• Aucun résultat trouvé

The center of quiver Hecke algebras

Assume thatkDL

n2Nknis Noetherian andN-graded and thatk0is a field. We may abbreviatekDk0, and we will identifykwith the quotientk=k>0without mentioning it explicitly.

3.1.1. Cartan datum

A Cartan datum consists of a finite-rank free abelian group P called the weight lattice whose dual lattice, called the coweight lattice, is denoted P_; a finite set of vectors ˆD ¹˛1; : : : ; ˛nº P called simple roots; and a finite set of vectors ˆ_D ¹˛1_; : : : ; ˛n_º P_ calledsimple coroots. Let QCDNˆP be the semi-group generated by the simple roots, and let PCP be the subset of dominant weights, that is, the set of weightsƒsuch thatƒiD h˛i_; ƒi 0for alli2I. We will call aBruhat orderthe partial order onP such thatwhenever2QC. Set I D ¹1; : : : ; nº, and let h;i be the canonical pairing on P_P. The .I I /-matrix Awith entries aij D h˛i_; ˛jiis assumed to be a generalized Car-tan matrix. We will assume that the CarCar-tan datum is nondegenerate (i.e., the simple roots are linearly independent) and symmetrizable (i.e., there exist nonzero integers di such thatdiaij Ddjaj i for alli; j). The integersdi are unique up to an overall common factor. They can be assumed positive and mutually prime.

Let.j/be the symmetric bilinear form onhDQ˝ZP given by.˛ij/D diaij. Letgbe the symmetrizable Kac–Moody algebra over kassociated with the generalized Cartan matrixAand the lattice of integral weightsP. Leth;nCgbe the Cartan subalgebra and the maximal nilpotent subalgebra spanned by the positive root vectors ofg. For any dominant weightƒ2PC, letV .ƒ/be the correspond-ing integrable simpleg-module. For each2P, let V .ƒ/ V .ƒ/be the weight subspace of weight.

3.1.2. Quiver Hecke algebras

Fix an element ci;j;p;q 2k for each i; j 2I, p; q 2N such that deg.ci;j;p;q/D 2di.aijCp/2djqandci;j;aij;0is invertible. Fix a matrixQD.Qij/i;j2I with entries inkŒu; vsuch that

Qij.u; v/DQj i.v; u/;

Qi i.u; v/D0;

Qij.u; v/D X

p;q0

ci;j;p;qupvq ifi¤j:

Definition 3.1

Thequiver Hecke algebra(or QHA) of rankn0associated withAandQ is the k-algebraR.nIQ;k/generated bye. /,xk,l with 2In,k; l2Œ1; n,l¤n, sat-isfying the following defining relations:

(a) e. /e. 0/Dı;0e. /,P

e. /D1, (b) xkxlDxlxk,xke. /De. /xk,

(c) le. /De.sl. //l,klDlkifjklj> 1, (d) l2e. /DQl;lC1.xl; xlC1/e. /,

(e) .kxlxsk.l/k/e. /Dık;kC1l;kC1ıl;k/e. /,

(f) .kC1kkC1kkC1k/e. /Dık;kC2@k;kC2Qk;kC1.xk; xkC1/e. /, where@k;l is the Demazure operator onkŒx1; x2; : : : ; xnwhich is defined by

@k;l.f /D.f .k; l/.f /=.xkxl/:

The algebraR.nIQ;k/is free as ak-module. It admits aZ-grading given by deg

e. /

D0; deg xke. /

D2dk; deg ke. /

D dkak;kC1: For˛2QCsuch that ht.˛/Dn, we set

I˛

D. 1; : : : ; n/2In1C C˛nD˛¯ : The idempotent e.˛/DP

2I˛e. / is central in R.nIQ;k/. Given 2In, 02 Im we write 02InCm for their concatenation. Sete.˛; 0/DP

2I˛e. 0/and e.n; 0/DP

2Ine. 0/. The quiver Hecke algebra of rank˛is the algebra R.˛IQ;k/De.˛/R.nIQ;k/e.˛/:

3.1.3. Cyclotomic quiver Hecke algebras Given a dominant weightƒ2PCwe set

Iƒ

.i; p/Ii2I; pD1; : : : ; ƒi

¯:

For a future use, let

Iƒ!I; t7!it (2)

denote the canonical map such that .i; p/7!i. Fix a family of commuting formal variables¹ctIt2Iƒº. Letkƒbe theN-graded ring given by

kƒDkŒctIt 2Iƒ; deg.cip/D2pdi: We will abbreviatekDkƒ, and we will writeci 0D1.

Now, fix aN-gradedk-algebrak. Letctdenote both the element inkabove and its image in kby the canonical map k!k(which is homogeneous of degree 0).

Then, set

aiƒ.u/D

ƒi

X

pD0

cipuƒip2kŒu: (3)

The monic polynomial aƒi .u/ is called the ith cyclotomic polynomial associated withk.

