Cusps in 3d gravity

Douglas Stanford · 2025

Read Cusps in 3d gravity by Douglas Stanford (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2510-19920.txt.

Introduction

Douglas Stanford and Cynthia Yan propose a fix for a divergence that plagues the gravitational sum over topologies in three dimensions. Hyperbolic three-manifolds have volumes that accumulate at limit points, which are cusped manifolds, making the sum over geometries diverge. Their proposal is to renormalize by including the cusped manifold itself as a counterterm.

The move reinterprets the zeta-function regularization that Maloney and Witten used for a sum over black hole geometries. For pure supergravity, cusps with even spin structure play the same canceling role. Cusps with odd spin structure are not needed against any divergence, the authors note, but they earn their place by making an index nonzero, a small fact with structural consequences.

Catalog record

CREATORDouglas Stanford
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE43 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2510-19920

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Cusps in 3d gravity Douglas Stanford Abstract Three dimensional hyperbolic manifolds have accumulation points in the spectrum of their volumes, leading to a divergence in the sum over topologies. The limit points are cusped hyperbolic manifolds, and we propose to renormalize the sum by including the cusped manifold as a counterterm. This gives a reinterpretation of the zeta-function regularization procedure used by Maloney and Witten in the sum over SL (2,\mathbb{Z}) black holes. For pure \mathcal{N}=1 supergravity, cusps with even spin structure can be used in a similar way. Cusps with odd spin structure are not needed to cancel any divergence, but they find an application by making the index nonzero. 1 Introduction In the path integral approach to quantum gravity, the goal is to sum over the topology and geometry of d -dimensional manifolds M with fixed boundary conditions \partial M : Z(\partial M)=\sum_{\text{topology}(M)}\int_{\text{geometry}(M)}e^{-I(M)}. (1.1) To what extent does this make sense, especially the sum over topologies? In d=2 dimensions, the sum makes sense as asymptotic series, because topologies have an ordering by their integer Euler characteristic \chi . If we include \chi in the action with a large negative coefficient, complicated topologies will be suppressed. We will focus on the case d=3 , and we will take I to be the Einstein-Hilbert action for gravity with a negative cosmological constant {}I=-\frac{1}{16\pi G}\left[\int_{M}\sqrt{g}\left(R+2\right)+2\int_{\partial M}\sqrt{h}(K-1)\right]. (1.2) For small G , the integral over geometries with a given topology will be dominated by a saddle point (if one exists) satisfying Einstein’s equations R_{\mu\nu}-\frac{1}{2}g_{\mu\nu}R=g_{\mu\nu} . Such a saddle point is a hyperbolic manifold, for which the action reduces to 1 1 1 For a closed manifold vol is literally the hyperbolic volume. For manifold with an asymptotic boundary, this is generalized to the renormalized volume, see e.g. [ 1 ] . I=\frac{\text{vol}(M)}{4\pi G}. (1.3) So saddle points are weighted by e^{-\#\text{vol}(M)} , and big complicated manifolds are suppressed just like in two dimensions. Indeed, (minus) the hyperbolic volume can be considered a type of generalization of the Euler characteristic to three dimensions. However, there are some problems. For one thing, there are infinitely many topologies with no hyperbolic metric. These “off shell topologies” should also contribute to ( 1.1 ), but tools to compute them are still under development. 2 2 2 See [ 2 , 3 , 4 , 5 , 6 , 7 , 8 , 9 , 10 ] for some ideas and puzzles. For a second thing, in three dimensions (and only three [ 11 ] ) the spectrum of hyperbolic volumes has accumulation points. In particular, there are infinitely many topologies with hyperbolic volume less than a certain finite value. 3 3 3 The first accumulation point for closed orientable manifiolds is at \text{vol}(\text{figure 8 complement})\approx 2.02988 . So the suppression by the volume is not enough to make ( 1.1 ) an asymptotic series. 