Career: Silicon Etching: Gas-Surface Dynamics and Profile Evolution

CTS-9623450这个项目讨论了非弹性散射和反应散射在半导体刻蚀中的重要性。本文提出了一种新发明的具有3-25 eV平动能的高超声速氟原子中性束的实验，用于硅稳态刻蚀过程中气体表面动力学的基础研究和硅中亚微米特征的刻蚀研究，以了解中性束是如何刻蚀的。这一努力是基于最近剖面演化模型和平动能量在3到18 eV之间的氟原子束实验的结果。用氟化学在室温下对硅进行各向异性刻蚀（无侧壁沉积），证明了这种光束的能力。这是可能的，因为直接反应发生在高平动能所使用的。碰撞诱导解吸（CID）已被认为是导致氟化硅部分在较高能量下从SiFx层解吸速率增加的机制。在CID模式下快速去除SiF3产物可以使暴露的表面快速氟化，从而增加氟的利用率，从而减少热解吸未反应的氟原子通量。减少分散到侧壁的氟显著改善了各向异性。在蚀刻过程中，对未反应的F原子和反应产物（SixFy）离开SiFx层的能量和通量进行了基本测量。此外，在二氧化硅表面进行了类似的测量。二氧化硅在今天的晶圆上无处不在（例如，作为硬掩膜，栅极氧化物等），其表面的散射也会影响轮廓的演变。除了提高对气体表面动力学的基本理解外，这些测量还提供了相互作用的物理和化学描述，这些描述被纳入到剖面演化的蒙特卡罗模拟器中。基于SiFx上详细散射实验的初步计算显示，可以捕捉到剖面的特性，例如所谓的“微沟”，即在蚀刻特征的底部出现凹槽。该模型得到了改进，可以描述与模式相关的蚀刻、三维效应和蚀刻表面上出现的粗糙度。通过包括微结构充电和离子偏转效应，增强了轮廓模拟器的适用性，使其能够与等离子体反应器模型（由其他公司开发）相结合，用于设计下一代超高密度半导体器件的蚀刻工具。蚀刻实验也进行了支持模型和验证预测的轮廓特性。了解预测和观测到的偏离理想各向异性剖面的原因，有望实现对最终剖面的控制。对MOS器件栅极氧化物的电损伤进行了研究，以证明该技术的潜力。最后，在技术不断完善的基础上，尝试制作量子约束和光子带隙结构。为了将化学工程的研究与教育结合起来，揭示了经典化学工程概念与现代电子材料加工技术之间的关系。例如，在本科热力学课上提供了特殊的主题，如等离子体平衡和表面热力学，这些传统上没有涉及。一个介绍性的等离子体实验计划在本科实验室与实践经验的诊断。扩展的研究项目也提供给本科生在P.I.以提高他们对实验科学的欣赏。在研究生阶段，除了开设电子材料加工课程外，还计划开设一门关于表面光束修饰的新课程，这是基于本研究的结果和理解。＊＊＊

ABSTRACT CTS-9623450 This project addresses the importance of inelastic and reactive scattering in the etching of semiconductors. Experiments are proposed with a recently invented hypersonic neutral beam of fluorine atoms with translational energies in the 3-25 eV regime for fundamental studies of the gas-surface dynamics occurring during steady-state etching of silicon and etching studies of submicron features in silicon in order to understand how neutral beams etch. The effort is based on recent results from profile evolution modeling and experiments with flourine atom beams with translational energies between 3 and 18 eV. Anisotropic etching of silicon at room temperature with fluorine chemistry (no sidewall deposition) is demonstrated as a proof of the capabilities of such beams. This has been possible because of direct reactions occurring at the high translational energies employed. Collision-induced desorption (CID) has been implicated as the mechanism responsible for the increased desorption rate of fluorinated silicon moieties from the SiFx layer at the higher energies. Rapid removal of the SiF3 product in the CID mode enables rapid fluorination of the exposed surface, which increases fluorine utilization and, in turn, results in a reduction of the thermally desorbing unreacted fluorine-atom flux. Reduction of scattered fluorine to the sidewalls dramatically improves anisotropy. Fundamental measurements of the energy and flux of unreacted F atoms and reaction products (SixFy) leaving the SiFx layer during etching are made. Moreover, similar measurements are performed on the silicon dioxide surface. Silicon dioxide is ubiquitous on today's wafers (e.g., as a hard mask, gate oxide, etc.) and scattering on its surface will also affect profile evolution. Besides improving fundamental understanding of gas-surface dynamics, these measurements provide for a description of the physics and chemistry of the interaction, which is incorporated into a Monte Carlo simulator of profile evolution. Preliminary calculations based on detailed scattering experiments on SiFx are shown to capture profile peculiarities such as the so-called "microtrenching", that is, the appearance of grooves at the bottom of etched features. The model is improved to describe pattern-dependent etching, three-dimensional effects, and the roughness appearing on etched surfaces. By including microstructure charging and ion deflection effects, the applicability of the profile simulator is enhanced, enabling its combination with plasma reactor models (developed by others) for designing the etch tools for the next generation of ultrahigh-density semiconductor devices. Etching experiments are also performed to support the model and verify predictions of profile peculiarities. Understanding the origin of predicted and observed deviations from the ideal anisotropic profile is expected to enable control over the final profile. Electrical damage studies on gate oxides of MOS devices are done as a demonstration of the potential of the technique. Finally, upon perfection of the technique, fabrication of quantum-confined and photonic bandgap structures is attempted. In order to integrate research with education in chemical engineering, the relation between classical chemical engineering concepts and modern electronic material processing techniques is exposed. As an example, special topics are offered in the undergraduate thermodynamics class, such as equilibrium in plasmas and surface thermodynamics, which are not traditionally covered. An introductory plasma experiment is planned for the undergraduate laboratory with hands-on experience in diagnostics. Extended research projects are also offered to undergraduates in the P.I.'s laboratory to enhance their appreciation of experimental science. At the graduate level, in addition to a course on electronic materials processing, plans are made to offer a new course on beam modification of surfaces, which is based on results and understanding obtained from this proposed research. ***