For each˛2QCand1kht.˛/, we set a˛ƒ.xk/D X

2I˛

aƒ

k.xk/e. /: (4)

Note thataƒ˛.xk/e. /is a homogeneous element ofR.˛IQ;k/with degree2dkƒk. Definition 3.2

The cyclotomic quiver Hecke algebra of rank ˛ and level ƒ is the quotient Rƒ.˛IQ;k/ofR.˛IQ;k/by the two-sided ideal generated bya˛ƒ.x1/.

To simplify notation, we write R.˛/DR.˛Ik/DR.˛IQ;k/ and Rƒ.˛/D Rƒ.˛Ik/DRƒ.˛IQ;k/. We may also write RDL

˛R.˛/,R.k/DL

˛R.˛Ik/, RƒDL

˛Rƒ.˛/, and so on. The following is proved in [16, Corollary 4.4, Theo-rem 4.5].

PROPOSITION3.3

Thek-algebraRƒ.˛Ik/is free of finite type as ak-module.

Remark 3.4

A morphism ofN-gradedk-algebras k!hyields canonical gradedh-algebra iso-morphismsh˝kR.˛Ik/!R.˛Ih/andh˝kRƒ.˛Ik/!Rƒ.˛Ih/.

Example 3.5

(a) SetRƒ.˛/DRƒ.˛Ik/. We callRƒ.˛/theglobal(oruniversal) cyclotomic quiver Hecke algebra.

(b) If kDk, then aƒi .u/ Duƒi for each i. We call Rƒ.˛Ik/ the local (or restricted) cyclotomic quiver Hecke algebra.

(c) For eachi2Iwe fix an elementci2kof degree2di. We define the following:

k0Dkas ak-algebra with the newk-algebra structure associated with the elementsc0ip given bycip0 Dep.yi1ci; yi 2ci; : : : /, where the yip’s are commuting formal variables such thatcipDep.yi1; yi 2; : : : / and ep is the pth elementary symmetric polynomial. Then, the ith cyclotomic polynomial associated withk0isaƒi .uci/for eachi2I.

Q0ij.u; v/DQij.uci; vcj/. In particular, we haveQij0 .u; v/D Q0j i.v; u/.

Then, the assignment e. /; xke. /; le. / 7!e. /; .xk Cck/e. /; le. / extends uniquely to ak-algebra isomorphismRƒ.˛IQ;k/! Rƒ.˛IQ0;k0/.

(d) Under the hypothesis above, assume that i2I and that ƒD!i is the ith fundamental weight. Assume also that the polynomialQij.u; v/satisfies con-dition (11) below. Ifaƒi .u/DuCci, then we have ak-algebra isomorphism

R!i.˛IQ;k/'k˝kR!i.˛IQ;k/: (5) Definition 3.6

For each k2Œ1; n1, the kth intertwiner operator is the element 'k 2Rƒ.n/

defined by 'ke. /Dke. / if k¤ kC1 and by the following formulas if kD

kC1:

'ke. /D.xkkkxk/e. /D.kxkC1xkC1k/e. / D

.xkxkC1/kC1

e. /D

k.xkC1xk/1 e. /:

We have the following facts (see [18, Section 5.1] for details):

xsk.`/'ke. /D'kx`e. /.

¹'kºsatisfies the braid relations.

Ifw2Snsatisfiesw.kC1/Dw.k/C1, then'wkDw.k/'w.

'k2e. /De. /if kD kC1 and'k2e. /DQk;kC1.xk; xkC1/e. /if k¤

kC1.

3.1.4. Induction and restriction

Leti2I and˛2QCbe of heightn. SetDƒ˛andiD h˛i_; i.

We have a Z-graded k-algebra embedding i WR.˛/ ,!R.˛ C˛i/ given by e. /; xk; l7!e. ; i /; xk; l for each 2I˛ with1knand 1ln1. It induces aZ-gradedk-algebra homomorphismiWRƒ.˛/!Rƒ.˛C˛i/.

The restriction and induction functors form an adjoint pair.Fi0; Ei0/with Ei0WRƒ.˛C˛i/-grmod!Rƒ.˛/-grmod; N7!e.˛; i /N;

Fi0WRƒ.˛/-grmod!Rƒ.˛C˛i/-grmod;

M7!Rƒ.˛C˛i/e.˛; i /˝Rƒ.˛/M:

The counit "0i;WFi0Ei01!1 and the unit 0i;W1!Ei0Fi01 are represented, respectively, by the multiplication mapand the mapi:

"0i;WRƒ.˛/e.˛˛i; i /˝Rƒ.˛˛i/e.˛˛i; i /Rƒ.˛/!Rƒ.˛/;

0i;WRƒ.˛/!e.˛; i /Rƒ.˛C˛i/e.˛; i /:

Finally, letij;0 WFi0Ej01!Ej0Fi01be the morphism represented by the linear map

Rƒ.˛˛ji/e.˛˛j; i /˝Rƒ.˛˛j/e.˛˛j; j /Rƒ.˛/

!e.˛˛ji; j /Rƒ.˛C˛i/e.˛; i /;

x˝y7!xny:

Forj Di, the elementn2Rƒ.˛C˛i/centralizes the subalgebrae.˛˛i; i2/ Rƒ.˛C˛i/e.˛˛i; i2/, so we havei i;0 Dn(see (1)).

