分子动力学

nullP-V phase diagram&MSD12 27 2012P-V phase diagram&MSDGroup Members : Jiang Qingdong ; Zheng Feipeng ; Hou Huasheng ; Shen Huaze; Duan Wenye ;Song Lingshuang;Two StoriesTwo StoriesFirst Story :P—V Phase Diagram(Expected )Second Story :MSD Simulation(unexpected)But all unexpected is reasonable!null The First StoryP—V Phase DiagramP—V Phase DiagramWe use the lammps to simulate a system with L-J potential, and we want to find out the relation between the pressure and the volume at different temp. => At first we tried to simulate the system by fixing NPT , however we realized that we are unable to draw the loop out ,because for one pressure there are two or more possible volume . So our method is to use NPT to draw the phase diagram roughly, then draw the detail of this diagram by fixing NVT. Phase DiagramPhase Diagram1. NPT diagram Phase DiagramPhase Diagram2.Fix NVT for drawing the loop Phase DiagramPhase Diagram Conclusion 1. We find that the critical point is at around 1.3 2. We find that the loop will appear under 1.3 This is what we expected! null The Second Story MSD of Cu Alloy SystemMSD of Cu Alloy SystemMSD Result Our theory on Diffusion CoefficientExplanation of the simulation result MSD Change With Time MSD Change With Time System: Cu Alloy Temp: 300K----2000K Time: scale 1fs/step The Configuration of the SystermThe Configuration of the Systerm First ,show some picture by using VMD 400K 1200K 2000K The Configuration of the SystermThe Configuration of the SystermSecond, we show the RDF of the system Theory on Diffusion CoefficientTheory on Diffusion CoefficientExplanation for the ResultExplanation for the Result 第N+1步和第N步的关系式 Explanation for the ResultExplanation for the Result于是会有 而温度与频率有关系 所以有 Explanation for the ResultExplanation for the Result 1. Experience Vs Simulation Our experience told us there should be phase transition at around 1200 K , but from the configuration we find there is no phase transition. So, which is right? 2 . Experience Vs Theory Our theory told us a should be different in different phase ,from the result of simulation we think there is no phase transition around 1200K . nullWho is the winner in the battle? Experience , Simulation or Theory?Explanation for the ResultExplanation for the ResultAll because that we cool down the system so fast that the system goes to an substable state at around 700K ---- glass state .We can find it from the volume-temp picture, as follows null Conclusion : Experience sometimes is misleading! The whole physical process is as follows(a is something like a parameter)The Whole ProcessThe Whole Process 1 Liquid (2000K) 2.Supercooled liquid(1200K) 3.Glass state (700K) That is the whole story .(unexpected but reasonable) null12 27 2012 Our special thanks goes to Pan Shaopeng, Sun Zhaoru and Sun Gang for their fruitful discussion and helpful advices! Thanks Zheng Feipeng for his attentive labor in simulating the phase diagram; Thanks Hou Huasheng for drawing a lot of pictures! Thanks our team members for helpful discussions!null# EAM potential for FeP system units metal boundary p p p atom_style atomic lattice fcc 3.62 region box block 0 12 0 12 0 12 create_box 1 box create_atoms 1 box pair_style eam/alloy pair_coeff * * Cu_zhou.eam.alloy Cu neighbor 2.0 bin neigh_modify delay 10 timestep 0.001 thermo_style custom step temp vol lx ly lz thermo 100 compute 1 all pe/atom compute 2 all ke/atom variable s equal stagger(10000,9999) dump 1 all custom 10000 *.lammpstrj id type x y z c_1 c_2 vx vy vz dump_modify 1 every v_s first yes velocity all create 2000.0 23456789 label loop variable a loop 18 variable temp0 equal "2000-(v_a-1)*100" compute 3 all msd fix 1 all npt temp ${temp0} ${temp0} 0.1 iso 0.000 0.000 0.75 fix 2 all ave/time 1 1 10 c_3[4] file ${temp0}K-msd.txt run 10000 write_restart ${temp0}K.restart unfix 1 unfix 2 uncompute 3 next a jump in.melt loop