4
$\begingroup$

I have a kind of vague question.

Two non-degenerate symplectic vector space of same dimension (say $\mathbb{R}^{2n}$) are isomorphic. Then why all noncommutative tori of same dimension aren't isomorphic?

$\endgroup$
1
  • 3
    $\begingroup$ The "semiclassical" analogue of a noncommutative torus is something like a symplectic torus, and here it's not true that two symplectic tori of the same dimension are isomorphic: they can be distinguished e.g. by their symplectic volume. $\endgroup$ Jun 19, 2016 at 19:16

1 Answer 1

7
$\begingroup$

If you view a noncommutative $2n$-torus not as a strict deformation quantisation along a $\mathbb{R}^{2n}$-action (that happens to be periodic) but as a strict deformation quantisation along a $\mathbb{T}^{2n}$-action, then what actually matters isn't the non-degenerate symplectic form $\theta : \mathbb{R}^{2n} \times \mathbb{R}^{2n} \to \mathbb{R}$ but rather the induced alternating bicharacter $\lambda_\theta : \mathbb{Z}^{2n} \times \mathbb{Z}^{2n} \to U(1)$, defined by $ \lambda_\theta(\mathbf{m},\mathbf{n}) := e^{2\pi i \theta(\mathbf{m},\mathbf{n})} $, which governs the commutation relations on the corresponding noncommutative $2n$-torus. Indeed, in general, the noncommutative $N$-tori are classified up to $\mathbb{T}^{N}$-equivariant $\ast$-isomorphism by the $U(1)$-valued alternating bicharacters on $\mathbb{Z}^{N}$, the Pontrjagin dual of $\mathbb{T}^{N}$. It's now easy to see that two non-degenerate symplectic forms $\theta$ and $\theta^\prime$ on $\mathbb{R}^{2n}$ needn't give isomorphic $U(1)$-valued alternating bicharacters $\lambda_\theta$ and $\lambda_{\theta^\prime}$ on $\mathbb{Z}^{2n}$ (in the sense that $\lambda_{\theta^\prime} = \Phi^\ast \lambda_\theta$ for some $\Phi \in GL(2n,\mathbb{Z})$), which suggests that one really shouldn't expect all noncommutative $2n$-tori to be isomorphic.

For example, let's take a closer look at the case when $n=1$. Given $\theta$, $\theta^\prime \in \mathbb{R} \setminus \{0\}$, one can check that the non-degenerate symplectic forms $\theta dx \wedge dy$ and $\theta^\prime dx \wedge dy$ on $\mathbb{R}^2$ give isomorphic $U(1)$-valued alternating bicharacters on $\mathbb{Z}^2$ if and only if $\theta^\prime = \pm \theta \bmod \mathbb{Z}$. As it turns out, when $\theta$ and $\theta^\prime$ are both irrational, a $K$-theoretic argument confirms that the irrational noncommutative $2$-tori $C(\mathbb{T}^2_\theta)$ and $C(\mathbb{T}^2_{\theta^\prime})$ are $\ast$-isomorphic if and only if $\theta^\prime = \pm \theta \bmod \mathbb{Z}$.

In the higher-rank case, $\ast$-isomorphism of $C^\ast$-algebraic noncommutative tori becomes a bit more delicate. Still, at the level of Frechet pre-$C^\ast$-algebras, at least, Cuntz--Elliott--Goodman--Jorgensen [1985] showed that for $\theta$ and $\theta^\prime$ in a certain dense subset of $\wedge^2 \mathbb{R}^N$, the smooth simple noncommutative $N$-tori $C^\infty(\mathbb{T}^N_\theta)$ and $C^\infty(\mathbb{T}^N_{\theta^\prime})$ are $\ast$-isomorphic if and only if the induced $U(1)$-valued alternating bicharacters on $\mathbb{Z}^N$ are isomorphic.

$\endgroup$

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.