You are on page 1of 2

Bradley j. nartowt; many body physics; dr.

Muttalib; Wednesday, November 21, 2012; assignment 12 416 - pr 19-1 - some useful expressions for boson-operators: the following three expressions are useful for further studies of the 1D transmission line model of the Luttinger liquid. a) let O and B be two operators whose commutator [O,B] is a c-number. Prove the commutator for the displacement-operator, e B is: [O, e B ] = [O, B]e B . Hint: consider the function f ( ) = e BOe B and differentiate with respect to . First: act on the left of the thing were trying to prove with e B , and denote [O, B] = z z* ,

[O, eB ] = [O, B]eB e B [O, B]eB = e B [O, eB ] = e B (OeB eB O) = e B OeB O e B OeB = O + z Another preliminary: some noteworthy values of the function f ( ) , and its Taylor series expansion,
F ( ) = e B Oe B = n =0
n n n F ( ) n! n =0

(1.1)

; F (0) = 1 O 1 = O; F (1) = e B Oe B ;

(1.2)

Now we heed the hint: first derivative, F = (e B Oe B ) = ( Be B )Oe B + e BO ( Be B ) = e B (OB BO)e B = e B [O, B ]e B = ze B e B = z (1.3) Because [O, B] = z z* , we see that all higher derivatives beyond (1.3) vanish. Evaluating F (1) then gives, F ( ) = n = 0
n n n F (0) n! n 1 1 = 0! 0 F (0) + 1! 1 z =0 + 0 + ... = O + z F (1) = O + z = e B Oe B [O, e B ] = [O, B]e B (1.4)

(b) prove the baker-Hausdorff formula: e A+ B = e Ae B e [ A, B ]/ 2 , in which [A,B] is a c-number. Hint: use the analogy with the time-evolution operator in the interaction-picture (e.g., by writing U ( ) = e Ae ( A+ B ) , differentiate with respect to , and just keep going. Let [ A, B] = z z* . In the same spirit as part-a, consider the function,
G ( ) e Ae B e
1 2z 2

G (1) = [RHS of equality we are trying to prove];

(1.5)

Computing dd G ( ) , and using the formula [ A, B] = z e B Ae B = A + z e ( A) B = ( B z)e ( A) , from1 (1.4) we get, 2 1 2 z 1 2 z 1 2 z A B 1 z dG 2 + e A Be B e 2 + e Ae B ( 21 2 z)e 2 = e A ( B + e A Ae A z ) e B e 2 d = Ae e e (1.6) 1 2 z 1 2 z 1 2 z = e A ( B + e A Ae A z ) e B e 2 = e A ( A + B z ) e B e 2 = ( A + B + z z ) e Ae B e 2 Thus: we have two equations: G (0) = 1 (the identity) and
dG d
dG d

= ( A + B )e Ae B e

1 2z 2

= ( A + B )G ( ) ,
1z 2

= ( A + B)G ( ) dG = ( A + B) d G ( ) = G0 e( A+ B ) ; G (0) = 1 = G0 G (1) = e A+ B = ( A + B)e Ae B e G

(1.7)

(c) consider coherent states of the form g = e gak 0 , where g is some constant. Show that the overlap
between two such states with different constants is: f | g = 0 | e fak e gak | 0 = e gf . Hint: expand the exponentials and show that a n (a ) n 0 = na n 1 (a ) n 1 0 . Note this then implies; a n (a )n 0 = n ! 0 . Doing a series-expansion of the exponentials and noting that the vacuum-kets introduce a nn (i.e., equal
number of hits from ak as from ak ), and using the 2nd and 3rd identities we have yet to prove, we get,

Caution: switching the roles of A and B introduces a minus sign, due to the appearance of A and B in a commutator.

f | g = 0
n

n =0

( fak ) n n ( gak ) n 0 = 0 n! n ! n = 0

n ,n

n , n = 0

n ( fg )n ( fa )n ( gak )n nn 0 = 0 (a )n (ak ) n 0 n! n ! n =0 n !n !

( fg )n n ( fg )n = 0 n! 0 = 00 = e fg n =0 n !n ! n =0 n ! But now we stop and prove what we just used: a n (a )n 0 = a n (a )n 1 0 + 1 1 = a n (a ) n 2 0 + 1 1 + 2 2 = a n n ! n = a n 1 (n 1)! n n n 1


= a n 1 ( n ) 2 (n 1)! n 1 = a n 1 ( n )2 (a )n 1 0 = na n 1 (a )n 1 0 ak n (ak )n 0 = nak n 1 (ak ) n 1 0

(1.8)

(1.9)

This expression (1.9) can be invoked repeatedly (e.g., self-consistency) in order to state, ak n (ak )n 0 = n (n 1)ak n 11 (ak )n 11 0 = n(n 1)(n 2)ak n 3 (ak )n 3 0 = n(n 1)...(2)(1)ak n n (ak ) n n 0 = n ! 0 (1.10) This justifies the step we took in (1.8).

You might also like