I was told that a strongly correlated system is such that Fermi liquid theory fails, or a single-particle picture doesn't work. So, there is no energy band for a strongly correlated system.

So, I would like to know:

  1. Whether it's a good definition (single-particle fails)? Considering we always use ARPES to test high-$T_c$ superconductors, where ARPES focuses on the energy band.
  2. Is a traditional BCS superconductor a strongly correlated system? I'm not sure about this.
  3. Why it is called a strongly correlated system? Is there any relation with the correlation function? (For, I know that at a phase transition, the correlation length tends to diverge, but is that related?)



(1) Your definition of strongly correlated system is correct "single-particle fails." We can still use ARPES to study strong correlated systems, we just do not see features that would be present in a weakly correlated system. The most prominent feature in a weakly correlated system is a sharp peak at certain energy and momentum. If you track this peak in energy as a function of momentum using ARPES you have essentially measured the energy band. In strongly correlated systems this peak is not sharp. The precise definition of sharp is that it has a delta-function component to it (at least in theory).

(2) BCS SC is not a strongly correlated system. There are still energy bands in a BCS SC it is just that the energy bands do not describe electrons. The energy bands of a BCS SC tell you something about the SC quasiparticles called Bogoliubov quasiparticles. One interesting thing about Bogoliubov quasiparticles is that they carry non-integer charge.

(3) The "strongly correlated" refers to the interacting nature of the system and the fact that there is no single-particle description. If you excite a strongly correlated system in two steps, the excitation you make in the first step will effect which excitations you can make in the second step in a highly non-trivial way. The excitation you add first has a strong influence on the system and rearranges everything. In contrast, in a band metal you can add an electron at momentum $\mathbf{k}$ and the bands do not shift. You can then add another electron at momentum $\mathbf{k}'$ and its properties can be understood in terms of the original bands, i.e., the bands that were there before you added the electron at momentum $\mathbf{k}$.

| cite | improve this answer | |

Now, based on my current understanding, I would like to answer my question as:

Any system described by Hamiltonian that is not quadratic is a strongly correlated system. Examples are: Fermi-Hubbard model with non-zero interaction:

$$H = J\sum_{j,\sigma} c_{j,\sigma}^\dagger c_{j+1,\sigma} + \text{h.c.} -\mu\sum_{j,\sigma} c_{j,\sigma}^\dagger c_{j,\sigma} + U\sum_j(2n_{\uparrow}-1)(2n_{\downarrow}-1)$$

where the last term is not quadratic so it yields a strongly correlating behavior.

| cite | improve this answer | |

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.