THEOREM3.7 (see [16])

For each˛2QCof heightn, we have the following.

(a) Ifi0, then the following morphism of endofunctors onRƒ.˛/-Modis an isomorphism:

i;0 Di i;0 C

i1

X

kD0

.Ei0xk0i;WFi0Ei01˚

i1

M

kD0

kxk˝1!Ei0Fi01: (b) Ifi0, then the following morphism of endofunctors onRƒ.˛/-Modis an

isomorphism:

0i;D

i i;0 ; "0i;ı.Fi0x0/; : : : ; "0i;ı.Fi0xi1/ W

Fi0Ei01!Ei0Fi01˚

i1

M

kD0

k.x1/k˝1:

The theorem can be rephrased as follows.

Assume thati0: for anyz2e.˛; i /Rƒ.˛C˛i/e.˛; i /there are unique ele-ments.z/2Rƒ.˛/e.˛˛i; i /˝Rƒ.˛˛i/e.˛˛i; i /Rƒ.˛/andpk.z/2 Rƒ.˛/such that

zDn

.z/

C

i1

X

kD0

pk.z/xnC1k : (6)

Assume thati 0: for any z2e.˛; i /Rƒ.˛C˛i/e.˛; i / and any z0; : : : ; zi12Rƒ.˛/, there is a unique elementy2Rƒ.˛/e.˛˛i; i /˝Rƒ.˛˛i/

e.˛˛i; i /Rƒ.˛/such that n.y/Dz; xk

n.y/Dzk; 8k2Œ0;i1: (7)

For a future use, let us introduce the following notation. Assume thati0and thatz2e.˛; i /Rƒ.˛C˛i/e.˛; i /. For each`2Œ0;i1, let

Q

z;Q`2Rƒ.˛/e.˛˛i; i /˝Rƒ.˛˛i/e.˛˛i; i /Rƒ.˛/ (8) be the unique elements such that

n.z/Q Dz; xk

n.Qz/D0; n.Q`/D0; xk

n.Q`/Dık;`:

THEOREM3.8 (see [18])

The pair.Ei0; Fi0/ is adjoint with the counit"O0i;WEi0Fi01!1 and the unitO0i;W 1!Fi0Ei01represented by the morphisms

O

"0i;We.˛; i /Rƒ.˛C˛i/e.˛; i /!Rƒ.˛/;

O

0i;WRƒ.˛/!Rƒ.˛/e.˛˛i; i /˝Rƒ.˛˛i/e.˛˛i; i /Rƒ.˛/

such that

"O0i;.z/Dpi1.z/ifi> 0and

xni.Qz/ifi0,

O0i;.1/D .xnC1i /ifi0andQi1ifi< 0.

We abbreviate"0iD"0i;,0i D0i;,"O0i D O"0i;,O0i D O0i;, and so on, when is clear from the context.

COROLLARY3.9

The linear maps "0i, 0i are homogeneous of degree 0. The linear maps "O0i, O0i are homogeneous of degrees2di.1i/,2di.1Ci/, respectively. The linear mapij0 is homogeneous of degreediaij.

3.1.5. The symmetrizing form For each˛2QCwe set

dƒ;˛D.ƒjƒ/.ƒ˛jƒ˛/:

We will need the following result from [40].

PROPOSITION3.10 ([40, Remark 3.19])

Thek-algebra Rƒ.˛/ is symmetric and admits a symmetrizing formtƒ;˛ which is homogeneous of degreedƒ;˛.

The definition oftƒ;˛is given in DefinitionA.6. We will abbreviatet˛Dtƒ;˛and tƒDP

˛t˛. Since we have not found any proof of the proposition in the literature, we have given one in AppendixA.

3.2. Categorical representations

Letkbe anN-graded commutative ring as in Section3.1. WritegkDk˝kg. Fix an integer`.

3.2.1. Definition

For each2P, letCbe an.`Z/-gradedk-category. SetCDL

Cand denote by 1the obvious functor1WC!C. For eachi; j 2I,2P, we fix

aZ-gradedk-algebra homomorphismkŒ`!Z.C=`Z/;

a functor 1˛iFi D Fi1 with a right adjoint 1EiŒ`di.1 i/ D Ei1˛iŒ`di.1i/;

morphisms of functors xi1W Fi1 !Fi1Œ2`di and ij1 WFiFj1 ! FjFi1Œ`diaij.

ThusC=`Zis ak-category, and the functorsFi1,Ei1arek-linear. Let

"i1WFiEi1!1

`di.1Ci/

; i1W1!EiFi1

`di.1i/ be the counit and the unit of the adjoint pair.1Fi; Ei1Œ`di.1Ci//. We will abbreviate

EiDM

Ei1; FiDM

Fi1; F˛DM

2I˛

F; and so on;

where FDF1F2 Fn for D. 1; 2; : : : ; n/. Next, we define the following morphisms:

ij D.EjFi"j/ı.Ejj iEi/ı.jFiEj/WFiEj !EjFi;

i1 D i i1 C Pi1

lD0 ."i1/ ı .xilEi1/ W FiEi1 ! EiFi1 ˚ Li1

lD0 1Œ`di.1C2lCi/ifi0;

i1Di i1CPi1

lD0 .Eixil1/ı.i1/WFiEi1˚Li1

lD0 1Œ`di.1C2l i/!EiFi1ifi0.

