13.2 The interaction picture
We now turn to time dependent problems in the general case whereH=H 0 +V(t). We shall
assume that the eigenvalues and eigenstates ofH 0 are known as follows,
H 0 |n〉=En|n〉 (13.14)
For simplicity, we shall assume here that the spectrum ofH 0 is discrete, but it is not hard to
generalize to the continuous case, something we shall do later on. We prepare an initial state|φa〉
att= 0, which we may decompose in the basis of eigenstates|n〉by
|φa〉=
∑
n
cn(0)|n〉 (13.15)
In the absence of interaction potential,V(t) = 0, this state would evolve as follows,
|φa;t〉=
∑
n
cn(0)e−itEn/ ̄h|n〉 (13.16)
On the other hand, once the interaction potential is turned on, the time evolution will be more
complicated, giving time dependence also to the coefficientscn. The corresponding state will be
denoted|ψa;t〉. It satisfies the full time dependent Schr ̈odinger equation,
i ̄h
d
dt
|ψa;t〉= (H 0 +V(t))|ψa;t〉 (13.17)
and has the following decomposition,
|ψa;t〉=
∑
n
cn(t)e−itEn/ ̄h|n〉 (13.18)
Notice that we can also write this time evolution as
|ψa;t〉=e−itEn/ ̄h|ψa;t〉I (13.19)
where
|ψa;t〉I=
∑
n
cn(t)|n〉 (13.20)
The subscriptI has been added to indicate that this state is now considered in theinteraction
picture, a method due to Dirac. The advantage is that the coefficientscnnow have time depen-
dence only through the effects of the time-dependent interaction potential, the effects of the free
Hamiltonian having been factored out.
More generally, if|ψa;t〉Sis a general state in the Schr ̈odinger picture, satisfying the Schr ̈odinger
equation,
i ̄h
d
dt
|ψa;t〉S=
(
H 0 +V(t)
)
|ψa;t〉S (13.21)