The Art of Branching:

Cobordism Junctions of 10d String Theories · 2026

Read The Art of Branching: by Cobordism Junctions of 10d String Theories (2026) 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-2603-24667.txt.

Introduction

Altavista and co-authors construct junctions where several ten-dimensional string theories meet at a nine-dimensional branch point. Such junctions give a dynamical realisation of cobordisms between string theories, as the Cobordism Conjecture predicts. The construction generalises recent work on interpolations between IIA and IIB domain walls, with a worldsheet description. Extra degrees of freedom stay gapped except at the branch point, where the gap closes and branching begins. Examples join heterotic, type 0, and type II theories and their orientifolds, including chiral theories whose fields flow between branches.

Catalog record

CREATORCobordism Junctions of 10d String Theories
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE149486
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.24667

Browse the full library · About the Tower of Babel

Full text

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

The Art of Branching: Cobordism Junctions of 10d String Theories Abstract IFT-UAM/CSIC-26-35 ZMP-HH/26-7 We describe the explicit construction of configurations of several 10d string theories joining at a 9d junction, providing a dynamical realization of cobordisms between multiple 10d string theories, predicted by the Cobordism Conjecture. We provide the microscopic worldsheet description of the configuration in a generalization of the ‘going up and down the RG flow’ interpolations recently used in the description of IIA/IIB domain wall. The interpolations involve additional degrees of freedom, which are gapped except at the branch point, at which the gap closes and triggers the branching transition. The extra degrees of freedom admit an interpretation in terms of additional dimensions in a supercritical string theory, which reduces to the 10d junction configuration upon closed tachyon condensation. Quantum corrections of the 2d worldsheet theory turn the junction into a strongly coupled lightlike core whose UV resolution lies beyond worldsheet techniques. We construct explicit examples of junctions of 10d heterotic string theories, type 0, and type II theories and orientifolds thereof. Our explicit examples include junctions of 10d chiral theories whose chiral fields flow between different branches. One particularly nice configuration is a 4-branch junction of the IIB theory, with type I, the non-supersymmetric USp(32) theory and the U(32) orientifold of 0B theory, thus assembling the four non-tachyonic descendants of type 0B theory. 1 Introduction One of the central ideas driving progress in string theory (both within the swampland program Vafa:2005ui and more broadly) is the cobordism conjecture McNamara:2019rup . It posits that the (suitably defined) cobordism charges of any consistent quantum gravity configuration should be trivial. This implies the existence of an extremely rich set of spacetime dynamical configurations connecting different theories, specially interesting and non-trivial when applied to string theories or M-theory in their maximal spacetime dimension. One instance is given by end of the world (ETW) boundaries at which spacetime terminates, like the Horava-Witten ETW wall in 11d M-theory Horava:1995qa ; Horava:1996ma (see also Montero:2025ayi ), or the type I’ O8-planes (plus D8-branes) Polchinski:1995df as ETW boundaries of (possibly massive) 10d type IIA theory 1 1 1 See Buratti:2021yia ; Buratti:2021fiv ; Angius:2022aeq ; Blumenhagen:2022mqw ; Angius:2022mgh ; Angius:2023xtu ; Blumenhagen:2023abk ; Angius:2023uqk ; Angius:2024zjv ; Delgado:2023uqk for other dynamical realizations of cobordism ETW boundaries in diverse contexts. . Another interesting class of configurations are 9d interfaces interpolating between 10d string theories, like the 10d type IIA/IIB domain wall (and its 10d 0A/0B cousin) whose properties and construction have been recently explored in Heckman:2025wqd ; Anastasi:2026cus ; Torres:2026vxx . There is however yet another kind of configurations which have not been considered in the literature so far (see however the early work Hellerman:2010dv ): configurations of more than two 10d string theories joining at a 9d junction, see Figure 1 a. The aim of this paper is to provide explicit constructions of such junctions for various sets of 10d string theories. At the risk of