【作者】 韩冬；

【导师】 张绳百； 孙洪波； 李贤斌；

【作者基本信息】 吉林大学 ， 微电子学与固体电子学， 2014， 博士

【摘要】 半导体的应用极大的促进了微电子工业的发展，改变了人类的生活方式。化合物半导体的种类极为丰富，对应的半导体参数性质也各不相同，这极大的满足各类应用领域对半导体性质提出的特定需求。化合物半导体的多样性也给对其系统研究带来了困难，不同的化合物半导体材料在应用过程中会涉及到不同的核心物理问题。二氧化硅（SiO2）、铜锌锡硫（Cu2ZnSnS4）、氮化硼（BN）和砷化铟（InAs）这四种半导体分别涉及到宽禁带半导体的非对称掺杂、影响器件性能的关键缺陷、材料生长中缺陷结构、表面电荷输运等半导体物理中基态性质和激发态动力学的关键问题，对这些半导体物理问题的理解研究可以拓展这些半导体材料在相关领域的应用前景。本文通过基于密度泛函理论的第一性原理计算和含时密度泛函理论的激发态动力学分别对它们进行了系统的研究，得到了以下四个方面的创新性结果，1.以二氧化硅（SiO2）为例研究宽禁带半导体的非对称掺杂问题。为了适应微电子器件在极端条件下和高性能领域的应用，人们对半导体材料提出了宽禁带、耐高温、高迁移率等新的要求。但是宽禁带半导体在掺杂过程中会遇到非对称掺杂问题。我们通过系统的研究SiO2的替位掺杂特性发现，并没有发现施主能级和受主能级位置的非对称掺杂性质，仅在替位掺杂缺陷的形成能中有稍许的非对称性。通过GGA+U方法，能够正确实现对SiO2中替位杂质缺陷局域结构的正确描述，符合电子顺磁共振实验的结果。SiO2可以实现浅（相对于SiO2的带隙值）的施主过渡能级和受主过渡能级。计算表明最好的p型和n型替位掺杂缺陷分别为AlSi（价带顶以上0.86eV）和PS（i导带底以下0.74eV），计算得到的过渡能级和实验测量吻合。这些发现坚实的证明了宽禁带半导体材料不一定会引入非对称掺杂问题。2.铜锌锡硫（CZTS）半导体中本征缺陷问题的研究。CZTS是铜铟硒（CIS）半导体在太阳能电池领域的完美替代品，组成CZTS的四种元素都是地壳中储量丰富、廉价、而且对环境无害的元素，它能够克服生产CIS太阳能电池遇到的原材料供应问题。尤其是近几年在光电转换效率的巨大提高，加速了CZTS的应用。对于四元化合物半导体来说，可能存在的本征缺陷类型很多，直接在实验中辨别对半导体性质有明显影响的缺陷是非常困难的。而且对CZTS的缺陷理论研究还不够充分，在涉及CZTS的导电类型和影响器件性能的关键缺陷仍然存在争论。我们通过杂化泛函的计算揭示了影响CZTS性能的关键本征缺陷。缺陷团簇(Cu3)Sn和反位缺陷SnZn被证明是CZTS中主要的少数载流子束缚缺陷。计算结果从微观上解释了CZTS的最优生长条件，即Cu缺乏和Zn富集的最优生长条件是为了抑制缺陷团簇(Cu3)Sn和反位缺陷SnZn的产生。在最优生长条件下，VCu是CZTS中提供空穴载流子最主要的缺陷，尽管VCu的缺陷浓度远远小于另外一个受主缺陷CuZn。除此以外，计算中没有发现负形成能的缺陷存在，表明CZTS是一种热力学稳定的材料。3．氢（H）在氮化硼（BN）生长过程中缺陷行为的研究。BN包括两种常见构型，六方BN（hBN）和立方BN（cBN），它们拥有良好的润滑性能和机械性能，还具有耐高温、耐腐蚀、绝缘性能好等特点。基于BN半导体广阔且重要的应用前景，全世界范围内的众多科研人员对其表现出了浓厚的研究兴趣。BN最大的问题来自于单晶的合成问题，BN在生长过程中存在大量H，直接影响了其合成质量，但是H在BN生长过程中的缺陷行为却鲜有讨论，是值得进一步探索的关键问题。我们通过研究H在BN晶体中的热力学行为和振动性质发现，H更容易以H2分子或H2**缺陷团簇形式存在于hBN中而不是cBN中。令人意外的是，这些H2分子或H2**缺陷团簇会自发形成团簇集合。在H存在的条件下，cBN的形成受到抑制。因此，Al在hBN中引入的sp3晶核会迅速的被钝化掉。解释了为什么Al的掺杂不会明显的促进hBN向cBN的转变的原因。4.砷化铟（InAs）表面量子photo-Dember效应的研究。飞秒激光在InAs半导体表面进行激发后可以产生THz辐射，这种现象可以用photo-Dember效应来解释，但对photo-Dember效应的理解还局限在半经典理论框架下。这种解释在量子力学角度框架下存在其固有的局限，比如表面载流子的动力学弛豫过程和降温过程都是有非绝热过程所主导，表面态载流子的弛豫过程会涉及到强烈的载流子-离子之间的相互作用。因此，从量子力学角度框架下重新理解在半导体表面产生THz辐射的电荷输运过程是十分必要的。我们通过基于含时密度泛函理论的激发态动力学揭示了半导体表面量子photo-Dember效应的存在。与当前理解相反的是，半导体表面有着无限大质量的载流子可以以高达106cm/s的速度向半导体内部传播。这种传播是由电子-声子耦合产生的。这种耦合不仅发生在表面态上，我们认为量子photo-Dember效应可以存在表面激发的任何态上，即使是在半经典部分占主导的区间。电子-声子耦合通常会以散射或者形成电荷偶极的方式降低载流子的扩散速度。我们的结果给出了一个相反的情况，即耦合可以促进载流子的扩散。更多还原

