Definition

Let 0<p<0<p<\infty. A (A,H,D)(\mathcal A,H,D) is pp-summable if

(1+D2)1/2Lp(H),(1+D^2)^{-1/2}\in\mathcal L^p(H),

where Lp(H)\mathcal L^p(H) is the . Equivalently,

Tr ⁣((1+D2)p/2)<.\operatorname{Tr}\!\left((1+D^2)^{-p/2}\right)<\infty.

This is a quantitative compact-resolvent condition: the singular values of the regularized inverse of DD are pp-summable, with multiplicity. The regularization by 1+D21+D^2 treats kerD\ker D harmlessly. For an invertible DD, it may be replaced by D1|D|^{-1} without changing membership in the relevant ideal.

Equivalent spectral tests

When DD has and eigenvalues λn\lambda_n, repeated with multiplicity, pp-summability is equivalent to

n(1+λn2)p/2<.\sum_n(1+\lambda_n^2)^{-p/2}<\infty.

It is also equivalent to finiteness of the spectral zeta function Tr(Dp)\operatorname{Tr}(|D|^{-p}) after omitting the finite-dimensional kernel, provided DD is invertible away from zero. These equivalences follow directly from functional calculus and the definition of Schatten ideals Gracia-Bondía–Várilly–Figueroa, §10.1.

If the triple is pp-summable, then it is qq-summable for every q>pq>p. The converse can fail at the endpoint, so the value pp must not be inferred merely from summability at all larger exponents.

Examples and consequences

The canonical spin spectral triple on a closed nn-dimensional manifold is pp-summable for every p>np>n, by Weyl eigenvalue asymptotics. At the critical exponent p=np=n, its inverse generally belongs only to a , while its nnth power supports a . Thus saying that the triple has dimension nn is not the same as saying it is strictly nn-summable.

implies because exponential decay dominates every negative power. It also makes sufficiently long products of Dirac or bounded-transform commutators trace class, enabling finite-degree .

Conventions and scope

For 0<p<10<p<1, the same trace formula defines membership in a quasi-Banach Schatten ideal rather than a Banach ideal. This extension is included in the core convention and in the linked Schatten-class convention. replace the by a specified faithful normal semifinite trace.

References
  1. A. Connes, Noncommutative Geometry, Academic Press, 1994. Author-hosted text. Relevant: Chapter IV, §§1–2 on summable Fredholm modules and trace ideals.
  2. J. M. Gracia-Bondía, J. C. Várilly, and H. Figueroa, Elements of Noncommutative Geometry, Birkhäuser, 2001. DOI record. Relevant: §10.1 on summability of spectral triples.