Definition 3.11

Acategorical representationofgkof degree`inC is a tupleC,Ei,Fi,"i,i,xi, ij as above such that the following hold:

the assignmente. /7!1F,xke. /7!xk1F,le. /7!l;lC11Ffor each 2 I˛ defines a Z-graded kŒ`-algebra homomorphism R.˛Ik/Œ` ! EndC=`Z.F˛/;

the morphismsi1,ij,i¤j, are isomorphisms.

Morphisms of categorical representations are defined in the obvious way.

We will call the mapR.˛Ik/Œ`!EndC=`Z.F˛/thecanonical homomorphism associated with the categorical representation ofgkinC.

Unless specified otherwise, a categorical representation will be of degree 1.

Degrees`¤1are used only in the nonsymmetric case and the reader interested only in symmetric ones may set`D1everywhere. Note that, given a categorical represen-tation ofgkinC, there is a canonical categorical representation ofgkof degree`in CŒ`called its`-twistsuch that theZ-gradedkŒ`-algebra homomorphism

R.˛Ik/Œ`!EndCŒ`=`Z.F˛/DEndC=Z.F˛/Œ`

is equal to the homomorphismR.˛Ik/!EndC=Z.F˛/associated with thegk-action onC.

We will also use the following definitions:

CisintegrableifEi,Fiare locally nilpotent for alli.

Cisbounded aboveif the set of weights ofC is contained in a finite union of cones of typeQCwith2P.

Thehighest weight subcategoryChwC is the full subcategory given by Chw

M 2CIEi.M /D0;8i2I¯ : Remark 3.12

(a) Taking the left transpose of the morphisms of functors

xi1WFi1!Fi1Œ2`di; ij1WFiFj1!FjFi1Œ`diaij we get the morphisms of functors

1_

xiW1Ei!1EiŒ2`di; 1_

ij1W1EiEj !1EjEiŒ`diaij: We will abbreviatexiD_xi andij D_ij.

(b) Forgetting the grading at each place we define as above a categorical repre-sentation ofgkin a (not graded)k-categoryC.

(c) For each short exact sequence of Z-graded k-categories 0!I !C ! B!0such thatI!C is a morphism of categorical representations ofgk, there is a unique categorical representation ofgkonBsuch thatC!Bis a morphism of categorical representations.

(d) Given a categorical representation ofgkon C, there is a unique categorical representation ofgkonCc such that the canonical fully faithful functorC !Cc is a morphism of categorical representations. Recall that the objects of the idempotent completionCc are the pairs.M; e/, whereM is an object of C ande is an idem-potent of EndC.M /, and that HomCc..M; e/; .N; f //DfHomC.M; N /e. Then, we haveFi.M; e/D.Fi.M /; Fi.e//,Ei.M; e/D.Ei.M /; Ei.e//, andxi1,ij1 are defined in a similar way.

3.2.2. The minimal categorical representation

Fix a dominant weightƒ2PCand anN-gradedk-algebrak. Given˛2QCwe write Dƒ˛. Recall that we abbreviateRƒ.˛/DRƒ.˛Ik/. LetkAƒ DRƒ.˛/-grmod be the Z-graded abelian k-category consisting of the finitely generated Z-graded Rƒ.˛/-modules, and let kVƒDRƒ.˛/-grproj be the full subcategory formed by the projective Z-graded modules. When there is no confusion, we will abbreviate Aƒ DkAƒ andVƒDkAƒ. LetAƒ,Vƒbe the categories

AƒDM

Aƒ; VƒDM

Vƒ:

Fix an integer`. LetVƒ;Œ`D.Vƒ/Œ` be the`-twist of Vƒ, and letRƒ.˛/Œ` be the`-twist ofRƒ.˛/. Thus,Rƒ.˛/Œ`is a.`Z/-gradedkŒ`-algebra andVƒ;Œ`is the category of finitely generated projective.`Z/-graded modules; that is,

Vƒ;Œ`DRƒ.˛/Œ`-grproj:

Definition 3.13

Theminimal categorical representationofgkof highest weightƒand degree`is the representation onVƒ;Œ`given by the following:

Ei1DEi0Œ`di.1Ci/;

Fi1DFi0;

"i1D"0i;andi1D0i;;

xi12Hom.Fi1; Fi1Œ2`di/is represented by the right multiplication by xnC1onRƒ.˛C˛i/e.˛; i /;

ij12Hom.FiFj1; FjFi1Œ`diaij/is represented by the right multipli-cation bynC1onRƒ.˛C˛ij/e.˛; j i /.

The categoryAƒ;Œ` is Krull–Schmidt with a finite number of indecomposable projective objects. The categoryVƒ;Œ`=.`Z/is the category of.`Z/-graded finitely generated projectiveRƒ;Œ`-modules with morphisms which are not necessarily homo-geneous. We will call it the category of all.`Z/-gradableprojective modules.

