1.Jiangsu Key Laboratory of Advanced Metallic Materials, School of Materials Science and Engineering, Southeast University, Nanjing 211189, China 2.School of Mechanical Engineering, Southeast University, Nanjing 211189, China
Fund Project:Project supported by the National Natural Science Foundation of China (Grant Nos. 51371051, 51771118) and the Key Laboratory of Advanced Metallic Materials of Jiangsu Province, China (Grant No. BM2007204).
Received Date:06 September 2018
Accepted Date:13 November 2018
Available Online:01 February 2019
Published Online:05 February 2019
Abstract:Superhydrophobic surfaces resulting from nanoarrays have good performance in anti-condensation. However, the study of droplet nucleation during water vapor condensation is a challenge because of the limitation of observation on a nanoscale, and therefore the fundamental understanding of the influence of geometrical parameters of nanoarrays on the condensation behavior is still less clear. In this work a three-dimensional (3D) multiphase lattice Boltzmann (LB) model is employed to simulate the phenomenon of droplet condensation on the superhydrophobic nanostructured surface. The model validation is carried out through the comparison of the simulations with the results from the Laplace's law and the intrinsic contact angle theory. The LB simulations accord well with the results from Laplace's law. The relative deviation between the simulated intrinsic contact angle and the theoretical value is less than 0.14%, demonstrating the validity of the LB model. Then, the 3D LB model is used to simulate the different preferential nucleation positions and final wetting states of condensate droplets by changing the geometrical parameters, including interpost space, post height and post width, and local wettability of the nanoarrays on superhydrophobic surfaces. It is found that for the nanostructured surfaces patterned with tall posts, the droplets nucleate in the upside interpost space and at the bottom of nanostructures simultaneously. By designing wider and thinner interpost spaces at the downside and upside of the tall nanostructures, respectively, the phenomenon of droplet nucleation at the bottom can be avoided. The simulation results show that the condensate droplets nucleated in the upside interpost space of tall nanostructures migrate upwards during growth, producing a Wenzel-to-Cassie wetting state transition. On the other hand, the condensate droplets nucleated at the bottom of nanostructured surface patterned with short posts produce the Wenzel state. However, by setting non-uniform hydrophilic and hydrophobic regions on the top of the short nanostructures, the condensate droplets are found to nucleate on the hydrophilic top and generate a Cassie state. The simulated final wetting states of condensate droplets on the nanostructures, having various geometrical parameters, compare reasonably well with the experimental observations reported in the literature. It is demonstrated that the migration of condensate droplets is correlated with the evolution of the statistical average force. If the direction of the statistical average force acting on the droplet is upward, the condensate droplets nucleated in the upside interpost space move upward during growth. The 3D LB simulations provide an insight into the physical mechanism of droplet nucleation, growth and wetting state transitions on superhydrophobic nanostructured surfaces. Keywords:nanostructure/ condensation/ superhydrophobic/ lattice Boltzmann method
全文HTML
--> --> -->
2.模型及算法本文采用D3Q19格式多相流伪势LBM模型[26]对固相表面的润湿现象进行模拟研究. 图1给出了D3Q19格式的离散速度示意图[27]. LBM演化方程可表示为 图 1 D3Q19 LBM模型离散速度示意图[27] Figure1. Schematic sketch of LBM discrete velocities in the D3Q19 scheme[27].
$\Delta P = {P_{{\rm{in}}}} - {P_{{\rm{out}}}} = \frac{{2\gamma }}{R}, $
式中Pin和Pout分别为液滴内、外的压力; $ \Delta P $为液滴内外压力差; $\gamma $为液滴的表面张力; R为液滴曲率半径. 设置计算区域为120 lu × 120 lu × 120 lu (lu为格子单位), 计算区域下边界设置为固相; 其他边界均采用周期性边界条件, 固体壁面采用反弹边界条件. 在区域正中心预置1个初始液滴, 在每次计算前改变初始液滴的半径R. 图2为分别设置流-流作用系数Gc = ?6.0, ?6.5和?7.0时, 三维多相流LBM模型模拟得到的液滴内外压力差$ \Delta P$与液滴曲率1/R之间的关系曲线. 由图2可见, 每条线的截距很小, 基本通过坐标原点, 与(11)式的形式相符, 表明本研究中采用的三维多相流LBM模型的模拟结果与Laplace定律预测结果符合良好. 图 2 模拟的液滴内外压力差与液滴曲率的关系(符号, 模拟值; 直线, 线性拟合) Figure2. Simulated pressure difference between inside and outside of a spherical droplet as a function of droplet curvature (symbols, simulated data; lines, linear fitting).
