3
$\begingroup$

This question refers to the paper Nonstandard symmetry classes in mesoscopic normal-superconducting hybrid structures by Altland and Zirnbauer. In the paper the authors give a classification of Bogoliubov de Gennes Hamiltonians. More precisely the authors state that

The aim of the current section is to classify systems according to their symmetries. Using the BdG formalism we will show that the presence or absence of time-reversal and/or spin-rotation invariance leads to four distinct symmetry classes.

While I do understand their classifications I am wondering if it is sufficient to restrict oneself to time-reversal and spin-rotation invariance only? I am wondering why the authors seem to disregard important symmetries such as inversion symmetry of the underlying lattice or other point group symmetries? Hence I am wondering if the classification given in the paper is in fact complete? In particular the presence and absence of inversion symmetry plays a key role in the theoy of superconductivity (centrosymmetric vs. non-centrosymmetric superconductors)

I would be very happy to hear thoughts and opinions on that.

$\endgroup$
  • 1
    $\begingroup$ A link to the paper that is not behind a paywall can be found here (for now, at least). $\endgroup$ – Mark Mitchison Apr 20 '15 at 15:56
1
$\begingroup$

The reason that space group is not considered by Altland and Zirnbauer is that they were interested in the implication of the symmetries for transport properties of disordered electrons, and once you have disorder, spatial symmetries are no longer there. But you are right that the classification can be extended to include inversion symmetry and similarly for point group symmetries. For a modern view on the classification, see http://arxiv.org/abs/0912.2157. For inversion symmetry, one recent work is http://arxiv.org/abs/1403.5558. For other point group symmetry you can just search "topological crystalline insulators". A recent review is http://arxiv.org/abs/1501.00531.

$\endgroup$
  • 1
    $\begingroup$ Thanks for the reply. Naturally the inclusion of point group symmetries increases the complexity of the discussion: Each of the 32 point group symmetries imposes its own restrictions on the BdG Hamiltonian. With sufficient patience I assume that these restrictions can be worked out. But what I am wondering is: Are there any simplifications known to this procedure? Are there certain symmetries that are (from a practical perspective) more important than others? $\endgroup$ – CMFT_guy Apr 20 '15 at 17:34

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.