1
$\begingroup$

A two-level system is governed by $\mathcal{H_0} = E_0 \left( {\begin{array}{cc} 2 & 0 \\ 0 & 4 \\ \end{array} } \right)$. A small perturbation $\mathcal{H^{'}} = \epsilon \left( \begin{array}{cc} a & b \\ c & d \end{array} \right)$ is applied. So what is the first order correction to the lowest unperturbed energy.

I have tried the following. The eigenvalues $\mathcal{H_0}$ will give me the unperturbed energy, $2$ and $4$. So the eigenvalues of the matrix, $\mathcal{H_0} + \mathcal{H_{'}}$ should give me the perturbed energy. Is that right?

$\endgroup$
3
  • 4
    $\begingroup$ I'm voting to close this question as off-topic because the answer to the question would be "Yes" - this shows that this is a check-my-work type of questions which is not a good fit for this site. Hence I'm voting to close. Also, this is a purely mathematical question. $\endgroup$
    – Martin
    Commented Jun 16, 2017 at 7:13
  • $\begingroup$ @Martin I disagree, but only because the OP has misunderstood the problem; although they're right that the exact perturbed energies are the eigenvalues of the full Hamiltonian, the quoted question actually seeks a first-order perturbation. I'll answer the question in a moment, because I think it's important to clarify. $\endgroup$
    – J.G.
    Commented Jun 16, 2017 at 7:48
  • $\begingroup$ @Martin Its definitely not a homework assignment. I was trying to solve a problem, and wanted to clarify my approach/understanding as rightly said by J.G. Any help regarding my understanding will be much appreciated! :) $\endgroup$
    – sbp
    Commented Jun 16, 2017 at 7:57

2 Answers 2

2
$\begingroup$

Additional Comment

The nice thing about small matrices is you can easily check the first computation using the pertubative formula $\langle \psi_n | \Delta \hat{H} | \psi_n \rangle$ with a direct expansion, and they should match!

Hint:

compute the eigenvalues directly using

$$\det [\left( {\begin{array}{cc} 2+ \epsilon a & \epsilon b \\ \epsilon c & 4+ \epsilon d \\ \end{array} } \right) - \lambda I] = 0$$

This should give you an exact quadratic for $\lambda$ which you can solve (exactly), and expand to first order in $\epsilon$. These values of $\lambda$ corrected to first order should match whatever you computed previously.

$\endgroup$
0
0
$\begingroup$

In theory you can obtain exactly the eigenvalues of any Hamiltonian - for a $2\times 2$ Hamiltonian, this is equivalent to solving a quadratic equation - but in practice this is computationally impossible. What perturbation theory does is to write the eigenvalues (and also eigenvectors) of a Hamiltonian of the form $\mathcal{H}_0+\epsilon\Delta\mathcal{H}$ with $\epsilon$ small as a power series in $\epsilon$.

Suppose $v$ is an eigenvector of the unperturbed Hamiltonian; then the first-order correction to the associated eigenenergy is $v\cdot\epsilon\Delta\mathcal{H}v$ (or, in bra-ket formalism, $\epsilon\left\langle v\right|\mathcal{\Delta\mathcal{H}}\left| v\right\rangle$). In this case the unperturbed energy of interest is $2E_0$ so $v=\left(\begin{array}{c} 1\\ 0 \end{array}\right)$. The first-order correction to the unperturbed energy is then $\epsilon a$.

$\endgroup$
1
  • $\begingroup$ Thanks a lot. My bad, should have approached the problem this way. $\endgroup$
    – sbp
    Commented Jun 16, 2017 at 8:06

Not the answer you're looking for? Browse other questions tagged or ask your own question.