ebook img

Representations of trigonometric Cherednik algebras of rank 1 in positive characteristic PDF

0.21 MB·English
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview Representations of trigonometric Cherednik algebras of rank 1 in positive characteristic

Representations of trigonometric Cherednik algebras of rank 1 in positive characteristic Fr´ed´eric Latour 4 0 0 February 1, 2008 2 n a 1 Introduction J 7 2 Cherednik’s double affine Hecke algebras are an important class of algebrasattached to root systems. They were introduced in [Che95] as a tool of proving Macdonald’s conjectures, but are also interesting by them- ] selves,sincetheyprovideuniversaldeformationsoftwistedgroupalgebrasofdoubleaffineWeylgroups. One T may distinguish rational, trigonometric,and elliptic Cherednik algebras,which contain 0, 1, and 2 copies of R the root lattice, respectively (rational and trigonometric algebras are degenerations of the elliptic ones; see . h [EG02]). t Development of representation theory of Cherednik algebras (in particular, description of all irreducible a m finite dimensionalrepresentations)isanimportantopenproblem. Inthe characteristiczerocase,itissolved completely only for type A, while in other types only partial results are available (see [EG02],[BEG03], and [ [CO] for the rank 1 case). In positive characteristic,the rank 1 case (in the more generalsetting of complex 1 reflectiongroups)issettledbytheauthorin[Lat],afterwhichthehigherrankcase(oftypeA)wasconsidered v in [FG]. 7 8 The goal of this paper is to extend the results of [Lat] to the trigonometric case. That is, we study the 3 representationtheory oftrigonometricCherednik algebrasin positive characteristicp in the simplestcase of 1 rank 1. Our main result is a complete description of irreducible representations of such algebras. 0 The paper is organized as follows. 4 In Section 2, we state the main results. 0 / In Section 3, we prove the results for the “classical” case, i.e. the case when the “Planck’s constant” t h is zero. In this case, generic irreducible representations have dimension 2; one-dimensional representations t a exist when the “coupling constant” k is zero. m In Section 4, we prove the results for the “quantum” case, i.e. the case when the “Planck’s constant” t : is nonzero. In this case, generic irreducible representations have dimension 2p; smaller representations exist v when the “coupling constant” k is an element of F ⊂k; namely, if k is an integer with 0≤k ≤p−1, then i p X there exist irreducible representations of dimensions p−k and p+k. r Acknowledgements. The author thanks his adviser Pavel Etingof for posing the problem and useful a discussions, as well as for helping to write an introduction. The work of the author was partially supported by the National Science Foundation (NSF) grant DMS-9988796and by a Natural Sciences and Engineering Research Council of Canada (NSERC) Julie Payette researchscholarship. 