氮化鎵V型缺陷成長機制與新氮化鎵磊晶層轉移技術

Abstract

藉由大尺寸V型缺陷的發現，本研究得以利用新證據透過不同觀點探討氮化鎵V型缺陷的生成原因並解釋大尺寸V型缺陷的出現機制。由總體自由能的評估可知，雖然大尺寸V型缺陷與點狀開口之氮化鎵側向覆蓋生長晶體具有相同的晶面組成與互補的晶體外型，但兩者在熱力學與動力學上的意義並不相同，大尺寸V型缺陷於熱力學上是處於一種不穩定的能量狀態，因此大尺寸V型缺陷的出現應是來自於磊晶層成長時動力學影響所造成的結果。 本研究把V型缺陷的出現分成”成核”與”生長”兩個階段。”成核”過程與氮化鎵磊晶層表面特定區域極性的改變有關，而”生長”過程則取決於氮化鎵不同晶面相對生長速度不同的影響。磊晶層表面極性改變來自於生長時磊晶缺陷的產生，極性的改變造成了該區域生長速度與周遭環境不同並”成核”了V型缺陷。在V型缺陷”成核”後，其生長會因磊晶成長時環境的不同造成此缺陷消滅或進一步變大，而消滅或成長的因素主要控制於氮化鎵 面與(0001)面（或 面）生長競爭的結果。透過前人的研究結果分析得知氮化鎵這兩個晶面生長之相對速度與生長時的溫度、壓力、V/III比、參雜等因素有關，而這些因子背後的控制因素據本文推測是鎵反應源於生長基準面（(0001)面）上是否具有足夠的能力或機會移動到能量上比較穩定的 面上並與氮反應源產生鍵結。利用V型缺陷”成核生長”模型並結合”晶面的生長速度差異”觀念來解釋V型缺陷的生成，不僅可以完全符合前人的相關研究結果，也可以合理解釋大尺寸V型缺陷的出現原因。 另外，將氮化鎵磊晶層與他種材料結合是實現光電整合積體電路與拓展氮化鎵材料應用範圍所需具備的關鍵技術。本論文提出一種新方法，透過氮化鎵異質側向生長技術、晶圓接合技術、與選擇性濕式蝕刻以達成高品質氮化鎵磊晶層與他種材料結合的目的。此方式經初步評估具有成本低廉、磊晶基板可回收、磊晶層品質良好、並適合於大量製造生產等潛在優點，值得進行更詳細的研究。

This thesis investigates and discusses the formation mechanism of V-defect on GaN epilayer based on the appearance of large V-defect. Although facets and crystal shape of large V-defect and dot patternd GaN ELOG are similar, they are not the same from the thermodynamic and kinetic point of view based on calculated total free energy. V-defect is a result of an unstable energy of state so the appearance of V-defect is mostly likely resulted from the kinetic control during epitaxy process. The discussion of the formation of V-defect is divided into two parts: nucleation and growth. Nucleation step is related to the change of polarity of certain area on GaN epitaxy film; growth step is influenced by the different growth rate of different facets. Change in polarity on epitaxy film is due to the defect that is generated during the deposition of epitaxy film. The polarity of a certain area changes the growth rate of that specific area and thus nucleates a V-defect. After nucleation, V-defect turns into growth up or elimination depend on comparing of (0001) and facets of GaN which affected by different growth conditions. By previous research, the facets growth rate may change by varying temperature, pressure, V/III ratio, and dopant. We suppose upper factors are associate with Ga diffuse and bonding abilities. Better Ga diffuse ability and lower bonding probability may cause facet growth faster. The “nucleation and grwoth” model can reasonable explain both previous experience evidences without any exception and the formation of large V-defect. This thesis also introduces a novel GaN epilayer transfer method. This method combines ELOG GaN, wafer bonding technology, and selective wet etching to acheieve the purpose of integration a high quality GaN film on a selective substrate. By preliminary evaluation, the novel method has many advantages and worth to advanced research.