Through the small vapor etching experiment, we know the relationship between the surface structure of low reflectivity and the etching mechanism. Compare Vapor etching and liquid phase etching, liquid etching is the use of wafer cutting damage layer to texturing, The VTE has different etching mechanisms from the liquid etching, and for some conditions we have found that it is still effective to a polished wafer. From the small experiment, the average reflectivity of the DW wafer after texturing is about 19% (400-1100nm), so we will scale up experiment and compare the wafer morphology with the vapor concentration, temperature, air flow rate, wafer spacing, additive and other variables. And then by the ratio of steam to air with the gas flow, we will be able to control the thickness of the etched etched and uniformity of the wafer.