The Research On The Modulation Of The Mechanical And Thermal Properties Of Low-dimentional Carbon Materials

Carbon nanomaterials have attracted enormous attention from researchers,because of their unique crystal structure,excellent various properties and promising applications in the fields of micro-nano electromechanical systems.Modulating the performance of carbon nanomaterials,expanding their application range and meeting the demands in new application have always been the research focus in nanoscale research.In particular,effectively modulating the mechanical properties and thermal transport of carbon nano-materials is of great significance for the application of carbon nanomaterials in flexible electronic devices,energy conversion,protective structures and reinforcement of com-posite material.Previous studies have shown that the mechanical properties and thermal transport of carbon nanomaterials can be modulated by defect engineering,functional modification and encapsulation.In this paper,molecular dynamics and machine learn-ing are used to further study the thermal transport and mechanical properties of carbon nanomaterials.We find that oxidation,the degree of crumpling,fullerene encapsulation and structural design have an important influence on the mechanical properties and thermal transport of carbon materials.The main contents are as follows,(1)The oxidation effect on the Poisson s ratio in graphene is studied by molecular dynamics simulation,demonstrating that the de-wrinkle effect is the underlying mecha-nism for the reduction of the Poisson's ratio in graphene,which provides a new strategy for modulating the Poisson's ratio.Specifically,we calculate the Poisson's ratio of graphene with different degrees of oxidation,and focus on identifying the relationship between the Poisson's ratio and the degree of oxidation.We find that the Poisson's ratio decreases linearly from positive to negative with the increase of the degree of oxidation in the range p ?[0,1.0],where the valley value of the Poisson's ratio is-0.567 for fully oxidized graphene(i.e.,p=1.0).The exact degree of oxidation at which the Poisson's ratio becomes negative is about p=0.27.The underlying mechanism for the reduction of the Poisson's ratio is the suppression of the oxidation-induced ripples during the stretching of the graphene,which is also called de-wrinkle effect.The temperature effect on the Poisson's ratio is also studied for graphene of various degrees of oxidation.(2)The effect of degree of crumpling on the stability of crumpled graphene is studied by molecular dynamics simulation,demonstrating that the critical degree of crumpling is closely related to the self-adsorption phenomenon,which provides a theo?retical reference for obtaining stable crumpled graphene.Specifically,we simulate the crumpling process of graphene under hydrostatic compression and biaxial compression.We find that in both cases,graphene exhibits a critical degree of crumpling of about 0.5 or 0.55 for hydrostatic and biaxial compression,above which graphene is irreversibly crumpled after the removal of external constraints.The critical degree of crumpling is closely related to the self-adhesion phenomenon of graphene,which leads to a step-like decrease in the adhesion energy.The MD simulation results are agreement with the analytic solutions of the model derived from the competition of balancing bending and adhesive energies to determine the critical degree of crumpling(3)The effect of the fullerene encapsulation on the thermal conductivity of the single-walled carbon nanotube(SWCNT)is studied by molecular dynamics simulation,revealins that the interaction between fullerene and SWCNT is the main factor affecting the thermal conductivity of SWCNT,which provides a new method for modulating the thermal conductivity of SWCNT.Specifically,we calculate the thermal conductivity of SWCNTs with different diameters after filling with fullerenes.We find that the fullerene encapsulation can greatly reduce the thermal conductivity of the narrower SWCNT(n,n)with n=8 and 9,which is attributed to the increased phonon-phonon scatterins and the decreased acoustic velocities due to the strong interaction between the outer nanotube shell and the encapsulated fullerene.In contrast,the fullerene encapsulation sliehtly enhanced the thermal conductivity of the thicker SWCNT(n,n)with n=10 and 11,owing to the additional phonon conduction channels contributed by the encapsulated fullerenes.Our simulations may assist in the clarification of the discrepancy between previous numerical predictions and experimental observations.(4)Inverse design for the nanoporous graphene structure with large thermal re-sistance is performed by applying the machine learning method,revealing that the convolutional neural network can quickly and efficiently extract the pore arrangement characteristics making the nanoporous graphene thermal resistance increase based on a small amount of data,which provides new ideas of applying the machine learning to the research of structural design and modulating the thermal conductiviy of nanoporous graphene.Specifically,we test the performance of different convolutional neural net-work modes in predicting the thermal conductivity of nanoporous graphene.The results show that a convolutional neural network with three convolutional layers and one fully-connected layer can effectively extract the features of nanoporous graphene,and make an accurate prediction of nanoporous graphene(compared with the results from MD).The inverse design of the convolutional neural network model with search algorithm on the system with small sample space shows that the convolutional neural network can find the optimal structure with few training data.The inverse design of the convolutional neural network model on the system with large sample space shows that main feature of nanoporous graphene with larger thermal resistance is that the pores are arranged in a row perpendicular to the direction of heat flow.Our research on the effect of oxidation,degree of crumpling,fullerene encapsula-tion and structural design on the mechanical properties and thermal transport of carbon nanomaterials provides not only new strategies for modulating the mechanical properties and thermal conductivity of carbon nanomaterials,but also a prototype for applying the convolutional neural network method to the structural design of nanomaterials,which provides a new idea for the research in the holey nanostructure.