Hermitian K K -theory and Milnor-Witt motivic cohomology over ℤ \mathbb{Z}

Håkon Kolderup, Oliver Röndigs, Paul Arne Østvær · 2025

Read Hermitian K K -theory and Milnor-Witt motivic cohomology over ℤ \mathbb{Z} by Håkon Kolderup, Oliver Röndigs, Paul Arne Østvær (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-2509-16404.txt.

Introduction

Hermitian K-theory sorts algebraic objects with a symmetry twist, and a newer cohomology theory may hold the key to computing it. Håkon Kolderup, Oliver Röndigs, and Paul Arne Østvær use a sliced spectral sequence inside motivic homotopy theory to link the two. They work out the sequence in depth for base schemes of arithmetic interest.

Results include a Grothendieck-Riemann-Roch theorem and the map between the two theories up to degree five for all fields. They also compute the hermitian K-groups and higher Witt groups of the integers. Such concrete values anchor a deep abstract theory.

Catalog record

CREATORHåkon Kolderup, Oliver Röndigs, Paul Arne Østvær
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE149 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2509-16404

Browse the full library · About the Tower of Babel

Full text

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

Hermitian K K -theory and Milnor-Witt motivic cohomology over ℤ \mathbb{Z} Håkon Kolderup, Oliver Röndigs, Paul Arne Østvær Abstract The theme of this paper is to compute hermitian K K -groups in terms of the recently developed theory of Milnor-Witt motivic cohomology. Our approach makes use of the very effective slice spectral sequence within the motivic stable homotopy category, which we analyze in detail for base schemes of arithmetic interest. We show a Grothendieck-Riemann-Roch theorem, determine the map between Milnor-Witt and hermitian K K -theory up to degree five for all fields, and compute the hermitian K K -groups and the higher Witt-groups of the ring of integers. 1 Introduction The effective and very effective slice filtrations are computational tools in motivic homotopy theory with close connections to classical stable homotopy; see [ 42 , 52 , 83 ] . The Adams–Novikov spectral sequence determines the effective slices of the motivic sphere [ 52 , Conjecture 8.1] , [ 71 , Theorem 2.12] . Effective slices are accessible, but the very effective slices yield stronger invariants [ 80 ] . The recognition principle for infinite motivic loop spaces [ 30 , 34 ] illustrates their strength. For the motivic sphere, only the zeroth and first very effective slices are known [ 10 , Theorem 3.30] . The very effective slice spectral sequence sacrifices transparency for robustness: its input is less explicit, but its use is often simpler. The case of hermitian K -theory considered here builds on [ 6 ] and simplifies parts of [ 49 ] . Hermitian K -groups approximate the first stable homotopy groups of motivic spheres [ 59 , 71 , 73 ] and have been studied extensively in recent years [ 21 , 22 , 23 , 27 , 38 , 41 ] . Explicit motivic computations appear in [ 13 , 14 , 15 , 16 , 17 , 31 , 44 , 49 , 75 ] . Let \mathsf{KQ} denote the motivic spectrum representing Karoubi’s hermitian K -groups in the stable motivic homotopy category [ 20 , 39 ] . We compute hermitian K -groups via the very effective slice spectral sequence [ 80 ] (see Section 3 ): \pi_{\star}\tilde{\mathsf{s}}_{\ast}\mathsf{KQ}\implies\pi_{\star}\mathsf{KQ} (1) The description of \tilde{\mathsf{s}}_{\ast}\mathsf{KQ} over fields of characteristic \neq 2 in [ 6 , Theorem 16] , together with our extension to schemes essentially smooth over Dedekind schemes ( Theorem 4.1 ), identifies the E_{1} -page with Milnor–Witt motivic cohomology, a quadratic form analogue of Suslin–Voevodsky motivic cohomology and of Bloch’s higher Chow groups [ 12 ] . The spectral sequence ( 1 ) degenerates rationally. Completeness of the very effective slice filtration yields a Grothendieck–Riemann–Roch theorem for rational hermitian K -groups (cf. [ 26 ] ). The first nontrivial differential in ( 1 ) is determined by motivic Steenrod operations, which allows identification of the image of Milnor–Witt K -theory in hermitian K -theory through degree five, extending [ 4 ] . We also compute the algebraic K -theory, hermitian K -theory, and higher Witt groups of the ring of integers. The higher Witt groups follow the periodic pattern \mathbb{Z},\mathbb{Z}/2,0,0 . Certain algebraic and hermitian K -groups depend on the Kummer–Vandiver conjecture, a phenomenon first observed for K(\mathbb{Z}) by Kurihara [ 48 ] . Outline of the paper Section 2 establishes fundamental facts about algebraic and