Example 3.14

We will abbreviateeDsl2. Assume thatgDe,ƒDk!1, and˛Dn˛1withk; n2N. In this case, we write VkDVƒ and Vk2nk DVƒ. Consider the polynomial ring ZkDkŒc1; : : : ; ck with deg.cp/D2p for all p. Let Hkn be the global cyclotomic affine nil Hecke algebra of rank n and level k, that is, the Z-graded Zk-algebra denoted by Hn;k in [35, Section 4.3.2]. Note that we have Hk0 DZk and kDZk. Given anN-graded Zk-algebrak, the cyclotomic quiver Hecke algebraRƒ.˛Ik/is isomorphic to k˝Zk Hkn as aZ-gradedk-algebra by [35, Lemma 4.27]. In partic-ular, we haveRƒ.˛/DHkn. For each integer`, we abbreviate Zk;Œ`D.Zk/Œ` and Hk;Œ`n D.Hkn/Œ`. We have

Vk;Œ`DM

n0

.kŒ`˝Zk;Œ`Hk;Œ`n /-grproj; (9) and the trace tr.Vk;Œ`/ is given in Proposition 3.30 below. We will also identify Rƒ.˛Ik/with thelocal cyclotomic affine nil Hecke algebraof rank nand level k, which is the quotient HknDk˝ZkHknof Hknby the ideal.c1; : : : ; ck/.

3.2.3. Factorization

Fix a dominant weightƒ2PC. Recall the mapIƒ!I,t7!it introduced in (2).

We will abbreviate !t D!it and dt Ddit. Consider the N-graded k-algebra hD kŒytIt2Iƒ, whereyt is a formal variable of degree deg.yt/D2dt. It has a natural structure of anN-gradedk-algebra such that the elementcip2his given bycipD ep.yi1; : : : ; yi ƒi/. The corresponding cyclotomic polynomials are

aiƒ.u/D

ƒi

Y

pD1

.uCyip/; 8i2I: (10)

Leth0 be the fraction field ofh. Then we have the algebrasRƒ.n;h/andRƒ.n;h0/ overhandh0such thatRƒ.n;h0/Dh0˝hRƒ.n;h/. Next, for eacht2Iƒ, we have

hDO

t2Iƒ

kŒytDO

t2Iƒ

k!t:

Therefore, we can viewhas ak!t-algebra, and hence theh-algebraR!t.n;h/ asso-ciated with the cyclotomic polynomialsa!it.u/DuCytis well defined. Note that by

Remark3.5(d), if condition (11) below is satisfied, then we may as well assume that a!it.u/Du. Recall that

Rƒ.n;h/Dh˝kRƒ.n/;

R!t.n;h/Dh˝k!t R!t.n/:

When no confusion is possible, we will abbreviate VƒDM

n

Rƒ.n/-grproj; V!t DM

n

R!t.n/-grproj:

Finally, we consider the condition

Qij.u; v/Drij.uv/aij for somerij inksuch that

rij D.1/aijrj i for alli¤j: (11)

THEOREM3.15

Letnbe a positive integer, and let˛2QC be of height n. If the condition (11) is satisfied, then the following hold:

(a) There is anh0-algebra isomorphism Rƒ.n;h0/!M

.nt/

MatSn=Q tSnt

O

t2Iƒ

R!t.nt;h0/

;

where.nt/runs over the set ofIƒ-tuples of nonnegative integers with sumn.

(b) There is an isomorphismh0˝kVƒ!h0˝hN

t2IƒV!t of (nongraded) cate-goricalgh0-representations taking the functorF˛toL

t/

N

tF˛t, where the sum runs over the set ofIƒ-tuplest/of elements ofQCwith sum˛. More precisely, the canonical homomorphism

O

t2Iƒ

R.˛tIh0/!End O

t2Iƒ

F˛t

is the composition of the inclusionN

t2IƒR.˛tIh0/R.˛Ih0/underlying (a) and of the canonical homomorphismR.˛Ih0/!End.F˛/.

Proof

FixM 2Rƒ.nIh0/-mod andg.u/2h0Œu. From [16, pp. 715–716], we get g.xa/e. /MD0

)Qa;aC1.xa; xaC1/g.xaC1/e sa. /

M D0; 8a2Œ1; n/;8 2In:

SetQ.u; v/DQ

i¤jQij.u; v/. We deduce that g.xa/e. /MD0)Q.xa; xaC1/g.xaC1/e

sa. /

M D0; 8a2Œ1; n/: (12) Now, assume that the polynomialQ.u; v/2h0Œu; vhas the following form,

Q.u; v/Dr Y

2S

.uv/; (13)

for some finite familySof elements ofh0and some elementr2h0. Let spe./M.xa/ h0be the set of2h0such that the operatorxaid2Endh0.e. /M /is not invert-ible. Since xa and xaC1 commute with each other, from (12) and (13) we deduce that

spe.sa.//M.xaC1/ spe./M.xa/t

spe.sa.//M.xa/S : Thus, switching andsa. /, we deduce that

spe./M.xaC1/ spe.sa.//M.xa/t

spe./M.xa/S : Next, recall thatg.x1/.e. /M /D0ifg.u/Daƒ

1.u/. We deduce that spe./M.x1/ ¹y1;pIpD1; : : : ; ƒ1º:

