• 検索結果がありません。

LatticeDefectsProblems Some Statics ApplicaitonsofMolecular Method

N/A
N/A
Protected

Academic year: 2021

シェア "LatticeDefectsProblems Some Statics ApplicaitonsofMolecular Method"

Copied!
6
0
0

読み込み中.... (全文を見る)

全文

(1)

OriginalPaper

Some Applicaitons of Molecular Statics Method to Lattice Defects Problems

Yutaka TAKAHASIlI

(Department of MechanicalEngineeI・ing)

(Received Novemberl,1993)

New Tnethod to calculate equilibrium state of a.tom system

bonded with pair‑POtential was developed. The met,hod was

applied to mechanical property problems such as uniaxial

t,enSion and,laしt,ice defect,prOblems such as dislocationin t,WO‑

dimensionall,ennard‑Jones crystals.These pI・Oblems were able to

be soIvable by the continuum elast,ic t,heory and t,he accordanc(ユ

and dit‑ference were discussed.

Key Words: equjlibrium sLructure,pajr‑pOLential,Principle of

minimum potentialenergy,Ⅰ,ennard‑Jones crystal

1.Introduction

Material̲S are COnVent′ionally tr・eated by conLinuum theory, howevel・,they

act,ually consist of a number of atoms. Because bonding among atoms doII】inates

macroscopic mechanical pr・Operties such as (チ1ast,ic behavioI・, plastic

deformation and fracture toughness,many Studies in the atomic scale has been

performed by means of computer simulation to understand and combine

macroscopic and microscopic worldin ma†.erials. As calculuses for such

simulation, many kinds of methods wer・e already developed such as 'molecular

dynamics methodl)'and'Monte Carlo method 2))' 'car‑Parrinello's method

3)'・is thelatest,trendin this field. Principles of these met,hod are based

on Newton's equation, Stat,istical mechanics and quantum mechanics,

respect.ively.

Recently, the present aut,hor has developed new method to calculat,e

(2)

equilibrium state at zero Kelvin based on the principle of minimum of

potentialenergy4)I Of which conceptis wellknownin structuralmechanics.

The method was termed

asIJnOlecular statics methodI.Although Isimulated

annealing method5)IIImodified MI)methodIandIsteepesrt‑descent

method,have

been frequently applied to such equilibrium state calculations,the new

methodis superior to them with respect to convergency and calculation time.

In this reportISOme applications were presented to show the validity of

the new method and the applicability to defect problems.

2.Theory

Formulation of the methodis briefly denoted,Since detailwas discussed

in my previous paper・In a word,itis quite similar to finite element

method.

⊥土⊥.̲̲Pri坤IpOtent主旦L̲旦旦邑E已又

Force balanceis equalto a state where potentialenergy His minimized;

H = Wext(◎:internalpotentialenergy,Wext:eXternalwork)(1).

The state corresponds to equilibrium state at OK thermodynamically because H

is enthalpy and Hisidenticalwith free energy at OK.Ifinteraction between

atomsis expressed by pair‑POtential¢・(r),Eq.(1)is written as

H = ∑¢(=r(k)+u(k)‑r(1)‑u(1)‖)‑ ∑(u(k)・F(k))

k〉1 k

(2)

Where r(k)I u(k) and F(k)areinitialpositionI displacement vector and

applied force of k‑th

atom,and(Ⅹ・y)isinner product of vectors x and y.

▼2

Talor ex ansion a roximat迦

Note that ¢is function depending on distance r only,gradient and Hesse

matrix of are written as

= ¢'‑

1txIl

llxtl¢''‑¢' ¢, (∂2¢/∂Ⅹi∂xj)=

【Ⅹ=

〔……j

llxll3 1lxll

,E:unit matrix].

