Molecular dynamics study of gas sorption by organic-coated water droplets and parametrization of liquid-liquid interfaces

Ph.D.Gas sorption by atmospheric water droplets has significant influences on the climate and air quality by altering both the physical and chemical properties of the atmosphere. To gain a better understanding of the gas sorption processes, we studied the sorption of water and three common atmospheric gases (N₂O₅, SO₂, and O₃) by water droplets/slabs in molecular dynamics (MD) simulations. We showed the calculated solvation free energies are linearly proportional to the reciprocal of the number of water molecules in both the slabs and the droplets. Moreover, we found all our droplet results are more accurate than the slab results in the limit of infinitely large system size when compared to the experimental estimates. Decomposition of the free energy changes ΔF into the energetic ΔU and entropic −TΔS contributions allowed us to analyze the intrinsic driving forces with molecular details. In particular, we found that the trends in ΔU are primarily determined by local effects surrounding the gas molecule, and the trends in −TΔS are strongly correlated with the mean residence time of water, which is spatially nonlocal and related to the mobility of water.With improved knowledge of the gas sorption by pure water droplets, we investigated the more complicated scenario with water droplets coated by organic matter, which is pervasively found in the atmosphere. To understand how the organic coatings affect gas sorption, we studied the sorption of H₂O, O₃, and SO₂ by a water droplet coated with either benzoic acid (BA) or lauric acid (LA) in both equilibrium and nonequilibrium MD simulations. ΔF for the gas sorption processes was again calculated and decomposed into the energetic and entropic components. By analyzing the intermolecular interactions, we found the changes in the potential energy ΔU are dominated by the changes in the Coulomb interactions in the BA-coated systems, while in the LA-coated systems, changes in the Lennard-Jones interactions are crucial due to the long alkyl chains. For the entropic components, we found a strong correlation between TΔS and the mobility of water and organic molecules. In the nonequilibrium simulations, we investigated the effects of the organic coatings on the mass accommodation coefficient α and managed to find a general connection between the change in α and the relative position of ΔF minimum.In my last project, I investigated a methodology for accurately parametrizing the molecular interactions in systems with liquid-liquid interfaces, which are pertinent to various phenomena such as separations, extractions, and phase transfer catalysis. Most of the common force fields are optimized for single phases without special consideration on the intermolecular interactions at the interfaces. Ab initio MD simulations are believed to describe the interactions accurately based on first principles, but the demanding computational cost makes it infeasible to simulate large system sizes and/or long timescales. To resolve this, we aimed to reproduce the results of ab initio MD simulations in classical MD simulations by a parametrization procedure based on the minimization of the relative entropy. To include polarizability, which can be important for interfacial properties, we adopted the Drude model in our force fields. As test examples, we studied the chloride and tetraethylammonium (TEA) ions with water-dichloromethane (DCM) or water-benzene interfaces due to their importance in phase transfer catalysis. By using Newton’s method, the value of relative entropy, which measures the difference between the probability distributions of states of the AIMD target and our model, decreased significantly compared to that of the original model from GAAMP and others. Our methodology of generating polarizable models based on ab initio data allows us to readily study different liquidliquid interfaces accurately and efficiently.大氣中懸浮的微小水滴對氣體的吸附/吸收可改變大氣的理化性質，從而影響氣候和空氣質量。為更好地理解氣體的吸附/吸收過程，我們使用分子動力學（MD）模擬研究了水蒸氣以及三種常見氣體（五氧化二氮，二氧化硫和臭氧）在水滴（曲面）和水板（平面）上的吸附/吸收。我們發現氣體溶解自由能的計算值與水滴/水板中的水分子數的倒數呈線性正相關。相較於水板模擬，水滴模擬得到的溶解自由能的外推結果與實驗值更吻合。通過分析勢能變化和熵變對自由能變化的貢獻，我們可以从分子层面分析气体吸附/吸收的內在驱动力。其中，氣體吸附/吸收導致的勢能變化主要由氣體分子周圍的局部效應決定，而其對熵變的影響則是非局部的。氣體吸附/吸收過程的熵變與水滴中水分子的移動性密切相關。在深入了解純水表面氣體吸附/吸收過程的基礎上，我們進一步探究了大氣中常見有機物在水滴表面形成的薄膜對氣體吸附/吸收的影響。為此，我們對氣體在有苯甲酸或月桂酸塗層的水滴上的吸附/吸收過程進行了平衡態及非平衡態的MD模擬，並計算了氣體吸附/吸收過程的自由能變化。通過分析分子間的相互作用，我們發現在苯甲酸塗覆的體系中，勢能變化主要受庫侖相互作用的變化影響。而在月桂酸塗覆的體系中，由於月桂酸具有長烷基鏈，蘭納-瓊斯相互作用的變化至關重要。在對熵變的分析中，我們發現其變化趨勢與水分子和有機分子的移動性之間有顯著的相關性。我們在非平衡態模擬中研究了有機塗層對氣體質量吸附係數的影響，並找到了質量吸附係數的變化與相應自由能極小值相對位置之間的聯繫。除了研究水-氣界面的氣體吸附/吸收之外，我們還探討了一種對液-液界面體系中的分子相互作用進行準確參數化的方法。液-液界面與許多現象都緊密相關，例如多組分體系的萃取與分離和相轉移催化。但目前常用的大多數力場都只針對單相物質進行了優化，而沒有考慮界面附近的分子相互作用。雖然從頭算分子動力學（AIMD）模擬可以基於第一性原理準確地描述分子相互作用，但其高昂的計算成本使其可模擬的體系尺寸及時間範圍都十分受限。因此，我們旨在通過相對熵的最小化對力場參數進行優化，從而在經典MD 模擬中複現AIMD 模擬的結果。為了更準確的描述界面性質，我們在力場中採用了德魯德模型來顯式地描述極化性。由於氯離子和四乙銨離子與水-二氯甲烷（或苯）界面在相轉移催化中的重要性，我們選取了它們作為測試體系。我們使用了GAAMP 等力場參數作為原始模型，經過牛頓法優化，相對熵顯著降低。這代表優化後的模型具有和AIMD目標更相似的狀態概率分佈。這種基於AIMD數據的可極化力場開發方法使我們能準確高效地研究各種液-液界面體系。Li, Wentao.Thesis Ph.D. Chinese University of Hong Kong 2020.Includes bibliographical references.Abstracts also in Chinese.Title from PDF title page (viewed on February 17, 2022)