Kinetic Multiscale Modeling and Simulation of Cluster Formation Processes in Supersonic Flows

星系团的形成过程涵盖了无数对社会非常重要的应用，如医学研究、新材料的创造和开发低成本的天基世界电信等等。最近有研究表明，沉积特定尺寸的团簇开辟了一类新的表面纳米结构，制造具有特定尺寸催化性能的材料。此外，为了减少脉冲激光烧蚀沉积薄膜中团簇的形成，必须控制靶面附近团簇的尺寸。这只是许多例子中的两个，这些例子表明了开发能够模拟团簇形成和演化的非平衡空间分布的理论和模拟工具以推进新的制造技术的重要性。预测超音速反应控制喷流中的星团大小分布和速度，将通过智能放置太阳能电池阵列等敏感表面来降低天基电信的成本。用于医疗外科手术的新材料正在通过簇形成过程开发，以在战略位置创建药物的纯微观植入物。显然，上述应用具有许多重要的全球和社会影响。在这些工程应用中，由于涉及团簇演化的非平衡物理化学过程的一些基本速率的不确定性和流动长度尺度的大变异性，对气相中团簇冷凝的预测建模目前是一项艰巨的任务。因此，开发模拟工具的一个合乎逻辑的步骤是考虑超音速喷流膨胀这一更简单的现象，在这种现象中，团簇的形成和增长过程只与非平衡气体动力学耦合。超音速喷流中集群行为的成功建模将作为未来涉及集群演化的更复杂技术的建模的原型。对非均质两相流系统的成功模拟还将使我们能够确定单个基本截面的相对重要性。本研究的主题是超音速射流中由均匀成核形成的团簇的定量表征，即团簇大小、内能和动能的非平衡空间分布。提出了用多尺度计算模型来模拟超音速射流中团簇的形成和演化过程，该模型基于一种适用于向稀薄流型过渡的动能粒子模拟方法--直接模拟蒙特卡罗方法。所提出的模型是多尺度的，因为除了DSMC之外，还将使用连续介质流体动力学(CFD)/纳维斯托克斯(NS)和分子动力学(MD)公式。采用CFD的Navier-Stokes方法来模拟稠密气体的初始膨胀阶段，在此阶段可以忽略凝结。分子动力学将主要被用来开发一个用于DSMC的团簇反应截面数据库，并进一步扩展DSMC的模拟能力，以解决与非二元碰撞引起的凝聚有关的问题。将开发一种统计上有效的方法来估计反应截面和碰撞后的结果。这项工作的主要重点涉及由成对势能描述的系统，因为必须首先为众所周知的简单气体建立多尺度建模和模拟可信度。瑞利散射数据集不仅提供了平均星系量，而且还提供了星系团分布，将用于模拟验证。这些数据一直未得到充分利用，是模型界的重要资源。该项目将促进发现和理解，同时促进教学。将开发一门新的研究生课程，内容是稀薄流动的计算建模及其在材料加工中的应用。该课程将涉及学习和使用MD-DSMMC方法，该方法将在拟议的研究中开发。更具体地说，在修完这门课程后，航空航天与机械工程专业和工程科学专业的学生将能够模拟具有已知相互作用势的化学体系的稀薄-冷凝耦合流动。这门课程是美国为数不多的学生完成课堂项目的课程之一，该课程要求学生使用DSMC来模拟稀薄气体流动对航天器的影响。本科生和研究生将通过宾夕法尼亚州立大学的各种项目和直接征求同事的意见，参与拟议的学术活动，扩大到主要专业学生的范围。这项拟议的研究结合了稀薄气体动力学、物理化学、大规模并行计算、材料科学和统计学的元素，因此是多学科的。

Cluster formation processes are all encompassing and found in countless applications important to societymedical research, creation of new and novel materials, and developing low cost space-based world widetelecommunications, to name a few. It has been shown recently that deposition of size-specific clustersopens up a new class of surface nanostructure fabrication of materials that have size-specific catalyticproperties. Additionally, to reduce cluster formation in pulsed laser ablation deposition of thin films thesize of clusters in the vicinity of a target surface must be controlled. These are just two of many examplesthat demonstrate the importance of developing a theory and simulation tool capable of modeling nonequilibrium spatial distributions of cluster formation and evolution to advance new fabrication technologies. Predicting cluster size distributions and velocities in supersonic reaction control jets will reduce the cost of space-based telecommunications by allowing intelligent placement of sensitive surfaces such as solar cell arrays. Novel materials for medical surgical procedures are being developed by cluster formation processes to create pure microscopic implants of pharmaceuticals in strategic locations. Clearly the above applications have many important global, societal implications. The predictive modeling of cluster condensation in the gas phase in these engineering applications currently presents a formidable task due to the uncertainties in some of the fundamental rates for nonequilibrium physical chemical processes involving the evolution of clusters and the large variability of flow length scales. Therefore, a logical step in the development of simulation tools is to consider the simpler phenomenon of supersonic jet expansion in which the process of cluster formation and growth is coupled only with non-equilibrium gas dynamics. Successful modeling of cluster behavior in supersonic jets will serve as a prototype for future modeling of more complex technologies involving the evolution of clusters. Successful modeling of a non-homogenous two-phase flow system will also allow us to establish the relative importance of individual fundamental cross sections.The subject of this proposed research is the quantitative characterization of clusters formedby homogeneous nucleation in a supersonic jet, in terms of non-equilibrium spatial distributionsof cluster size and internal and kinetic energies. We propose to model the processof cluster formation and evolution in supersonic jets by a multiscale computational model.The proposed model is based on a kinetic particle simulation method, the direct simulation Monte Carlo(DSMC), which is applicable in the transitional to rarefied flow regime. The proposed model is multiscalebecause, in addition to DSMC, continuum fluid dynamics (CFD)/Navier-Stokes (NS) and molecular dynamics(MD) formulations will also be used. The CFD Navier-Stokes approach will be used to simulate theinitial expansion stage of the dense gas, during which the condensation can be neglected. Molecular dynamicswill be used primarily to develop a cluster reaction cross section data base for DSMC and to extend theDSMC simulation capabilities further to address issues related to condensation due to non-binary collisions.A statistically efficient approach will be developed for estimation of both the reaction cross sections andpost-collisional outcomes. The primary emphasis of this work involves systems described by a pair-wisepotential, since multi-scale modeling and simulation credibility must first be established for well known,simple gases. Rayleigh scattering data sets providing not only average cluster quantities, but cluster distributions as well, will be used for simulation validation. These data have been under-utilized and represent an important resource for the modeling community.The project will advance discovery and understanding while promoting teaching and learning. A newgraduate course on computational modeling of rarefied flows with applications to materials processing willbe developed. The course will involve the learning and usage of the MD-DSMC methodology to be developedin the proposed research. More specifically, after taking this course Aerospace and Mechanical Engineeringand Engineering Science students will be able to simulate coupled rarefied-condensation flows for chemicalsystems with known interaction potentials. This course is one of the very few in the nation where studentsperform a class project requiring the use of DSMC for modeling the effects of rarefied gas flows on spacecraft.Undergraduate and graduate students will participate in the proposed schololarly activities with outreach tominority students through various programs at Penn State and direct solicitation of colleagues. The proposedresearch combines elements of rarefied gas dynamics, physical chemistry, massively parallel computing,materials science, and statistics and is, therefore, multidisciplinary.