DOCSLIB.ORG

Integral Bilinear Forms and Dynkin Diagrams

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Integral Bilinear Forms and Dynkin Diagrams APPENDIX A Integral Bilinear Forms and Dynkin Diagrams (AI) Definition. A lattice (M, ( , )) is a pair consisting of a finitely generated free abelian group M together with a bilinear form ( , ): M x M -+ 7l. which is either: (i) symmetric, i.e., (x, y) = (y, x) for all x, y EM; (ii) skew-symmetric, i.e. (x, y) = -(y, x) for all x, y E M. For simplicity, we refer to M as a lattice when the bilinear form ( , ) is clear. The main example which we have in mind is the Milnor lattice Lx = (iin(F), < , » of an n-dimensional IHS X: f = 0 as defined in (3.3.6). (A2) Definition. A symmetric lattice M is even if (x, x) == 0 (mod 2) for any xEM. A symmetric lattic M is odd if it is not even. Note that the Milnor lattice Lx of an even dimensional IHS X is an even lattice by (3.3.7). (A3) Definition. The lattice M is nondegenerate if it satisfies the following two equivalent conditions: (i) Rad(M):= {x E M; (x, y) = 0 for all y EM} = 0; (ii) the natural group homomorphism iM : M -+ M' := Hom(M, 7l.), x 1--+ (x, .) is injective. The lattice M is unimodular if iM is an isomorphism. When the lattice M is non degenerate, we call the finite group D(M) = coker(iM ) 220 Appendix A. Integral Bilinear Forms and Dynkin Diagrams the discriminant group of M and denote its order ID(M)I by det(M}. The quo­ tient lattice M = M/Rad M is called the reduced lattice associated to M. (A4) Exercise. (i) Show that D(M) = D( -M), D(Ml Ef) M 2 ) = D(Ml) ® D(M2 ). Here - M is the lattice (M, - ( , » obtained by changing to signs of all the products (x, y) and Ml Ef) M2 is the direct sum lattice, i.e., (Xl + X2' Yl + Y2) = (Xl' yd + (x2, Y2) for all Xl' Yl E Ml and X2' Y2 E M 2· (ii) Show that det(Lx} = IA(I)I, for a nondegenerate Milnor lattice Lx, where A denotes the characteristic polynomial of the monodromy operator of X. Hint. Recall the proof of (3.4.7). (A5) Lemma. If N c M is a sub lattice in the nondegenerate lattice M such that rk M = rk M, then the quotient group M/N is finite and its order satisfies the relation IM/NI2 det M = det N. Proof Use the structure of the subgroup N as described in [La], p. 393. 0 (A6) Definition. Let M and N be two lattices. A group homomorphism qJ: M ~ N is called a lattice morphism if (x, y) = (qJ(x), cp(y)) for all x, Y E M. A lattice morphism qJ is called an embedding (resp. an isomorphism) if qJ is a group monomorphism (resp. isomorphism). (A 7) Theorem (Structure of Skew-Symmetric Lattices). Any skew-symmetric lattice M is isomorphic to a direct sum of "elementary" skew-symmetric lattices (Z2, ( )d,) 61'" EB (Z2, ( }dk ) Ef) (zm, ( )0) where Z2 = Ze l + Ze2 and (e l , e2 )d = d; for some integers d; > 0 and (x, y)o = 0 for all x, y E zm. Moreove'r the positive integers d; are uniquely determined if we ask in addition that dl ld 2 •• ·Idk • For a proof, see [La], p. 3S0. In particular, M is nondegenerate if and only if m = 0 and then det M = d i ... dt. And M is unimodular if and only if m = 0 and d 1 = ... = dt = 1. We use the notation 12k for this unimodular skew-symmetric lattice of rank 2k. (AS) Exercise. Let X: f = 0 be a reduced plane curve singularity with Milnor number Il and number of irreducible branches r. Then the Milnor lattice Lx is isomorphic to the direct sum Appendix A. Integral Bilinear Forms and Dynkin Diagrams 221 Hint. Use the Wang exact sequence (3.1.18) associated to the Milnor fibration for X. When M is a symmetric lattice we can obtain a real bilinear form tensoring by IKt All the usual terminology for the latter applies to M and hence we can speak about the signature sign M = (m_, mo, m+). For instance, M is nega­ tive definite when mo = m+ = 0, and M is indefinite when m_ > 0 and m+ > O. The difference m+ - m_ is called the index of the lattice. The classification of the indefinite unimodular lattices is due to Milnor [M2] and we do not recall it here since it is more complicated than (A 7). Moreover, the Milnor lattices Lx for the most familiar singularities X are neither indefinite nor unimodular, so the best way to introduce them to the reader is just by listing them. As remarked in (3.3.23), to each IHS X there are two naturally associated Milnor lattices, a symmetric one Lx and a skew­ symmetric one V;. In what follows, we list the Dynkin diagrams for several important classes of singularities, namely