• Aucun résultat trouvé

SOME PROPERTIES OF WOLFF'S MODEL OF A MAGNETIC IMPURITY IN A METAL

N/A
N/A
Protected

Academic year: 2021

Partager "SOME PROPERTIES OF WOLFF'S MODEL OF A MAGNETIC IMPURITY IN A METAL"

Copied!
3
0
0

Texte intégral

(1)

HAL Id: jpa-00217842

https://hal.archives-ouvertes.fr/jpa-00217842

Submitted on 1 Jan 1978

HAL is a multi-disciplinary open access

archive for the deposit and dissemination of

sci-entific research documents, whether they are

pub-lished or not. The documents may come from

teaching and research institutions in France or

abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est

destinée au dépôt et à la diffusion de documents

scientifiques de niveau recherche, publiés ou non,

émanant des établissements d’enseignement et de

recherche français ou étrangers, des laboratoires

publics ou privés.

SOME PROPERTIES OF WOLFF’S MODEL OF A

MAGNETIC IMPURITY IN A METAL

P. Schlottmann

To cite this version:

(2)

JOURNAL DE PHYSIQUE

Colloque C6, supplément au n° 8, Tome 39, août 1978, page C6-844

SOME PROPERTIES OF WOLFF'S MODEL OF A MAGNETIC IMPURITY IN A METAL

P. Schlottmann

Institut fur Theoretisohe Physik, Freie Universitdt Berlin

1000 Berlin 33, Arnimallee 3, Germany

Résumé.- Nous obtenons l'identité de Ward pour la fonction de vertex et la fonction de Green d'un électron à l'aide de la méthode des équations de mouvement. Cette relation nous donne une équation intégrale linéaire par rapport à la fonction de Green et nous permet d'obtenir la solution du problème à la limite des temps longs.

Abstract.- Using an equation of motion method we obtained the Ward relation between the vertex function and the one—electron Green's function. This relation leads to a linear in-tegral equation for the Green's function and yields the solution of the problem in the long-time limit.

Magnetic impurities having a defined spin

in a nonmagnetic metal have been usually described

by the Anderson model and the Kondo-Hamiltonian.

At low temperatures these models can be described

by a Fermi liquid theory

l\-1j,

i.e. the magnetic

moment (defined at high T) is screened by the

elec-tron gas for sufficiently small T. The Wolff model

/3/ provides an alternative, but much simpler,

des-cription of this low temperature behavior. The

model Hamiltonian for an one-orbital impurity at

the origin is

H = J e, C

+

C, - J E' n + U(n, - <n >)

f k Tea tea

L

<r a ^ + t

ko

a

(n

+

- <n

+

>) (1)

where C. is the creation operator for an electron

with momentum k and spin a, n is the number

ope-rator for an electron of spin

a

at the impurity

site and U is the Coulomb repulsion of electrons

on the impurity. We will consider a kinetic energy

linear in the momentum near the Fermi level, i.e.

e = v (k-kp) where k_ and v_ are the Fermi

momen-tum and velocity Tespectively. E' = E - U <n >+

0 * J- a B ~a

2

characterizes the energy difference of the impurity

levels with respect to the Fermi energy. Here, B

is the external magnetic field and a takes the

va-lues + 1.

The localized impurity electrons may

dif-fuse into the metal due to the overlap between the

Wannier states which is the same as on the regular

lattice. This is the main difference with respect

to Anderson's model where the overlap with the

im-purity electrons is considered as an additional

adjustable parameter.

We derive here the Ward identity relating the vertex function and the one-electron propaga-tor within the long-time limit. We define the following auxiliary three-time correlation func-tions

Faa,(t;p,T) =-i<T{CCT(t)ppa,(T)C*(0)}> (2) where Co = Y, C, and p „ = Y, C, C, . is

essen-ce ka pa '•kTca k+pa

tially a boson-operator /4/. Considering the time evolution with respect to T we obtain

E

•£

-

V

F 3

F

ac

(t

5P'

T)= u

^q W ^ ' ^

+

+ i

\_S(i) -

6(-t - T ) H G

a

(t)

Li 37 " V O Vo

(t;

P'

T)

"

V

*l

qP

a

ff(t

"'

T> 0 )

Here Ga(t) = -i< T{Cff(t)C*(o)}> . Laplace

transfor-ming equation (3) with respect to T we have an algebraic system of equations, which solved for F„ ,(t;Z..) = Y F ,(t;q,Z.) yields era' m '•q aa' n m

' ^ . v - v ^ o ^ - a

(I

P

Z~F]

L ~

U

l> Zm-v

F

P J J

and F (t;Z„) = F (t;Z- ) U Y - (5) a-aK ' m' acr m' [L ^-Vpp J v ' Here, Z = 2H imT are the thermodynamical Bose poles.

The vertex function can be obtained by La place transforming equation (4) with respect to the time t (6 = 1/1)

-0 J dtc"*1^,:;^) = %(Z„.>UeffCVlX'VV

"WW" <

6

>

(3)

Defining a n e f f e c t i v e ~ o t e n t i a l The s u s c e p t i b i l i t y i s given by t h e e f f e c t i v e p o t e n t i a l s Duo, and i s i d e n t i c a l t o the RPA (random phase approximation) r e s u l t . Following Yamada's / I / we o b t a i n f o r t h e v e r t e x a n a l y s i s of t h e Anderson model we s e p a r a t e even

1 ( 8 ) and odd p e r t u r b a t i o n o r d e r s i n U and w r i t e ( z n , z n + ~ ; - z a ) =

I

-

1

Xo = 1

F ( 'even + 'odd ) ' 'even

-

I - ( U / V ~ ) Z '