Thus,◎in Eq.(1)is approximately expressed by

r(k)‑r(1)

◎=∑【¢(11r(k)‑r(1)‖)+¢(‖r(k)‑r(1)‖)(

k〉1 llr(k)‑r(1)ll

1

+ ‑(u(k)‑u(1)・Skl(u(k)‑u(1)))】,

2

u(k)‑u(1))

(3)

llr(k)‑r(1)‡l¢‑'‑¢

(Skl= ((ri(k)‑ri(1))(rj(k)‑rj(1)))

ltr(k)‑r(1)tl3

using Taylor series of second order.Eq・(2)is simply rewritten as

I H = α一(b・u)+‑(u・Su)

2

in the quadratic form of u =

望遠R相通画届空軍 圃醐通日冨岡

(5)

(6)

In Eq.(6),finding u such that His minimizedis equivalent to aH/∂u=0・

Therefore,following equationis obtained

Su = b. (7)

The linear

equation is soIvable for given boundary condition using

conventional calculuses such as Gauss's sweep‑Out method or conJugate

gradient method.

Because above explanation was derived from energy principle,

relationship between the present method and FEMis not clear・ According to

formulation base on force balance, b and u correspond to array vectors of

force and displacement and Sis overa11stiffness matrix combining b and u

linearly.

3.Examples

In following calculations, tWO‑dimensionalLennard‑Jones crystal was

treated.The pair‑pOtentialis given by

¢(r)= Ⅴ【(a/r)12‑2(a/r)6】 (8)

where two parameters(V, a)are potentialdepth and bonding distance between

two at,OmS,reSPeCtively.Elastic constants such as YoungIs modulus E,Poisson

ration v and shear modulus G are

E = 95.983 Va‑2,V = 0.3333,G = 35.842 Va 2

in the two‑dimensionalsystem fromlattice statics theory4)・

(1)■Uniaxial‑Lensile prou旦担

Results of elastic deformation when one‑directional tensile load is

applied is described first. Because finite rectangular‑Shape specimen was

treated in the previous paper4)andlocal inhomogeneous deformation and

stress concentration near surface and corners wasin problem,finite specimen

was calculated to prevent surface effect and edge effectin this calculation・

(1) Atoms were arrangedin hexagonalshape. Periodic boundary condition was

(4)

x‑and y‑directions of rectangular unit cell(size of ce」1:1JXXLy,

number of atoms‥ n=576,inset of Fig.1).

(2)Equilibrium state was calculated for severalsets of(Lx,I,y).Intemal

potential◎is also calculated foI・the equilibrium staLe.Noload condiLion

COrreSpOnds to a state where¢is minimized for Lx and Ly.

(3)◎is approximately expressed by quaTtLic function oflongitudinalsLrain

Eyy and transversaユ strain Exx usingユeast‑Squar・e method for severalsets of

(fxx,ぎyy)=ぎ★ぃく3%,!ぎ,y蓼く3%ト

①=¢(どxx,どyy)

(9)

(4)Relations among tensile stress oyy aIld strains Exx,fyy WaS Obtained from

pt・inciple of vil・tualwork・If displacement6Exx and6Eyy are given,increment

Of q)is equalt,0 Virtualwork.

(Lx〔アy)(Lyy6ttyy)=(D(Exx+6Exx,Eyy+̀了Eyy)‑①(Exx,Eyy)

∂◎/∂.Fxx = 0

0yy =(∂◎/∂Eyy)/LxLy.

tn Fig.l,Open a11d closed

Ci上・Cles show calculated results

()f‑ uyy‑Eyy and Exx‑Eyy

T'elaLions frol11Eqs.(11) and

(12)・They are perfectlyin good

agreement with theoretical

Cu上●V(‑S Oflattice statics t.heory

draw‖ by solid lines in the

figure‑ Although Lhe exampleis

Very primary and notinteresting

for a problem of structural

englneering, it is confirmed

from the results that the

present method is valid for

Statistical equilibrium

Calculation.

z・ロ○>、丈b、ss2)SP巴コPむ∝

…U

ご馬上SO 5 0

0

∩)

∩〜

0

0

Fig・1 Stress‑Strain(oyy‑ぎyy)and

transversalstrain‑longitudinal

Strain(Exx‑Uyy)curvesin

iniaxialloading two‑dimensional

Lennard‑JorleS CryStal.The Number

Of atomsin a rectangular cell

is 576.

坤生臼旦旦̲豊里旦£gヱ

Dislocations are the origin of plastic flow・Macroscopi,C deformationis

naturally achieved in consequence of collective motion andinteractions of

(5)

to know deformation.

