Moving contact line problem: Advances and perspectives

The solid-liquid interface, which is ubiquitous in nature and our daily life, plays fundamental roles in a variety of physical-chemical-biological- mechanical phenomena, for example in lubrication, crystal growth, and many biological reactions that govern the building of human body and the functioning of brain. A surge of interests in the moving contact line （MCL） problem, which is still going on today, can be traced back to 1970s primarily because of the exis- tence of the ＂Huh-Scriven paradox＂. This paper, mainly from a solid mechanics perspective, describes very briefly the multidisciplinary nature of the MCL problem, then summarizes some major advances in this exciting research area, and some future directions are presented.

Molecular dynamics simulation of peeling a DNA molecule on substrate

Molecular dynamics （MD） simulations are performed to study adhesion and peeling of a short fragment of single strand DNA （ssDNA） molecule from a graphite surface. The critical peel-off force is found to depend on both the peeling angle and the elasticity of ssDNA. For the short ssDNA strand under investigation, we show that the simulation results can be explained by a continuum model of an adhesive elastic band on substrate. The analysis suggests that it is often the peak value, rather than the mean value, of adhesion energy which determines the peeling of a nanoscale material.

Estimating the thickness of diffusive solid electrolyte interface

electrolyte. The properties of lithium-ion (Li-ion) battery, such as cycle life, irreversible capacity loss, self-discharge rate, electrode corrosion and safety are usually ascribed to the quality of the SEI, which are highly dependent on the thickness. Thus, understanding the formation mechanism and the SEI thickness is of prime interest. First, we apply dimensional analysis to obtain an explicit relation between the thickness and the number density in this study. Then the SEI thickness in the initial charge-discharge cycle is analyzed and estimated for the first time using the Cahn-Hilliard phase-field model. In addition, the SEI thickness by molecular dynamics simulation validates the theoretical results. It has been shown that the established model and the simulation in this paper estimate the SEI thickness concisely within order-of-magnitude of nanometers. Our results may help in evaluating the performance of SEI and assist the future design of Li-ion battery.

Bridging length and time scales in moving contact line problems

Physical,mechanical,chemical,and biological processes for various problems span length and time scales of many orders of magnitude[1-6].As a typical dynamic process,moving contact line(MCL)problem or droplet spreading consists of7-8 length and time scales from the atomistic to the continuum[2],as shown in Figure 1.A grand challenge in MCL problems is to link these vastly different length and time scales.All these length and time scales origin from the competition among governing forces[7],which indicates that they

Solute transport and interface evolution in dissolutive wetting

Dissolutive wetting, i.e., droplet wetting on dissolvable surfaces, is essential for various natural phenomena and industrial applications such as the formation of sinkholes, enhancing shale gas recovery, drug design, MEMS, and so on. It is difficult to predict the evolution of concentration field and solid-liquid interface owing to the coupled effects of wetting, diffusion, and convection. This study makes substantial progress by proposing a new theory based on Onsager’s variational principle and finding two modes of solute transport, i.e., shifting and lifting modes. Furthermore, we investigate the influence of wetting and dissolution coupling on the interface shape using a phase diagram. Using our theory, we can predict and inversely predict the interface evolution.

Preliminary Study on Adapting ProtoPNet to Few-Shot Learning Using MAML

ProtoPNet proposed by Chen et al.is able to provide interpretability that conforms to human intuition,but it requiresmany iterations of training to learn class-specific prototypes and does not support few-shot learning.We propose the few-shot learning version of ProtoPNet by using MAML,enabling it to converge quickly on different classification tasks.We test our model on the Omniglot and MiniImagenet datasets and evaluate their prototype interpretability.Our experiments showthatMAML-ProtoPNet is a transparent model that can achieve or even exceed the baseline accuracy,and its prototype can learn class-specific features,which are consistent with our human recognition.

Analyzing Interpretability Semantically via CNN Visualization

Deep convolutional neural networks are widely used in image recognition,but the black box property is always perplexing.In this paper,a method is proposed using visual annotation to interpret the internal structure of CNN from the semantic perspective.First,filters are screened in the high layers of the CNN.For a certain category,the important filters are selected by their activation values,frequencies and classification contribution.Then,deconvolution is used to visualize the filters,and semantic interpretations of the filters are labelled by referring to the visualized activation region in the original image.Thus,the CNN model is interpreted and analyzed through these filters.Finally,the visualization results of some important filters are shown,and the semantic accuracy of filters are verified with reference to the expert feature image sets.In addition,the results verify the semantic consistency of the same important filters under similar categories,which indicates the stability of semantic annotation of these filters.