优化温度相关力场预测正构烷烃热力学性质

为了实现不同温度下正构烷烃及其混合物热力学性质的准确预测,以正构烷烃(n-C4~C10)为训练集,通过对全原子OPLS-AA力场中非键相互作用参数(ε)的模拟优化,得到了ε与对比温度(Tr)以及正构烷烃碳原子个数(NC)的经验关系式。利用该关系式计算出不同温度不同种类的正构烷烃的ε值,预测了正构烷烃纯物质及其混合物的黏度、密度、扩散系数等物性,并将新力场模拟计算值与理论估算值以及实验值进行比较。结果表明,采用优化温度相关力场预测烷烃及其混合物的物性与实验值最为吻合。密度、黏度和扩散系数的预测值与实验值的相对偏差分别小于2%、5%和10%,显著优于目前的理论方法和原OPLS-AA力场模拟计算的预测值。上述温度相关力场参数的确立,对于利用分子动力学模拟方法准确地预测正构烷烃及其混合物的热力学性质具有重要的实际应用价值。     

采用泰勒分散法测量蜡分子扩散系数

蜡分子扩散系数是蜡沉积预测模型中非常重要的物性参数。为测量蜡分子（高碳数正构烷烃）在液相体系中的分子扩散系数,建立了基于泰勒分散法的扩散系数测量装置,并对装置操作注意事项进行了详细探讨。采用该装置测量了丙酮水溶液、n-C₆+n-C₇溶液以及甲醇水溶液中溶质在溶剂中的分子扩散系数,以文献值对测量值进行校验,验证了装置的可靠性。分别以n-C₁₈、n-C₂₀、n-C₂₂、n-C₂₄、n-C₂₆为溶质,以n-C₇为溶剂,在不同温度、浓度下测量蜡分子在液相体系中的分子扩散系数,并将实验测量值与Hayduck-Minhas关系式计算值进行比较。结果表明：蜡分子扩散系数随温度升高而线性增大,随溶液中蜡分子摩尔分数的增大而以指数形式减小;在相同条件下,高碳数正构烷烃的分子扩散系数低于低碳数正构烷烃的分子扩散系数。采用Hayduck-Minhas关系式的计算结果比实验测量结果平均小50%,应用于蜡沉积预测时,将低估蜡分子扩散质量流。     

纯CO2体系扩散性质的分子动力学模拟

使用基于COMPASS力场的分子动力学（MD）模拟方法计算了CO2在气相、液相以及超临界区的自扩散系数。计算结果表明温度对扩散系数的影响在密度较低时比较明显,随着密度的升高温度的影响逐渐减弱;密度较高时密度的影响对扩散系数起主导作用。通过研究温度和密度对扩散系数的影响规律,提出了预测CO2自扩散系数的新方程,该方程与模拟值和文献试验值吻合良好。     

分子动力学模拟预测壳聚糖的玻璃化转变温度

为了预测壳聚糖的玻璃化转变温度,在COMPASS力场和恒温恒压（NPT）系综条件下,利用分子动力学模拟方法,在343～543 K温度范围内研究了壳聚糖的玻璃化转变行为,通过模拟体系在不同温度下的比体积、回转半径和能量参数,获得了壳聚糖的玻璃化转变温度(T_g)。研究结果表明,壳聚糖的比体积、回转半径、内能随温度有规律的变化并在T_g处发生转折。模拟计算得到的壳聚糖的T_g与实验方法获得的值基本相符,分子动力学方法可用于壳聚糖玻璃化转变温度的预测。其中,通过回转半径-温度曲线获得的T_g与实验值最相符,回转半径是影响玻璃化转变的一个重要因素,可用于预测聚合物的玻璃化转变温度。     

基于ASOG-VISCO模型的醚类混合溶液黏度预测

基于44种二元含醚类物质混合溶液的运动黏度实验数据,采用ASOG-VISCO黏度模型回归得到了新的基团对O-CH₂,O-CyCH,O-ArCH,O-OH,O-COO,O-CCl₃,O-CCl₄的交互作用参数,所研究溶液体系为醚类与正烷烃、异烷烃、环烷烃、芳香族、醇类、氯仿、四氯化碳和酯类的混合物,黏度计算值与实验值的绝对平均偏差为3.31%。同时,为了验证回归得到的新交互作用参数对醚类混合溶液黏度预测的通用性,对11种其他醚类混合溶液的黏度进行了推算,计算值与实验值的绝对平均偏差为5.95%,可以满足实际工程需要。     