设置计算区域大小为123 lu × 123 lu × 200 lu, 纳米阵列的宽度、高度和间隙为W = 5 lu × 5 lu, H = 100 lu和S = 4 lu × 4 lu. 根据前期的二维模拟结果, 如果采用完全均匀的纳米结构几何尺寸, 冷凝的液相不是以液滴的形式而是以液膜的形式在纳米阵列上均匀析出[24]. 此外, 考虑到实际中纳米结构会有局部不均匀性, 因此在本模拟中, 在中部区域取较窄的间隙S = 3 lu × 3 lu. 在初始时刻, 模拟区域内充满密度均匀的水蒸气. 当LBM计算开始后, 由于模型上边界不断供应的水蒸气, 模拟区域内水蒸气的密度不断升高. 水蒸气粒子与固相表面产生流-固作用力使得水蒸气在纳米阵列的间隙处富集, 并在过饱和区域形核析出液滴. 图5为液滴在纳米结构粗糙表面上侧面和底部同时形核和生长过程的模拟结果. 约在21200 ts时, 水蒸气在较窄间隙上部和底部达到过饱和状态而形核. 由于阵列间隙上端的液滴更靠近上边界处的水蒸气来源, 其长大速度快于底部形核生长的液滴. 随后, 液滴吸收周围区域的水蒸气粒子, 位于上部的液滴同时向上和向下生长. 如图5所示, 大约到22000 ts时, 液滴上端从间隙中溢出并与周围相邻液滴合并. 之后, 位于间隙上部的液相随纳米阵列上方液相的生长而向上运动, 直到液滴完全位于纳米阵列的上方呈现Cassie态并继续长大, 即液滴的润湿状态由Wenzel态逐渐转变为Cassie态. 然而, 处于间隙底部的液相生长速度较慢, 并始终与间隙底部接触而呈现Wenzel态. 图 5 模拟的液滴在纳米阵列上部侧面和底部同时形核、生长及合并演化过程 Figure5. Simulated evolution of droplet nucleation, growth, and coalescence for the droplets that nucleate simultaneously in the upside space and the bottom corners between the posts of nanoarrays.
Zhang等[24,25]采用二维多相流LBM模型对液滴冷凝现象的模拟研究中, 当冷凝液滴在纳米阵列的侧面形核时, 没有观察到在阵列底部出现液滴核心.关于二维和三维模拟结果的这一差别, 我们分析认为三维模拟得到纳米阵列上部侧面和底部同时出现液滴核心这一现象与三维空间中的阵列间隙相互连通有关, 水蒸气粒子可在间隙的底部聚集并形核长大.根据水蒸气易在较窄的纳米阵列间隙聚集形核的现象, 为了避免液滴在底部形核, 将模拟区域中部较窄的纳米阵列间隙值进行重新设置: 当0 ≤ H < 20 lu时, 阵列间隙尺寸取为4 lu × 4 lu; 当20 lu ≤ H ≤ 100 lu时, 中部阵列间隙尺寸仍保持3 lu × 3 lu. 其他模拟条件与图5一致. 图6为液滴在上述尺寸的纳米结构表面上部侧面形核和生长过程的模拟结果. 可以看出, 改变纳米阵列的几何尺寸后, 阵列底部没有发生液滴的形核. 而上部侧面液滴的冷凝过程与图5中的相似: 水蒸气在中部上侧面间隙处聚集、形核、长大、合并, 间隙中的液相向上运动使其润湿状态发生改变, 即从Wenzel态逐渐转变为Cassie态. 将图6中在36000 ts时的冷凝液滴形貌与Lau等[17]在液滴冷凝实验中观察到的液滴形貌进行对比, 可以看出模拟得到的液滴的最终润湿状态与实验结果符合良好. 图 6 液滴在纳米阵列上部侧面形核、生长及合并演化过程的LBM模拟结果和实验结果[17]对比 Figure6. Comparison of LBM simulation and experiment[17] regarding the evolution of droplet nucleation, growth, and coalescence for the droplets that nucleate in the upside space between the posts of nanoarrays.
