氮极性 GaN 的几何控制微尺度图形化与外延横向过生长
Geometrically-Controlled Microscale Patterning and Epitaxial Lateral Overgrowth of Nitrogen-Polar GaN
AI总结：
提出一种结合 KOH 各向异性湿法刻蚀几何图形化与保向外延再生长的两步法，实现氮极性 GaN 平滑合并并降低穿位错密度。
摘要：
在本研究中，我们报道了一种新型两步外延生长技术，该技术能够显著改善氮极性 GaN 的晶体质量。起始材料通过金属有机气相外延在 4° 邻晶面蓝宝石衬底上生长，并采用初始高温蓝宝石氮化处理以控制极性。随后，通过在 KOH 溶液中进行湿法刻蚀，将材料转化为规则排列的六方锥阵列，接着进行再生长，使这些锥合并并恢复为晶体质量得到改善的光滑层。使该技术得以实现的关键在于利用湿法刻蚀步骤的各向异性行为来控制阵列几何形状，以及采用能够保持锥侧壁取向的再生长条件。相比之下，若生长条件导致锥顶部残余 (000-1) 晶面过度扩展，则在完全合并时会引发非常粗糙的表面形貌。X 射线衍射研究证实，随着再生长步骤的推进，穿位错密度降低。对 000-2 GaN 峰相对于 0006 蓝宝石峰的相对位置分析揭示了锥的宏观倾斜，这可能由较大的离轴衬底取向诱导。该倾斜与再生长步骤开始时 000-2 衍射峰的异常展宽具有很好的相关性。
英文摘要：
In this study we report on a novel two-step epitaxial growth technique that enables a significant improvement of the crystal quality of nitrogen-polar GaN. The starting material is grown on 4° vicinal sapphire substrates by metal organic vapour phase epitaxy, with an initial high-temperature sapphire nitridation to control polarity. The material is then converted into a regular array of hexagonal pyramids by wet-etch in a KOH solution, and subsequently regrown to coalesce the pyramids back into a smooth layer of improved crystal quality. The key points that enable this technique are the control of the array geometry, obtained by exploiting the anisotropic behaviour of the wet-etch step, and the use of regrowth conditions that preserve the orientation of the pyramids' sidewalls. In contrast, growth conditions that cause an excessive expansion of the residual (000-1) facets on the pyramids' tops cause the onset of a very rough surface morphology upon full coalescence. An X-ray diffraction study confirms the reduction of the threading dislocation density as the regrowth step develops. The analysis of the relative position of the 000-2 GaN peak with respect to the 0006 sapphire peak reveals a macroscopic tilt of the pyramids, probably induced by the large off-axis substrate orientation. This tilt correlates very well with an anomalous broadening of the 000-2 diffraction peaks at the beginning of the regrowth step.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