for all simple and unimodular sin­ gularities in Arnold's lists [AGVl]. These Dynkin diagrams determine at once the symmetric lattices L"x and they determine also the skew-symmetric lattices Lx via Gabrielov's result (3.3.22'). Recall that each vertex in a Dynkin diagram D corresponds to a vanishing cycle 11, with <l1i, l1i> = -2. Two distinct vertices, corresponding to vanishing cycles l1i and I1j , respectively, are joined by kedges (resp. k dotted edges) if their intersection number (l1 i, I1j ) is k (resp. -k). Hence j • • j • • Note also that the vertices in a Dynkin diagram are numbered (corresponding to the order of the vanishing cycles 111 ... , 11/l in the associated distinguished basis 11), unless any order is good (i.e., any order of l1/s corresponds to a distinguished basis). All the singularities discussed in what follows are stably equivalent to sur­ face singularities in (:3. The reader has already encountered the notations for (and the equations of) these singularities in Chapter 2, §4, so here we describe only their Dynkin diagrams and some of their properties. 222 Appendix A. Integral Bilinear Forms and Dynkin Diagrams (A9) Dynkin Diagrams for the Simple Singularities. Type Dykin diagram Ak,k~l --... - (kpoints) Dk,k~4 -- ... ---< (kpoints) • • I • • • • • I • • • • • • I • • We let All: denote the (symmetric) Milnor lattice of the singularity A" and use the same convention for all the other singularities discussed. A first striking fact about the Dynkin diagrams ofthe simple singularities is that they coincide with their resolution graphs, see (2.4.3). This remark gives us the first part ofthe following result (for a complete proof we refer to [Df4J). (AIO) Proposition. (i) The Milnor lattices A", Dk , E6 , E7 and Es are negative definite. (ii) Any IHS X whose symmetric Milnor lattice is negative definite is a simple singularity. (iii) det All: = k + 1, det D" = 4, det E, = 9 - I fori = 6,7,8. (All) Dynkin Diagrams for the Simple-Elliptic Singularities £6, £7' and Ea. First we associate to each of these simple-elliptic singularities £ a triple of positive integers (p, q, r) as follows £6 1--+ (3, 3, 3), £7 1--+ (2,4,4), £8 1--+ (2, 3, 6). A look at the equations for these singularities given in (2.4.9) will explain to the reader where these triples come from! See also (A13). Next to each triple Appendix A. Integral Bilinear Forms and Dynkin Diagrams 223 (p, q, r) we associate the following diagram: p-l q - 1 T(p,q,r) - ... >--..... - with a numbering such that the vertices denoted by i and i + 1 get consecutive indices. (AI2) Proposition. (i) The correspondence E 1--+ (p, q, r) 1--+ T(p, q, r) associates to each simple­ elliptic singularity E,,(k = 6, 7, 8) a Dynkin diagram. (ii) The lattice E" is negative semidefinite, its radical has rank 2 and the cor­ responding reduced Milnor lattice E" is exactly the Milnor lattice E". (iii) Any IHS X whose symmetric Milnor lattice is negative semidefinite is a simple-elliptic singularity. For a proof, see, for instance, [Df4]. For a proof of the following more complicated results we refer to Ebeling [EI] and [E2]. (A 13) Proposition. (i) A Dynkin diagram for the cusp singularity T",q,. is given by the diagram T(p, q, r) described above. (ii) sign(T",q,.) = (p + q + r - 3, 1, 1). (iii) det(J;"q,.) = pqr(l - lip - llq - llr). This is not a surprise, since we can regard the simple-elliptic singularities E" as special cases of the T",q,.-singularities for 1 1 1 -+-+-=1. p q r (AI4) Proposition. (i) A Dynkin diagram for the triangle singularity Dp,q,. is given by the follow­ ing diagram: 224 Appendix A. Integral Bilinear Forms and Dynkin Diagrams p' - 1 q' - 1 r T(p,q,r) _ ... ~-..... - with a numbering such that the vertices denoted by i and i + I get consecu­ tive indices and (p', q', r') are the Gabrielov numbers given in (2.4.7). (ii) sign(Dp,q,r) = (p' + q' + r' - 2,0, 2). (iii) det(Dp,q,r) = p' q'r'(l - lip' - l/q' - l/r'). Recall that Dp.q,r denotes in fact two singularities: one weighted homoge­ neous and the other one semiweighted homogeneous. By our discussion in (3.1.19) it follows that the two singularities have the same topological in­ variants, in particular, the same Dynkin diagrams. (A15) Exercise. Check that the Dolgacev numbers (p, q, r) and the Gabrielov numbers (p', q', r') associated to a given triangle singularity Dp,q,r in (2.4.7) satisfy the following relation pqr(l- ~ - ~ -~) = p'q'r'(l- ~ - ~ - ~).
