Categories:
Perturbation,
Physics,
Quantum mechanics.
Kubo formula
Consider the following quantum Hamiltonian,
split into a main time-independent term H^0,S
and a small time-dependent perturbation H^1,S,
which is turned on at t=t0:
H^S(t)=H^0,S+H^1,S(t)
And let ∣ψS(t)⟩ be the corresponding solutions to the Schrödinger equation.
Then, given a time-independent observable A^,
its expectation value ⟨A^⟩ evolves like so,
where the subscripts S and I
respectively refer to the Schrödinger
and interaction pictures:
⟨A^⟩(t)=⟨ψS(t)∣A^S∣ψS(t)⟩=⟨ψI(t)∣A^I(t)∣ψI(t)⟩=⟨ψI(t0)∣K^I†(t,t0)A^I(t)K^I(t,t0)∣ψI(t0)⟩
Where the time evolution operator K^I(t,t0) is as follows,
which we Taylor-expand:
K^I(t,t0)=T{exp(iℏ1∫t0tH^1,I(t′)dt′)}≈1−ℏi∫t0tH^1,I(t′)dt′
With this, the following product of operators (as encountered earlier) can be written as:
K^I†A^IK^I≈(1+ℏi∫t0tH^1,I(t′)dt′)A^I(t)(1−ℏi∫t0tH^1,I(t′)dt′)≈A^I(t)−ℏi∫t0tA^I(t)H^1,I(t′)dt′+ℏi∫t0tH^1,I(t′)A^I(t)dt′
Where we have dropped the last term,
because H^1 is assumed to be so small
that it only matters to first order.
Here, we notice a commutator, so we can rewrite:
K^I†A^IK^I=A^I(t)−ℏi∫t0t[A^I(t),H^1,I(t′)]dt′
Returning to ⟨A^⟩,
we have the following formula,
where ⟨⟩ is the expectation value for ∣ψ(t)⟩,
and ⟨⟩0 is the expectation value for ∣ψI(t0)⟩:
⟨A^⟩(t)=⟨K^I†A^IK^I⟩0=⟨A^I(t)⟩0−ℏi∫t0t⟨[A^I(t),H^1,I(t′)]⟩0dt′
Now we define δ⟨A^⟩(t)
as the change of ⟨A^⟩ due to the perturbation H^1,
and insert ⟨A^⟩(t):
δ⟨A^⟩(t)≡⟨A^⟩(t)−⟨A^I⟩0=−ℏi∫t0t⟨[A^I(t),H^1,I(t′)]⟩0dt′
Finally, we introduce
a Heaviside step function Θ
and change the integration limit accordingly,
leading to the Kubo formula
describing the response of ⟨A^⟩ to first order in H^1:
δ⟨A^⟩(t)=∫t0∞CAH1R(t,t′)dt′
Where we have defined the retarded correlation function CAH1R(t,t′) as follows:
CAH1R(t,t′)≡−ℏiΘ(t−t′)⟨[A^I(t),H^1,I(t′)]⟩0
Note that observables are bosonic,
because in the second quantization
they consist of products of even numbers
of particle creation/annihiliation operators.
Therefore, this correlation function
is a two-particle Green’s function.
A common situation is that H^1 consists of
a time-independent operator B^
and a time-dependent function f(t),
allowing us to split CAH1R as follows:
H^1,S(t)=B^Sf(t)⟹CAH1R(t,t′)=CABR(t,t′)f(t′)
Since CABR is a Green’s function,
we know that it only depends on the difference t−t′,
as long as the system was initially in thermodynamic equilibrium,
and H^0,S is time-independent:
CABR(t,t′)=CABR(t−t′)
With this, the Kubo formula can be written as follows,
where we have set t0=−∞:
δ⟨A⟩(t)=∫−∞∞CABR(t−t′)f(t′)dt′=(CABR∗f)(t)
This is a convolution,
so the convolution theorem
states that the Fourier transform
of δ⟨A^⟩(t) is simply the product
of the transforms of CABR and f:
δ⟨A^⟩(ω)=C~ABR(ω)f~(ω)
References
- H. Bruus, K. Flensberg,
Many-body quantum theory in condensed matter physics,
2016, Oxford.
- K.S. Thygesen,
Advanced solid state physics: linear response theory,
2013, unpublished.