giving away the punchline, we advance that we construct explicit junctions of (up to six) 10d type 0A or 0B theories; junctions of various non-supersymmetric heterotic theories, including a junction of three 10d non-supersymmetric (but chiral!) E_{8}\times SO(16) heterotic theories; junctions of four type IIA or IIB theories; and a beautiful 4-branch junction of one 10d type IIB theory, one 10d type I theory, one 10d non-supersymmetric non-tachyonic USp(32) theory, and one 10d non-tachyonic U(32) orientifold of 0B theory, precisely the four 10d non-tachyonic theories related to type 0B theory. Figure 1: a) Sketch of a 3-branch junction of 10d string theories. b) Depiction of the interplay between multi-branch junctions and ETW boundaries for a single branch. The main reason why multi-branch junction configurations have not been considered in the literature is that they can be regarded as a cobordism between one 10d string theory in one branch and the disconnected sum of the 10d string theories in the remaining branches. It is non-trivial to make sense of this disconnected sum, and it is even less clear what tools one could use to even formulate them consistently. We advocate that such junction configurations can be appropriately defined in terms of a worldsheet formulation, which encompasses all the string theories in the junction. The worldsheet approach is analogous to that successfully employed in Anastasi:2026cus to describe the 10d type IIA/IIB domain wall (and its 0A/0B cousin). In the context of 9d interfaces connecting 10d string theories, the key concept is regarding the 2d worldsheet CFTs of the theories as the IR limit of non-trivial RG flows from UV 2d QFTs, which are connected via an interpolating parameter deforming the UV lagrangian. The worldsheet version of the cobordism thus corresponds to a transition along a one-parameter trajectory moving up and down the RG flow in the sense of Gaiotto:2019asa (analogous to the application in the heterotic context with the notion of cobordism charges formalized in terms of elliptic genera or topological modular form classes Witten:1986bf ; Gukov:2018iiq ; Gaiotto:2019asa ; Tachikawa:2021mvw ; Tachikawa:2021mby ; Tachikawa:2023nne ; Tachikawa:2025awi ; Moore:2025tmt ). Actually, the dynamical realization in Anastasi:2026cus motivated enlarging the class of interpolations by allowing them to cross non-compact CFTs, thus permitting non-trivial changes in the topological order of the 2d theory. The dynamical realization of these interpolations in Anastasi:2026cus simply corresponds to letting the parameter depend on a 10d spacetime coordinate, \lambda(X) , and explicit examples of such dynamical realizations were provided, in which the additional degrees of freedom admit an interpretation in terms of supercritical strings. Our worldsheet construction of junctions shares a similar spirit, with one extra ingredient, namely branching . A simplified description of how this works is as follows. We start with one 2d CFT 1 describing a 10d string theory, and move up the RG flow to a UV theory QFT 1 with a parameter \lambda_{1} . As we move this parameter in some coupling space, we hit a branching point, at which two extra parameters \lambda_{2} , \lambda_{3} appear, which are mutually incompatible (among themselves and with \lambda_{1} ): turning on one of them makes the others disappear. We may hence continue the interpolation by turning on \lambda_{2} , with \lambda_{3} (and \lambda_{1} ) disappearing, or viceversa. By letting the corresponding theories QFT 2 and QFT 3 run down the RG flow, we end up with two theories CFT 1 , CFT 2 , describing two 10d disconnected string theories. This is the worldsheet version of a spacetime junction. By considering a 3-branch junction of spacetime and letting the parameters depend on the corresponding spacetime coordinates, we have a dynamical realization of the junction. Our actual realizations include an extra subtlety: given that in the physical realization the parameters depend on the coordinates themselves, they effectively turn into 2d fields, so that the mutual incompatibility of the two parameters/fields \lambda_{2}\sim X and \lambda_{3}\sim Y translates into the fact that a vev for one makes the other massive, and viceversa. A simple realization, arising as a subsector in our configurations, is the following: consider (a sector) of our CFT 1 to be a 2d (1,1) theory with one free real superfield Z . For generic Z , this can be seen as the IR limit of

Read the complete text