Load more

Recommended publications
  • Automorphic Forms for Some Even Unimodular Lattices
    AUTOMORPHIC FORMS FOR SOME EVEN UNIMODULAR LATTICES NEIL DUMMIGAN AND DAN FRETWELL Abstract. We lookp at genera of even unimodular lattices of rank 12p over the ring of integers of Q( 5) and of rank 8 over the ring of integers of Q( 3), us- ing Kneser neighbours to diagonalise spaces of scalar-valued algebraic modular forms. We conjecture most of the global Arthur parameters, and prove several of them using theta series, in the manner of Ikeda and Yamana. We find in- stances of congruences for non-parallel weight Hilbert modular forms. Turning to the genus of Hermitian lattices of rank 12 over the Eisenstein integers, even and unimodular over Z, we prove a conjecture of Hentschel, Krieg and Nebe, identifying a certain linear combination of theta series as an Hermitian Ikeda lift, and we prove that another is an Hermitian Miyawaki lift. 1. Introduction Nebe and Venkov [54] looked at formal linear combinations of the 24 Niemeier lattices, which represent classes in the genus of even, unimodular, Euclidean lattices of rank 24. They found a set of 24 eigenvectors for the action of an adjacency oper- ator for Kneser 2-neighbours, with distinct integer eigenvalues. This is equivalent to computing a set of Hecke eigenforms in a space of scalar-valued modular forms for a definite orthogonal group O24. They conjectured the degrees gi in which the (gi) Siegel theta series Θ (vi) of these eigenvectors are first non-vanishing, and proved them in 22 out of the 24 cases. (gi) Ikeda [37, §7] identified Θ (vi) in terms of Ikeda lifts and Miyawaki lifts, in 20 out of the 24 cases, exploiting his integral construction of Miyawaki lifts.
    [Show full text]
  • Genus of Vertex Algebras and Mass Formula
    Genus of vertex algebras and mass formula Yuto Moriwaki * Graduate School of Mathematical Science, The University of Tokyo, 3-8-1 Komaba, Meguro-ku, Tokyo 153-8914, Japan Abstract. We introduce the notion of a genus and its mass for vertex algebras. For lattice vertex algebras, their genera are the same as those of lattices, which play an important role in the classification of lattices. We derive a formula relating the mass for vertex algebras to that for lattices, and then give a new characterization of some holomorphic vertex operator algebras. Introduction Vertex operator algebras (VOAs) are algebraic structures that play an important role in two-dimensional conformal field theory. The classification of conformal field theories is an extremely interesting problem, of importance in mathematics, statistical mechanics, and string theory; Mathematically, it is related to the classification of vertex operator algebras, which is a main theme of this work. An important class of vertex operator algebras can be constructed from positive-definite even integral lattices, called lattice vertex algebras [Bo1, LL, FLM]. In this paper, we propose a new method to construct and classify vertex algebras by using a method developed in the study of lattices. A lattice is a finite rank free abelian group equipped with a Z-valued symmetric bilinear form. Two such lattices are equivalent (or in the same genus) if they are isomorphic over the ring of p-adic integers Zp for each prime p and also equivalent over the real number arXiv:2004.01441v2 [math.QA] 5 Mar 2021 field R. Lattices in the same genus are not always isomorphic over Z.
