【作者】 郑江海；

【导师】 康俊勇；

【作者基本信息】 厦门大学 ， 微电子学与固体电子学， 2006， 硕士

【摘要】 InGaN半导体的带隙基本上覆盖了整个可见光波段,还包含了部分红外波段,在光电器件和光存储器件方面都有着广泛的应用。目前,影响其未来发展的几个难题,包括晶体质量、发光机制和相分离现象,深受大家的关注并需要迫切解决。本论文通过计算模拟和实验研究,并结合前人的研究,全面深入地考察了InGaN中相分离的有关问题,分析其性质、阐明其物理机制,进而讨论其抑制方法及其对InGaN发光的影响等。首先从热力学基础出发,着重分析了相分离与自由能变化量的关系,得出自由能变化量是否大于0,由混合焓和熵共同确定;当?2G / ?x2 < 0时,亚稳态分解自发产生,从而导致了相分离的发生,反之,?2G / ?x2 < 0时则为成核和生长过程,要求大的组分涨落。接着,简述了InGaN半导体相分离的形式主要有三种,即团聚、有序和分离成两个或多个不同组分InGaN相,实验上主要以X射线衍射、透射电镜、选区电子衍射、阴极荧光、拉曼谱等技术检测相分离是否发生以及存在的形式。只要InN成分较高和外延层较厚,相分离就会产生于InGaN外延层,其抑制方法是受到衬底应力作用的外延层必须薄于临界厚度,在此基础上重点讨论影响临界厚度的主要参数;相反地,一定组分InGaN层薄于临界厚度时,诱发相分离的产生以形成富InN量子点且不产生晶格缺陷的方法主要有在粗糙GaN表面上生长InGaN等。在综合前人研究的基础上详细论述了相分离对InGaN外延层光学性质的影响,不仅表现在它导致了单色发光效率的下降和不利于器件的性能控制,也表现在因它将形成富In量子点而提高了LED的发光强度,但同时导致了出射光的红移。这可能就是为什么高缺陷的InGaN材料在室温下能够高效率发光的原因。在模拟计算方面,通过第一性原理和正规溶液模型计算了InGaN的自由能和混溶隙,结果发现,当晶格均匀分布且无受应力时,绝对零度下的InxGa1-xN趋于分离成GaN和InN两相,临界温度超过1825K;若引入GaN/InGaN或InN/InGaN界面失配应力时,自由能变化量变成负的,系统将分离成其它相,混溶区的范围向富InN方向移动,系统的临界温度分下降。晶格微结构计算表明,在没有应力的作用下,系统确实趋向于形成InN团簇,而在GaN/InGaN失配应力作用下In原子分布从团簇到两相分离的变化,另一方面,更多还原

【Abstract】 InGaN has been widely used in the fabrication of optoelectronic devices due to its wide bandgap ranges (0.7– 3.4 eV). The stringent requirement for the growth of high-quality InGaN films, which has become a major obstacle for the further developments of high-performance devices, and studies on mechanisms of light emission and phase separation (PS) have attracted intense interests. The thesis has focused on the PS-related problems of InGaN, i.e., properties of InGaN, PS mechanisms, effects of PS on light emission properties, and suppression of PS. The polarity of the Mixing Free Energies (MFEs) is determined by the mixing enthalpy and entropy. For ?2G/?x2<0, the spinodal decomposition (SD) will automatically occur, which results in PS; Inversely for ?2G/?x2<0, nucleation and growth will occur when there is a large composition fluctuation. The PS can occur in several ways including clustering, ordering and several phases with different InN mole fraction. Experimental techniques such as X-ray Diffraction (XRD), Transmission Electron Microscopy (TEM), Selected Area Diffraction (SAD), Cathodoluminescence, and Raman have been applied to determine whether PS has occurred and the way in existence. Generally, high InN mole fraction and thick InGaN film will result in PS, which can be suppressed when the thickness of InGaN film under the substrate’s strain is smaller than the critical layer thickness (CLT). On the other hand, values for CLT of InxGa1-xN within GaN/InxGa1-xN can be experimentally determined using electrical data (mobility and conductivity) and optical emission energy of the InxGa1-xN films, and can be estimated as a function of x using both the PB model and Fischer model, which both were as a function of the lattice mismatch and the film structural properties. Self-assembled In-rich quantum dots (QDs) by enhanced PS in the InGaN layer less than CLT, can be grown on a GaN layer with a rough surface. If PS happens, though it reduces the efficiency of unique photoluminescence (PL) and has bad influence on the physical properties and controls of the epi-films, the phase-separated QDs are responsible for the enhanced更多还原