Therefore, an easy induction implies that

spe./M.xa/ ¹yb;pIpD1; : : : ; ƒb; bD1; : : : ; aº NS; 8a2Œ1; n:

Assume further that condition (11) holds. Then we haveSD ¹0º, and hence

spe./M.xa/ ¹yb;pIpD1; : : : ; ƒb; bD1; : : : ; aº; 8a2Œ1; n: (14) In the rest of the proof we writeIQDIƒto simplify the notation. For eachn-tuple Q2 QIn, set

MQ

m2MI.xkCyQk/DmD0;8k2Œ1; n;8D0¯ :

Considering the decomposition of the regular module, we deduce from (14) that there is a complete collection of orthogonal idempotents¹e. /I QQ 2 QInºinRƒ.n;h0/such thate. /MQ DMQ. The groupSnacts onIQnin the obvious way. The following hold:

e. /e. /Q De. /e.Q /,

xle. /Q De. /xQ l,

'ke. /Q De.sk.Q //'k,

ke. /Q De. /Q kif QkD QkC1,

where k, l, , andQ run over the sets Œ1; n/,Œ1; n,IQn, and In, respectively. The relations above are immediate, except the last one. To prove it, note that ifQkD QkC1, then the operatorsxkCyQk andxkC1CyQk commute to each other and their sum and product commute withk and act nilpotently onke. /MQ . ThereforexkCyQk and xkC1CyQkalso act nilpotently onke. /MQ . Note that the first relation above implies that we may view¹e. /I QQ 2 QInºas a complete collection of orthogonal idempotents inRƒ.˛;h0/for each˛2QCof heightn.

LEMMA3.16

For eachM 2Rƒ.nIh0/-mod, the map'ke. /Q We. /MQ !e.sk.Q //M is invertible whenever Qk¤ QkC1.

Proof

The lemma is an immediate consequence of the following relations, for each 2In, 'k2e. /D1 if kD kC1;

'k2e. /DQk;kC1.xk; xkC1/e. / if k¤ kC1;

because (11) implies thatQk;kC1.xk; xkC1/is invertible if k¤ kC1 and Qk¤ QkC1.

SetQQCDNIQ. For any element˛Q DP

t2Iƒntt inQQCwe set ht.˛/Q DP

tat. Assume that ht.˛/Q Dn; then we consider the setIQ˛Q D ¹ Q 2 QInIP

kQk D Q˛º. We will say that twon-tuples ;Q Q02 QIn are equivalent, and we write Q Q0 if we have Q

;Q02 QI˛Q for some˛Q 2 QQC. We define the idempotente.˛/Q inRƒ.n;h0/bye.˛/Q D P

Q

2 QI˛Qe. /.Q

Next, fix a total order onIQ, and setIQCn D ¹ Q2 QInIi < j ) Qi Qjº. For any tuple Q 2 QIn there is a unique element QC2 QICn such that Q QC. Let ˛QC be the unique element inIQCn\ QI˛Q. The theorem is an easy consequence of the following lemma.

LEMMA3.17

Let˛2QCand˛Q2 QQCbe of heightn, with˛QDP

t2 QIntt. The following hold:

(a) e. /RQ ƒ.˛;h0/e.Q0/D0unless Q Q0;

(b) e.˛/RQ ƒ.˛;h0/e.˛/Q 'MatIQ˛Q.e.˛QC/Rƒ.˛;h0/e.˛QC//ash0-algebras;

(c) e.˛QC/Rƒ.˛;h0/e.˛QC/'L

t/

N

t2 QIR!tt;h0/ash0-algebras, wheret/ runs over the set of allIQ-tuples of elements inQCsuch thatP

t˛tand ht.˛t/Dnt.

Proof

Part (a) of the lemma follows from the following.

CLAIM3.18

Theh0-algebraRƒ.n;h0/is generated by the subset

®e. /e.Q /; he. /; 'Q ke. /; xQ le. /IQ h; k2Œ1; n/; l2Œ1; n;

2In;Q2 QIn;QhD QhC1;Qk¤ QkC1¯ :

Let us concentrate on part (b). For each tupleQ2 QI˛Q, we fix a sequence of simple reflectionssl1; sl2; : : : ; slj inSnsuch that

QDsl1sl2 slj.˛QC/;

thelhth entry ofslhC1slhC2 slj.˛QC/inIQis smaller than the.lhC1/th one for eachh2Œ1; j .

In particular,wDsl1sl2 slj is a reduced decomposition andwis minimal in its left coset inSnrelatively to the stabilizer of˛QC. Hence, we have

'l1'l2 'lj D'w; (15) and the elementQine. /RQ ƒ.˛;h0/e.˛QC/given by

QDe. /'Q we.˛QC/ (16) depends only onQand not on the choice ofl1; l2; : : : ; lj. It is invertible by Lemma3.16.