    [Show full text]
  • 1 Positive Definite Even Unimodular Lattices
    Algebraic automorphic forms and Hilbert-Siegel modular forms Tamotsu Ikeda (Kyoto university) Sep. 17, 2018, at Hakuba 1 Positive definite even unimodular lattices Let F be a totally real number field of degree d. Let n > 0 be an integer such that dn is even. Then, by Minkowski-Hasse theorem, there exists a quadratic space (Vn;Q) of rank 4n with the following properties: (1) (V; Q) is unramified at any non-archimedean place. (2) (V; Q) is positive definite ai any archimedean place. Definition 1. An algebraic automorphic form on the orthogonal group OQ is a locally constant function on OQ(F )nOQ(A). Put 1 (x; y) = (Q(x + y) − Q(x) − Q(y)) x; y 2 V: Q 2 Let L be a o-lattie in V (F ). The dual lattice L∗ is defined by ∗ L = fx 2 Vn(F ) j (x; L)Q ⊂ og: Then L is an integral lattice if and only if L ⊂ L∗. An integral lattice L is called an even lattice if Q(x) ⊂ 2o 8x 2 L: Moreover, an integral lattice L is unimodular if L = L∗. By the assumption (1) and (2), there exists a positive definite even unimodular latice L0 in V (F ). Two integral lattices L1 and L2 are equivalent if there exists an element g 2 OQ(F ) such that g · L1 = L2. Two integral lattices L1 and L2 are in the same genus if there 1 exists an element gv 2 OQ(Fv) such that g · L1;v = L2;v for any finite place v. The set of positive definite even unimodular lattices in V (F ) form a genus.
    [Show full text]
  • Lattice Theory, K3 Surfaces, and the Torelli Theorem
    UMass, Reading Seminar in Algebraic Geometry, Lattice theory, K3 surfaces, and the Torelli theorem Luca Schaffler October 4, 2019 1 What is a lattice? Definition 1.1. A lattice is a pair (L; ·), where L is a finitely generated free abelian group and ·: L × L ! Z such that (v; w) 7! v · w is a symmetric bilinear form. • L is even if v2 := v · v is an even number for all v 2 L. • The Gram matrix of L with respect to a basis fe1; : : : ; eng for L is the matrix of intersection (ei · ej)1≤i;j≤n. • The rank of L is defined to be the rank of one of its Gram matrices. • A lattice L is unimodular if the determinant of one of its Gram matrices is ±1. Equivalently, L of maximal rank is unimodular provided the discriminant group ∗ ∗ L =L is trivial (we let L = HomZ(L; Z)). • We can define the signature of L as the signature of one of its Gram matrices. By saying signature (a; b), we mean that a (resp. b) is the number of positive (resp. negative) eigenvalues. If L has signature (a; b), then we say that L is indefinite provided a; b > 0. 2 Example 1.2. U = (Z ; ·) with (v1; v2) · (w1; w2) = v1w2 + v2w1 is called the hyperbolic plane. Check that this is even, unimodular, and of signature (1; 1). 1 8 Example 1.3. E8 = (Z ;B), where 0 −2 1 0 0 0 0 0 0 1 B 1 −2 1 0 0 0 0 0 C B C B 0 1 −2 1 0 0 0 0 C B C B 0 0 1 −2 1 0 0 0 C B = B C : B 0 0 0 1 −2 1 0 1 C B C B 0 0 0 0 1 −2 1 0 C B C @ 0 0 0 0 0 1 −2 0 A 0 0 0 0 1 0 0 −2 This matrix can be reconstructed from the E8 Dynkin diagram.
