0
$\begingroup$

I am working with a Hamiltonian which looks like this (Heisenberg model) $$ \hat{H} = -\frac{1}{2}\sum_{j=1}^N \left( J_x\sigma_j^x\sigma_{j+1}^x +J_y\sigma_j^y\sigma_{j+1}^y +J_z\sigma_j^z\sigma_{j+1}^z + h\sigma_j^z \right). $$

I have made a program which computes this Hamiltonian using Pauli matrices (spin 1/2). My working space is then the tensor product ($N$ times) of $\mathbb{C}^2$. I know that the canonical base of my space can be expressed as a tensor product of base vectors of $\mathbb{C}^2$, for example: $(1,0,0,0) = (1,0)\otimes (1,0)$

This works fine when I am working with only $J_z$ not null (classical Ising model) because all the eigenstates can be expressed this way (all eigenstates are vectors of the canonical base). When I work with, for example, only $J_x$ not null (quantum Ising) I get eigenstates which are a bit more messy, for example $(0,1/\sqrt{2},-1/\sqrt{2},0)$.

This eigenstate can be expressed as a linear combination of canonical base vectors and those as a tensor product of the spin 1/2 Z base.

My problem is that I seek a "visual representation" of all states (or eigenspaces), I believe that any two level system can be represented in polar coordinates in $\mathbb{R}^3$ (Bloch sphere) but I fail at doing so, how should I proceed? Let's say I wanted to represent the state I used as a example before, which is non-degenerate, in a visual way, that is, in polar coordinates (it corresponds to the case of two 1/2 spins, so two points in polar coordinates would be required).

$\endgroup$

1 Answer 1

0
$\begingroup$

Maybe you would like to trace over other base. For example, your interest eigenstate is $N-$ body state $|\psi\rangle$, then you have density matrix $\rho=|\psi\rangle\langle\psi|$, and to derive a single site's information you only have to trace over other state, i.e., $\rho_j=\prod_{i\neq j}\sum_{\sigma_i}\langle \sigma_i|\rho|\sigma_i\rangle$, and get a $2\times2$ matrix.

For your case as an example, say $|\psi\rangle = (0, 1/\sqrt2,-1\sqrt2,0)^T$,

$\rho=\left(\begin{matrix}0&0&0&0\\0&1/2&-1/2&0\\0&-1/2&1/2&0\\0&0&0&0\end{matrix}\right)$,

and to get information about the first site, we trace over the second by $|2_\uparrow\rangle=(1,0)^T,|2_\downarrow\rangle=(0,1)^T$, then every matrix element of $\rho_1$ is given by (take $\langle1_\uparrow|\rho_1|1_\uparrow\rangle$ as example)

$\langle1_\uparrow|\rho_1|1_\uparrow\rangle=(|1_\uparrow\rangle\otimes|2_\uparrow\rangle)^\dagger|\psi\rangle\cdot\langle\psi|(|1_\uparrow\rangle\otimes|2_\uparrow\rangle)+(|1_\uparrow\rangle\otimes|2_\downarrow\rangle)^\dagger|\psi\rangle\cdot\langle\psi|(|1_\uparrow\rangle\otimes|2_\downarrow\rangle)$

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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