2 Statement of Results Let k be an algebraically closed field of characteristic p, where p 6= 2. Let t,k ∈ k, and let H(t,k) be the algebra (over k) generated by X,X−1,s and y, subject to the following relations: 1 sX = X−1s (1) s2 = 1 (2) sy+ys = −k (3) XyX−1 = y−t+ks. (4) We will classify the irreducible representations of H(t,k). Now, for t 6= 0, H(t,k) is clearly isomorphic to H(1,k) under the map t 1 X7→X,s7→s,y7→ y. t Thus itis sufficientto classify irreduciblerepresentationsofH(0,k)andH(1,k). For brevitywe willuse the def def notation H = H(0,k) and H = H(1,k), assuming that k has been fixed once and for all. 0 1 2.1 Irreducible representations of H 0 Proposition 2.1. Let k6=0. Then the irreducible representations of H are the following: 0 • For a,β ∈ k, a,β 6= 0, we have a two-dimensional representation Vβ,a with basis {v ,v }, defined by 0,1 0 1 the following: yv = βv , 0 0 yv = −βv , 1 1 k2 Xv = av − v , 0 0 4β2 1 1 k2 Xv = v + − v , 1 0 a 4aβ2 1 (cid:18) (cid:19) k k3−4kβ2 sv = − v + v , 0 2β 0 8aβ3 1 2aβ k sv = − v + v ; 1 0 1 k 2β • For a = ±1,b ∈ k, we have a two-dimensional representation Va,b with basis {v ,v }, defined by the 0,2 0 1 following: yv = 0, 0 yv = v , 1 0 sv = v −kv , 0 0 1 sv = −v , 1 1 Xv = a(v −kv ), 0 0 1 Xv = bv +(a−kb)v . 1 0 1 Vβ,a and Vβ′,a′ are isomorphic if and only if β′ = β,a′ = a or β′ = −β,a′ = 4β2−k2. Va,b and Va′,b′ are 0,1 0,1 4aβ2 0,2 0,2 isomorphic if and only if a=a′ and b=b′. Furthermore, representations with different subscripts are never isomorphic. Proposition 2.2. Let k=0. Then the irreducible representations of H are the following: 0 2 • For a,β ∈ k, a,β 6= 0, we have a two-dimensional representation Vβ,a with basis {v ,v }, defined by 0,3 0 1 the following: yv = βv , 0 0 yv = −βv , 1 1 Xv = av , 0 0 1 Xv = v , 1 1 a sv = v , 0 1 sv = v ; 1 0 • For a ∈ k, a ∈/ {0,±1}, we have a two-dimensional representation Va with basis {v ,v }, defined by 0,4 0 1 the following: yv = 0, 0 yv = 0, 1 Xv = av , 0 0 1 Xv = v , 1 1 a sv = v , 0 1 sv = v ; 1 0 • For a = ±1,b = ±1, we have a one-dimensional representation Va,b on which y,X and s act as 0,a 0,5 and b respectively. Vβ,a and Vβ′,a′ are isomorphic if and only if β′ =β,a′ =a or β′ =−β,a′ = 1. Va and Va′ are isomorphic 0,3 0,3 a 0,4 0,4 if and only if a′ = a or a′ = 1. Va,b and Va′,b′ are isomorphic if and only if a′ = a,b′ = b. Furthermore, a 0,5 0,5 representations with different subscripts are never isomorphic. 2.2 Irreducible representations of H 1 Proposition 2.3. Let k∈/ F . Then the irreducible representations of H are the following: p 1 • For µ,d∈k,d6=0,b=(µp−µ)2 with k not a root of f(y)=(yp−y)2−b, and also for µ=±k,d6=0, 2 2 we have a 2p-dimensional representation V1µ,1,d with basis {vµ+j,v−µ+j,j = 0,1,...,p−1}, defined by the following: yv = βv , β =±µ,±µ+1,...,±µ+p−1; (5) β β 1 k sv−µ−j = − vµ+j + v−µ−j, j =1,2,...,p−1; (6) µ+j 2(µ+j) k2 k svµ+j = −(µ+j) v−µ−j − vµ+j, j =1,2,...,p−1; (7) 4(µ+j) 2(µ+j) (cid:18) (cid:19) k d sv−µ = v−µ− vµ; (8) 2µ µ k2 µ k svµ = − v−µ− vµ; (9) 4dµ d 2µ (cid:18) (cid:19) Xvβ = sv−β−1, β =±µ,±µ+1,...,±µ+p−1; (10) 3 • Forθ =±1,wehavea2p-dimensionalrepresentationVθ withbasis {v ,w ,j =0,1,...,p−1},defined 1,2 j j by the following: yv = jv , j =0,1,...,p−1; (11) j j yw = jw +v , j =0,1,...,p−1; (12) j j j sv = −kw ; (13) 0 0 1 sw = − v ; (14) 0 0 k 1 k p−1 sv−j = vj + v−j, j =1,2,..., ; (15) j 2j 2 k2 k p−1 svj = j− v−j − vj, j =1,2,..., ; (16) 4j 2j 2 (cid:18) (cid:19) 1 k 1 k p−1 sw−j = j2vj + 2j2v−j − jwj + 2jw−j j =1,2,..., 2 ; (17) k2 k k k2 p−1 swj = 1+ 4j2 v−j + 2j2vj − 2jwj + 4j −j w−j, j =1,2,..., 2 ; (18) (cid:18) (cid:19) (cid:18) (cid:19) p−1 Xvj = −sv−j−1, j 6= ; (19) 2 Xvp−1 = θsvp−1; (20) 2 2 p−1 Xwj = sw−j−1, j 6= ; (21) 2 Xwp−1 = −θswp−1. (22) 2 2 Vµ,d and Vµ′,d′ are isomorphic if and only if 1,1 1,1 k2 (µ′−µ∈F and d′ =d) or (µ′+µ∈F and dd′ = −(µ+c)2 ). p p 4 cY∈Fp(cid:18) (cid:19) Vθ and Vθ′ are isomorphic if and only if θ =θ′ Furthermore, representations with different subscripts are 1,2 1,2 never isomorphic. Now, in the case where k ∈F , note that there is an isomorphism between H(1,k) and H(1,−k), given p by y7→y,s7→−s,X7→X,k 7→−k. So we may assume that k is an even integer with 0≤k ≤p−1. Proposition 2.4. Let k be even with 2 ≤ k ≤ p−1. Then the irreducible representations of H are the 1 following: • For µ,d∈k,d6=0, we have Vµ,d, defined as in Proposition 2.3. 1,1 • For θ = ±1, we have a (p − k)-dimensional representation Vθ with basis {v ,v ,...,v }, 1,3 k k+1 −k−1 2 2 2 defined by k k k yv = jv , j = , +1...,− −1; (23) j j 2 2 2 k 1 k p−1 sv−j = v−j − vj, j = +1,..., ; (24) 2j j 2 2 k2 k k p−1 svj = −jv−j + v−j − vj, j = +1,..., ; (25) 4j 2j 2 2 sv = −v ; (26) k k 2 2 4 p−1 Xvj = −sv−j−1, j 6= ; (27) 2 Xvp−1 = θsvp−1. (28) 2 2 • For θ =±1, we have a (p+k)-dimensional representation Vθ with basis {v ,w ,j =0,...,p−1,i= 1,4 j i −k,...,k −1}, defined by 2 2 yv = jv , j =0,1,...,p−1; (29) j j k k k yw = jw +v , j =− ,− +1,..., −1; (30) j j j 2 2 2 sv = −kw ; (31) 0 0 1 sw = − v ; (32) 0 0 k 1 k k k p−1 sv−j = vj + v−j, j =1,..., −1, +1, ; (33) j 2j 2 2 2 k2 k k k p−1 svj = j− v−j − vj, j =1,..., −1, +1, ; (34) 4j 2j 2 2 2 (cid:18) (cid:19) sv = v ; (35) −k −k 2 2 sv = 2v −v ; (36) k −k k 2 2 2 1 k 1 k k sw−j = j2vj + 2j2v−j − jwj + 2jw−j j =1,..., 2 −1; (37) k2 k k k2 k swj = 1+ 4j2 v−j + 2j2vj − 2jwj + 4j −j w−j, j =1,..., 2 −1; (38) (cid:18) (cid:19) (cid:18) (cid:19) 2 2 sw = − v + v +w ; (39) −k k −k −k 2 k 2 k 2 2 p−1 Xvj = −sv−j−1, j 6= ; (40) 2 Xvp−1 = θsvp−1; (41) 2 2 k k Xwj = =sw−j−1, j =− ,..., −1. (42) 2 2 • For c ∈ k, we have a 2p-dimensional representation Vc with basis {v ,w ,u ,j = 0,...,p−1,i = 1,5 j i l −k,...,k −1,l= k,...,−k −1}, defined by 2 2 2 2 yv = jv , j =0,1,...,p−1; (43) j j k k k yw = jw +v , j =− ,− +1,..., −1; (44) j j j 2 2 2 sv = w ; (45) 0 0 sw = v ; (46) 0 0 1 k k k p−1 sv−j = vj + v−j, j =1,..., −1, +1, ; (47) j 2j 2 2 2 k2 k k k p−1 svj = j− v−j − vj, j =1,..., −1, +1, ; (48) 4j 2j 2 2 2 (cid:18) (cid:19) sv = v ; (49) −k −k 2 2 sv = 2v −v ; (50) k −k k 2 2 2 5 1 k 1 k k sw−j = j2vj + 2j2v−j − jwj + 2jw−j j =1,..., 2 −1; (51) k2 k k k2 k swj = 1+ 4j2 v−j + 2j2vj − 2jwj + 4j −j w−j, j =1,..., 2 −1; (52) (cid:18) (cid:19) (cid:18) (cid:19) 2 2 sw = − v + v +w ; (53) −k k −k −k 2 k 2 k 2 2 1 k k p−1 suj = − u−j − uj, j = +1,..., ; (54) j 2j 2 2 k2 k k p−1 su−j = −j uj + u−j, j = +1,..., ; (55) 4j 2j 2 2 (cid:18) (cid:19) 2c su = v −u ; (56) k −k k 2 k 2 2 p−1 Xvj = −sv−j−1, j 6= ; (57) 2 Xvp−1 = sup−1; (58) 2 2 k p−3 p+1 k Xuj = −su−j−1, j = ,..., , ,...,− −1; (59) 2 2 2 2 Xup−1 = svp−1. (60) 2 2 Vµ,d and Vµ′,d′ are isomorphic if and only if 1,1 1,1 k2 (µ′−µ∈F and d′ =d) or (µ′+µ∈F and dd′ = −(µ+c)2 ). p p 4 cY∈Fp(cid:18) (cid:19) Vθ and Vθ′ are isomorphic if and only if θ =θ′. Vc and Vc′ areisomorphic if and only if c=c′. Vc and 1,3 1,3 1,4 1,4 1,5 Vc′ are isomorphic if and only if c = c′. Furthermore, representations with different subscripts are never 1,5 isomorphic. Proposition 2.5. Let k=0. Then the representations of H are the following: 1 • For µ,d∈k,d6=0,b=(µp−µ)2, we have Vµ,d, defined as in Proposition 2.3. 1,1 • For c,θ =±1, we have a p-dimensional representation Vc,θ with basis {v ,j =0,1,...,p−1}, defined 1,6 j by yv = jv , j =0,1,...,p−1; (61) j j sv = cv ; (62) 0 0 p−1 svj = −jv−j, j =1,..., ; (63) 2 1 p−1 sv−j = − vj, j =1,..., ; (64) j 2 p−1 Xvj = sv−j−1, j 6= ; (65) 2 Xvp−1 = θsvp−1. (66) 2 2 • For c=±1,a∈k, we have a 2p-dimensional representation Vc,a with basis {v ,u ,j =0,1,...,p−1}, 1,7 j j defined by yv = jv , j =0,1,...,p−1; (67) j j yu = ju , j =0,1,...,p−1; (68) j j sv = v ; (69) 0 0 su = av −u ; (70) 0 0 0 6 1 p−1 sv−j = − vj j =1,2,..., ; (71) j 2 p−1 svj = −jv−j j =1,2,..., ; (72) 2 1 p−1 suj = u−j, j =1,2,..., ; (73) j 2 p−1 su−j = juj, j =1,2,..., ; (74) 2 p−1 Xvj = sv−j−1, j 6= ; (75) 2 Xvp−1 = sup−1; (76) 2 2 p−1 Xuj = su−j−1, j 6= ; (77) 2 Xup−1 = svp−1. (78) 2 2 Vµ,d and Vµ′,d′ are isomorphic if and only if 1,1 1,1 k2 (µ′−µ∈F and d′ =d) or (µ′+µ∈F and dd′ = −(µ+c)2 ). p p 4 cY∈Fp(cid:18) (cid:19) Vc,θ and Vc′,θ′ are isomorphic if and only if θ = θ′. Va and Va′ are isomorphic if and only if a = a′. 1,6 1,6 1,7 1,7 Furthermore, representations with different subscripts are never isomorphic. 3 Proof of Propositions 2.1 and 2.2 Lemma 3.1 (PBW for H , easy direction). The elements 0 siXjyl, j,l ∈Z,l ≥0,i∈{0,1} span H over k. 0 Proof. Given a product of X,y,s,X−1 in any order,one can ensure that the y’s are to the rightof all the X’s by using yX=Xy−ksX repeatedly, and one can also ensure that the s’s are to the left of all the X’s and y’s by using Xs=sX−1 and ys=−k−sy repeatedly. Lemma 3.2. X+X−1,y2 and Xy−yX−1 belong to the center Z(H ) of H . 