    [Show full text]
  • Arxiv:2001.07094V3 [Math.NT] 10 May 2021 Aeo Nirdcbeplnma)Alclgoa Principl a Local-Global Conditions, a Local Polynomial) I Are Irreducible Mcmullen 20]
    ISOMETRIES OF LATTICES AND HASSE PRINCIPLES EVA BAYER-FLUCKIGER Abstract. We give necessary and sufficient conditions for an integral polynomial without linear factors to be the characteristic polynomial of an isometry of some even, unimodular lattice of given signature. This gives rise to Hasse principle questions, which we answer in a more general setting. As an application, we prove a Hasse principle for signatures of knots. 0. Introduction In [GM 02], Gross and McMullen give necessary conditions for a monic, irreducible polynomial to be the characteristic polynomial of an isometry of some even, unimodular lattice of prescribed signature. They speculate that these conditions may be sufficient; this is proved in [BT 20]. It turns out that the conditions of Gross and McMullen are local conditions, and that (in the case of an irreducible polynomial) a local-global principle holds. More generally, if F Z[X] is a monic polynomial without linear factors, the conditions of Gross∈ and McMullen are still necessary. Moreover, they are also sufficient everywhere locally (see Theorem 25.6). However, when F is reducible, the local-global principle no longer holds in general, as shown by the following example Example. Let F be the polynomial (X6 + X5 + X4 + X3 + X2 + X + 1)2(X6 X5 + X4 X3 + X2 X + 1)2, arXiv:2001.07094v3 [math.NT] 10 May 2021 − − − and let L be an even, unimodular, positive definite lattice of rank 24. The lattice L has an isometry with characteristic polynomial F everywhere locally, but not globally (see Example 25.17). In particular, the Leech lattice has no isometry with characteristic polynomial F , in spite of having such an isometry everywhere locally.
    [Show full text]
  • Notes on 4-Manifolds
    ALGEBRAIC SURFACES AND FOUR-MANIFOLDS PHILIP ENGEL Contents 1. Structures on manifolds 1 2. Sheaves, bundles, and connections 4 3. Cohomology 8 4. Characteristic classes 12 5. The Riemann-Roch theorem and its generalizations 16 6. The Hodge theorems 19 7. Introduction to algebraic surfaces 25 8. Blow-ups and blow-downs 30 9. Spin and Spinc groups 33 10. Spin bundles and Dirac operators 38 11. Rokhlin's theorem 42 12. Freedman's theorems 45 13. Classification of unimodular lattices 49 4 14. An exotic R 49 15. The Seiberg-Witten equations 49 1. Structures on manifolds Definition 1.1. A topological or C0-manifold of dimension n is a Hausdorff n topological space X, which admits local charts to φU : U ! R . It follows that the transition functions −1 tUV := φV ◦ φU : φU (U \ V ) ! φV (U \ V ) are homeomorphisms. Definition 1.2. By requiring tUV to respect extra geometric structures on n R , we produce many other classes of manifolds. In increasing order of restrictiveness, we say X is additionally a (1) PL manifold if tUV are piecewise-linear. k 1 k (2) C - or C -manifold if tUV 2 C are k-differentiable or smooth. (3) real-analytic manifold if tUV are real-analytic. (4) complex-analytic manifold if tUV is holomorphic: i.e. has a complex- n n=2 linear differential with respect the identification R = C . 1 2 PHILIP ENGEL (5) projective variety if there is a homeomorphism X ! V (f1; : : : ; fm) ⊂ N CP to a smooth vanishing locus of homogenous polynomials. N N+1 ∗ CP := (C n f0g)= ∼ with x ∼ λx for all λ 2 C : There is a natural notion of isomorphism for each class of manifold/variety, by declaring that a homeomorphism f : X ! Y defines an isomorphism if it preserves the given structure.
    [Show full text]
  • Weil's Conjecture for Function Fields
    Weil’s Conjecture for Function Fields February 12, 2019 1 First Lecture: The Mass Formula and Weil’s Conjecture Let R be a commutative ring and let V be an R-module. A quadratic form on V is a map q : V Ñ R satisfying the following conditions: paq The construction pv, wq ÞÑ qpv ` wq ´ qpvq ´ qpwq determines an R-bilinear map V ˆ V Ñ R. pbq For every element λ P R and every v P V , we have qpλvq “ λ2qpvq. A quadratic space over R is a pair pV, qq, where V is a finitely generated projective R-module and q is a quadratic form on V . One of the basic problems in the theory of quadratic forms can be formulated as follows: Question 1.1. Let R be a commutative ring. Can one classify quadratic spaces over R (up to isomorphism)? Let us begin by describing some cases where Question 1.1 admits a complete answer. Example 1.2 (Quadratic Forms over C). Let R “ C be the field of complex numbers (or, more generally, any algebraically closed field of characteristic different from 2). Then every quadratic space over R is isomorphic to pCn, qq, where q is given by the formula 2 2 qpx1, . , xnq “ x1 ` ¨ ¨ ¨ ` xr for some 0 ¤ r ¤ n. Moreover, the integer r is uniquely determined: it is an isomorphism- invariant called the rank of the quadratic form q. Example 1.3 (Quadratic Forms over the Real Numbers). Let R “ R be the field of real numbers. Then every quadratic space over R is isomorphic to pRn, qq, where the quadratic form q is given by 2 2 2 2 2 qpx1, .
