Real space electrostatics and non-equilibrium molecular dynamics for nanoscale transport

圣母大学的J.Daniel Gezelter获得了化学理论、模型和计算方法(CTMC)计划以及计算和数据科学与工程(CDS&Amp；E)计划颁发的奖项，该计划的重点是了解与水和其他溶剂接触的金属纳米颗粒(直径几十亿分之一米的小颗粒)。格泽尔特博士正在开发计算方法和软件，以便在粒子受到附着分子保护的重要情况下，研究粒子表面的热传递。这一主题在能源、材料研究和健康科学领域具有重要意义，特别是对于已被提议用于光热疗法的粒子，在光热疗法中，激光被用于非常局部的加热。这一界面上的溶剂结构和动力学的细节决定了能量如何从金属颗粒中流出。为了使这些颗粒在医学上有用，重要的是要确定激光加热后周围环境温度升高的机制。这使得计算方法可以预测表面的导热系数，这可能很难通过实验确定。这项提议的目标是：(1)有效地模拟纳米颗粒表面之间的能量传递的算法，以及(2)非常有效地模拟像水这样的极性分子的方法，即在相反的两端具有主要的正电荷和负电荷。该项目支持的所有算法和软件将在允许的开源许可下向公众和其他研究人员发布。PI组织了开放科学项目，这是一个网站，突出并提供有用的科学工具和研究代码的例子，以及可供更广泛的受众(倾向于科学但不是专家)使用的工具。这项工作的两个重要的分子模拟方法开发项目是这项工作的一部分。该项目将有效的实空间静电模型扩展到涵盖点-多极子，并将实空间模型与已有的(但更昂贵的)仿真技术进行了比较。这些模型使用两种截然不同的方法来截断泰勒级数在真实空间截止半径处对静电核的逼近。蛋白质、脂肪和液体的粗粒度模型都有望从真实空间方法的效率收益中受益。这项工作还涉及非平衡技术的算法开发，以研究纳米颗粒弯曲界面上的界面属性，如热导和界面摩擦。这些量是界面的动态属性，是对非平衡条件或极少数偏离平衡行为的波动的响应。正因为如此，用标准的平衡分子动力学或蒙特卡罗方法几乎不可能测量它们。为了研究界面输运，在材料之间产生热流或角动量流的方法是必要的。算法开发和这些算法在纳米颗粒现实模型中的应用相结合，将为我们提供对强烈科学兴趣的问题的洞察。

J. Daniel Gezelter of the Uninversity of Notre Dame is supported by an award from the Chemical Theory, Models and Computational Methods (CTMC) program and the Computational and Data-Enabled Science and Engineering (CDS&E) program focused onunderstanding of metallic nanoparticles (small particles a few billionths of a meter in diameter) that are in contact with water and other solvents. Dr. Gezelter is developing computational methods and software to study heat transfer at the surfaces of the particles in the important case that they are protected by attached molecules. This topic is of significant interest in the areas of energy, materials research, and the health sciences, particularly for the particles that have been proposed for use in photothermal therapies where laser light is used to cause very localized heating. The details of the solvent structure and dynamics at this interface determine how energy flows away from the metal particles. In order to make these particles medically useful, it is important to determine the mechanism for the increase of the temperature in the surrounding environment following laser heating. This allows the computational methods to predict heat conductance at the surface, which can be difficult to determine experimentally. The goals of this proposal are to develop: (1) algorithms to efficiently model the transfer of energy across nanoparticle surfaces, and (2) methods to very efficiently model molecules like water that are polar, i.e., have positive and negative charges dominate at opposite ends. All of the algorithms and software supported by this project will be released to the public and other researchers under a permissive open source license. The PI organizes the OpenScience Project, a website which highlights and makes available examples of useful scientific tools and research codes as well as tools that can be used by a broader audience (scientifically-inclined but non-expert).Two significant method-development projects for molecular simulation are part of this work. This project expands efficient real-space electrostatic models to cover point-multipoles, and the real-space models are compared with established (but more expensive) simulation techniques. These models use two distinct approaches to truncated Taylor series approximations to the electrostatic kernel at a real-space cutoff radius. Coarse-grained models for proteins, lipids, and liquids are all expected to benefit from the efficiency gains of a real-space approach. The work also involves algorithmic development of non-equilibrium techniques to study interfacial properties like thermal conductance and interfacial friction at the curved interfaces of nanoparticles. These quantities are dynamic properties of interfaces, and are responses to non-equilibrium conditions or very rare fluctuations away from equilibrium behavior. Because of this, they are nearly impossible to measure using standard equilibrium molecular dynamics or Monte Carlo methods. To study interfacial transport, methods that create heat or angular momentum fluxes between materials are necessary. The combination of algorithmic development and application of these algorithms to realistic models for nanoparticles will provide insight into problems of intense scientific interest.