Categories:
Physics,
Quantum mechanics.
Imaginary time
Let A^S and B^S be time-independent in the Schrödinger picture.
Then, in the Heisenberg picture,
consider the following expectation value
with respect to thermodynamic equilibium
(as found in Green’s functions for example):
⟨A^H(t)B^H(t′)⟩=Z1Tr(exp(−βH^0,S(t))A^H(t)B^H(t′))
Where the “simple” Hamiltonian H^0,S is time-independent.
Suppose a (maybe time-dependent) “difficult” H^1,S is added,
so that the total Hamiltonian is H^S=H^0,S+H^1,S.
Then it is easier to consider the expectation value
in the interaction picture:
⟨A^H(t)B^H(t′)⟩=Z1Tr(exp(−βH^S(t))K^I(0,t)A^I(t)K^I(t,t′)B^I(t′)K^I(t′,0))
Where K^I(t,t0) is the time evolution operator of H^1,S.
In front, we have exp(−βH^S(t)),
while K^I is an exponential of an integral of H^1,I, so we are stuck.
Keep in mind that exponentials of operators
cannot just be factorized, i.e. in general
exp(A^+B^)=exp(A^)exp(B^)
To get around this, a useful mathematical trick is
to use an imaginary time variable τ instead of the real time t.
Fixing a t, we “redefine” the interaction picture along the imaginary axis:
A^I(τ)≡exp(ℏτH^0,S)A^Sexp(−ℏτH^0,S)
Ironically, τ is real; the point is that this formula
comes from the real-time definition by replacing t→−iτ.
The Heisenberg and Schrödinger pictures can be redefined in the same way.
In fact, by substituting t→−iτ,
all the key results of the interaction picture can be updated,
for example the Schrödinger equation for ∣ψS(τ)⟩ becomes:
ℏdtd∣ψS(τ)⟩=−H^S∣ψS(τ)⟩⟹∣ψS(τ)⟩=exp(−ℏτH^S)∣ψH⟩
And the interaction picture’s time evolution operator K^I
turns out to be given by:
K^I(τ,τ0)=T{exp(−ℏ1∫τ0τH^1,I(τ′)dτ′)}
Where T is the
time-ordered product
with respect to τ.
This operator works as expected:
∣ψI(τ)⟩=K^I(τ,τ0)∣ψI(τ0)⟩
Where ∣ψI(τ)⟩ is related to
the Schrödinger and Heisenberg pictures as follows:
∣ψI(τ)⟩≡exp(ℏτH^0,S)∣ψS(τ)⟩=exp(ℏτH^0,S)exp(−ℏτH^S)∣ψH⟩
It is interesting to combine this definition
with the action of time evolution K^I(τ,τ0):
∣ψI(τ)⟩exp(ℏτH^0,S)exp(−ℏτH^S)∣ψH⟩=K^I(τ,τ0)∣ψI(τ0)⟩=K^I(τ,τ0)exp(ℏτ0H^0,S)exp(−ℏτ0H^S)∣ψH⟩
Rearranging this leads to the following useful
alternative expression for K^I(τ,τ0):
K^I(τ,τ0)=exp(ℏτH^0,S)exp(−ℏ(τ−τ0)H^S)exp(−ℏτ0H^0,S)
Returning to our initial example,
we can set τ=ℏβ and τ0=0,
so K^I(τ,τ0) becomes:
K^I(ℏβ,0)⟹exp(−βH^S)=exp(βH^0,S)exp(−βH^S)=exp(−βH^0,S)K^I(ℏβ,0)
Using the easily-shown fact that
K^I(ℏβ,0)K^I(0,τ)=K^I(ℏβ,τ),
we can therefore rewrite the thermodynamic expectation value like so:
⟨A^H(τ)B^H(τ′)⟩=Z1Tr(exp(−βH^0,S)K^I(ℏβ,τ)A^I(τ)K^I(τ,τ′)B^I(τ′)K^I(τ′,0))
We now introduce a time-ordering T,
letting us reorder the (bosonic) K^I-operators inside,
and thereby reduce the expression considerably:
⟨T{A^HB^H}⟩=Z1Tr(T{K^I(ℏβ,τ)K^I(τ,τ′)K^I(τ′,0)A^I(τ)B^I(τ′)}exp(−βH^0,S))=Z1Tr(T{K^I(ℏβ,0)A^I(τ)B^I(τ′)}exp(−βH^0,S))
Where Z=Tr(exp(−βH^S))=Tr(K^I(ℏβ,0)exp(−βH^0,S)).
If we now define ⟨⟩0 as the expectation value with respect
to the unperturbed equilibrium involving only H^0,S,
we arrive at the following way of writing this time-ordered expectation:
⟨T{A^HB^H}⟩=⟨K^I(ℏβ,0)⟩0⟨T{K^I(ℏβ,0)A^I(τ)B^I(τ′)}⟩0
For another application of imaginary time,
see e.g. the Matsubara Green’s function.
References
- H. Bruus, K. Flensberg,
Many-body quantum theory in condensed matter physics,
2016, Oxford.