Here, Z =IIi (2n+l)T a r e thermodynamical fermion n (U/vF) (18)

p o l e s . 'odd' 1 -(U/vF) 2

The t h r e e - l e g v e r t e x e q u a t i o n

(8)

c o r r e s - We have now t h a t f o r b o t h t h e Anderson and t h e ponds t o t h e s c a t t e r i n g of one fermion q u a s i p a r t i c l e Wolff models t h e s p e c i f i c h e a t can be w r i t t e n a s w i t h the emission of a Bose e x c i t a t i o n . I n analogy (yo = 2n2/3vF)

t o t h e electron-phonon i n t e r a c t i o n we can d e f i n e 6Cv = Yo 'even (19) t h e p o l a r i z a t i o n through t h e Dyson e q u a t i o n f o r t h e and t h e s u s c e p t i b i l i t y t o s p e c i f i c h e a t r a t i o a s e l e c t r o n - h o l e e x c i t a t i o n s . The p o l a r i z a t i o n i s gi- yoxT/6CV = 1 +

g

ven by t h e e x p r e s s i o n odd 'even

For t h e c a s e of t h e Wolff model i t reduces t o na (z,) = T

In

G~ ( Z J ~ ( z n , z n + % ; - ~ ) G ~ ( Z ~ + Z , ) =

I 1

+

U/v,.

= ( 'P zm-vFP ) T

In

C G ~ ( Z ~ ) - G ~ ( Z ~ + Z . ~ ) ~ (9)

The thermodynamic p r o p e r t i e s were a l r e a d y only the asymptotic part of Go(Z ) f o r large will o b t a i n e d by M a t t i s 141 through t h e e x c i t a t i o n spec- s u r v i v e t h e summation over t h e thermodynamical p o l e s . trum of t h e bosonized Hamiltonian. The p r e s e n t r e - Hence the dressed polarization is equal to the bare s u l t ~ a r e com~lementary t o t h o s e o b t a i n e d by M a t t i s one / 4 / and Fogedby 151. The d e t a i l s of t h e c a l c u l a t i o n

C 1

no

( z $

=

(

1

-=---

)

T

1

{ G ~ ~ ( Z )-G,,("~Q) = a r e published elsewhere 161.

P 'm V ~ P n

References Here, we have a complete c a n c e l l a t i o n between v e r t e x

and s e l f - e n e r g y c o n t r i b u t i o n s . / I / Yamada, K., Prog. Theor. Phys.

53

(1975) 970 We write D y s o n ~ s equation f o r the effective 121 NoziPres, P . , J. Low Temp. Phys.

17

(1974) 31 p o t e n t i a l s Duo, between e l e c t r o n s w i t h s p i n o=

2

131

Wolff, P.A., Phys. Rev.

124

(1961) 1030 D++ = UII-D-+

,

D-- = UII+D+-

,

D+-=D-+=U+UII-D-- (1 1) / 4 / M a t t i s , D.C., Ann. Phys. (N.Y.)

89

(1975) 45 They y i e l d t h e f o l l o w i n g e f f e c t i v e i n t e r a c t i o n / 5 / Fogedby, H . C . , J. Phys. C

10

(1977) 2869

( 1 2 ) / 6 / S c h l o t t m a n n , P . , P h y s . R e v . B ~ ( 1 9 7 8 ) i n D++(Z.~) = D--(ZJ = UII(Z.dUeff(Z$ pr-int

.

and D+-(Z.m> = D-+(Z,$ = Ueff

(Zd

(1 31 where Ueff(Zm) i s d e f i n e d by e q u a t i o n ( 7 ) . - With t h e a i d of t h e v e r t e x and t h e e f f e c t i v e p o t e n t i a l we can c a l c u l a t e t h e e l e c t r o n s e l f e n e r g y

1

= - T

lm

Go(ZQ++Z.$ru(Z,, Zn+zm;-ZJDao<ZJ (14) Using Dyson's e q u a t i o n we o b t a i n a n i n t e g r a l equa- t i o n f o r

I

G i A ( z d Go(z,) = 1

-

T ~ , ( z ~ ( f + O ~ ( Z , ~ Z , , J m (15) where Goq (Z,d i s given by

Goo(Za> = Goo(Zn.)/C 1 + E'oGuo(Zn)) (16) Goo i s t h e Green's gunction corresponding t o t h e one-body s c a t t e r i n g problem o f f t h e p o t e n t i a l E ' For small r e a l e n e r g i e s and zero temperature we have Go(.) = G ~ ~ ( . ) exp { i n ( U / v ~ ) 2

x w + ~ : ,

Références

Documents relatifs

No essential change is observed for the specimen from that with [I101 along z axis, except that the domain width for the former is a little larger than for the latter

Following the transfer matrix method for the pure case [I] we first introduce a dual lattice on which a dual spin is defined as a relative angle between adjacent

intra-chain interaction is assumed, this is in agreement with the magnetic structure results (overall antiferro- magnetic structure, but ferromagnet~c chains [3,

the impurity concentration ni : in the absence of other friction mechanisms the electrons sit still in the frame of reference moving along with the impurities,. whatever

Quenching implies a ground state for the sys- tem which has a random structure - a random walk in the simplest case of a spin glass in the local random mean field

This model can be used to study ferromagnetic relaxation, because the probability of finding theeffective ma- gnetic field &#34;up&#34; does not have to equal that of finding

In this work, the the- oretical study of a 2D spin-1 BEG model under the constraint of a time-dependent oscillating magnetic-field energy is performed to investigate SCO and

2014 Using the theory of Lacroix-Lyon-Caen et al., who have calculated the magnetic susceptibility of a cerium Kondo system in the Anderson model, including the