I'm studying the quantum Ising model, i.e. with Hamiltonian $H= -h\sum_{i}X_i-\sum_{\langle i,j\rangle}Z_iZ_j$.

I know conceptually how to compute the ground state of the Ising model at zero temperature - just find the lowest eigenvalue $E$ and eigenstate $|\psi\rangle$ of $H$. This is equivalent to minimising $\langle \psi|H|\psi\rangle$ over states $|\psi\rangle$.

But I don't know how to do the same computation when we have non-zero temperature, I think it has something to do with minimising the free energy $F=\langle E \rangle - TS$, where think $\langle E \rangle$ is just $\langle \psi|H|\psi\rangle$, however looking at the answers on here, it would seem that the entropy $S$ of any pure state $| \psi \rangle$ is zero so if I were to minimise $F$, it's just the same as minimising the energy. But I thought it should be different, e.g. at very high termperatures shouldn't the "ground state" be a more uniform superposition over the spin configurations.

I feel like I'm missing something here and may be using wrong terminology for things. I would really appreciate a practical method of finding the ground state (and energy) at non-zero temperature.


1 Answer 1


A system is not usually in its ground state at finite temperature. Indeed the notion of ``state'' is not really applicable unless you use a microcanonical distribution. In that case the system us in whatever excited energy state that you chose to give it. A microcanonical system only has a "temperature" if the system/state satisfies the "eigenstate thermalization" condition.

For the canonocal distribution, on the other hand, you have a temperature but must describe the system by a density matrix $\rho=e^{-\beta \hat H}$ rather than a pure state.

  • $\begingroup$ could you say a bit more about what the "eigenstate thermalization" condition is please? $\endgroup$
    – wdc
    Commented Nov 25, 2018 at 0:29
  • $\begingroup$ @Daochen The "eigenstate thermalization hypothesis" is topic of current research. If true for a particular system/state, it means that the finite, isolated, quantum system in an excited state of energy $E$ has the same local properties as a the same system coupled to a heat bath at the temperature that gives the system the same average energy $<E>$. It's a the quantum analogue of the ergodicity that makes a microcanonical ensemble equivalent to a canonical one. For more details Google "eigenstate thermalization hypothesis." $\endgroup$
    – mike stone
    Commented Nov 25, 2018 at 13:19
  • $\begingroup$ Hi Mike, I wonder if you could please comment on my new related question physics.stackexchange.com/questions/444366/…? Thanks a lot :) $\endgroup$
    – wdc
    Commented Dec 1, 2018 at 0:19

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.