Process Development of Zone-Leveling Pulling Growth of Stoichiometric Lithium Niobate Single Crystals

鈮酸鋰晶體為一相當有潛力的光電材料，其具有多樣化的性能，而隨著晶體中組成的改變，例如晶體中Li/Nb比的提高、或是摻雜的使用，更可加強或改善鈮酸鋰晶體中，特定的某些性能，例如鈮酸鋰晶體的光折性能、或是其光阻抗強度等。然而，對於等化學計量比、或是含有摻雜之鈮酸鋰單晶的生長，晶體生長過程中所發生的偏析現象，則常常造成所生長晶體中的組成偏移、或是均勻性不佳。目前，用於生長等化學計量比鈮酸鋰(SLN)單晶的方法，以日本Oxide公司的雙坩堝柴氏法(DCCz)為主，然而其所使用的粉末進料裝置並不適用於一般的商業規格長晶爐。於是在本論文中，進行連續進料之區熔提拉法(zone-leveling Czochralski method，簡稱為ZLCz法)晶體生長程序之開發，此方法得以直接使用於商業規格長晶爐，無須改裝。並首次將ZLCz法用於生長組成均勻的鎂摻或是鋅摻之SLN單晶。在論文中，我們以直徑一吋的小尺寸鎂摻或鋅摻SLN晶體的生長，探討ZLCz法之生長特性及摻雜的偏析行為，而更進一步的進行生長程序之改良，以生長尺寸接近兩吋的高品質SLN單晶。 區熔提拉法所存在的幾項問題，主要有晶體中的氣泡包覆；熔區變化所造成晶體之組成改變；及晶體劈裂的問題。其中晶體中氣泡包覆的問題，得藉由內坩堝的使用加以解決。而熔區變化及晶體劈裂的問題，則由熱場設計進一步改進。藉由適當熔區深度及熔區周圍之熱場設計，可避免系統中散熱條件的改變，所造成熔區長度於晶體生長中發生變動。同樣由熱場設計，可將生長系統中之溫度梯度降至20℃/cm以下，以避免晶體的劈裂。在此熱場下，由熔區組成為59∼60 mol%鋰過量之熔區，配合SLN組成之固體進料，所生長之1 mol%鎂摻之SLN晶體中，其Li/Nb比約為0.975，而在晶體軸向上的變動則約在1%以內。若進一步調整熔區(含16 mol% K2O之熔區)或進料(51 mol% Li-excess)之組成，則可進一步提升晶體的軸向性質及組成均勻性，其中晶體軸向之Li/Nb比變動可降至0.5%以下，然而以51 mol% Li-excess進料所生長之1 mol%鎂摻SLN晶體之Li/Nb比可提升至近0.98。由於隨著晶體中摻雜濃度的增加，SLN晶體中的Li/Nb比會隨著降低，對於1 mol%鎂摻SLN晶體，其理論上可達到的Li/Nb比極限亦約為0.98，我們所生長1 mol%鎂摻SLN晶體中之Li/Nb比，已接近理論上可達到的極限。此外，對於鎂摻晶體中摻雜之分佈，晶體中之軸向摻雜分佈變動可控制在約5%以下。而對於晶體之徑向組成分佈，則由量測結果看來，其分佈變動約在0.1%以下。 對於鋅摻SLN晶體，由於其偏析係數約為0.58，則相對於鎂摻晶體中的摻雜濃度分佈均勻，在鋅摻晶體中之摻雜分佈變動會大上許多。而當晶體由具有較高之預設預摻濃度之熔區提拉，則由區熔的效果，可有效降低晶體中摻雜之變動，其鋅摻濃度沿晶體軸向之變動幅度約與鎂摻晶體相近，可控制在5%以下。此外，由晶體性質對應於鋅摻濃度之變化，我們也可發現當鋅摻濃度約為1 mol%時，性質的變化趨勢會出現一明顯的轉折。由晶格缺陷模型分析，在低濃度摻雜下，摻雜主要取代鈮酸鋰晶格中的 對位缺陷；在高濃度摻雜下，摻雜則開始佔據晶格中的正常鈮原子及鋰原子位置。所以對於SLN晶體，晶體中之對位缺陷濃度約為1 mol%。 整體而言，以ZLCz法生長鎂摻或鋅摻SLN晶體之程序開發非常成功。在初步的應用測試上，已可成功製作出PPSLN晶片，而在晶體的性質及組成均勻性控制上亦有相當不錯的成果。Lithium niobate (LN) single crystal is a potential opto-electro material with various properties. The usage of the doping or the increase of the Li/Nb ratio in the LN crystal could be highly enhanced the properties for different applications, such as the photorefractive enhancement for holographic storage devices, or the larger photoresistance for the laser application. However, the crystal having non-uniformity along the grown boules was found to be a serious problem due to the segregation. The conventional technique for growing stoichiometric lithium niobate (SLN) crystals was the double crucible Czochralski (DCCz) method (Oxide Inc., Japan). However, because the usage of the powder feeding requires a complicate design, this is not convenient for a conventional Czochralski