標題: 金屬誘發側向結晶之底部閘極多晶矽薄膜非揮發性記憶體研究
Bottom-gate Poly-Si Thin Film Transistor Nonvolatile Memories with
作者: 盧瑞琪
Lu, Ray-Chi
趙天生
Chao, Tien-Sheng
理學院應用科技學程
關鍵字: 金屬誘發側向結晶;鎳捉聚;固相結晶法;MILC;Ni-Gettering;SPC
公開日期: 2013
摘要: 低溫多晶矽(LTPS)薄膜電晶體提升效能所製作的方法中,利用鎳金屬誘發非晶矽薄膜側向結晶法(NILC),能在較低退火溫度與較短退火時間下,結晶變大顆粒,而廣受矚目。對於鎳金屬誘發多晶矽薄膜來說,在誘發結晶過程中所殘留的鎳金屬,會被晶界及缺陷所捕捉並產生深層能階,進而造成漏電流與元件效能衰退,因此發展有效的捉聚鎳方法來降低鎳金屬誘發多晶矽薄膜中的鎳金屬殘留,與降低多晶晶界處的懸鍵及缺陷,是提升高效能及高品質薄膜電晶體的重要課題。 我們選擇了SPC與MILC作比較,並針對不同磷離子佈植濃度(1x1015cm-2,5x1014cm-2)做不同製程條件上比較,是否磷真的能幫助金屬鎳結晶。我們會分別從物性以及電性做討論,包括XRD 分析、電導值、載子寫入、漏電流比較、源極和汲極交換量測之可靠度探討。 在此研究中我們得到在記憶體通道中,摻雜高濃度的磷之樣品具有較小的晶粒、載子遷移率沒有多大差異與寫入速度快,但就資料保存時間看來是沒有差異。由於B、C元件無法順利的進行抹除的特性,我們判斷可能是因為B、C元件製作時先行進行源、汲極通道層的磷離子佈植後再進行金屬誘導結晶(MILC)多晶矽造成金屬汙染,導電性不佳,漏電流大,導致B、C元件在進行抹除的時候,順利將捕捉電荷進入氮化矽層,但是無法順利的將電荷保存在氮化矽層,電荷會經由汲極洩漏掉,所以導致元件幾乎完全不具有記憶體特性。本論文中我們將逐步的檢驗並驗證磷對鎳金屬誘導側向結晶的影響。
TFTs of Low-Temperature Polycrystalline Silicon, LTPS have good characteristic by Metal Induced Lateral Crystallization (MILC) technique. It has been purposed due to metal can be used to cause crystallization at controlled locations with low temperature. More than large crystal grain size can reduce defects and enhance carriers mobility, many research focus on how to increase grain size. Nickel Induced Lateral Crystallization will residual nickel in crystallization, and it can be trapped by grain boundary and damage. It cause current leakage of device and degradation. So, it is important to develop the Ni-gettering to reduce the residue nickel and damage in grain boundary in NILC, and further to increase the high quality and high effective TFT transistor. We choice the SPC and MILC for our study , and to force the different phosphorus concentration (1x1015cm-2 and 5x1014cm-2) compare by the different process. We want to demonstrate whether the phosphorus will reduce nickel induced crystallization. I will express the character of the fundamental physics and electric. The item included the XRD、program、current leakage, and exchanged the source and drain. In this study, we found the high doping of phosphors who will get the small grain size、no apparent different in mobility and retention, but the program is better for it. Due to device B/C can not easy to erase, it could be the process sequence issue. S/D then MILC would cause the metal residual in channel, and it would cause the leakage of current. So, we cannot erase the device. We will describe the detail to this thesis.
URI: http://140.113.39.130/cdrfb3/record/nctu/#GT070052917
http://hdl.handle.net/11536/73366
顯示於類別:畢業論文