1
$\begingroup$

I am trying to find the system of eigenvalues and corresponding normalized eigenvectors for the following Hermitian matrix:

$$\mathbf{H}=\begin{pmatrix}10 & 3i \\ -3i & 2\end{pmatrix}$$

I began by solving the equation $\det(\mathbf{H}-\lambda\mathbf{I}_{2\times2})$, giving:

$$\det(\mathbf{H}-\lambda\mathbf{I}_{2\times2})=\lambda^2-12\lambda+11 \implies \lambda_{1}=1,\, \lambda_{2}=11$$

We then find our corresponding eigenvectors as follows:

$$\begin{pmatrix}10 & 3i \\ -3i & 2\end{pmatrix}\begin{pmatrix}x_{1} \\ y_{1}\end{pmatrix}=\begin{pmatrix}x_{1} \\ y_{1}\end{pmatrix} \implies x_1=k,\, y_1=3ik \\ \begin{pmatrix}10 & 3i \\ -3i & 2\end{pmatrix}\begin{pmatrix}x_{2} \\ y_{2}\end{pmatrix}=11\begin{pmatrix}x_2 \\ y_2\end{pmatrix} \implies x_2 = k,\, y_2=-\frac{i k}{3}$$

Normalizing this we get:

$$\vec{v}^{1}=\frac{1}{\sqrt{10}}\begin{pmatrix}1 \\ 3i\end{pmatrix} \implies \langle\vec{v}^{1},\vec{v}^{1}\rangle = 1\\ \vec{v}^{2}=\frac{1}{\sqrt{2}}\begin{pmatrix}1 \\ -i\end{pmatrix} \implies \langle\vec{v}^{2},\vec{v}^{2}\rangle=1$$

However, this would imply that $\mathbf{U}^{\dagger}\mathbf{H}\mathbf{U}=\mathbf{\Lambda}$, with:

$$\mathbf{U}=\begin{pmatrix}\frac{1}{\sqrt{10}} & \frac{1}{\sqrt{2}} \\ \frac{3i}{\sqrt{10}} & -\frac{i}{\sqrt{2}}\end{pmatrix} \text{ and } \mathbf{\Lambda}=\begin{pmatrix}1 & 0 \\ 0 & 11\end{pmatrix}$$

However, when actually computing $\mathbf{U}^{\dagger}\mathbf{H}\mathbf{U}$ I get something very different to $\mathbf{\Lambda}$ (it's not even diagonal). So I have clearly done something very wrong, but I'm unsure where the error lies?

|cite|improve this question
$\endgroup$
  • $\begingroup$ $$\vec{v}^{2}=\frac{1}{\sqrt{2}}\color{red}{\begin{pmatrix}1 \\ -i\end{pmatrix}} \implies \langle\vec{v}^{2},\vec{v}^{2}\rangle=1$$ Where did the red come from? $\endgroup$ – Git Gud Dec 16 '13 at 19:46
  • $\begingroup$ Something that I always find helps in situations like this is checking that I have the right eigenvalues/eigenvectors using wolfram!alpha. $\endgroup$ – Avi Steiner Dec 16 '13 at 19:51
1
$\begingroup$

Looks to me like $v_1=(3i,1)\implies \|v_1\|=\sqrt{10}$ so $\displaystyle{v_1\over \|v_1\|}={1\over \sqrt{10}}(3i,1)$.

And $v_2=(-i,3)\implies \|v_2\|=\sqrt{10}$ so $\displaystyle{v_2\over \|v_2\|}={1\over \sqrt{10}}(-i,3)$.

Then $$U=\begin{bmatrix} v_1 & v_2\end{bmatrix}={1\over \sqrt{10}}\begin{bmatrix} 3i & -i\\ 1 & 3\end{bmatrix}$$ and $$U^*HU=\begin{bmatrix} 11 & 0\\ 0 & 1\end{bmatrix}.$$

|cite|improve this answer
$\endgroup$
  • $\begingroup$ This implies my $\vec{v}^{1}$ is wrong as well? $\endgroup$ – Thomas Russell Dec 16 '13 at 21:04
  • $\begingroup$ Looks that way. My quick scratch work on that part yields $x_1+{i\over 3}y_1=0$ so take $x_1=1\implies y_1=-i/3$. Mathematica yields $x_1=3i$, $y_1=1$ which is consistent with what I had since the latter is just the former times $3i$. $\endgroup$ – JohnD Dec 16 '13 at 21:21
  • $\begingroup$ Did this resolve your question? $\endgroup$ – JohnD Dec 17 '13 at 14:33

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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