We deduce that there is anh0-algebra isomorphism MatIQ˛Q

e.˛QC/Rƒ.˛;h0/e.˛QC/

!e.˛/RQ ƒ.˛;h0/e.˛/;Q

E;QQ0.m/7!Qm1Q0 : (17) HereE;QQ0.m/is the elementary matrix with anmat the spot.Q ;Q0/and0s elsewhere.

Part (b) is proved.

Now, we prove part (c). Set¹t2 QIInt¤0º D ¹t1; t2; : : : ; tmºwitht1< t2< <

tm. Fix anIQ-tuple.˛t/of elements inQCsuch thatP

t˛tD˛and ht.˛t/Dnt for eacht. There is a canonical inclusion

O

t2 QI

R.˛t;h0/DR.˛t1;h0/˝R.˛t2;h0/˝ ˝R.˛tm;h0/!R.˛;h0/:

Composing it with the multiplication by the idempotente.˛QC/, we get a map O

t2 QI

R.˛t;h0/!e.˛QC/R.˛;h0/e.˛QC/ (18)

such that, for eachIQ-tuple. t/with t2I˛t, we have

e. t1/˝e. t2/˝ ˝e. tm/7!e. t1 t2 tm/e.˛QC/:

To prove (c) we must check that the direct sum of all maps (18), where.˛t/runs over the set of all tuples as above, gives anh0-algebra isomorphism

M

t/

O

t2 QI

R!tt;h0/! e.˛QC/Rƒ.˛;h0/e.˛QC/: (19) First, we will prove that theh0-algebra homomorphism (19) is well defined. To do that, it is enough to check that for each tuple.˛t/the map (18) factors to an algebra homomorphism

O

t2 QI

R!tt;h0/!e.˛QC/Rƒ.˛;h0/e.˛QC/; (20)

because (20) maps into orthogonal direct summands of e.˛QC/Rƒ.˛;h0/e.˛QC/ for different tuples.˛t/. To simplify the notation, we will check this in a particular case.

The proof of the general case is very similar. Assume that

ƒD2!iC!j; IQD ¹t1; t2; t3º; t1D.i; 1/; t2D.i; 2/; t3D.j; 1/;

withi¤j andt1< t2< t3. Thus, we have

aƒi .u/D.uCyt1/.uCyt2/; aƒj .u/DuCyt3; akƒ.u/D1; 8k¤i; j;

!t1D!i; !t2D!i; !t3D!j:

Next, fix a positive integernand an element˛Q 2 QQCof heightn. We have Q

˛Dnt1t1Cnt2t2Cnt3t3; Q

˛CD.˛Q1C;˛Q2C; : : : ;˛QCn/D

.t1/nt1.t2/nt2.t3/nt3

;

withnt1Cnt2Cnt3Dn. To simplify we will assume that the integersnt1; nt2; nt3 are positive.

CLAIM3.19

For any 2In, the following relations hold ine.˛QC/Rƒ.n;h0/e.˛QC/:

.x1Cyt1/e.˛QC/e. /D0if 1Di, ande.˛QC/e. /D0otherwise;

.x1Cnt1Cyt2/e.˛QC/e. /D0if 1Cnt1 Di, ande.˛QC/e. /D0otherwise;

.x1Cnt1Cnt2Cyt3/e.˛QC/e. /D0if 1Cnt1Cnt2Dj, ande.˛QC/e. /D0 oth-erwise.

The first relation is obvious, because

.x1Cyt1/.x1Cyt2/e. /D0if 1Di;

.x1Cyt3/e. /D0if 1Dj;

e. /D0if 1¤i; j;

.x1Cyt2/e.˛QC/,.x1Cyt3/e.˛QC/are invertible ine.˛QC/Rƒ.n;h0/e.˛QC/.

To prove the second relation, note that

'1'2 'nt1.x1Cnt1Cyt2/e.˛QC/e. /D.x1Cyt2/e. /e. /'Q 1'2 'nt1; where QDs1s2 snt1.˛QC/andDs1s2 snt1. /. SinceQ1Dt2and1D 1Cnt1, we deduce that

'1'2 'nt1.x1Cnt1Cyt2/e.˛QC/e. /D0if 1Cnt1Di,

'1'2 'nt1e.˛QC/e. /D0otherwise.

Further, by Lemma3.16, the operator

'1'2 'nt1e.˛QC/We.˛QC/Rƒ.˛;h0/e.˛QC/!e. /RQ ƒ.˛;h0/e.˛QC/ is invertible. Thus, we have

.x1Cnt1Cyt2/e.˛QC/e. /D0if 1Cnt1 Di,

e.˛QC/e. /D0otherwise,

proving the second relation. The third relation is proved in a similar way, using the product of intertwiners'1'2 'nt1Cnt2 instead of'1'2 'nt1. The claim is proved.

The claim implies that the homomorphism (20) is well defined, and so (19) is also well defined. We must check that it is invertible. To prove the surjectivity, we must check that theh0-algebra isomorphism (19) yields a surjective map

O

t2 QI

R!t.nt;h0/! e.˛QC/Rƒ.n;h0/e.˛QC/:

This is a consequence of the following fact.

