Characterizing Rock Microstructure Using A Gridbased Electron Trajectory Simulation Approach

Three-dimensional characterization of materials' microstructure using electron trajectory Monte Carlo simulation has attracted wide interest in several research fields.Comparing with traditional characterization methods,this method resolves the microstructure information in electron scattering signals by electron trajectories simulation,which has the advantages of high imaging accuracy,fast imaging speed,non-destructive characterization and low cost.However,due to the obvious shortcomings of traditional electron trajectory Monte Carlo simulations in dealing with complex heterogeneous microstructures of rocks,this method is still not applicable to rock microstructure characterization.In order to reveal the complex relationship between electron scattering signal and rock microstructure,and promote the application of this method in rock microstructure characterization,this paper develops a grid-based electron trajectory Monte Carlo simulation program and uses it to study the influence of common heterogeneous features on the electron scattering signal,building the foundation for further development of this method.(1)The grid concept is introduced into the traditional electron trajectory Monte Carlo simulation theory,and a grid-based electron trajectory Monte Carlo simulation program with efficient parallel operations on a graphics processing unit is developed.The program can simulate the motion of electrons in the complex heterogeneous microstructure of rocks,as the continuous variation of rock inhomogeneous microstructure in space are described with a highdimensional Cartesian sampling grid,and the motion of electrons in rock microstructure is described with a grid of electron free path.The simulation results are compared with those of the CASINO,showing that the theoretical and programming implementation of the grid-based electron trajectory Monte Carlo simulation is correct.(2)Electron interaction volume sizes and scattering signal intensities in rock-forming minerals are important for quantitatively interpreting characterization signals of rock in scanning electron microscopy.By simulating electron trajectories in homogeneous microstructures,the size of electron interaction volume and sampling depth of backscattered electron in common sedimentary rock-forming minerals are investigated.We found the nonlinear relationship between the maximum electron penetration depth and accelerating voltage.It is also found that backscattered electrons have a large sampling depth in rockforming minerals and are suitable for internal rock microstructure characterization.(3)By simulating the electron trajectories in heterogeneous microstructures,the effects of common inhomogeneous features on the electron interaction volume and backscatterred electron signal are investigated.It is found that both the diffusion zone between rock-mineral interface and complex microscopic pore structure can laterally expand the electron interaction volume and change the backscattering signal intensity,while the traditional Monte Carlo simulation cannot investigate the effects of these inhomogeneous features.(4)Based on different numerical microstructures models,electron trajectory simulations are used to reveal the complex coupling relationship between the microscopic pore structure of sandstone and the backscattered electron signal,and a probabilistic model describing this coupling relationship is established.We characterize the porosity of sandstone model and predicts uniaxial compressive strength using this probabilistic model.Comparing the probability model obtained from the homogenized equivalent porous sandstone model and the non-homogeneous porous sandstone model,it is found that the expectation values of porosity given by the traditional homogenized equivalent model and the non-homogeneous model are different,due to the inability of tradition model in considering the non-homogeneous features such as the spatial morphology of the pore structure.The paper has 45 pairs of figures,18 tables and 134 references.