式中Fz(x, t)为液滴内每个流体粒子受到的竖直方向的作用力, 其值为(7)和(9)式计算得到粒子所受流-流作用力和流-固作用力在z方向的矢量和; 式中求和符号$\displaystyle\sum $的下标${x} \in S$表示统计平均作用力的计算只针对液滴内部的流体粒子. 由(15)式计算得到的${{{\bar{ F}}_z}} $为矢量, 正值和负值分别表示液滴所受到的统计平均作用力方向为向上和向下. 图7为图6中位于上部间隙的液相在21200—36000 ts所受到的统计平均作用力, 图7中标号c, d, e, f, g分别对应图6中的不同时间步. 可以发现, 在上述润湿状态转变时段, 液相所受的统计平均作用力的方向均向上, 使得间隙上部的液相向上运动直至液滴完全位于纳米阵列的上方, 完成由Wenzel态向Cassie态的转变. 图7的统计平均作用力曲线开始均呈缓慢增大趋势, 此时位于间隙上部的液相向上运动缓慢; 随着位于纳米阵列上方的液滴不断长大, 在约30800 ts时统计平均作用力开始急剧增大, 此时位于纳米阵列顶端的液相发生合并, 间隙中的液相加速向上运动完成润湿状态的转变, 使液滴呈现Cassie态并继续长大. 图 7 对应于图6中间隙上部的液相在润湿状态转变阶段所受的统计平均作用力随时间的变化 Figure7. Statistical average force of the condensate liquid in the upside space between the posts of nanoarrays in Fig. 6 during wetting state transition as a function of time.
随后, 对冷凝液滴在纳米阵列底端形核和生长过程进行模拟研究. 设置计算区域大小为144 lu × 144 lu × 100 lu, 纳米阵列的宽度、高度和间隙为W = 5 lu × 5 lu, H = 21 lu和S = 7 lu × 7 lu, 但中部区域纳米阵列的间隙设为S = 3 lu × 3 lu. 其他模拟条件与图5相同. 图8为液滴在纳米结构表面底部形核和生长过程的模拟结果. 从图8可见, 在冷凝初期, 水蒸气在相邻纳米阵列的间隙处富集. 约在6800 ts, 液滴在较窄的间隙底部优先形核. 随着液滴的不断长大, 相邻小液滴在底部开始合并. 随后, 小液滴吸收周围区域的水蒸气粒子沿纵向和横向生长, 横向生长方向受相邻液滴合并前“竞争生长”的显著影响. 大约到8400 ts时, 液滴从纳米阵列上端溢出. 之后, 随着液滴的生长, 相邻液滴间相互接触而发生多次合并. 约在20000 ts, 所有小液滴合并成一个大液滴, 随后继续长大. 同时, 在冷凝初期向周围间隙处横向生长的液相逐渐向液滴下方处汇聚, 形成一个稳定的Wenzel态液滴并继续长大. 将图8中液滴在冷凝过程中的演化形貌与Chen等[30]在液滴冷凝实验中观察到的液滴形貌进行对比, 可以看出模拟得到的液滴最终润湿状态与实验结果符合良好. 图 8 液滴在纳米阵列底部形核、生长及合并演化过程的LBM模拟结果和实验结果[30]对比 Figure8. Comparison of LBM simulation and experiment[30] regarding the evolution of droplet nucleation, growth, and coalescence for the droplets that nucleate in the bottom corners between the posts of nanoarrays.