Si基SiC薄膜和低维SiO₂的生长及其光致发光机理研究

【作者】 陈征；

【导师】 王玉霞；

【作者基本信息】 中国科学技术大学 ， 凝聚态物理， 2007， 博士

【摘要】 SiC半导体材料是自第一代半导体材料（Si、Ge）和第二代半导体材料（GaAs、InP、GaP、InAs和AlAs等）之后发展起来的第三代宽带隙半导体。由于制备SiC体单晶非常困难而且价格昂贵，因此SiC薄膜的异质外延生长是重要的。目前，人们研究较多的是在Si衬底上异质外延生长SiC薄膜。但由于Si和SiC的晶格常数及热膨胀系数的不匹配，在SiC／Si界面处易形成空洞缺陷。这些空洞缺陷使薄膜的电学及光学性质变差，严重影响SiC／Si器件性能及其集成的发展。所以，有必要寻求生长无界面空洞缺陷的SiC／Si薄膜的工艺和方法。此外，SiC的自然氧化物和Si一样都是SiO₂。SiO₂不仅被应用于光学器件，还被广泛的应用于Si基、SiC基的微电子器件。因此，它的光学活性缺陷对于这些器件是十分重要的。本论文的内容可分为两大部分：一是探索了无界面空洞缺陷的SiC／Si薄膜的生长，并研究了其界面微结构与物性等；二是研究了低维SiO₂材料的生长及其光致发光的机理。主要内容有：●PS／Si叠层热解反应法生长SiC薄膜的制备、结构和物性：我们在1 atm的Ar气氛下用PS／Si叠层热解反应法生长了SiC薄膜，优化了工艺条件，给出了SiC薄膜的生长过程。在1250℃时，优化生长条件，获得了2.51 （?）密排面取向的晶态SiC薄膜。薄膜表面平整，无明显的界面空洞缺陷。我们用空位聚集理论解释了空洞缺陷形成的机理，以及抑制空洞缺陷的原因。利用Mie的散射理论能很好的解释SiC薄膜样品的TO～LO之间宽的红外吸收，这与其他研究者给出解释不同。结合薄膜的生长过程，这样的红外吸收说明少量的C扩散进入了Si衬底的深处并形成了分散的SiC小颗粒。我们用红外光谱的方法研究了SiC薄膜的生长动力学问题，得出在1250℃以下的温度薄膜生长是2D模式，高于1250℃温度时薄膜生长是3D模式，在1270℃以上，SiC的挥发明显，并计算了不同温度下生长的SiC薄膜的厚度。对处理后样品的XPS分析得出在薄膜表面上的SiO₂层是在SiC生长末期的降温过程中形成的，样品实际结构为SiO₂／SiC／Si。此外，我们还表征了薄膜的电学性质，优化生长的SiC薄膜呈现出了较好的Ⅰ-Ⅴ特性。对SiC薄膜样品的光致发光进行了详细讨论，首次发现了SiO₂／SiC的界面的NITs的发光。NITs是在SiO₂／SiC界面处存在的一个重要的势阱缺陷，它严重的影响了SiC基MOS结构器件的电学性质。目前NITs的构型还是未知的，而关于NITs的信息还多来源于电学测量，这是首次用光学方法探测出NITs缺陷，因此我们的这一发现对光学探测NITs、研究NITs和制作SiC基MOS器件有非常重要的意义。●用PS／OCS／Si叠层热解反应法制备SiC薄膜。在真空(10^-3 Pa)条件下最佳生长温度是1050℃，获得单晶的6H-SiC；在5×10⁴ Pa的Ar气氛下最佳生长温度是1300℃，主要获得单晶的4H-SiC薄膜。探讨了不同条件下生长的SiC薄膜晶型不同的原因。SEM表面和界面分析表明，两种条件下都可以生长无空洞缺陷的SiC薄膜，这是由于致密的SiO₂层阻止了Si原子的外扩散，同时本身又参加反应提供Si形成SiC。●我们制备了非晶SiO₂线并研究了它的生长机制和强蓝光发射的发光机制。在非晶SiO₂线中观察到了一个峰值约为2.84 eV的宽蓝色发光峰，并伴有3.0 eV的肩峰。这个宽发光峰从2.84到3.0 eV的区间发光强度变化不大。大多数文献认为这个发光是来自SiO₂体系中熟知的ODC（Ⅱ）。但是这样的结论缺乏发光峰的本征性质的支持。针对这个问题，我们研究了这个蓝光发射的衰减、激发峰、发光峰形等本征性质，结果表明这个发光峰不是来源于SiO₂体系中的ODC（Ⅱ），也没发现有已知的SiO₂点缺陷的发光能匹配它。对这个蓝光发射的发光性质进行了更深入的分析后，我们认为这个宽蓝色发光峰是由相关联的2.8 eV和3.0 eV发光峰因为“非晶宽化效应”加宽后重叠形成的；260 nm激发下的衰减谱符合双曲线衰减规律，说明这个发光可能与SiO₂的导带或价带有关，拟合的结果还表明，电子可能是从势阱中激发出来的，并且移动范围比较大。通过对比发射光谱和激发光谱，我们发现SiO₂线的这个蓝色发光与PS／Si叠层热解反应法生长的SiC薄膜的蓝光发射是同一类型发光，即这个蓝光发射来自于SiO₂／Si界面处的NITs。更多还原

