Department of Chemistry, University of Calgary, 2500 University Drive NW, Calgary, Alberta T2N 1N4, Canada.
J Chem Phys. 2012 Sep 7;137(9):094702. doi: 10.1063/1.4748377.
Growth of ice crystals has attracted attention because ice and water are ubiquitous in the environment and play critical roles in natural processes. Hexagonal ice, I(h), is the most common form of ice among 15 known crystalline phases of ice. In this work we report the results of an extensive and systematic molecular dynamics study of the temperature dependence of the crystal growth on the three primary crystal faces of hexagonal ice, the basal {0001} face, the prism {1010} face, and the secondary prism {1120} face, utilizing the TIP4P-2005 water model. New insights into the nature of its anisotropic growth are uncovered. It is demonstrated that the ice growth is indeed anisotropic; the growth and melting of the basal face are the slowest of the three faces, its maximum growth rates being 31% and 43% slower, respectively, than those of the prism and the secondary prism faces. It is also shown that application of periodic boundary conditions can lead to varying size effect for different orientations of an ice crystal caused by the anisotropic physical properties of the crystal, and results in measurably different thermodynamic melting temperatures in three systems of similar, yet moderate, size. Evidence obtained here provides the grounds on which to clarify the current understanding of ice growth on the secondary prism face of ice. We also revisit the effect of the integration time step on the crystal growth of ice in a more thorough and systematic way. Careful evaluation demonstrates that increasing the integration time step size measurably affects the free energy of the bulk phases and shifts the temperature dependence of the growth rate curve to lower temperatures by approximately 1 K when the step is changed from 1 fs to 2 fs, and by 3 K when 3 fs steps are used. A thorough investigation of the numerical aspects of the simulations exposes important consequences of the simulation parameter choices upon the delicate dynamic balance that is involved in ice crystal growth.
冰晶体的生长引起了人们的关注,因为冰和水在环境中无处不在,对自然过程起着至关重要的作用。六方冰,I(h),是 15 种已知冰的结晶相中最常见的冰的形式。在这项工作中,我们报告了利用 TIP4P-2005 水模型对六方冰的三个主要晶面,即基面{0001}面、棱柱面{1010}面和次棱柱面{1120}面的晶体生长随温度变化的广泛而系统的分子动力学研究结果。揭示了其各向异性生长的本质。结果表明,冰的生长确实是各向异性的;基面的生长和熔化是三个面中最慢的,其最大生长速率分别比棱柱面和次棱柱面慢 31%和 43%。还表明,周期性边界条件的应用会导致由于晶体各向异性的物理性质,不同取向的冰晶的尺寸效应不同,从而导致三个相似但中等尺寸的系统中热力学熔化温度可测量地不同。这里获得的证据为澄清当前对冰晶次棱柱面生长的理解提供了依据。我们还以更彻底和系统的方式重新审视了积分时间步长对冰晶体生长的影响。仔细评估表明,当积分时间步长从 1 fs 增加到 2 fs 时,冰的体相自由能会显著变化,并将生长速率曲线的温度依赖性降低到大约 1 K,当时间步长增加到 3 fs 时,会降低到 3 K。对模拟的数值方面进行彻底调查,揭示了模拟参数选择对冰晶生长所涉及的微妙动态平衡的重要影响。