A dislocation accommodates strain field characteri2;ed by direction of

dislocat.ion line and Burgers vector.Isotropic elastic theory has been well

used because the strain fieldis analytica11y soluble. The solution is,

however,incorrect physically because strainisinversely proportional to

distance r from dislocationline andit diverges at r=0. Thus,nOt COntinuum

approximation but discrete treatment ofindividualatoms are required for

discussion of dislocation core reglOn.

(a) Unit

「Cell「

ny=7[

ny

=x十1 甜壬弓Fauけed

盛葦灘迂葦丑葦PLane

ミミ票請甑だごご;::::ミ:ミm翳諮;岩

000C=⊃000=====000̀⊃(⊃00 000(⊃000=====00(⊃00000 00000000=====0000000

(⊃000000=====00000000 00000000=====0000000

0000〔)00=====000CI0000

nx=15

1∴.̲」

Lx=15.Oq

(b)

毒頚蓋蒙

Lx=15.6G

Fig.2 Procedure of dislocation core structure calculation;(a)initial

atom arrangement. Two crystals with differentlattice planes are

adhered.(b)Shrinkage of core occurred after severalit,erations.

(1) First, COre StruCture WaS Calculated by the present method.It is

explainedin Fig. 2 using smallspecimen.In Fig.2(a),tWO perfect,CryStals

were adhered where the number oflattice planes(nx)was not same. Because

extra planes wereinserted from top surfacein this case and Burgers vector b

is parallel to x‑direction, a dislocationlays on horizontal plane. Core

region was, however, Widely extended on the faulted plane and not clear in

the figure.

Periodic boundary condition wasimposed to x‑direction to avoid free

Surface effect.If not, because mirror force was applied to core from

Surface,dislocation slipped andit is sunk at the surface.

Asiterations were repeated untilforce balance was achieved, Shrinkage

of faulted plane occurred.In converged state of Fig. 2(b),PentagOnalatom

arrangement appeared which was dislocation core. One can always see such a

picture in a text book of dislocation theory as Bragg's bubble mode16).

Therefore,bubble bath was reproduciblein computer.Additionally,itis very

advantageous that any pair‑pOtential and any boundary condition were

(6)

availablein the present method.