超临界CO2在聚苯乙烯熔体中的扩散系数

微孔发泡材料的质量与超临界CO₂在聚合物熔体中的溶解量和溶解速率紧密相关,采用有限元模拟的方法得出不同条件下的等效扩散系数来表征溶解速率。首先,通过基于体积法的实验装置测出不同条件下超临界CO₂在聚合物熔体中的溶解量;其次,通过COMSOL软件计算得出不同的扩散系数下溶解量的变化,最后通过实验值与模拟计算值之间的对比,得出不同条件下,超临界CO₂在聚合物中的等效扩散系数,并分析了误差来源。结果表明,超临界CO₂在聚苯乙烯（PS）熔体中的溶解量随着压力的增加而增加,相反,随着温度的升高呈下降的趋势;模拟计算所得等效扩散系数与实验保持一致,温度与压力的增大都会提高扩散系数的数值;模拟的溶解平衡时间与实验的溶解平衡时间之间的误差不超过20 %。实验结果与模拟计算的结果吻合较好。因此,计算溶解量及等效扩散系数具有重要意义,可为微孔发泡工艺提供理论指导。     

R1234yf+CO2和R1234ze (E)+CO2二元混合物的比容平移Soave-Redlich-Kwong方程

采用温度相关比容平移项的比容平移Soave-Redlich-Kwong（VTSRK）方程计算新工质R1234yf和R1234ze（E）的热力学性质以及两种物质与CO₂的二元混合物性质,混合物计算采用van der Waals混合规则,二元交互作用系数由密度数据拟合得到。对纯净物计算与专用状态方程进行对比,VTSRK方程比SRK方程显著改善了液相密度表征效果。对混合物的密度计算结果与实验数据进行对比,对于R1234yf+CO₂二元混合体系方程与实验数据相对均方根偏差为1.17%,对于R1234ze（E）+CO₂二元混合体系相对均方根偏差为0.82%。结果显示,采用温度相关比容平移项的VTSRK方程应用于R1234yf和R1234ze（E）纯流体以及R1234yf+CO₂和R1234ze（E）+CO₂密度性质计算,可获得较高精度。     

长链正构烷烃（n-C16~C20）脱氢制直链单烯烃反应热力学分析

对长链正构烷烃（n-C₁₆~C₂₀）一次脱氢制直链单烯烃反应进行了详细的热力学分析,考察反应温度、反应压力、氢烃物质的量比和原料碳数对热力学平衡的影响。结果表明,n-C₁₆~C₂₀脱氢反应的平衡常数及平衡转化率随烷烃碳数的增大而增加;提高温度和降低压力均可显著提高平衡转化率,降低氢烃物质的量比在一定程度上提高平衡转化率。适宜的工艺条件为：反应温度>430 ℃,反应压力(0.10~0.20) MPa,氢烃物质的量比为6~8。     

无定形PET中小分子扩散系数的分子动力学模拟

采用分子动力学模拟研究了相对分子质量在32～339范围内小分子在无定形PET中的扩散过程。基于Einstein关系式计算了扩散系数,讨论了模拟时间、密度对扩散系数的影响。结果表明：在较高的温度下,较短的模拟时间内就能观察到均方位移曲线的线性区;而在较低的温度下,需更长的模拟时间。扩散系数随密度增加而降低,聚合物密度越大,所需的模拟时间越长。通过比较扩散系数的计算值与实验值,发现两者的比值在一个数量级范围内,表明建立的聚合物模型可接受且能正确描述小分子在无定形PET中的扩散过程, 为获得迁移模型中关键参数——扩散系数提供了一种近似的计算方法。     

R600a/矿物油的黏度和密度

利用振动弦黏度/密度实验测量系统对R600a/SUNISO 3GS矿物油混合物5种配比(R600a质量分数为100%、95%、90%、85%、80%)下的黏度和密度进行了实验研究,温度范围为253～333 K。实验系统黏度和密度测量的不确定度分别为±2%和±0.2%,温度测量的不确定度小于±10 mK。在获得实验数据的基础上得到了黏度和密度的关联式。黏度实验数据与方程计算值的最大绝对偏差和平均绝对偏差分别为5.73%和1.75%,密度实验数据与方程计算值的最大绝对偏差和平均绝对偏差分别为0.23%和0.07%。为实际制冷系统的优化设计提供了更为可靠的热物性数据。     