【Abstract】 The application of semiconductor promotes the development and innovation ofmicroelectronics, and changed the human life. The species of compound semiconductorsis extremely rich. The semiconductor property varies from compound to compound,which greatly meet the different kinds of applications in the semiconductor field.However, the diversity of compound semiconductors makes them difficult to investigate.Each compound semiconductor has its own core physical problem in the using process.Silicon dioxide (SiO2), Copper zinc tin sulfide (Cu2ZnSnS4), Born nitride (BN) andIndium arsenide (InAs), these four semiconductors relate to four key semiconductorissues of ground state properties and excited state dynamics, including dopingasymmetry in the wide bandgap semiconductors, critical defects in semiconductorswhich affect the device’s performance, defects’ behavior in semiconductors’ growth andsurface charge transport. Understand these physical issues could expand thesemiconductors’ application prospects in the related field. In this dissertation, using thefirst-principles calculation based on the density functional theory (DFT) and excited statedynamics based on the time-dependent density functional theory (TDDFT), we havesystematically studied these four compound semiconductors, and got four innovativeresults as following,1. Investigate the doping asymmetry in the wide bandgap semiconductor by using SiO2. In order to accommodate the application of microelectronic devices with the highperformance under extreme condition, people require the semiconductors with the widebandgap, high temperature and high mobility. However, dope in wide bandgapsemiconductor would meet the issue of doping asymmetry. Our systematic investigationof the doping property of SiO2reveals no obvious doping asymmetry in the donor andacceptor level position, and only modest asymmetry in the impurities’ formation energy.With an advanced approach of GGA+U, we achieve an accurate description on the localconfiguration of substituting impurities in SiO2, which is quite consistent with EPRspectroscopy. The SiO2can have both relatively shallow acceptor transition level anddonor transition level. Our predicted best candidates for p-type and n-type doping areAlSi(0.86eV above the VBM) and PSi(0.74eV below the CBM), respectively. Thefindings strongly proved the wide band gap semiconductor material may not introducethe issue of doping asymmetry.2. Investigate the native defects in the CZTS semiconductor. The CZTSsemiconductor is the perfect substitute of CuInSe2(CIS) in the application of solar cell.all the four constituent elements in CZTS are earth-abundant, cheap and nontoxic. CZTShas the potential to overcome the limitations of CIS in production capacity. In recentyears, the solar-energy conversion efficiency of CZTS has a huge raise, which acceleratesits application. For the quaternary compound semiconductor, it has a great number ofintrinsic defects. Therefore, in the experiment, it is very different to directly identify thenative defects, which significantly affect the properties. However, the correspondingdefect physics is not sufficient, the critical defects which affect the conductivity type anddevice performance is still under debate. Hybrid functional calculation reveals thephysical identity of native defects in CZTS. The SnZnantisite and the defect complex(Cu3)Snare found to be the main deep minority traps. Our results provide a microscopicunderstanding on the optimal growth condition of CZTS, e.g. the optimal Cu-poor andZn-rich growth condition suppress the formation of SnZnantisite and the defect complex(Cu3)Sn. Under the optimal growth condition, VCucould contribute the majority of thehole carriers, even though its concentration is much lower than another acceptor CuZn. In addition, no defects in our calculation show negative formation energy, indicating thatCZTS is a thermodynamically stable material.3． Investigate hydrogen behavior in the growth of BN semiconductor. BNincludes two common crystalline forms: hexagonal BN (hBN) and cubic BN (cBN).They have not only the excellent lubricating and mechanical properties, but also hightemperature, corrosion resistance and high insulation. Based on the broad and importantapplication prospect, lots of researchers over the world show a strong interest onstudying BN. The biggest problem in using BN is to synthesize the BN with single phase.A large number of hydrogens (H) exist during the BN growth and directly affect thequality of synthesizing BN material. But the H’s behavior during the BN growth is rarelydiscussed, which is worthy of further studying. We have systemically studied thethermodynamic and vibration properties of H in BN crystal. Essentially, H prefers toreside in hBN with forms of H2and H2**rather than in cBN. Unexpectedly, these H2andH2**can automatically gather to form clusters. In present of H, cBN phase tends to besuppressed. Therefore, Al-induced sp3nucleus in hBN is readily passivated. Thatexplains the puzzle why Al has no significant improvement to grow cBN from hBN.4. The study of quantum mechanical photo-Dember effect on InAs surface. TheTHz radiation will generate on the semiconductor by the femtosecond laser absorption.This phenomenon can be explained by the photo-Dember effect, but the photo-Dembereffect descripted by the semi-classical theory, has its innate limitations. For example, thedynamical relaxation and cool-down of the excited carriers in these states are governedby non-adiabatic processes involving strongly coupled carrier and ion motions. Therefore,it is very important to understand the surface charge transport, generating THz radiationfrom the perspective of quantum mechanics. First-principles MD simulation of iondynamics, coupled with TDDFT for electron dynamics, reveals the possible existence ofa non-classical pDe at semiconductor surface. Contrary to current understanding, surfaceexcitations with infinite mass can propagate into bulk with a speed as high as106cm/s.The propagation is caused by electron-phonon coupling. Because such a coupling is not reserved only to SSs, one can expect the quantum mechanical pDe to apply to any stateexcited at the surface, even when the semi-classical component dominates.Electron-phonon coupling usually slows down carrier diffusion via scattering and/orpolaron formation. Our findings thus represent a counter example where the couplingmay enable carrier diffusion.更多还原