(2) Secondly, dislocation energy was

discussed.Similarly,equilibrium of a

dislocation was calculated for n=840

atoms (Fig. 3). On condition that

Periodic boundary condition wasimposed

to x‑direction, Calculations were

Performed making ce111ength (Lx) a

●●●●●●●●l

●●●●●●●●●

l■■■■l■l■●l==●1

==●●●●●●●●

====●1

====●l

●●●●●●●●●●●●●

====●l

●●●●●●●●●●●●●

====●l

●●t■■●●●●●●●●

●●●●●●●●●●●●■

●●●●●●●●●●●●

====l

■■■■■■■■●●●●●●

■●●●●●●●●●●●l

==■●●●●●●

====l

==1■1■■l■■l■■ll====l

●●●●●●●●●●●●

l●●●●●●●●●●●l

free parameter and minimizing ¢.Because 〟㍊㌫〃㌫

X‑directional tensile load was applied

in t,his case, unloading condition was a

●●●

l●●1

●●●

====l

●●==■●●●●●

====1

==●●●●●●●

====1

==■●●●●●●●

●●●●●●●●●●●●l

Unit Cell

l=====●l

●●●●●●●●●●●●●●●●●

l====●●●●1 l======●1

●●■t●●●●●●●●●●●●●●●●

l======●l

====●●●●●●●●

l======●l

●●●●●●●●●●●●●●●●●●●●

l======●l ●1■l■■===l l==●●●●●●

====l l●●●●●●●●●●●●

===●●●●l l==●●■●●●

==l■●

l==●1

■●●●●●●●

■●●●●●●●●●●●●●●●●●●●

l======●l

======■●

l●●●l==l■●l■●l■t■■■■■■■■■■■l===●●●●●●●●●●

l==■●■●■●●■=●1

=●●●●■●■●■●●■●■●●●

l●■=====●●●1

●●●●●●●●●●●●●●●●●●●●

l●●●●●●●●● ●●●●●●

==●■==●l l●●●●●●●●●●●●●●●

●●●l■l■1■1■■●●●l■l■■l■■■

l●●●●●●●

==●l l●●●●●●●

●●●●●●●

l●●●●●●l

●●●●●●●●

l●●●●●●●l

●■●●■●●●

l●●●●●●●●●●●●●

====●l l====●

===●」■l■l■l====●l l●●●●●●●●●●●●●

====■l●●●●●●●●●●●●●●●

l●●●●●●●●●●●●●●■

●●●●●●●●●●●●●●●

l●●●●●●●●●●●●●●l

●●●・===●●●

●●●●●●●●

l■==1

==・●●

==●1■●●●●●

l===●●●●

●■●●」■●●●●●●●●

ll==●●●●●●●

=●●●●

l●●●●●●

●●●●●●●

l●●●●●●

l●●●●●●●

==●l

lt●●●●●●●●●l

●●●●●●●●●●●

l===●1 l●●●●●●■■■■■■●●

●●●●●●●●●●●●

l●●●●●●●●●●●●

●●●●●●●● ●●●

l=●■●■●■ ●●●

●●●●●●●●●●●●

l●●●●●●●●●●●●

●●●●●●●●●●●●

l●■==●■●

====1 1●●●●●●●●●●●

State Where ◎is minimi2;ed from Eq.(11).

Fig.3 Equilibrium atom

By comparing of perfect crystal with

arrangement of

the minimum value, dislocation energy r

dislocation in

was 25.1V or O.70Gb2 because the

n=840 Lennard‑

magnitude of Burgers vector b was a and

Jones crystal.

shear modulus G was 35.842Va‑2.

The result comparatively agreed with a value calculated by elastic

theory.According toisotroplC elastic theory,Strain energy of a dislocation

is estimated to be r=Gb2 7).

4.Conclusion

New method to calculate a static state in discrete atom systems was

applied to some probleTnS Of elastic deformation andlattice defect.The state

COrreSpOnds to force‑balancingin structural engineering and equilibrium

State in thermodynamics. Above examples may prove the usefulness of the

method for the study of atomic‑SCale simulations of static behavior.

References

l)L.Verlet:Phys.Rev.,.巳坦,98‑104(1967).

2)N.Metropolis,A.Rosenbluth,M.Rosenbluth,A.Teller and E・Teller:J・CheT・Phys・・Bi・831‑836(1953)・

3)R.Car and M.Parrlnello:Phys.Rev.Lett.,屋旦,2471‑2479 (1985)̀.

4)Y.Takahashi:J.Jpn.Inst.Metall.,51,597‑603(1993)【inJapanese】.

5)S.Kirpatrick,C.D.GelattJr.and M.P.Vecchi:IBM Res.

Rep.,RC‑9335(1982).

6)W.L.Bragg: Proc.Roy.Soc.London,全土旦旦,474‑478(1947), 皇皇旦旦,171‑174(1949).

7)H.Si2:uki:Introduction of Dislocation THeory,Agne,Tokyo(1980), pp.69‑77【inJapanese】.

参照

関連したドキュメント

In this work, we have applied Feng’s first-integral method to the two-component generalization of the reduced Ostrovsky equation, and found some new traveling wave solutions,

A variety of powerful methods, such as the inverse scattering method [1, 13], bilinear transforma- tion [7], tanh-sech method [10, 11], extended tanh method [5, 10], homogeneous

In recent years, several methods have been developed to obtain traveling wave solutions for many NLEEs, such as the theta function method 1, the Jacobi elliptic function

More specifically, we will study the extended Kantorovich method for the case n = 2, which has been used extensively in the analysis of stress on rectangular plates... This

It is the aim of this paper to continue these investigations and to present some new inequalities for the gamma function and some polygamma functions. Our results also lead to two

To derive a weak formulation of (1.1)–(1.8), we first assume that the functions v, p, θ and c are a classical solution of our problem. 33]) and substitute the Neumann boundary

There is a robust collection of local existence results, including [7], in which Kato proves the existence of local solutions to the Navier-Stokes equation with initial data in L n (

Variational iteration method is a powerful and efficient technique in finding exact and approximate solutions for one-dimensional fractional hyperbolic partial differential equations..