Integrated Computational Research Of Molecular Dynamics For Hardening-softening Of Al_0.3CoCrFeNi High Entropy Alloy During Nanoindentation

The Al_0.3CoCrFeNi high entropy alloy has garnered significant interest owing to its stable phase structure and exceptional mechanical properties.Experimental studies have been conducted to explore its nanoindentation behavior.However,the competition and balance mechanism between material softening caused by plastic deformation and work hardening resulting from dislocation surge has not been fully clarified in the experiment,and this problem needs further research and exploration.Although nanoindentation can quickly and accurately measure the hardness of high-entropy alloys,and molecular dynamics can be used to assist in the design of high entropy alloys with high hardness,it is still a great challenge to screen high-entropy alloys with high hardness in the huge composition space of high-entropy alloys.Although high-throughput experiments,detection,and calculation methods greatly simplify this process,high cost is still an important problem.Therefore,how to quickly,accurately,and low-cost screening of high hardness high entropy alloy is also an urgent problem to be solved.In this paper,the dependence of nanoindentation behavior on crystallographic orientation and the hardening-softening mechanism of Al_0.3CoCrFeNi single crystal high entropy alloy has been studied through molecular dynamics.The impact of twin boundary spacings on the hardening-softening mechanism of Al_0.3CoCrFeNi nanocrystalline high entropy alloy was studied binding phase-field-crystal method.The coalescence kinetics and nanoindentation response of Al_0.3CoCrFeNi polycrystalline high entropy alloy during sintering were studied by binding sintering phase field method.High entropy alloys with high hardness are predicted based on composition by machine learning combined with experiments,and phase formation is predicted and verified.The main conclusions are as follows:（1）The prismatic dislocation loop with independent nucleation is found in the[111]orientation,while the formation of stacking fault tetrahedra without the participation of Frank dislocation loops can be observed in the[101]orientation.Dislocation rosette caused by dislocation crossing inhibits dislocation motion and leads to work hardening.Free dislocation slip and dislocation annihilation can accommodate plastic strain and reduce indentation resistance,resulting in plastic softening.（2）In the Al_0.3CoCrFeNi high-entropy alloy with different twin boundary spacings,the relationship between hardness and twin boundary spacing can be expressed by Hall-Petch and reverse Hall-Petch relationship.There are two reasons for the softening effect caused by twin boundaries in the later stages of plastic deformation:（i）the formation of steps and the partial slip where the slip plane and Burgers vector are parallel to the twin boundary;（ii）steps and local damage zones in twin boundaries become new nucleation sites for dislocations.Phase-field-crystal simulations confirm that this alloy has the highest hardness when the twin spacing is 2.467 nm.（3）In the polycrystalline high entropy alloy Al_0.3CoCrFeNi,the maximum indentation force decreases as the grain size reduces.Grain boundaries are destroyed in the process of nanoindentation,and the indentation stress can assist grain boundary migration and grain growth.The entanglement of the dislocation under the indenter and at the grain boundaries can effectively prevent dislocation propagation,thus promoting hardening.After plastic deformation,grain boundary will become the main factor leading to softening effect:（i）the dislocation is blocked by grain boundary and slides along the grain boundary;（ii）The grain boundary absorbs the dislocation without requiring dislocation transport;（iii）The dislocation is transferred to adjacent grains;（iv）The absorbed dislocation in the grain boundary leads to the re-emission of the dislocation in the adjacent grain;（v）The dislocation that is re-emitted glides along the grain boundary and is once again absorbed by the grain boundary.（4）The pressure accelerates the sintering kinetics and makes the particle contact zone further evolve into a power exponential creep zone.The phase field simulation shows that the pores disappear and the grains fuse,and the regular and stable grain boundaries are formed under the action of chemical free energy and elastic energy,and the stress distribution is relatively uniform.Uniaxial tensile test results show that Young’s modulus of the sintered sample is 214.11±1.03 GPa,and the deviation from the experimental value is only 0.82%.In the process of nanoindentation,dislocation entanglement under the indenter and at the grain boundary can effectively prevent dislocation propagation and promote hardening.After plastic deformation,the twin and layer faults present in the grain boundary and sintered body will emerge as the primary factors contributing to the softening effect:（i）The newly formed dislocation under the indenter and the sintered body will be cross-linked and annihilate;（ii）The dislocation is absorbed by the grain boundary;（iii）The dislocation emitted by the grain boundary annihilates by cross-linking with the sintered body itself;（iv）The dislocation that is re-emitted glides along the grain boundary and is once again absorbed by the grain boundary.（5）To accelerate the discovery of high hardness and high entropy alloys,an alloy design system based on high entropy alloy composition prediction is proposed.The model’s coefficient of determination on both the training and validation sets exceeds 0.9,with a Pearson correlation coefficient R of 0.95,suggesting excellent performance.Inverse projection design and high throughput screening were used to guide the composition optimization of high entropy alloys.The hardness of Al_1.2Cr_17.42Fe_25.42Ni_28.32Ti_27.62 high-entropy alloy is up to 869.88 HV,which is21.15%higher than the highest hardness in the original dataset of the Al-Cr-Fe-Ni-Ti system.（6）The high entropy alloy phase formation prediction is realized by composition.Compared with traditional machine learning,this strategy omits the tedious feature calculation and screening process.Through training,the prediction accuracy of intermetallic compounds is more than 90%,the prediction accuracy of solid solution and amorphous phase is more than98%,and there is no underfitting and overfitting phenomenon.