Explicit Brauer-Manin obstructions on plane quartics

Nils Bruin, Brendan Creutz · 2026

Read Explicit Brauer-Manin obstructions on plane quartics by Nils Bruin, Brendan Creutz (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-2601-16975.txt.

Introduction

Nils Bruin and Brendan Creutz give a way to prove a curve has no whole points. Flat quartic curves are set by fourth-power rules. Some hold no points with whole or ratio marks. Past proofs for that fact were ad hoc. Their 2026 preprint builds a set path based on Brauer-Manin blocks.

Those blocks bar points when local facts clash in sum. The path can also bar splits of low depth. It spreads to smooth curves of all forms. Whole-point hunts sit at the core of number work. A set proof path saves years of case-by-case work.

Catalog record

CREATORNils Bruin, Brendan Creutz
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE55 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2601-16975

Browse the full library · About the Tower of Babel

Full text

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

Explicit Brauer-Manin obstructions on plane quartics Nils Bruin, Brendan Creutz Abstract . We describe a method to show a plane quartic over a number field has no rational points. The method can be adapted to show that a curve does not have divisors of degree 1 or 2 and can be generalized to arbitrary smooth projective curves. Our approach significantly improves on the applicability over previous 2-cover descent methods by not requiring the computation of the full S S -unit group of the étale algebras involved. We illustrate the practicality with several examples, including examples where we determine plane quartics to be of index 2 or 4 when the maximum local index is strictly smaller. 1. Introduction We describe a practical method to show that a plane quartic curve X over a number field k has no rational points, even if it has points everywhere locally. The strategy can also be used to prove that X has no k -rational divisors of degree 1 or 2 . More specifically, we compute obstructions that arise from 2 -cover descent. The main innovation over previous work [ 4 , 3 , 8 , 9 ] is that we do not need to fully determine S -unit groups of number fields to obtain unconditional results: we are already able to obtain nontrivial results by knowing some explicit S -units. Our algorithm generalizes the one developed in [ 6 ] for hyperelliptic curves, and we present a much simplified proof of correctness. Like [ 3 ] , our approach can be readily generalized to apply to arbitrary smooth projective curves, although computational costs rise exponentially with the genus for non-hyperelliptic curves. In Section 2 we consider an étale algebra L over a number field k and a finite set of primes S . In Theorem 2.2 we use Hilbert symbols to establish a pairing which imposes constraints on the image of the localization map \left(\frac{L^{\times}}{k^{\times}L^{\times 2}}\right)_{S}\to\prod_{v\in S}\frac{L_{v}^{\times}}{k_{v}^{\times}L_{v}^{\times 2}}\,. coming from S -units of square norm. In Section 3 we explain how to use this to obtain an obstruction to rational points and/or divisors. We illustrate the argument here for rational points on a plane quartic X over a number field k . The 28 bitangents give rise to an L and a map C_{f}\colon X(k)\to L^{\times}/k^{\times}L^{\times 2} as well as local versions. For a suitable finite set of primes