Calculating eigenstates of an operator

Join the discussion
Ask a follow-up here, or get your own question answered by working scientists, mathematicians and engineers — people, not an autocomplete.
Real named experts · corrections over time · the nuance an AI answer skips
4 replies · 2K views
barefeet
Messages
58
Reaction score
2

Homework Statement


Consider a two-dimensional space spanned by two orthonormal state vectors [itex]\mid \alpha \rangle[/itex] and [itex]\mid \beta \rangle[/itex]. An operator is expressed in terms of these vectors as
[tex]A = \mid \alpha \rangle \langle \alpha \mid + \lambda \mid \beta \rangle \langle \alpha \mid + \lambda^* \mid \alpha \rangle \langle \beta \mid + \mu \mid \beta \rangle \langle \beta \mid[/tex]

Determine the eigenstates of A for the case where (i) [itex]\lambda = 1, \mu = \pm 1[/itex], (ii) [itex]\lambda = i, \mu = \pm 1[/itex]. Do this problem also by expressing A as a 2 X 2 matrix with eigenstates as the column vectors.

Homework Equations


Just linear algebra rules.

The Attempt at a Solution


I started with [itex]\lambda = 1, \mu = 1[/itex]. Then A is:
[tex]A = \mid \alpha \rangle \langle \alpha \mid + \mid \beta \rangle \langle \alpha \mid + \mid \alpha \rangle \langle \beta \mid + \mid \beta \rangle \langle \beta \mid[/tex]

[tex]A \mid \alpha \rangle = \mid \alpha \rangle \langle \alpha \mid \alpha \rangle + \mid \beta \rangle \langle \alpha \mid \alpha \rangle + \mid \alpha \rangle \langle \beta \mid \alpha \rangle + \mid \beta \rangle \langle \beta \mid \alpha \rangle = \mid \alpha \rangle + \mid \beta \rangle[/tex]

[tex]A \mid \beta \rangle = \mid \alpha \rangle \langle \alpha \mid \beta \rangle + \mid \beta \rangle \langle \alpha \mid \beta \rangle + \mid \alpha \rangle \langle \beta \mid \beta \rangle + \mid \beta \rangle \langle \beta \mid \beta \rangle = \mid \alpha \rangle + \mid \beta \rangle[/tex]

The eigenstate is [itex]\mid a_n \rangle[/itex] with eigenvalue [itex]a_n[/itex]. Then the following holds:
[tex]A \mid a_n \rangle = a_n \mid a_n \rangle[/tex]

The eigenstate [itex]\mid a_n \rangle[/itex] can be expressed in the basis vectors [itex]\mid \alpha \rangle[/itex] and [itex]\mid \beta \rangle[/itex]:

[tex]\mid a_n \rangle = c_1 \mid \alpha \rangle + c_2 \mid \beta \rangle[/tex]

Then the earlier equation becomes:
[tex]A \mid a_n \rangle = A( c_1 \mid \alpha \rangle + c_2 \mid \beta \rangle ) = a_n (c_1 \mid \alpha \rangle + c_2 \mid \beta \rangle[/tex]

But this is also:
[tex]A( c_1 \mid \alpha \rangle + c_2 \mid \beta \rangle ) = c_1 A \mid \alpha \rangle + c_2 A \mid \beta \rangle = c_1 (\mid \alpha \rangle + \mid \beta \rangle) + c_2 (\mid \alpha \rangle + \mid \beta \rangle) \\ = (c_1 + c_2) \mid \alpha \rangle + (c_1 + c_2) \mid \beta \rangle[/tex]

This gives the equations :
[tex]a_n c_1 = c_1 + c_2[/tex]
[tex]a_n c_2 = c_1 + c_2[/tex]

The only solution is if [itex]c_1 = c_2 = 0[/itex]. Obviously I am doing something wrong but I can't see it.
 
on Phys.org
barefeet said:
This gives the equations :
[tex]a_n c_1 = c_1 + c_2[/tex]
[tex]a_n c_2 = c_1 + c_2[/tex]
The only solution is if [itex]c_1 = c_2 = 0[/itex]. Obviously I am doing something wrong but I can't see it.
Are you sure that's the only solution?
 
I see, but I still get one non trivial solution out of it.
If one is 0, then both are 0. So they are either both nonzero or both 0. If they are nonzero, I can write:
[tex]a_n = \frac{c_1 + c_2}{c_1} = \frac{c_1 + c_2}{c_2}[/tex]
If [itex]c_1 + c_2 =0[/itex] then [itex]a_n =0[/itex] or the c's are 0. If [itex]c_1 + c_2[/itex] is nonzero, then I can divide by it and gives me [itex]c_1 = c_2[/itex] with [itex]a_n = 2[/itex] Unnormalized I can take as eigenstate [itex]\mid \alpha \rangle + \mid \beta \rangle[/itex].
Now I can't find the other eigenstate. [itex]\mid \alpha \rangle - \mid \beta \rangle[/itex] is orthogonal to this eigenstate but letting A operate on it gives me 0. Or does this just mean that it is an eigenstate with eigenvalue 0?
 
barefeet said:
Or does this just mean that it is an eigenstate with eigenvalue 0?
That's it.

Another approach is to subtract the two equations:
##a_nc_1 = c_1+c_2##
##a_nc_2 = c_1+c_2##