puller. In this study, the zone-leveling Czochalski method (ZLCz) was developed, and could be convenient for the conventional Czochralski puller. The ZLCz method was then the first time using on the growth of SLN crystals with uniform MgO or ZnO doping. The 1”-diameter MgO or ZnO-doped SLN crystals was grown to investigate the characteristics of the ZLCz method and the segregation behavior of the doping. The ZLCz process was then further improved to grow near 2”-diameter high quality SLN crystals. There are some problems in the ZLCz technique, such as the bubble incorporation, volume change of the solution zone, and the cracking problem of the grown crystal. The bubble formation from the solid feed was found inevitable for the ZLCz method with continuous fed from below of the solution zone. The usage of the inner crucible could be used to avoid the bubble incorporation. The volume change of the solution zone during growth was found to be another serious problem which can cause the axial non-uniformity of the grown crystal. A proper hot zone design could crucially reduce the change of the heat lose condition, and thus the zone change during crystal growth. In addition, with a suitable thermal configuration, the temperature gradient could be controlled under 20℃/cm, and this was found necessary to avoid crystal cracking. 1 mol% MgO-doped SLN crystal pulled from a Li-rich (59-60 mol% Li2O) solution with SLN feeding was found to be 0.975 in Li/Nb ratio within 1% in deviation. Moreover, the uniformity of the as-grown crystal could be enhanced if we pulled the crystal from a K2O-added (16 mol%) solution zone, or fed by using Li-excess (51 mol% Li2O) solid feed. The deviation of the Li/Nb ratio was reduced to less than 0.5%. However, with a Li-excess solid feed, the grown crystal was found with a higher Li/Nb ratio being about 0.98. With the increase of the doping concentration, the Li/Nb ratio of the SLN crystal decreased. And the theoretical limit of the Li/Nb ratio in the 1 mol% MgO-doped SLN crystal is also about 0.98. The stoichiometry of the as-grown MgO-doped crystals was close to the theoretical limit value. In addition, the MgO concentration along the grown crystal was varied less than 5%. And the compositional change in radial direction was found within only 0.1% deviations from the absorption edge measurement. On the other hand, the