S , the map factors through the part unramified outside S , C_{f}(X(k))\to(L^{\times}/k^{\times}L^{\times 2})_{S}\to\prod_{v\in S}C_{f}(X(k_{v})). Given suitable S -units of square norm we may be able to show that the image of the localization map does not intersect \prod_{v\in S}C_{f}(X(k_{v}) , in which case we can conclude that C_{f}(X(k)) is empty and, hence, that X(k) is empty. This approach contrasts that of [ 3 ] where one must compute generators for all of (L^{\times}/k^{\times}L^{\times 2})_{S} to obtain the image in the localization. Both approaches require computing the local images C_{f}(X(k_{v})) . In Remark 3.7 we describe several ways to reduce the number of primes where this is required. In Section 4 we use our Magma [ 1 ] implementation [ 2 ] on several plane quartic curves over {\mathbb{Q}} occurring in the literature to prove they have no rational points. In Section 5 we describe how the methods can be used to determine the index of a plane quartic and give examples of curves of index 2 or 4 where the maximum local index is strictly smaller. Finally, in Section 6 we describe how our obstruction can be interpreted in terms of finite abelian descent and Brauer-Manin obstructions. Acknowledgements: The first author acknowledges support from Canada’s NSERC, RGPIN-2024-04233. The second author was supported in part by the Marsden Fund administered by Royal Society Te Apārangi. 2. Étale algebras Let L/k be an étale algebra over a number field k . Equivalently, L\simeq\prod_{i=1}^{m}K_{i} is a product of finite field extensions K_{i}/k . Let v be a prime of k with completion k_{v} . Then L_{v}:=L\otimes_{k}k_{v} decomposes as the product L_{v}\simeq\prod_{i=1}^{m}\prod_{w\mid v}K_{i,w} of the completions of the K_{i} at the primes of K_{i} above v . We write {\overline{k}} for an algebraic closure of k and {\overline{L}}=L\otimes{\overline{k}} . Definition 2.1 . A class in L_{v}^{\times}/L_{v}^{\times 2} is unramified if it is represented by some \ell_{v}=(\ell_{i,w})\in L_{v}\simeq\prod K_{i,w} such that all of the extensions K_{i,w}(\sqrt{\ell_{i,w}})/K_{i,w} are unramified. A class in L_{v}^{\times}/k_{v}^{\times}L_{v}^{\times 2} is unramified if it can be represented by a class in L_{v}^{\times}/L_{v}^{\times 2} that is unramified. For a finite set of primes S of k , we write (L^{\times}/k^{\times}L^{\times 2})_{S} and (L^{\times}/L^{\times 2})_{S} for the subgroups of classes whose images in L_{v}/k_{v}^{\times}L_{v}^{\times 2} (resp. in L_{v}^{\times}/L_{v}^{\times 2} ) are unramified at all v\not\in S . The norm map N_{L/k}\colon L\to k induces a homomorphism L^{\times}/L^{\times 2}\to k^{\times}/k^{\times 2} . We write (L^{\times}/L^{\times 2})_{N=1} to denote the kernel of this homomorphism and (L^{\times}/L^{\times 2})_{S,N=1} for its intersection with (L^{\times}/L^{\times 2})_{S} . On each factor of L_{v}\simeq\prod_{i=1}^{m}\prod_{w\mid v}K_{i,w} , the Hilbert symbol defines a nondegenerate bilinear pairing on the finite dimensional {\mathbb{F}}_{2} -vector space K_{i,w}^{\times}/K_{i,w}^{\times 2} taking values in \{0,\nicefrac{{1\!