CDS&E: Rigorous formulas for industrial supercritical-fluid mixture properties via systematic evaluation of molecular virial coefficients, and methods to expand their applicati

了解材料特性对技术发展至关重要：为了设计、控制或优化设备或制造过程，必须了解物理对象或材料在加热或机械力作用下的行为。这种行为是由材料的热物理性质决定的。虽然这些性质通常可以通过实验测量，但由于涉及大量变量，这种方法通常被发现是不切实际的。为了解决这个问题，科学家和工程师依靠数学模型来预测材料的性能，通过将基于物理的模型拟合到现有数据中，从而计算出测量数据值之外的条件下的热物理性能。虽然这种方法很有用，但由于缺乏验证数据，它的预测能力受到限制。一种更好的方法是基于分子计算材料特性，使用描述构成材料的分子如何相互作用的模型。然而，发展分子动力学的准确描述是具有挑战性的，正如在宏观尺度上解释这些模型预测一样。为了应对这些限制，本研究计划将扩展一种严格但被忽视的方法来计算热物理常数（维里系数），该方法可以模拟热力学行为与理想表示的偏差，从而可以预测现实情况下的性质。这将通过使用最先进的分子水平模拟来确定31种代表性化合物的维里系数值来完成，这些化合物跨越一系列重要的化学加工和能源应用。这项研究有能力改变分子模型开发者如何制定和测试他们的方法和模型的思维，以及属性模型开发者如何制定和参数化模型以匹配热力学数据的思维。研究小组正在开发和传播计算工具，这些工具可以帮助其他开发人员执行他们自己的病毒系数计算，从而扩大混合数据的范围，远远超出本项目的目标。还将通过培训学生，重点关注代表性不足的群体、社区外展和分发教育材料，产生积极影响。在这个项目中，将寻求几种途径将计算集群积分方法的最新进展带入实际应用。首先，研究小组将进行系统的努力，根据经验和第一性原理的分子间势来评估31个物种的温度依赖病毒系数，以及由它们形成的所有可能的混合物。这些结果将形成一个数据库，该数据库将是一个可供公众访问的几十万个系数资源，可用于计算蒸汽和超临界流体相中这些混合物的所有热力学性质。其次，将通过与文献中现有实验数据的综合比较来评估系数。这些数据包括报道的超临界混合物的维里系数、体积性质和热性质。这种比较对于评估分子模型的质量和找出它们的弱点是有价值的。然后，该数据库将用于进一步研究混合物的行为，包括：(a)如何利用混合物中临界点奇点的知识来加速vrial级数的收敛；(b)研究混合物中的焦耳-汤姆逊效应。最后，将研究将这些系数应用于工程实践中广泛使用的热力学模型参数化的可行性，以扩大这些数据的应用范围。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。

Understanding material properties is vital to technological development: to design, control, or optimize a device or manufacturing process, how that physical object or material will behave when heat or mechanical forces are applied to it must be understood. This behavior is governed by the thermophysical properties of the material. While these properties generally can be measured experimentally, such an approach is typically found to be impractical due to the large number of variables involved. To remedy this, scientists and engineers rely on mathematical models to predict material properties by fitting physically based models to available data so that the thermophysical properties at conditions outside the values of the measured data can be computed. While useful, this approach can be limited in its predictive capabilities by a lack of validation data. A better approach would be based on computing material properties from molecular considerations using models describing how the molecules that make up the material interact. Developing accurate descriptions of molecular dynamics, however, is challenging, as is the interpretation of these model predictions at the macroscopic scale. In response to these limitations, this research program will expand upon a rigorous but neglected approach to the computation of thermophysical constants (virial coefficients) that model thermodynamic behavior deviations from idealized representations, allowing for the prediction of properties under realistic situations. This will be accomplished by using state-of-the-art molecular-level simulations to identify virial coefficient values for 31 representative chemical compounds that span a range of important chemical processing and energy applications. This research has the capability to transform the thinking of molecular-model developers on how they formulate and test their methods and models as well as property-model developers on how they formulate and parameterize models to match thermodynamic data. The research team is developing and disseminating computational tools that can assist other developers to perform their own calculations of virial coefficients, thereby broadening the scope of mixture data well beyond what is targeted in this project. Positive impact also will be made through training of students, with a focus on underrepresented groups, community outreach, and dissemination of educational materials.In this project, several avenues will be pursued to bring recent advances in computational cluster-integral methods toward practical applications. First, the research team will engage in a systematic effort to evaluate temperature-dependent virial coefficients from empirical and first-principles intermolecular potentials for a collection of 31 species, and all possible mixtures formed from them. These results will form a database that will be a publicly accessible resource of hundreds of thousands of coefficients that can be used to compute all thermodynamic properties of these mixtures in the vapor and supercritical-fluid phases. Second, the coefficients will be evaluated through a comprehensive comparison to available experimental data from the literature. Such data encompass reported virial coefficients, volumetric properties, and thermal properties of supercritical mixtures. This comparison is valuable in assessing the quality of the molecular models and to pinpoint their weaknesses. The database then will be used to further study the behavior of mixtures including: (a) how knowledge of critical-point singularities in mixtures can be used to accelerate convergence of the virial series; and (b) studying the Joule-Thomson effect in mixtures. Finally, the feasibility of applying these coefficients toward parameterization of thermodynamic models used widely in engineering practice will be investigated to broaden the range of application of these data.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.