    [Show full text]
  • Dynamics on K3 Surfaces: Salem Numbers and Siegel Disks
    Dynamics on K3 surfaces: Salem numbers and Siegel disks Curtis T. McMullen 19 January, 2001 Abstract This paper presents the first examples of K3 surface automorphisms f : X X with Siegel disks (domains on which f acts by an irrational rotation).→ The set of such examples is countable, and the surface X must be non-projective to carry a Siegel disk. These automorphisms are synthesized from Salem numbers of de- gree 22 and trace 1, which play the role of the leading eigenvalue for f ∗ H2(X). The construction− uses the Torelli theorem, the Atiyah-Bott fixed-point| theorem and results from transcendence theory. Contents 1 Introduction............................ 1 2 K3surfaces ............................ 7 3 AutomorphismsofK3surfaces . 11 4 ErgodicdynamicsonKummersurfaces. 14 5 Siegel disks and transcendence theory . 18 6 HolomorphicLefschetznumbers. 20 7 SiegeldisksonK3surfaces. 22 8 Latticesinnumberfields. 24 9 From Salem numbers to automorphisms . 29 10 ExamplesofSiegeldisks . 30 11 Limits of K¨ahler-Einstein metrics . 34 Research supported in part by the NSF. 2000 Mathematics Subject Classification: 37F50 (11R06, 14J50, 32H50). 1 Introduction The first dynamically interesting automorphisms of compact complex man- ifolds arise on K3 surfaces. Indeed, automorphisms of curves are linear (genus 0 or 1) or of finite order (genus 2 or more). Similarly, automorphisms of most surfaces (includ- ing P2, surfaces of general type and ruled surfaces) are either linear, finite order or skew-products over automorphisms of curves. Only K3 surfaces, Enriques surfaces, complex tori and certain non-minimal rational surfaces admit automorphisms of positive topological entropy [Ca2]. The automor- phisms of tori are linear, and the Enriques examples are double-covered by K3 examples.
    [Show full text]
  • Kneser Neighbors and Orthogonal Galois Representations In
    Kneser neighbours and orthogonal Galois representations in dimensions 16 and 24 Gaetan¨ Chenevier (joint work with Jean Lannes) Let n ≥ 1 be an integer. Recall that an even unimodular lattice in the standard euclidean space Rn is a lattice L ⊂ Rn of covolume 1 with x · x ∈ 2Z for all x ∈ L. n Let Xn denote the set of isometry classes of even unimodular lattices in R . As is well-known, Xn is a finite set which is non-empty if and only if n ≡ 0 mod 8. For example, the lattice e + ... + e E = D + 1 n , n n Z 2 n {e1, . , en} denoting the canonical basis of R and Dn the sublattice of index n P 2 in Z whose elements (xi) satisfy i xi ≡ 0 mod 2, is even unimodular for n ≡ 0 mod 8. The set Xn has been determined in only three cases. One has X8 = {E8} (Mordell), X16 = {E8 ⊕ E8, E16} (Witt) and Niemeier showed that X24 has 24 explicit elements (see [V]). The number of numerical coincidences related to Niemeier’s lists is quite extraordinary and makes that list still mysterious. For the other values of n the Minkowski-Siegel-Smith mass formula shows that Xn is huge, perhaps impossible to describe. For instance, X32 already has more than 80.106 elements ([S]). Let L ⊂ Rn be an even unimodular lattice, and let p be a prime; Kneser defines a p-neighbour of L as an even unimodular lattice M ⊂ Rn such that M ∩ L has index p in L (hence in M).