hermitian K -theory over qcqs (=quasi-compact and quasi-separated) base schemes. The Wood motivic cofiber sequence relates \mathsf{KQ} to algebraic K -theory \mathsf{KGL} through the motivic Hopf map \eta . Section 3 shows strong convergence for the very effective slice spectral sequence relative to any motivic spectrum over a qcqs base scheme of finite Krull dimension. This is in contrast to the situation for the effective slice spectral sequence. We use the slice filtrations to define the Milnor-Witt motivic cohomology spectrum \widetilde{{\mathsf{M}}}{\mathbb{Z}} for schemes which are essentially smooth over a Dedekind scheme. Due to [ 6 ] , the very effective slice spectral sequence relates hermitian K -groups to the homotopy groups of \widetilde{{\mathsf{M}}}{\mathbb{Z}} for fields of characteristic unequal to 2 in [ 12 ] . Section 4 deepens this result by extending it to base schemes of arithmetic interest, such as the integers \mathbb{Z} and other rings of integers in number fields. In Section 5 , we work rationally and deduce rational splittings for both algebraic and hermitian K -theory. Moreover, in Section 6 , we compute examples of hermitian K -groups. These applications are elementary and practical in the sense that there are no obstructions in the form of nontrivial differentials in the very effective slice spectral sequence for \mathsf{KQ} . Section 7 analyzes the very effective zero slice of \mathsf{KQ} . With this in hand, we compute the first differentials in the very effective slice spectral sequence of \mathsf{KQ} in Section 8 . As one would hope based on the proof of Milnor’s conjecture on quadratic forms in [ 67 ] , the very effective slice differentials can be expressed in terms of motivic Steenrod operations. In Section 9 , we compute up to degree five the image of Milnor-Witt K -theory in hermitian K -theory for any field F . Our approach gives a quick proof of the isomorphism in degrees up to three established by Asok-Fasel [ 4 ] . The calculations in degrees four and five are more interesting as they involve certain very effective slice differentials, which are shown to vanish by means of stable motivic cohomology operations and quadratic forms. For example, we establish and use that for the integral motivic cohomology spectrum \mathsf{M}\mathbb{Z} , the group [\mathsf{M}\mathbb{Z},\Sigma^{3+(2)}\mathsf{M}\mathbb{Z}] is cyclic of order six when viewed in the stable motivic homotopy category of the integers \mathbf{SH}(\mathbb{Z}) . In Section 10 , we explicitly compute the homotopy groups \pi_{\star}\mathsf{KQ} over the integers \mathbb{Z} using the very effective slice spectral sequence ( 1 ). As input, we use the Milnor and Bloch-Kato conjectures together with knowledge of \pi_{\star}\tilde{\mathsf{s}}_{\ast}\mathsf{KQ} and very effective slice differentials. We show that the very effective slice spectral sequence for \mathsf{KQ} over \mathbb{Z} collapses at its second page, and deduce the hermitian K -groups of \mathbb{Z} , as given in the table n>0 \mathsf{KQ}_{n,0}(\mathbb{Z}) \mathsf{KQ}_{n+2,1}(\mathbb{Z}) \mathsf{KQ}_{n+4,2}(\mathbb{Z}) \mathsf{KQ}_{n+6,3}(\mathbb{Z}) 8k \mathbb{Z}\oplus\mathbb{Z}/2 \mathbb{Z}/2\oplus K_{n}(\mathbb{Z}) 0 K_{n}(\mathbb{Z}) 8k+1 (\mathbb{Z}/2)^{3} \mathbb{Z}\oplus\mathbb{Z}/2 0 \mathbb{Z}^{2}\oplus\mathbb{Z}/2 (for n>1 ) 8k+2 (\mathbb{Z}/2)^{2}\oplus K_{n}(\mathbb{Z})_{\text{odd}} 0 \mathbb{Z}\oplus K_{n}(\mathbb{Z})_{\mathrm{odd}} (\mathbb{Z}/2)^{2} 8k+3 \mathbb{Z}/w_{4k+2} \mathbb{Z} \mathbb{Z}/2w_{4k+2} (\mathbb{Z}/2)^{2} 8k+4 \mathbb{Z} K_{n}(\mathbb{Z}) \mathbb{Z}/2 K_{n}(\mathbb{Z}) 8k+5 0 \mathbb{Z} \mathbb{Z}/2 \mathbb{Z}^{2} 8k+6 K_{n}(\mathbb{Z})_{\text{odd}} \mathbb{Z}/2 \mathbb{Z}\oplus K_{n}(\mathbb{Z})_{\mathrm{odd}} 0 8k+7 \mathbb{Z}/w_{4k+4} \mathbb{Z}\oplus\mathbb{Z}/2 \mathbb{Z}/w_{4k+4} 0 We refer to Theorem 10.12 for a formulation using motivic cohomology groups instead of denominators of Bernoulli numbers and algebraic K -groups of the integers. The hermitian K -groups of the integers \mathbb{Z} in weights 0 and 2 are in agreement with [ 23 , Theorem 3.2.1] . (This reference does not discuss weights 1 and 3 .) The (8,4) -periodicity of \mathsf{KQ} implies that it suffices to consider these four weights, but fewer weights do not give the complete picture. The group \pi_{7,3}\mathsf{KQ}(\mathbb{Z}) is infinite cyclic, a low-dimensional exception to the pronounced 8 -periodicity in the table. Assuming the Kummer–Vandiver conjecture, stating that a prime p does not divide the class number of the maximal real subfield of the p -th cyclotomic field,

Read the complete text