}}{{2}}\}\subset{\mathbb{Q}}/{\mathbb{Z}} . Summing these over the factors yields a bilinear pairing of finite dimensional {\mathbb{F}}_{2} -vector spaces: (2.1) (\,,\,)_{v}\colon\frac{L_{v}^{\times}}{L_{v}^{\times 2}}\times\frac{L_{v}^{\times}}{L_{v}^{\times 2}}\to{\mathbb{Q}}/{\mathbb{Z}}\,. Theorem 2.2 . Let L/k be an étale algebra over a number field k and let S be a finite set of primes of k . (1) The pairings ( 2.1 ) induce a bilinear pairing of finite dimensional {\mathbb{F}}_{2} -vector spaces \langle\,,\rangle_{S}\quad\colon\quad\prod_{v\in S}\frac{L_{v}^{\times}}{k_{v}^{\times}L_{v}^{\times 2}}\times\left(\frac{L^{\times}}{L^{\times 2}}\right)_{S,N=1}\to\;{\mathbb{Q}}/{\mathbb{Z}} by \langle(\ell_{v})_{v\in S},m\rangle_{S}=\sum_{v\in S}(\ell_{v},\operatorname{res}_{v}(m))_{v}\in{\mathbb{Q}}/{\mathbb{Z}}\,. (2) The left kernel of the pairing \langle\,,\rangle_{S} is equal to the image of the restriction map \left(\frac{L^{\times}}{k^{\times}L^{\times 2}}\right)_{S}\to\prod_{v\in S}\frac{L_{v}^{\times}}{k_{v}^{\times}L_{v}^{\times 2}}\,. Proof. We first prove ( 1 ). Let v be a prime of k . For \ell,m\in L_{v}^{\times} we write {\mathcal{A}}(\ell,m) for the corresponding quaternion algebra over L_{v} . The pairing ( 2.1 ) equals the invariant of the corestriction to k_{v} , so (\ell,m)_{v}=\operatorname{inv}_{v}\Cor_{L_{v}/k_{v}}({\mathcal{A}}(\ell,m)) . For a\in k_{v}^{\times} , the corestriction \Cor_{L_{v}/k_{v}}({\mathcal{A}}(a,m)) is the class of the quaternion algebra {\mathcal{A}}(a,N_{L_{v}/k_{v}}(m)) over k_{v} . If N_{L_{v}/k_{v}}(m)\in k_{v}^{\times 2} , then {\mathcal{A}}(a,N_{L_{v}/k_{v}}(m)) has trivial invariant. Thus, the pairing ( 2.1 ) induces a pairing (2.2) \langle\,,\,\rangle_{v}\colon L_{v}^{\times}/k_{v}^{\times}L_{v}^{\times 2}\times(L_{v}^{\times}/L_{v}^{\times 2})_{N=1}\to{\mathbb{Q}}/{\mathbb{Z}}\,. Summing these over the primes in S gives the pairing in the statement. Now we prove ( 2 ). Let k_{S}/k be the maximal extension unramified outside S . We consider Galois cohomology of the self-dual \Gal(k_{S}/k) -modules \mu_{2}({\overline{k}}) and \mu_{2}({\overline{L}}) . Their respective Kummer sequences together with Hilbert’s Theorem 90 give isomorphisms \HH^{1}(k_{S}/k,\mu_{2})\simeq(k^{\times}/k^{\times 2})_{S} , and \HH^{1}(k_{S}/k,\mu_{2}({\overline{L}}))\simeq(L^{\times}/L^{\times 2})_{S} . Under this identification the Poitou-Tate exact sequences for these \Gal(k_{S}/k) -modules give rise to a commutative diagram of linear maps of {\mathbb{F}}_{2} -vector spaces with exact rows: (2.3) {\lx@inpgf@ignorespaces(k^{\times}/k^{\times 2})_{S}} {\lx@inpgf@ignorespaces\prod_{v\in S}k_{v}^{\times}/k_{v}^{\times 2}} {\lx@inpgf@ignorespaces\left(k^{\times}/k^{\times 2}\right)_{S}^{*}} {\lx@inpgf@ignorespaces 1} {\lx@inpgf@ignorespaces\left(L^{\times}/L^{\times 2}\right)_{S}} {\lx@inpgf@ignorespaces\prod_{v\in S}L_{v}^{\times}/L_{v}^{\times 2}} {\lx@inpgf@ignorespaces\left(L^{\times}/L^{\times 2}\right)_{S}^{*}\,,} \scriptstyle{\lx@inpgf@ignorespaces i_{*}} \scriptstyle{\lx@inpgf@ignorespaces i_{*}} \scriptstyle{\lx@inpgf@ignorespaces N_{L/k}^{*}} where the ∗ denotes the dual vector space, the vertical maps are induced by either the inclusion i\colon\mu_{2}({\overline{k}})\to\mu_{2}({\overline{L}}) or the norm N_{L/k}\colon\mu_{2}({\overline{L}})\to\mu_{2} , and the second map in the bottom row is given by (\ell_{v})_{v\in S}\mapsto\left[m\mapsto\sum_{v\in S}(\ell_{v},m)_{v}\right]\,. The claimed surjectivity of the map to \left(k^{\times}/k^{\times 2}\right)_{S}^{*} follows from the local-global principle for \HH^{2}(k,\mu_{2}({\overline{k}}))\simeq\Br(k)[2] . For the reader less familiar with duality theorems in Galois cohomology we note that exactness

Read the complete text