4 4 4 It might be that the first problem resolves the second: maybe the contribution of off-shell manifolds cancels the divergence from the sum over accumulation points of hyperbolic manifolds. That would be wonderful, but in this paper we will propose a way to make sense of ( 1.1 ) if this miracle does not occur. This problem was encountered in [ 12 , 13 , 14 ] , and in [ 14 ] Maloney and Witten defined the divergent sum near an accumulation point using a version of zeta function regularization. In this paper we will reinterpret the Maloney/Witten prescription slightly. In particular, we introduce new contributions to ( 1.1 ) from manifolds with cusps, and we adjust the fugacity for cusps to cancel the divergence from the accumulation points. Our analysis is conditional on an assumption that the partition function of a single cusp is one-loop exact. A separate motivation for including cusps (in supergravity) is discussed in section 5 . 2 Universal origin of accumulation points There is a nice fact about three-manifolds that makes it possible to resolve all accumulation points at once. Basically, accumulation points have a universal origin. 5 5 5 This follows from Jorgensen’s theorem presented as theorem 5.12.1 of [ 15 ] , see also [ 11 ] . In each case, what is happening is the formation of an infinite cusp, \rightarrow\infty (2.1) The cross section that one gets by cutting through the cusp with the plane shown above is conformal to a torus with some modular parameter \tau : Re Im 0 1 \tau (2.2) This cusped manifold is the limit of an accumulating set of smooth hyperbolic manifolds, constructed by cutting the cusp at some location and gluing in a solid torus in such a way that the cycle c\tau+d becomes contractible. Here c,d should be relatively prime integers, and Thurston’s hyperbolic Dehn surgery theorem (theorem 5.8.2 of [ 15 ] ) shows that for all but a finite number of pairs \{c,d\} , this topology admits a hyperbolic metric. When c,d are large, the resulting manifold M_{c,d} resembles the original cusped manifold out to a distance \log|c\tau+d| “along the cusp,” after which point the c\tau+d cycle smoothly contracts. The volume approaches that of the original cusped manifold from below [ 16 ] \text{vol}(M_{c,d})\approx\text{vol}(M_{\infty})-\pi^{2}\frac{\text{Im}(\tau)}{|c\tau+d|^{2}}. (2.3) For the purposes of quantum gravity, these accumulation points are bad because they make the sum over the saddle points M_{c,d} divergent. 6 6 6 Based on what we’ve said so far, this is actually not completely clear. It is clear that the suppression factor e^{-\text{vol}(M)/(4\pi G)} coming from the Einstein-Hilbert action will not make the sum converge, since the volume approaches a limiting value as c,d go to infinity. But one could imagine that the integral over small fluctuations around the saddle point could give a correction that makes the sum converge. Unfortunately, the one-loop determinant does not decay rapidly enough [ 14 , 17 ] as we will review below. This point was discussed by Collier, Eberhardt and Zhang [ 18 ] , who pointed out that precisely this divergent sum was already encountered in quantum gravity work [ 12 , 13 , 14 ] . Maloney and Witten assigned the sum a finite value using zeta function regularization [ 14 ] , implicitly subtracting an infinite contribution. In this paper we treat the sum using a related procedure in which the divergence is made explicit. We then include a new contribution from the limiting cusp, as a counterterm that cancels the divergence. 3 Cancellation in a simple case In this section, we show that a single cusp can be used to cancel the divergence in the simplest setting in which an accumulation point arises: the sum over so-called SL (2,\mathbb{Z}) black holes. The procedure is as follows: 1. Introduce a convergence factor into the sum. To preserve modular invariance, this factor should depend on an intrinsic feature of the geometry, and not the modular frame in which we consider it. A convenient choice is to insert e^{-\epsilon/\ell_{c,d}} where \ell_{c,d}=2\pi\text{Im}(\tau)/|c\tau+d|^{2} is the length of the closed geodesic where c\tau+d contracts, as measured by the metric of the classical solution. 2. Include a new contribution from the limiting cusp, with a coefficient \mu<0 . 3. Take \epsilon\to 0 , adjusting \mu(\epsilon) so that the

Read the complete text