Theory And Simulation Study On The Interface Behaviors Of Nano-fluids

The interface, which develops when two different phases of matter contact, exists extensively within both nature and industry. An interface often has one fluid on one side (e.g. the solid-liquid interface), or two different fluids on two sides (e.g. the liquid-gas interface). Microscopically,the forces acting on the fluid particles within the interface are different from that in the bulk, so the interface fluid has different properties than the bulk fluid. In many classical theories, the interface is treated as a mathematical surface at where the fluid properties have sudden changes.Based on this simplification, physical concepts such as the surface tension,the surface free energy and the boundary slip length are defined to rule the interface effect on the fluid behaviors. The important of the interface depends on the size of the system: the smaller size, the bigger surface-to-volume ratio the system has, the more effect the interface will have.Meanwhile, many theoretical and experimental studies suggest that the microscopic properties of the interface cannot be ignored for the fluid at nanoscale, such as the nucleus in the vapor-liquid nucleation and the fluid in the micro-/nano-fluid system. For example, the interface is a finite-thickness region in which the fluid properties change smoothly. The surface tension significantly depends on the curvature of the interface. In the vapor-liquid-solid system, the line tension effect often cannot be ignored.The nanoscale heterogeneities on the solid surface may have significant influences on the fluid behaviors such as the vapor-liquid nucleation and the nanobubble dynamics. In recent years, with the development of experimental techniques, many new microscopic interface behaviors (e.g.stable surface nanobubbles) have been observed, but have not received reasonable theoretical explanations. Moreover, at present it is still difficult to explore such interface behaviors in experiments, especially the dynamic behaviors.Therefore, in this thesis, the mean field lattice density functional theory (LDFT) and molecular dynamics (MD) simulation method are applied to study serveral interface behaviors of nano-fluids. The main contents are as follows.1. The LDFT is applied to investigate the heterogeneous vapor-liquid nucleation and the line tension of the nanodroplet on the flat solid surface.The LDFT calculations show: firstly, for the vapor-to-liquid nucleation induced by the nanoscale solid particles, there exist three different structures for the critical nucleus: the spherical structure for partial drying situations, the spherical cap structure for partial wetting situations, and the core-shell structure for complete wetting situations. Meanwhile, the nucleation barrier in general decreases with increasing the particle size or/and increasing the particle wettability. Whereas, for complete wetting situations, the pre-existing liquid film weakens the effect of increasing particle wettability. Besides, particles that deviate from the sphere can decrease the nucleation barrier. Secondly, for the vapor-to-liquid nucleation induced by the nanopore with one open end, the presence of the nanopore results in an intermediate state, which divides the whole phase transition process into two sequential sub-processes, i.e., the pore filling and the phase transition outside the pore. Depending on the chemical potential. (i.e. the supersaturation), the solid-fluid interaction strength and the pore size, there exist six different phase transition mechanisms: the vapor spinodal, single barrier with nucleation outside the pore, double barriers controlled by nucleation outside the pore, double barriers controlled by nucleation inside the pore, single barrier with nucleation inside the pore, and homogeneous nucleation. In addition, the presence of the nanopore may change the morphology of the critical nucleus from the counterpart on the smooth substrate. Lastly, the line tension of the nanodroplet on the homogeneous solid surface is always negative with maximum absolute value appearing at the moderate wettability. The chemical potential dependence (or equivalently, the droplet curvature dependence) of the interface tensions and line tension effect are crucial for the viability of the Young’s equation at nanoscale. In particular, the linear relationship between the cosine of the contact angle and the curvature of the contact line is incorrect at nanoscale.2. The LDFT and MD simulation method are applied to study the stability mechanism and the formation process of the surface nanobubble.Firstly, for the stability mechanism, the LDFT calculations show: the three-phase contact line pinning effect, which results from the intrinsic nanoscale physical roughness or chemical heterogeneities on the solid surface, leads to stable surface nanobubbles. The surface nanobubble with pinned contact line stays at a thermodynamical metastable state. The classical nucleation theory can explain the stability of the surface nanobubble, and predict their size and contant angle. The solid surface chemistry and the local feature of the heterogeneities together affect the substrate’s ability to pin the contact line, and therefore determine whether or not surface nanobubbles are stable.However, for stable nanobubbles, their morphologies (e.g. the contact angle) are independent of the solid surface chemistry and the local feature of the heterogeneities, but depend on the nanobubble size. On the other hand, MD simulations show: for the the solid surface with a nanopore immered in the liquid, the system with undersaturated and saturated liquids stays at Wenzel (liquid being in full contact with the solid surface) or Cassie (liquid being in contact with the peaks of the rough surface) wetting state, respectilvey; the system with the moderate supersaturated liquids contains stable surface nanobubbles, and their curvature radius and contact angle decrease with the increasing of the supersaturation; the liquid-to-vapor/gas phase transition occurs in the system with high supersaturated liquids. Therefore, stabilizing nanobubbles requires both the contact line pinning effect and the supersaturation (superheating or gas supersaturation).Secondly, for the formation of the surface nanobubble on the rough hydrophobic surface experiences a two-step nucleation route involving an intermediate state: the system transforms from the Wenzel state to the Cassie state after gas cavities occur in the grooves; then, the gas cavities coalesce and form a stable surface nanobubble with a pinned contact line.Additionally, the free energy barriers for the two transitions show opposing dependencies on the degree of surface roughness, indicating that the surfaces with moderate roughness are favorable for forming stable surface nanobubbles.3. The MD simulation method is applied to reveal the microscopic mechnisims for the solutal Marangoni effect and the diffusio-osmosis.Firstly, for the solutal Marangoni effect, the statistical mechanics suggests that the surface force, which drags the fluid and induces the flow, involves the local pressure gradient from the nonuniform pressure tensor distribution in the system, while the thermodynamics suggests that the surface force comes from the chemical potential gradient due to the concentration variation. Simulation results show that both two expressions satisfy that the total surface force is equal to the surface tension gradient,but they lead to significantly different force distributions within the interface. The flow profiles obtained by applying the chemical potential gradient surface force on the fluid are consistent with the profiles measured directly in the non-equilibrium simulations. However, when the surface force is calculated by computing the local pressure tensor distribution, the flow profiles are very different. These facts demonstrate that the surface force should be calculated through computing the chemical potential gradient of each species instead of computing the local pressure tensor distribution. Secondly, for the diffusio-osmosis on a flat solid surface, the Gibbs-Duhem equation shows that when the concentration gradient exists,each fluid particle experiences a force equalling to the negative of the corresponding chemical potential gradient. These forces develop a pressure difference in the interfacial region, and therefore lead to a macroscopic flow. On the other hand, this pressure difference can be obtained by computing the local pressure tensor distribution in the system. MD simulations still show that the driving force should be calculated through computing the chemical potential gradient of each species instead of computing the local pressure tensor distribution.