標題: 三五族晶片與矽基光柵之晶片低溫直接接合研究
Low temperature direct bonding of III-V semiconductor on Si-based grating arrays
作者: 陳百麒
Chen, Bai-Ci
林建中
郭浩中
Lin, Chien-Chung
Kuo, Hao-Chung
光電系統研究所
關鍵字: 矽光子學;積體光路;直接晶片接合;氧電漿輔助晶片接合;磷化銦;奈米光柵;鍵合推力測試;鍵合介面分析;光柵反射率;光致螢光光譜;Silicon photonics;Integrated optics;Direct wafer bonding;Oxygen plasma-activated wafer bonding;Indium phosphide;Nanometer scale gratings;Bonding die shear test;Bonding interface analysis;Grating reflectivity;Photoluminescence
公開日期: 2015
摘要: 在光積體電路(optical integrated circuits)的發展下,晶圓接合(Wafer bonding)的方法能夠達到直接在矽(Si)材料上,將三五族磊晶片(III-V epitaxial layer)接合的製程。如此可以藉由三五族半導體元件,將光傳遞到矽材料所設計的波導或是其他結構。在我們實驗裡,研究的便是如何處理鍵合過程使鍵結強度達到之後製程所需的要求。在低溫鍵結的過程中,藉由氧電漿處理完(O2 plasma treatment)的樣品可以得到平整與潔淨度高鍵合表面,其RMS值小於0.2 nm。藉由實驗矽基板與磷化銦(InP)基板的低溫鍵合,可以得到高於標準規範(MIL-STD 883)強度的鍵合,鍵合強度為29.62 kg/cm2。 對於鍵合的介面而言,可以由穿透式電子顯微鏡(TEM)觀測到小於10 nm以下的SiO2接合層與其元素的定性濃度分布,這對於之後製程的光耦合(Optical coupling)或是光傳遞(Optical transfer),不會因為SiO2層的厚度而造成太大的影響。因此在目前鍵合的結果中,比較適合之後製程的條件如下:退火溫度為220℃、鍵合壓力為96 kg/cm2 、腔體壓力為 0.3 torr、退火時間為8小時。 用一定比例的鹽酸與水(HCL:H2O)溶液將磷化銦基板蝕刻完後,可以明顯觀察到露出的InGaAsP epitaxial layer情形,也可藉由量測光激發螢光頻譜(PL spectrum)的訊號來判斷InGaAsP epitaxial layer是否成功留下。 我們也設計了不同類型的矽基光柵,並在光柵間隙中填入了二氧化矽(SiO2),去觀察反射率產生的相對應變化。並將1550nm MQW InP epitaxial layer以低溫鍵合的方式,接合在所設計的矽基光柵上,量測其光激發螢光頻譜,找尋光柵反射率(Grating reflectivity)及相對應螢光頻譜間的關係。我們得到不同類型的矽基光柵,其反射率與光激發螢光頻譜輸出能量的關係呈現正相關。希望以上研究能應用在之後光耦合與光通訊(Optical communication)的元件設計上,並對其未來之發展有益。
In the development of optical integrated circuits, there has been a process that can directly bond III-V epitaxial layer on the silicon materials. This process can transfer light to silicon waveguide or other structures by the III-V semiconductor device. Therefore, the aim of our research attempts to explore how to attain proper bonding strength for our bonding process. This research involved AFM, Bonding die shear, TEM, EDS. During the low temperature bonding process, we got a smoother and cleaner surface of the samples cleaned with O2 plasma and the RMS value was less than 0.2 nm. By low-temperature bonding for silicon substrate and InP substrate, we gained the bonding strength of 29.62 kg/cm2 higher than standard specification (MIL-STD 883). As for bonding interface, we can use TEM to observe the intermediate bonding layer and the qualitative concentration distribution of the elements. The intermediate bonding layer is less than 10 nm and the thickness doesn’t have too much impact on optical coupling and optical transfer in the following process. Consequently, the more proper bonding conditions we’ve measured so far are as following. The annealing temperature is 220°C; the bonding pressure is 96 kg/cm2; the chamber pressure is 0.3 torr and the annealing time is 8 hours. After etching InP substrate with HCL: H2O solution in certain ratio, we observed the condition of InGaAsP epitaxial layer obviously. Also, we judged whether the InP epitaxial layer is successful or not by measuring the PL signal. We also designed different types of gratings on silicon substrate and filled SiO2 in the interspace of grating to observe the relative change of reflectivity. Besides, we bonded the 1510nm MQW InP epitaxial layer on silicon dioxide grating under a low temperature and measured PL spectrum to figure out the relation between grating reflectivity and PL spectrum. To sum up, the relationship between reflectivity and PL output power in different types of gratings shows positive correlation. We sincerely hope our research will be applied to the design of optical coupling and optical communication device in the near future.
URI: http://140.113.39.130/cdrfb3/record/nctu/#GT070258008
http://hdl.handle.net/11536/127626
Appears in Collections:Thesis