0 0 Proof. First, let us show that y2 ∈Z(H ). We have 0 Xy2 = (yX+ksX)y = yXy+ksXy = y(yX+ksX)+ks(yX+ksX) = y2X+k(ys+sy)X+k2s2X = y2X+−k2X+k2X = y2X; thus [X,y2]=0. We also have sy2 = (−ys−k)y=−ysy−ky=−y(−ys−k)−ky=y2s; 7 thus [s,y2]=0. It follows that y2 ∈Z(H ). Next, we show that X+X−1 ∈Z(H ). We have 0 0 y(X+X−1)=Xy−ksX+X−1y+kX−1s=(X+X−1)y, and s(X+X−1)=X−1s+Xs=(X+X−1)s. Thus [y,X+X−1]=[s,X+X−1]=0, and so X+X−1 ∈Z(H ). Finally, we show that Xy−yX−1 ∈Z(H ). 0 0 First we note that yX−X−1y=yX+Xy−(X+X−1)y=yX+Xy−y(X+X−1))=Xy−yX−1, and thus X(Xy−yX−1)=X(yX−X−1y)=(Xy−yX−1)X, y(Xy−yX−1)=y(yX−X−1y)=y2X−yX−1y=Xy2−yX−1y=(Xy−yX−1)y, and s(Xy−yX−1)=s(yX−X−1y)=−(ys+k)X−Xsy=−yX−1s−kX+X(ys+k)=(Xy−yX−1)s. Thus Xy−yX−1 ∈Z(H ). 0 Corollary 3.3. H is finitely generated as a module over its center. 0 Proof. From Lemmas 3.1 and 3.2, we see that H is generated over its center by 0 siXjyl, i,j,l∈{0,1}. Corollary 3.4. Every irreducible H -module is finite-dimensional over k. 0 Proof. Standard. Thus Schur’s lemma implies that central elements of H act as scalars in any irreducible H -module. 0 0 Fromthis point,wewillusethe followingnotation: theeigenspaceofy witheigenvalueβ willbe denoted V[β]. Corollary 3.5. Let V be an irreducible H -module, and let β be an eigenvalue of y. Suppose β 6=0. Then, 0 V =V[β]⊕V[−β], and dimV[β]=dimV[−β]=1. Proof. Suppose V[β]6= 0, and let v ∈V[β] be nonzero. From the proof of corollary 3.3, we know that V is spanned by {v,Xv,sv,sXv}. Now let w =sXv; then yw =ysXv =−syXv−kXv =−sXyv+ks2Xv−kXv =−βsXv =−βw. Thus, w ∈V[−β]. Clearly, w 6=0, and thus V[−β]6=0. Now let v′ =2βXv−kw. Then, yv′ =2βyXv+βkw =2βXyv−2kβsXv+βkw =2β2Xv−βkw =βv′. Hence, v′ ∈V[β]. Also, if w′ =kv+2βsv, then yw′ =kyv+2βysv =βkv−2βsyv−2βkv =−βkv−2βsyv =−βw′, 8 Hence, w′ ∈V[−β]. From this it follows that V =V[β]⊕V[−β]. Now let H be the subalgebra of H generated by Z(H ) and 2βX−ksX. It is clear that V[β] = Hv. 0 0 0 Since this is true for all nonzero v ∈V[β], it follows that V[β] is an irreducible representation of H . Since 0 H is commutative, we see that V[β] is one dimensional. The same holds for V[−β], and the corollary is 0 proved. Corollary 3.6. Assume k 6= 0. Let V be an irreducible H -module, and suppose 0 is an eigenvalue of y. 0 Then, V =V [0], the generalized eigenspace of 0. We also have dimV =2 and dimV[0]=1. gen Proof. Let v ∈V[0] be nonzero. From the proof of corollary 3.3, we know that V is spanned by {v,Xv,sv,sXv}. Let w=−sv; then, yw =−ysv =syv+kv =kv. Let v′ =sXv =X−1sv; then, yv′ =yX−1sv =X−1ysv+kX−1s2v =−X−1syv =0. Let w′ =−Xv =−sv′; then, as above, we have yw′ =v′. So we have yv =0, yw =kv, yv′ =0, yw′ =kv′; therefore, V = V [0] and V[0] is spanned by v and v′. Now let H be the subalgebra of H generated by gen 0 0 Z(H ) and sX. It is clear that V[0] = Hv. Since this is true for all nonzero v ∈ V[0], it follows that V[0] 0 is an irreducible representation of H . Since H is commutative, we see that V[0] is one dimensional. The 0 0 corollary follows from this. Corollary 3.7. Assume k = 0. Let V