0% found this document useful (0 votes)
31 views

Problem 4 2

The document discusses an electron in a rotating magnetic field B and describes transforming into a rotating frame where the Hamiltonian is time-independent. It shows that in the adiabatic limit, the wavefunction reduces to a dynamic phase and Berry's geometric phase, which is equal to half the solid angle subtended by the tip of the magnetic field vector. It also presents two methods for calculating the geometric phase directly from the eigenstates and Berry connection.

Uploaded by

Naipunnya Raj
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
31 views

Problem 4 2

The document discusses an electron in a rotating magnetic field B and describes transforming into a rotating frame where the Hamiltonian is time-independent. It shows that in the adiabatic limit, the wavefunction reduces to a dynamic phase and Berry's geometric phase, which is equal to half the solid angle subtended by the tip of the magnetic field vector. It also presents two methods for calculating the geometric phase directly from the eigenstates and Berry connection.

Uploaded by

Naipunnya Raj
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

May 3, 2008

Electron in a rotating magnetic field


Rotating frame
Suppose we have an electron at rest in a magnetic field

B = B0 (sin α cos ωtî + sin α sin ωtĵ + cos αk̂)

So the field vector points along the polar angle α and rotates about the z-
axis with frequency ω. It is convenient to transform into the rotating frame.
At t = 0, the magnetic field is B~ = B0 cos αk̂ + B0 sin αî. At a later time
the field has rotated about the z-axis by an angle θ = ωt. If we rotate the
spinor about the z-axis we can move to a frame in which the hamiltonian is
independent of time. A rotation about z is accomplished with
iωt/2
e 0

iθσz /2 i(ωt)σz /2
R(θ) = e =e =
0 e−iωt/2
And since
ω1 h̄ cos α sin αeiωt

H= B0 (1)
2 sin αe−iωt − cos α
then

R−1 HR = H0
−iωt/2
sin αeiωt
iωt/2
e 0 ω1 h̄ cos α e 0

=
0 eiωt/2 2 sin αe−iωt − cos α 0 e−iωt/2
−iωt/2
ω1 h̄ e 0 cos αeiωt/2 sin αeiωt/2

=
2 0 eiωt/2 sin αe−iωt/2 − cos αe−iωt/2
ω1 h̄ cos α sin α

=
2 sin α − cos α

and H0 is the time independent hamiltonian in the rotating frame.


Then

Hχ = RH0 R−1 χ = ih̄ χ
∂t
1
and

H0 R−1 χ = ih̄R−1 χ
∂t
Since R−1 does not commute with ∂t

we need to be careful in the next step.
We find that
∂ −1 ∂R−1 ∂
(R χ) = χ + R−1 χ
∂t ∂t ∂t
and also
∂R−1 iω e−iωt/2 0 ω

=− iωt/2 = −i σz R−1
∂t 2 0 −e 2
With this in mind we can write
∂ ∂ h̄ω
H0 R−1 χ = ih̄R−1 χ = ih̄ (R−1 χ) − σz R−1 χ
∂t ∂t 2
and
∂ 0 ∂
χ = ih̄ χ0
H 0 χ0 = ih̄
∂t ∂t
0 −1 0 h̄ω
where χ = R χ and H = H0 + 2 σz . Now we have a time independent
hamiltonian. To solve we can compute eigenvalues and eigenvectors to get
χ0 (t) and then transform back to the lab frame.
0
Another strategy is to construct the time translation operator e−iH t/h̄ .
First write H 0 /h̄ in the form n̂ · σλ/2
h̄ ω1 cos α + ω ω1 sin α h̄

H0 = = n̂ · σλ
2 ω1 sin α −ω1 cos α − ω 2
where
ω1 cos α + ω
nz = 1
(ω12 + 2ω1 ω cos α + ω 2 ) 2
ω1 sin α
nx = 1
(ω12 + 2ω1 ω cos α + ω 2 ) 2
ny = 0

and
1
λ = (ω12 + 2ω1 ω cos α + ω 2 ) 2
Then
0
Q(t) = e−iH t/h̄ = e−in̂·σλt/2 = cos(λt/2) − in̂ · σ sin(λt/2)

2
cos(λt/2) − i(ω1 cos α + ω) sin(λt/2)/λ −iω1 sin α sin(λt/2)/λ

=
iω1 sin α sin(λt/2)/λ cos(λt/2) + i(ω1 cos α + ω) sin(λt/2)/λ
Finally
1

0 0
χ(t) = R(t)χ (t) = R(t)Q(t)χ (t = 0) = R(t)Q(t)
0

Adiabatic approximation
If we return to the original hamiltonian H, and define ψn such that

H(t)ψn (t) = En (t)ψn (t)

then
cos(α/2) sin(α/2)eiωt

ψ1 (t) = χ+ (t) = and ψ2 (t) = χ− (t) =
sin(α/2)e−iωt − cos(α/2)
are eigenvectors of the hamiltonian(Equation1 ) with eigenvalues E± =
cos(α/2)
±h̄ω1 /2. If our initial state is χ+ = , then
sin(α/2)
χ(t) = R(t)Q(t)χ+
cos(λt/2) − i(ω1 cos α + ω) sin(λt/2)/λ −iω1 sin α sin(λt/2)/λ

= R(t)
−iω1 sin α sin(λt/2)/λ cos(λt/2) + i(ω1 cos α + ω) sin(λt/2)/λ
cos(α/2)

×
sin(α/2)
 h i
(cos(λt/2) − i ω1λ+ω sin(λt/2)) cos(α/2) eiωt/2

= 
h i


(cos(λt/2) − i ω1λ−ω sin(λt/2)) sin(α/2) e−iωt/2
ω1 + ω cos α ω sin α
= (cos(λt/2) − i sin(λt/2))eiωt/2 χ+ (t) − i sin(λt/2)e−iωt/2 χ− (t)
λ λ
In the adiabatic limit, ω ω1
ω
χ(t) = (cos(λt/2) − i(1 − cos α) sin(λt/2))eiωt/2 χ+ (t)
ω1
= (cos(λt/2) − i sin(λt/2))eiωt/2 χ+ (t)

= e−iλt/2 eiωt/2 χ+ (t)

3
and

χ(t) → e−iλt/2 eiωt/2 χ+ (t)


∼ e−i(ω1 +ω cos α)t/2 eiωt/2 χ+ (t)

Berry phase
The dynamic phase is θ+ = −ω1 t/2. The remaining phase is geometric
γ+ = (ω/2)(− cos α +1). Unfortunately I started out rotating in the negative
φ direction. If ω changes sign then γ+ = (ω/2)(cos α − 1) and Berry’s phase
is (ω/2) 2π
ω
(cos α − 1) = π(cos α − 1). 1/2 of the solid angle subtended by the
tip of magnetic field vector.
We could also get the geometric phase by
Z
∂χ+ 0
γ = i hχ+ | idt
∂t0
0
Z
0
= i ( cos(α/2) sin(α/2)eiωt ) 0 dt0
−iω sin(α/2)e−iωt
Z
= ω sin2 (α/2)dt0
= −π(1 − cos α)

That last − comes from integrating backwards.

cos(λt/2) + i cos α sin(λt/2) i sin α sin(λt/2)



Q(t) →=
i sin α sin(λt/2) cos(λt/2) − i cos α sin(λt/2)

You might also like