Probability for a non-hermitian hamiltonian

shinobi20
Messages
277
Reaction score
20

Homework Statement


Given a non-hermitian hamiltonian with V = (Re)V -i(Im)V. By deriving the conservation of probability, it can be shown that the total probability of finding a system/particle decreases exponentially as e(-2*ImV*t)/ħ

Homework Equations


Schrodinger Eqn, conservation of probability (∂P/∂t = -∇⋅j), P is the probability and j is the probability current density, Re V is the real part of V while I am V is the imaginary part of V

The Attempt at a Solution


It is known that ∫∫∫ P d3r = constant, that is ,the total probability for finding the particle should be constant throughout the universe.

iħ ∂ψ/∂t = -ħ2/2m ∇2ψ + Vψ = -ħ2/2m ∇2ψ + (Re)Vψ -i(Im)Vψ
By taking the complex conjugate, we have
-iħ ∂ψ*/∂t = -ħ2/2m ∇2ψ* +V*ψ* = -ħ2/2m ∇2ψ* + (Re)Vψ* +i(Im)Vψ*
By multiplying by ψ* the first equation and by ψ the second equation then subtracting the second from the first, we have
iħ ∂(ψψ*)/∂t = -ħ2/2m (ψ*∇2ψ-ψ∇2ψ*) - 2i (Im)V ψψ* = -∇⋅j - 2i (Im)V ψψ*

I don't know where to go already from here. Any suggestions or hints?
 
Physics news on Phys.org
Just use the definition of total probability in quantum mechanics: ##\langle \psi(t) | \psi(t) \rangle##. Now ##|\psi(t)\rangle = e^{-iVt/\hbar} |\psi(0)\rangle##. What will you get then when calculating ##\langle \psi(t) | \psi(t) \rangle##?
 
blue_leaf77 said:
Just use the definition of total probability in quantum mechanics: ##\langle \psi(t) | \psi(t) \rangle##. Now ##|\psi(t)\rangle = e^{-iVt} |\psi(0)\rangle##. What will you get then when calculating ##\langle \psi(t) | \psi(t) \rangle##?
I will get ⟨ψ(t)|ψ(t)⟩ = ⟨ψ(0)|ψ(0)⟩ but what does this tell me? Also, I need to use the conservation of probability equation to derive the question.
 
shinobi20 said:
I will get ⟨ψ(t)|ψ(t)⟩ = ⟨ψ(0)|ψ(0)⟩ but what does this tell me?
No, you won't because in this problem ##V## is not Hermitian. Try plugging in ##V = \textrm{Re}[V] - i \textrm{Im}[V]##.
 
blue_leaf77 said:
No, you won't because in this problem ##V## is not Hermitian. Try plugging in ##V = \textrm{Re}[V] - i \textrm{Im}[V]##.
Sorry, I forgot that it is not hermitian, so we will get e2i(Im)Vt
 
It will be ##\langle \psi(0) |e^{-i\textrm{Im}[V]t/\hbar}| \psi(0) \rangle##, at this point we should know how ##\textrm{Im}[V]## acts on ##| \psi(0) \rangle##, i.e. whether it is a function of space at all. But since the question asks you to prove that "decreases exponentially as e(-2*ImV*t)/ħ", I think it is implied that ##\textrm{Im}[V]## is a constant and thus the sandwiched exponential operator may be taken out.
 
Last edited:
blue_leaf77 said:
It will be ##\langle \psi(0) |e^{-i\textrm{Im}[V]t/\hbar}| \psi(0) \rangle##, at this point we should how ##\textrm{Im}[V]## acts on ##| \psi(0) \rangle##, i.e. whether it is a function of space at all. But since the question asks you to prove that "decreases exponentially as e(-2*ImV*t)/ħ", I think it is implied that ##\textrm{Im}[V]## is a constant and thus the sandwiched exponential operator may be taken out.
Oh! I did something wrong when taking the complex conjugate of V, but anyways, you really cleared everything. Thanks!
 
Back
Top