Fern Jared T, Keffer David J, Steele William V
Physical Properties Research Facility, Chemical Engineering Department, 327 Dougherty Engineering Building, University of Tennessee, 1512 Middle Drive, Knoxville, Tennessee 37996-2200, USA.
J Phys Chem B. 2007 Apr 5;111(13):3469-75. doi: 10.1021/jp0674470. Epub 2007 Mar 9.
A new algorithm is presented that allows for the determination of bulk liquid and vapor densities from a two-phase Molecular Dynamics (2phiMD) simulation. This new method does not use any arbitrary cutoffs for phase definitions; rather it uses single-phase simulations as a self-consistency check. The method does not use any spatial bins for generating histograms of local properties, thereby avoiding the statistical issues associated with bins. Finally, it allows one to approach very close to the critical point. The new method utilizes Voronoi tessellations to determine the molecular volume of every point at every instance in a molecular dynamics simulation. Since the molecular volume is calculated throughout the simulation, statistical parameters such as the average molecular volume and average molecular variance are easy to obtain. To define the phases, the normalized variance of the molecular volume from 1phiMD and 2phiMD is used as a self-consistency check. The new method gives new insight into the nature of the near-subcritical fluid. The critical properties from this analysis are T(c) = 1.293 and rho(c) = 0.313. Direct simulation of the two-phase system was performed up to a temperature of 1.292. The results show excellent agreement to experimental results and Gibbs Ensemble Monte Carlo for coexisting densities. We see that well below the critical temperature, some particles are neither liquid nor vapor. These interfacial particles are primarily, but not exclusively, concentrated at the bulk interface. However, as we approach the critical point, some particles are considered both liquid and vapor. These interfacial particles are distributed through the system.
本文提出了一种新算法，可通过两相分子动力学（2phiMD）模拟来确定体相液体和蒸汽的密度。这种新方法在相定义中不使用任何任意截止值；而是使用单相模拟作为自洽性检验。该方法在生成局部性质直方图时不使用任何空间箱，从而避免了与箱相关的统计问题。最后，它允许人们非常接近临界点。新方法利用Voronoi镶嵌来确定分子动力学模拟中每个时刻每个点的分子体积。由于在整个模拟过程中都计算分子体积，因此诸如平均分子体积和平均分子方差等统计参数很容易获得。为了定义相，将1phiMD和2phiMD中分子体积的归一化方差用作自洽性检验。新方法为近亚临界流体的性质提供了新的见解。该分析得到的临界性质为T(c)=1.293和rho(c)=0.313。对两相系统进行了直接模拟，最高温度达到1.292。结果表明，与共存密度的实验结果和吉布斯系综蒙特卡罗方法具有极好的一致性。我们发现，在远低于临界温度时，一些粒子既不是液体也不是蒸汽。这些界面粒子主要但并非仅集中在体相界面处。然而，当我们接近临界点时，一些粒子被认为既是液体又是蒸汽。这些界面粒子分布在整个系统中。
