Study Of Solid-liquid Phase Transition Of Nanoparticle And Bulk Fe With Molecular Dynamics Simulation

In this thesis, the thermaldynamic properties and structural transformation during solid-liquid phase transition of nanoscaled Fe have been studied. For Fe nanoparticles, previous studies mainly focused on the variations of thermaldynamic properties and structural transformation in solidification, melting and sintering processes. However, to our knowledge, few researchers have systematically investigated (i) the relaitonship between thermaldynamic properties of nanoparticles and bulk materials and (ii) quantitative analysis of structural evolution. Moreover, different with bcc-Fe, less attention have been paid on fcc-Fe, which is proved stable at room temperature in nanoscaled material in previous studies.On the basis of above considerations, the detailed solid-liquid transition process of nanoscaled Fe has been studied by molecular dynamics (MD) in this thesis. Cluster type index method (CTIM) has been employed for structure analysis. The variation of thermaldynamic properties and detailed structural evolution during liquid-solid transition process of Fe nanoparticles have been discussed. Furthermore, the detailed structural evolution during solidification process of bulk Fe have been studied.The major contents of this dissertation are epitomized as follows:1. Thermodynamics analysis of Fe nanoparticles during solidification and meltingIt is well-known that size effects may significantly affect thermodynamic properties of nanoparticles. The melting point of small nanoparticles is lower than that of larger ones due to the existance of surface energy. In1909, Pawlow investigated the relationship between melting point and particle size and modeled it The follow-up studies modified this model several times and two main versions, and have been derived. Both of these two models describe the relation between the difference of melting points of nanoparticle and bulk Fe and the size of nanoparticle. In some specific assumptions the difference is caused by the two variables1/R and N-1/3of the linear relationship. Our results indicate that the latter model is more accurate in predicting the melting points Tm/cluster, solidification points Ts/cluster and equilibrium points Tequilibrium/cluster of Fe nanoparticles in the range of Fe113～Fe9577. The equilibrium point of bulk material Tequilibrium(∞)=1833.3K obtained by fitting procedure is strikingly close to the experimental data of bulk Fe (1811K) and is much better than previous results.Moreover, the cooling/heating rates (0.05-25K/ps) also affected the thermaldynamics properties of nanoparticles. In low cooling/heating rates (0.05～0.25K/ps), nanoparticles have enough time to relax their structures and thus the melting points, solidification points and equilibrium points fluctuate in a small range. The melting point obtained in this condition is closer to experimental results.2. Structural evolution of Fe nanoparticles during solidification and meltingCTIM-2has been employed to analyze the structural evolution of Fe nanoparticles during nucleation and melting process. The configurations obtained after solidification indicate that the solid phase in nanoparticles is mainly consisted of fcc and hcp atoms. Two main structure, fivefold twins and lamellar structure, have been found. In common belief, Fe belongs to bcc structures in ambient pressure and temperature and transforms to fcc structures at the temperature higher than1185K. Actually, researches in resent years showed that nanoscaled fcc-Fe could be stable in ambient conditions, which is in accordance with our results.At the early stage of nucleation of fivefold twins, we found that they are built by continuous growth of several hcp layers with the included angle fixed. Severaladjacent fivefold twins appeared and accompanied with formation of dislocations. Atthe stage of relaxation, the twinning boundaries and axes adjusted with each other andthe dislocations disappeared gradually while few internal defects remained due to itsstructural characteristics. These defects will significantly affect the melting behavior.With temperature increasing, there is an increasing number of defects insidenanoparticles, meanwhile surface premelting takes place. By the influence of bothexternal and internal defects, the nanoparticle becomes meltdown finally.For lamellar structures, a layer by layer growth mechanism has been found atearly stage of solidification. After the formation of nucleus with hcp layers, solid-likeatoms born from liquid attached to nucleus. This will form the new hcp layer parallelto the old ones. Few defects are observed in final configuration. For this reason,lamellar structure is more stable in heating process than fivefold twins. With theincreasing temperature, inner defects appeared accompanying to surface premeltingand both of them accelerated the melting of nanoparticles.The structural heredity of nanoparticles during nucleation has been found. Inother words, the structures of embryos in early stage determine the final structures toa great extent, e.g. whether it forms fivefold twins and lamellar structures. For smallnanoparticles, the magic number determines their shape and structures, and thus theirenergy minimized. However, in our simulations, completely different structuresmaybe caused by a slight change of simulation conditions. This is related with therandomness of choosing structures. With the increasing sizes of nanoparticles, therandomness enhances and thus the same simulation conditions may produce differentresults.3. Structural evolution of bulk Fe during solidification processIn the simulations of bulk Fe materials, monocrystal, fivefold twins and lamellarstructure have been identified during nucleation process and their structural evolutions have been discussed in detail. Significant dependences are lacking for final atomicconfiguration on the simulation temperature, which indicates the existence ofrandomness of structure choosing. The growth mechanisms of fivefold twins andlamellar structure are similar with that in nanoparticle, while monocrystal experiencetwo obvious stages. At the early stage of nucleation several nucleus are formed andgrow, when they coalesce with each other through lots of hcp planes. At the stage ofstructure optimization, most hcp planes disappear and monocrystal finally form.Inside the irregular shape of the nucleus of fivefold twins, the twinningboundaries, consisting of hcp atoms, joint with each other. Several tetrahedra joinedtogether to form fivefold twins, which are formed by9twinning axes. Due to theproportion of amorphous atoms is higher than67%, temperature disturbance andcontinuous cooling process have been applied. Further study found it is difficult tosignificantly affect the atomic percentage of fivefold twins with small temperaturedisturbance (±10K), while continuous cooling process (cooling rate0.25K/ps) willlead to a remarkable change of structure and after that nearly one fourth atomsremained amorphous.