CLAIM3.20

Theh0-algebrae.˛QC/Rƒ.n;h0/e.˛QC/is generated by the subset

®e. /e.˛QC/; he.˛QC/; xle.˛QC/Ih2Œ1; n/; l2Œ1; n; 2In;˛QhCD Q˛hC1C ¯

: (21) Let us concentrate on the injectivity. We will construct a left inverse to the map (19). To do that, recall that parts (a) and (b) yield anh0-algebra isomorphism

M

Q

˛

MatIQ˛Q

e.˛QC/Rƒ.˛;h0/e.˛QC/

!Rƒ.˛;h0/; (22) where the sum is over the set of all elements˛Q2 QQCof heightn.

CLAIM3.21

(a) FixQ2 QI˛Qandw2Snas in (16). For eachh; k; l2Œ1; nsuch thath; k¤n, QhD QhC1, andQk¤ QkC1, the map (22) is such that

E;QQ

xw1.l/e.˛QC/

7!xle. /;Q Esk./;QQ

e.˛QC/

7!'ke. /;Q E;QQ

w1.h/e.˛QC/

7!he. /:Q (b) Consider the assignment

xle. /Q 7!E;QQ.xw1.l//;

'ke. /Q 7!Esk./;QQ .1/;

he. /Q 7!E;QQ.w1.h//;

for eachh; k; l2Œ1; nand Q2 QI˛Q such thath; k¤n, QhD QhC1, and Qk¤ QkC1, wherew2Snas in (16). It extends uniquely to anh0-algebra homo-morphism

Rƒ.˛;h0/!M

Q

˛

M

t/

MatIQ˛Q

O

t2 QI

R!tt;h0/

: (23)

Proof

Part (a) follows from (17) and from the following computations:

xle. /Q DxlQe.˛QC/1Q DQxw1.l/e.˛QC/1Q ; 'ke. /Q D'kQe.˛QC/1Q

D'ke.Q /'we.˛QC/Q1 De

sk.Q /

'k'we.˛QC/Q1 Dsk.Q/e.˛QC/1Q ; he. /Q DhQe.˛QC/Q1

Dhe. /'Q we.˛QC/1Q De. /Q h'we.˛QC/1Q De. /'Q ww1.h/e.˛QC/1Q DQw1.h/e.˛QC/1Q :

Here, we used the equalityw1.hC1/Dw1.h/C1, which follows from the defini-tion ofwin (15), and the equalitiessh.Q /D Q andsw1.h/.˛QC/D Q˛C, which follow from the identities QhD QhC1and QDw.˛QC/.

Now, let us concentrate on (b). Since the elements e. /e.Q /, xle. /,Q 'ke.Q /, he. /Q generate Rƒ.n;h0/ by Claim 3.18, it is enough to check that the defining relations ofRƒ.˛;h0/given in Section3.1.3are satisfied. This is obvious.

Note that the element w1.h/ belongs indeed to N

t2 QIR!tt;h0/ because sw1.h/.˛QC/D Q˛C.

Composing (22) and (23), we get anh0-algebra homomorphism M

Q

˛

MatIQ˛Q

e.˛QC/Rƒ.˛;h0/e.˛QC/

!M

Q

˛

M

t/

MatIQ˛Q

O

t2 QI

R!tt;h0/ :

For each˛Q it restricts to anh0-algebra homomorphism e.˛QC/Rƒ.˛;h0/e.˛QC/!M

t/

O

t2 QI

R!tt;h0/;

which is a left inverse to (19).

3.2.4. The isotypic filtration

Recall thateDsl2. A.P Z/-gradedk-categoryC is a direct sum of categories of the formCDL

2PC, where eachCis aZ-gradedk-category. We will callthe P-degreeofC.

Giveni2I there is ansl2-tripleeig. In this section, we study the restriction of a categoricalgk-representation to such ansl2-triple. To simplify the notation, in this section we fix an elementi2I and identifyeDei.

Recall that for each weight2P we set iD h˛i_; i. By a .P Z/-graded categoricalek-representation onC, we will mean a representation such that for each 2P thek-subcategoryChas the weightirelative to thee-action. In particular, if C is an integrable categorical representation ofgk, restricting theg-action onCtoei

yields an integrable.PZ/-graded categoricale-representation onCof degreedi. Now, fix an integerk 2N. Given a .P Z/-graded k-categoryM, such that M D0 whenever i ¤k, and a Z-graded k-algebra homomorphism Zk;Œdi! Z.M=Z/, we equip the tensor productVk;Œdi˝Zk;Œdi Mwith the.P Z/-graded categoricale-representation of degreedi such that

eacts on the left factor;

the summandVnk;Œdi˝Zk;Œdi Mhas theP-degreen˛ifor eachn2N.

The vanishing M D0 whenever i ¤k is imposed so that the tensor product Vk;Œdi˝Zk;Œdi Mis a.PZ/-graded categoricale-representation as defined above.

PROPOSITION3.22

Fix an integrable categorical representation ofgkonC of degree1which is bounded above. For eachi2I there is a decreasing filtration C1 C0DC ofC by

Fix an integrable categorical representation ofgkonC of degree1which is bounded above. For eachi2I there is a decreasing filtration C1 C0DC ofC by

Documents relatifs