Combining Coarse-Grained Protein Models with Replica-Exchange All-Atom Molecular Dynamics

We describe a combination of all-atom simulations with CABS, a well-established coarse-grained protein modeling tool, into a single multiscale protocol. The simulation method has been tested on the C-terminal beta hairpin of protein G, a model system of protein folding. After reconstructing atomistic details, conformations derived from the CABS simulation were subjected to replica-exchange molecular dynamics simulations with OPLS-AA and AMBER99sb force fields in explicit solvent. Such a combination accelerates system convergence several times in comparison with all-atom simulations starting from the extended chain conformation, demonstrated by the analysis of melting curves, the number of native-like conformations as a function of time and secondary structure propagation. The results strongly suggest that the proposed multiscale method could be an efficient and accurate tool for high-resolution studies of protein folding dynamics in larger systems.     

Study of the molecular weight dependence of glass transition temperature for amorphous poly(l-lactide) by molecular dynamics simulation

Molecular dynamics simulation has been used to investigate the molecular weight dependence of glass transition temperature for amorphous poly(l-lactide). Amorphous PLLA systems were created using molecular modeling and NPT ensemble MD simulations were carried out using the modified OPLS-AA force field. The fractal dimension of the PLA systems was 1.62. The molecular weight dependence of glass transition temperature, self-diffusion coefficient and shear viscosity were studied and the good agreement between the simulation results and experiments was obtained.     

Machine learning for molecular thermodynamics

Thermodynamic properties of complex systems play an essential role in developing chemical engineering processes. It remains a challenge to predict the thermodynamic properties of complex systems in a wide range and describe the behavior of ions and molecules in complex systems. Machine learning emerges as a powerful tool to resolve this issue because it can describe complex relationships beyond the capacity of traditional mathematical functions. This minireview will summarize some fundamental concepts of machine learning methods and their applications in three aspects of the molecular thermodynamics using several examples. The first aspect is to apply machine learning methods to predict the thermodynamic properties of a broad spectrum of systems based on known data. The second aspect is to integer machine learning and molecular simulations to accelerate the discovery of materials. The third aspect is to develop machine learning force field that can eliminate the barrier between quantum mechanics and all-atom molecular dynamics simulations. The applications in these three aspects illustrate the potential of machine learning in molecular thermodynamics of chemical engineering. We will also discuss the perspective of the broad applications of machine learning in chemical engineering.     

基于第一原理力场预测热力学参数的探讨

讨论用第一原理全原子力场计算热力学参数的可行性.新开发的TEAM力场,在用气、液相基本性质验证后,不加任何参数调整,直接用于计算具有代表性的几个分子液体在不同热力学状态下的密度、蒸发焓、混合焓、亨利常数、饱和蒸气压等热力学参数.初步结果表明：液体密度和蒸发焓的预测结果良好,在较大范围内可获得与实验数据相吻合的结果;而混合焓的准确预测需要纯液体内聚能的高精度模拟数据;亨利常数的计算对液体的密度极为敏感.后两种性质的计算均可用全原子力场但需要高质量的力场参数.为了采用全原子力场计算气液相平衡数据,提出了一种半经验的、基于积分Clausius-Clapeyron方程的循环自洽方法计算液体的饱和蒸气压.     

Molecular dynamics simulation for the binary mixtures of high pressure carbon dioxide and ionic liquids简

Molecular dynamics simulation with an all-atom force field has been carded out on the two binary sys- tems of [bmim][PF6]-CO2 and [bmim][NO3]-CO2 to study the transport properties, volume expansion and micro- structures. It was found that addition of CO2 in the liquid phase can greatly decrease the viscosity of ionic liquids （ILs） and increase their diffusion coefficient obviously. Furthermore, the volume expansion of ionic liquids was found to increase with the increase of the mole fraction of CO2 in the liquid phase but less than 35% for the two simulated systems, which had a significant difference with CO2 expanded organic solvents. The main reason was that there were some void spaces inter and intra the molecules of ionic liquids. Finally, site to site radial distribution functions and corresponding number integrals were investigated and it was found that the change of microstructures of ILs bv addition CO2 had a great influence on the orooerties of ILs.     