【Abstract】 SiC is one of the third-generation semiconductor materials, which were developed later than the first （Si and Ge） and the second generation semiconductor materials （GaAs, InP, GaP, InAs, AlAs etc）. The heteroepitaxy of the SiC film is important since high quality SiC wafer is expensive and hardly to be achieved. To this day, the heteroepitaxy of SiC is mostly on Si substrate. However, cavities are prone to formatting at the Si/SiC interface due to the large lattice mismatch and thermal expansion coefficient difference between Si and SiC. These cavities degrade film’s electric/optical properties and inhibits applications for devices. Oxide of SiC is SiO₂ as that of Si. The optically active defects in SiO₂ influence not only the optic devices but also the Si- and SiC-based micro-electronic devices. Photoluminescence （PL） is an approach to investigate these optically active defects.This thesis covers mainly two areas: one is the works for suppression of interfacial cavity; the other is PL investigation of optically active defects in low-dimension SiO₂.Preparaton, structure and optic/electric properties of SiC film by PS/Si pyrolysis:SiC films are grown on Si substrates by PS/Si pyrolysis in 1 atm ambient Ar, and the growth conditions are optimized. The film growth process is proposed as well. The planar, compact and oriented SiC film was grown at 1250°C, optimized temperature, in quartz crucible. No cavity was observed in the SiC film. The mechanism of cavity formation and suppression was discussed based on the theory of vacancy aggregating. A broad IR absorption band between TO and LO phonon frequency of SiC were explained by Mie’s scattering theory. Based on the film growth process, the IR absorption indicated that some carbon atoms diffused into Si substrate and formed dispersed small SiC particles. The thicknesses of SiC films grown at different temperatures were estimated from the IR absorption of these films. It was suggested that the film growth is in 2D mode below 1250°C, in 3D mode above 1250°C. The volatilization of Si and C atom is obvious above 1270°C. A silica layer formatted on the SiC film at the late stages of SiC growth. The sample is actually SiO₂/SiC/Si. Moreover, the SiC film grown at optimized condition shows goodish I-V character. The mechanism of the blue PL from the SiC film was also discussed. The PL was attributed to NITs （near-interfacial traps） at the SiO₂/SiC interface. As far as know, it is first time to found the NITs’ PL. The NITs are important traps at SiO₂/SiC interface, with a still unknown nature. NITs is usually observed at SiO₂/SiC and SiO₂/Si interface by electric measurement. The trapping of electrons by NITs results in a degradation of the channel mobility, particularly at the SiO₂/4H-SiC interface. Our conclusion will help to detect NITs by optical methods, the study of NITs and is important to SiC-based MOS devices. High quality SiC film without interfacial cavity were prepared in vacuum (10^-3 Pa) and low-pressure （5×10⁴ Pa） ambient Ar by PS/OCS/Si pyrolysis. The optimized growth temperature is 1050°C and the product is 6H-SiC in vacuum condition, while the optimized growth temperature is 1300°C and the main product is 4H-SiC in low-pressure ambient Ar. The reason of growing different SiC polytypes is discussed. It was suggested that suppression of cavity was due to SiO₂ layer, which prevent Si atom diffusion and provide Si atom for SiC formation.Investigation of the growth and the PL of the amorphous silica wires.An intense broad blue PL band was observed from the silica wires. The blue PL band is made up of two relative PL peaks at 2.8eV and 3.0eV. Both the PL peaks are broadened by " inhomogeneous broadening effects" due to the glassy state. So the PL band intensity does not almost change from 2.84 to 3.0eV photon energy. The blue PL had been reported in some literatures and were usually attributed to ODC （II） centers of SiO₂. However, this attribution is unreliable for lack of the PL’s intrinsic properties. We investigated the mechanism and origin of this PL by its intrinsic properties, such as PL decay, PL excitation etc. The results negated that the blue PL is due to ODC （II）, and the PL can not be due to any familiar point defect in SiO₂. The PL decay excited with 260 nm light can be fitted well using hyperbola. Hyperbolic PL decay suggests the PL relates to the conduction band or the valence band of silica. The fitted parameter implies that the electrons are excited from traps and have large displacement. Based on the location and shape of emission and excitation spectra, it was concluded that the blue PL of SiO₂ wires had the same origin as that of SiC film grown by pyrolysis of PS/Si sol-gel coatings. The blue PL of SiO₂ wires is from NITs at the SiO₂/Si interface.更多还原