    [Show full text]
  • On Extremal Even Unimodular 72-Dimensional Lattices 1491
    MATHEMATICS OF COMPUTATION Volume 83, Number 287, May 2014, Pages 1489–1494 S 0025-5718(2013)02744-5 Article electronically published on July 16, 2013 ON EXTREMAL EVEN UNIMODULAR 72-DIMENSIONAL LATTICES GABRIELE NEBE AND RICHARD PARKER Abstract. By computer search we show that the lattice Γ of Nebe (2012) is the unique extremal even unimodular 72-dimensional lattices that can be constructed as proposed by Griess (2010). 1. Introduction In this paper a lattice (L, Q) is always an even unimodular positive definite lattice, i.e., a free Z-module L equipped with an integral positive definite quadratic form Q : L → Z of determinant 1. The minimum of L is the minimum of the quadratic form on the non-zero vectors of L, min(L)=min(L, Q)=min{Q() | 0 = ∈ L}1. From the theory of modular forms it is known that the minimum of an even unimod- n ular lattice of dimension n is at most 24 +1. Lattices achieving equality are called extremal. Of particular interest are extremal unimodular lattices in the “jump di- mensions” — the multiples of 24. There are five extremal even unimodular lattices known in the jump dimensions, the Leech lattice Λ24, the unique even unimodular 2 3 lattice in dimension 24 without roots, three lattices called P48p, P48q, P48n, of dimension 48 which have minimum 3 ([2], [8]) and one lattice Γ in dimension 72 ([9]). If (L, Q) is an even unimodular lattice, then L/2L becomes a non-degenerate quadratic space over F2 with quadratic form q( +2L):=Q()+2Z.
    [Show full text]
  • Even Unimodular 12-Dimensional Quadratic Forms Over C!(D)
    ADVANCES IN MATHEMATICS 64, 241-278 (1987) Even Unimodular 12-Dimensional Quadratic Forms over C!(d) PATRICK J. COSTELLO Department of Mathematics, Eastern Kentucky University, Richmond, Kentucky 40475 AND JOHN S. HSIA* Department of Mathematics, Ohio State University, Columbus, Ohio 43210 A complete list of all 1Zdimensional even unimodular lattices over Q(3) is given. Theta series are used to verify that the list of those with 2-vectors is complete. It is proved that there is a unique lattice with no 2-vectors and the Siegel mass for- mula gives the order of its automorphism group. 0 1987 Academic Press, Inc. Positive definite integral quadratic forms or, in the modern terminology, positive definite integral lattices have a long and interesting history with applications and connections to many branches of mathematics including modular form theory, group theory, Lie theory, and coding theory. One fundamental problem in the study of lattices is the complete classification of all lattices of a given discriminant and rank. The problem of classifying these lattices over Z has been studied by many people. It is well known that even unimodular Z-lattices exist only in dimensions which are multiples of 8. A complete classification of even unimodular Z-lattices is only known for dimensions 8, 16, and 24. There is precisely one 8-dimensional even unimodular Z-lattice which is commonly denoted by E8 (in the notation for exceptional Lie groups). It was first dis- covered by Korkine and Zolotareff [KZ] in 1873 although its existence was already nonconstructively known to Smith [S] in 1867. Mordell * This research was supported in part by the National Science Foundation under Grants MCS-8202065 and DMS-8503326.
    [Show full text]
  • Lecture 3. the Intersection Form 1 the Cup Product in Middle Dimensional
    Lecture 3. The intersection form 1 The cup product in middle dimensional cohomology Suppose that M is a closed, oriented manifold of even dimension 2n. Its middle-degree cohomology group Hn(M) then carries a bilinear form, the cup-product form, n n H (M) × H (M) ! Z; (x; y) 7! x · y := eval (x [ y; [M]): The cup product x [ y lies in H2n(M), and eval denotes the evaluation of a cohomology class on a homology class. The cup-product form is skew-symmetric when n is odd, and symmetric when n is even. Lemma 1.1 x · y is equal to the evaluation of x on DXy. ∼ Proof Under the isomorphism H0(X) = Z sending the homology class of a point to 1, one has x · y = (x [ y) \ [X] = x \ (y \ [X]) = hx; DXyi. As we saw last time, when interpreted as a pairing on homology, Hn(M) × Hn(M) ! Z, by applying Poincare´ duality to both factors, the cup product is an intersection product. Concretely, if S and S0 are closed, oriented submanifolds of M, of dimension n and intersecting transversely, and s = DM[S], s0 = DM[S0], then 0 X s · s = "x; x2S\S0 0 0 0 where "x = 1 if, given oriented bases (e1;:::; en) of TxS and (e1;:::; en) of TxS , the basis 0 0 (e1;:::; en; e1;:::; en) for TxM is also oriented; otherwise, "x = −1. 0 Notation: For an abelian group A, let A denotes its torsion-free quotient A=Ators . The cup product form necessarily descends to a form on the free abelian group Hn(M)0 .
    [Show full text]