用分子动力学模拟研究阳离子链长对离子液体输运特性的影响

阐述了用分子动力学（molecular dynamics,MD）模拟方法研究在T=303K条件下[mmim][TFSA]、[emim][TFSA]、[bmim][TFSA]、[C6 mim] [TFSA]和[C8mim] [TFSA]5种离子液体的输运特性.模拟力场采用修正后的OPLS力场.根据模拟轨迹计算得到5种离子液体的密度值.根据均方位移（MSD）的斜率计算得到的离子液体阴阳离子自扩散系数.采用Nernst-Einstein （NE）方程计算得到离子液体摩尔导电率.这些模拟结果与实验值很吻合.离子液体的自扩散系数和电导率随着阳离子链长的增长而变小,主要原因是阳离子链长增长使离子液体中的氢键作用和范德华作用变强.计算所得离子液体摩尔导电率略大于实验测量值则归因于离子关联运动的结果.     

Molecular understanding of pyridinium ionic liquids as absorbents with water as refrigerant for use in heat pumps

Aiming at developing new absorbent/refrigerant working pairs for heat pumps, thermodynamic and transport properties of two pyridinium ionic liquids (ILs), N‐ethylpyridinium bis(trifluoromethanesulfonyl)amide and N‐ethylpyridinium trifluoromethanesulfonate were studied using molecular simulation and nuclear magnetic resonance techniques. The microscopic structure of the ILs and the solvation environment of water, including hydrogen bonding, were studied. Free‐energies of solvation of water were obtained using perturbation methods, and the values agree with experimental observations. Self‐diffusion coefficients and viscosity were computed and compared with nuclear magnetic resonance measurements and literature. Simulations predict slower dynamics when compared with experiment: diffusion coefficients are underpredicted, whereas viscosity is overpredicted. As such, simulation is consistent in a Stokes‐Einstein sense. The trends in transport properties due to changing anion, to the presence of water and the effect of temperature are well predicted. © 2017 American Institute of Chemical Engineers AIChE J, 63: 3523–3531, 2017     

氧气和一氧化碳在人血红蛋白迁移过程研究

为了揭示CO和O₂竞争性结合人血红蛋白血红素位点的机制及其与人血红蛋白结构转换之间的关系，本文采用全原子分子动力学模拟（MD）结合马尔科夫状态模型（MSMs）研究氧气（O₂）和一氧化碳（CO）分子从水溶液迁移进入人血红蛋白四聚体α链和β链的全过程。分子动力学模拟揭示了O₂和CO结合α链和β链的稳态结合位点和瞬态结合位点、迁移通道以及α链的结构变化。结果显示，分子模拟不仅仅能够再现全部实验中所观察到的离散Xe结合位点和分子扩散通道，而且揭示了实验中无法观测的瞬态结合位点和多重气体迁移途径。上述结果表明人血红蛋白因其结构柔性所形成的瞬态通道对于气体分子迁移过程的重要性。除此之外，利用MSM和过渡路径理论（TPT）构建了人血红蛋白α链结构变化与气体分子迁移之间的关系，阐释了血红蛋白中影响气体迁移的关键结构及其微观机制。     

Properties of supercritical N2, O2, CO2, and NH3 mixtures from the virial equation of state

We report virial coefficients up to sixth order in density for N₂ , O₂ , NH₃ , and CO₂ , covering temperatures from 50 to 1,000 K. The reported values include coefficients and their first three temperature derivatives, for the pure species as well as all of those needed to evaluate full composition dependence of mixtures formed from any or all of these compounds. The values are obtained by calculation of appropriate cluster integrals using Mayer sampling Monte Carlo, with intermolecular interactions described by the Transferable Potential for Phase Equilibria (TraPPE) force field. All coefficients are fit as a function of temperature, yielding a thermodynamic model with analytic dependence on temperature, density, and composition. The coefficients and properties computed from them are compared to experimental data where available.     

分子动力学模拟基本原理及研究进展

阐述了分子动力学模拟的基本原理,介绍了分子动力学的牛顿运动方程、相关有限差分算法、边界条件的选取以及热力学性质的计算.最后还指出了分子动力学模拟方法进一步研究方向.