zinc-doped SLN crystal with a larger doping variation was found due to the segregation of ZnO. The segregation coefficient was about 0.58. With a higher pre-doped concentration in the solution zone, the doping variation in grown crystal could be significantly reduced to less than 5%. In addition, for ZnO-doped SLN, the properties were varied with the doping concentration, and showed a transition with the zinc concentration at about 1 mol%. Based on the defect model, at a low doping level, the doping was mainly replaced the anti-site defect. At a higher doping level, the Zn ions replaced the normal Nb and Li sites. Thus, the anti-site defects of the grown crystals were about at 1 mol%. The ZLCz method was successfully developed for MgO or ZnO doped SLN growth. The crystal could be used for the preparation of the periodically poled SLN (PPSLN) chip. The properties and the homogeneity of the grown crystals show good results in this study.中文摘要 I 英文摘要 III 目錄 VI 圖目錄 IX 表目錄 XIV 第一章 緒論 1 1.1 簡介 1 1.2 文獻回顧 4 1.2.1 鈮酸鋰晶體 4 1.2.2 SLN晶體之生長 14 1.2.3 區熔提拉法 20 1.3 研究動機 25 第二章 實驗藥品、設備及流程 26 2.1 實驗藥品 26 2.1.1 晶體生長用藥品 26 2.1.2 鈮酸鋰溶液配製用藥品 28 2.1.3 晶片研磨用藥品 29 2.1.4 清洗用藥品 30 2.2 晶體生長實驗之設備及流程 31 2.2.1 原料配製 31 2.2.1.1 原料配製設備 31 2.2.1.2 原料配製步驟 32 2.2.2 原料填裝、晶體生長及晶體回火 35 2.2.2.1 晶體生長設備 35 2.2.2.2 原料填裝步驟 46 2.2.2.3 晶體生長及晶體回火步驟 47 2.2.3 晶體定向、切割及研磨 49 2.2.3.1 晶體定向、切割及研磨拋光設備 49 2.2.3.2 晶體定向步驟 52 2.2.3.3 晶體切割步驟 53 2.2.3.4 晶體研磨拋光(lapping & polishing)步驟 53 2.2.4 晶體評價 55 2.2.4.1 量測設備 55 2.2.4.2 晶片穿透度、吸收波長量測步驟 59 2.2.4.3 晶體折射率量測步驟 59 2.2.4.4 晶體居禮溫度量測步驟 60 2.2.4.5 晶體晶格常數量測步驟 60 2.2.4.6 晶體組成量測步驟 61 2.2.4.7 矯頑電場量測步驟 61 2.3 鈮酸鋰晶片進行VTE實驗之流程 63 第三章 ZLCz晶體生長及晶體評價 65 3.1 ZLCz晶體生長之特性及熱場分析 65 3.1.1 ZLCz晶體生長特性 66 3.1.2 晶體生長系統之熱場分析 71 3.1.2.1 基本熱場 71 3.1.2.2 白金遮罩(Pt cover) 72 3.1.2.3 內坩堝遮蔽效應 74 3.1.2.4 石英遮罩 75 3.1.2.5 白金桶 75 3.2 SLN晶體之生長 77 3.2.1 鎂摻SLN晶體 77 3.2.2 鋅摻SLN晶體 80 3.3 晶體評價 84 3.3.1 鎂摻SLN晶體 85 3.3.1.1 晶體穿透度量測 85 3.3.1.2 晶體吸收波長量測 86 3.3.1.3 晶體居禮溫度量測 88 3.3.1.4 晶體折射率量測 89 3.3.1.5 晶體組成量測 91 3.3.1.6 晶體矯頑電場量測 93 3.3.1.7 晶體徑向均勻度分析 95 3.3.1.8 粉末X-ray繞射圖譜 97 3.3.2 鋅摻SLN晶體 99 3.3.2.1 晶體穿透度量測 99 3.3.2.2 晶體組成量測 99 3.3.2.3 晶體吸收波長及居禮溫度量測 102 3.3.2.4 晶體折射率及矯頑電場量測 106 第四章 大尺寸SLN生長與製程改良 108 4.1 晶體生長系統之改良 108 4.1.1 熱場設計 108 4.1.2 大尺寸SLN晶體之生長 111 4.1.3 晶體均勻性及組成分析 113 4.2 晶體生長組成之調整 118 4.2.1 Li2O過量進料組成之溶液生長(solution growth) 118 4.2.2 K2O助溶劑生長(flux growth) 123 4.3 結果討論 129 第五章 結論 133 參考文獻 13