be an irreducible H -module, and suppose 0 is an eigenvalue of y. 0 Then, V =V[0], the eigenspace of 0. We also have 1 if 1 or −1 is an eigenvalue of X dimV = (2 otherwise. Proof. From the proofof corollary3.6, we see that y acts on V as the zero operator. Let λ be an eigenvalue of X, let VX[λ] denote the associated eigenspace and let v ∈ VX[λ] be nonzero. From the proof of corollary 3.3, we know that V is spanned by {v,sv}. Now Xsv =sX−1v =λ−1sv, so sv ∈VX[λ−1]. Clearly, sv 6=0; thus, if λ6=±1, then V =VX[λ]⊕VX[λ−1] and dimV =2. If λ = ±1, it follows that X and s commute as operators on V; since V is irreducible, this implies that dimV =1. 9 Proof of Proposition 2.1. Letβ 6=0,andletV beatwo-dimensionalrepresentationofH inwhichV[β]and 0 V[−β] both have dimension 1. Let v ∈ V[β], v ∈ V[−β] be nonzero. Let the matrices representing s and 0 1 X with respect to the basis {v ,v } be as follows: 0 1 γ δ θ ω s7→ 0 0 , X7→ 0 0 . γ δ θ ω 1 1 1 1 (cid:18) (cid:19) (cid:18) (cid:19) First, we note that X and y cannot have a common eigenvector; for if Xw = γw and yw = β′w, then ksw = XyX−1w−yw = 0, and combining this with s2 = 1 gives w = 0. Hence, by scaling, we can assume that ω =1. 0 Now the central element Xy−yX−1 acts on V as a scalar. The matrix representation of Xy−yX−1 is β (θ detX−ω ) − β (detX−1) detX 0 1 detX . βθ1 (detX−1) − β (θ detX−θ ) (cid:18) detX detX 0 0 (cid:19) Thus, 0=− β (detX−1), which means that detX=1. Hence, detX θ =θ ω −1. (79) 1 0 1 Using (4), we see that XyX−1 − y − ks = 0. Using (79), we see that the matrix yrepresentation of XyX−1−y−ks is 2βθ ω −2β−kγ −2βθ −kδ 0 1 0 0 0 . 2βθ ω2−2βω −kγ −2βθ ω +2β−kδ (cid:18) 0 1 1 1 0 1 1 (cid:19) Hence, 2β γ = (θ ω −1), (80) 0 0 1 k 2β γ = θ ω2−ω , (81) 1 k 0 1 1 2β(cid:0) (cid:1) δ = − θ (82) 0 0 k 2β δ = (1−θ ω ). (83) 1 0 1 k Using (2), we see that s2 =1. Using (80)–(83), we see that the matrix representation of s2 is 4β2(1−θ ω ) 0 k2 0 1 . 0 4kβ22(1−θ0ω1) ! Thus, 1 k2 ω = 1− . (84) 1 θ 4β2 0 (cid:18) (cid:19) Using (79)–(84), we see that V is isomorphic to Vβ,θ0. Furthermore, it is easy to see that for all a,β ∈ 0,1 k\{0}, Vβ,a is a representation of H ; furthermore, each eigenvector of y clearly generates H , and thus 0,1 0 0 Vβ,θ0 is irreducible. 0,1 Now the eigenvalues of y in Vβ,a are β and −β, and 2β(Xy−yX−1)−2β2(X+X−1) acts on Vβ,a as 0,1 0,1 k2−4β2 Id, while for β = k, X+X−1 acts on Vβ,a as aId. From this it follows that Vβ,a and Vβ′,a′ are a 2 0,1 0,1 0,1 isomorphic if and only if β′ =β,a′ =a or β′ =−β,a′ = 4β2−k2. 4aβ2 Now let V be a two-dimensional representation of H in which V[0] has dimension 1 and V [0] has 0 gen dimension 2. Let v ,v ∈V be nonzero elements such that yv =0,yv =v . Let the matrices representing 0 1 0 1 0 s and X with respect to the basis {v ,v } be as follows: 0 1 γ δ θ ω s7→ 0 0 , X7→ 0 0 . γ δ θ ω 1 1 1 1 (cid:18) (cid:19) (cid:18) (cid:19) 10

See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.