Self-limited etching for atomic scale surface engineering of metals: understanding and design

所有的固体都有一个表面。为了控制表面的组成和结构，需要具有原子级精度的化学选择性和空间精确的修饰过程。半导体行业尤其如此，器件特征尺寸已进入个位数纳米尺度。为了实现未来纳米器件的制造，该项目寻求开发选择性和定向蚀刻工业上重要金属的方法。传统上，金属蚀刻和图案是通过将表面暴露在酸中来完成的。然而，这种方法不能用于纳米制造，因为它的选择性差，非方向性，以及较差的蚀刻深度控制。等离子蚀刻工艺在方向控制方面有了显著的改进，但这些工艺可能会受到金属再沉积的影响。原子层蚀刻（ALE）提供了一个很有前途的替代方案。ALE是一种两步自限循环工艺，首先通过等离子体工艺修饰金属表面，然后将其暴露在前驱体中，该前驱体选择性地蚀刻修饰表面。虽然有可能解决上述所有蚀刻工艺的缺点，但需要彻底了解重要金属（如Ni和Cu）的等离子体和表面反应机制。本提案试图用一种全面的计算化学方法来填补这一知识空白。该项目整合了理论建模和表面工程工艺设计与发现前沿的博士和本科生的研究和培训。这些学生将通过电子结构计算、人工智能算法和数据科学、表面化学实验以及对实验数据、能力和局限性的理解等方面的广泛培训做好准备。这个计算/实验研究项目的重点是开发原子层蚀刻（ALE）工艺，用于逐层去除金属薄膜。要模拟的反应性ALE过程从等离子体条件下的金属表面修饰步骤开始，将表面金属原子转化为表面化合物，当暴露于蚀刻剂时，形成挥发性金属络合物解吸，暴露蚀刻金属表面。修正步骤将使用具有神经网络电位（NNPs）的分子动力学（MD）模拟来模拟，并考虑轨迹的实际初始动能。NNP参数将使用基于密度泛函理论（DFT）计算生成的训练数据集来识别。等离子体修饰机器学习MD模拟的结果将在电感耦合等离子体（ICP）室的表面修饰实验中得到验证。对于蚀刻反应步骤，将首先构建一个能量的热力学数据库，作为蚀刻剂、衬底、改性剂和工艺条件的函数。热力学数据库将用于提出可行的蚀刻化学。对于实验验证的情况，蚀刻反应基本步骤的详细计算机制探索将确定关键低能途径的动力学势垒大小。总的来说，这种方法将增强对自限制表面反应的理解，以及精确设计金属表面的最佳反应物的定义。该项目支持的教育和推广活动包括与加州大学洛杉矶分校工程与多样性卓越中心（CEED）合作，确定未被充分代表的少数族裔学生（URM）参加该项目（第一年为本科生，第二年为高中生）。非正式的科学传播将通过教育门户进行，例如原子尺度设计网络（ASDN.net）和NanoHUB （nanohub.org）。研究小组将创建教育页面，提出原子尺度表面工程领域的新概念和想法。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

All solids are terminated by a surface. To control the composition and structure of a surface, chemically selective and spatially accurate modification processes with atomic-level precision are required. This is particularly true for the semiconductor industry where device feature sizes have entered the single-digit nanometer scale. To enable the fabrication of future nanodevices, this project seeks to develop methods to selectively and directionally etch industrially significant metals. Traditionally, metal etching and patterning has been performed by expositing the surface to an acid. However, such an approach cannot be used for nanofabrication because of its poor selectivity, non-directionality, and poor etch-depth control. Plasma etch processes constitute a significant improvement in directional control, but these processes can suffer from metal redeposition. Atomic layer etching (ALE) offers a promising alternative. ALE is a two-step self-limiting cyclic process where the metal surface is first modified by a plasma process and then is exposed to a precursor that selectively etches the modified surface. While having the potential to address all the aforementioned drawbacks of etch processes, a thorough understanding of the plasma and surface reaction mechanisms for important metals such as Ni and Cu is needed. This proposal seeks to fill this knowledge gap with a comprehensive computational chemistry approach. The project integrates research and training of Ph.D. and undergraduate students at the frontier of theoretical modeling and surface engineering process design and discovery. These students will be well prepared by the broad training on electronic structure calculations, algorithms of artificial intelligence and data science, surface chemistry experiments, and understanding of experimental data, capabilities, and limitations.This computational/experimental research program focuses on developing atomic layer etch (ALE) processes for the layer-by-layer removal of metal films. The reactive ALE process to be modelled starts with a metal surface modification step under plasma conditions to convert surface metal atoms to a surface compound that, when exposed to an etching agent, forms volatile metal-complexes that desorb, exposing the etched metal surface. The modification step will be modelled using molecular dynamics (MD) simulations with neural network potentials (NNPs), considering realistic initial kinetic energy for the trajectories. The NNP parameters will be identified using a training data set generated using density functional theory (DFT) based calculations. The results of plasma modification machine-learning MD simulations will be verified with surface modification experiments in an Inductively Coupled Plasma (ICP) chamber. For the etching reaction step, a thermodynamic database of energy will first be constructed as a function of etchant, substrate, modifiers, and process conditions. The thermodynamic database will be used to propose feasible etching chemistries. For experimentally validated cases, a detailed computational mechanistic exploration of reaction elementary steps for the etching reaction will determine the size of kinetic barriers for key low-energy pathways. Collectively, this methodology will result in an enhanced understanding of self-limiting surface reactions as well as the definition of optimal reactants to accurately engineer metallic surfaces. Educational and outreach activities supported by this program include partnering with the UCLA Center for Excellence in Engineering and Diversity (CEED) to identify top underrepresented minority (URM) students to work on this program (undergraduate student in the first year and high-school student in the second year). Informal science communication will be performed using the educational portals, such as Atomic Scale Design Network (ASDN.net) and NanoHUB (nanohub.org). The research team will create educational pages that bring forth the novel concepts and ideas in the field of atomic scale surface engineering.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.