scieee Open visual document viewer

On abelian subalgebras and ideals of maximal dimension in supersolvable Lie algebras

Ceballos González, Manuel; Towers, David A.

Abstract

In this paper, the main objective is to compare the abelian subalgebras and ideals of maximal dimension for finite-dimensional supersolvable Lie algebras. We characterise the maximal abelian subalgebras of solvable Lie algebras and study solvable Lie algebras containing an abelian subalgebra of codimension 2. Finally, we prove that nilpotent Lie algebras with an abelian subalgebra of codimension 3 contain an abelian ideal with the same dimension, provided that the characteristic of the underlying field is not two. Throughout the paper, we also give several examples to clarify some results.

Full text

a Xi :1110.2389 3 [ma h.RA] 5 Jun 2013 ON ABELIAN SUBALGEBRAS AND IDEALS OF MAXIMAL DIMENSION IN SUPERSOLVABLE LIE ALGEBRAS Manuel Ceballos 1 Depa men o de Geome ia y Topologia, Uni e sidad de Se illa Apa ado 1160, 41080, Se ille, Spain and Da id A. Towe s Depa men o Ma hema ics, Lancas e Uni e si y Lancas e LA1 4YF, England Abs ac In his pape , he main objec i e is o compa e he abelian subalge- b as and ideals o maximal dimension o ini e-dimensional supe sol - able Lie algeb as. We cha ac e ise he maximal abelian subalgeb as o sol able Lie algeb as and s udy sol able Lie algeb as con aining an abelian subalgeb a o codimension 2. Finally, we p o e ha nilpo en Lie algeb as wi h an abelian subalgeb a o codimension 3 con ain an abelian ideal wi h he same dimension, p o ided ha he cha ac e is ic o he unde lying ield is no wo. Th oughou he pape , we also gi e se e al examples o cla i y some esul s. Ma hema ics Subjec Classi ica ion 2010: 17B05, 17B20, 17B30, 17B50. Key Wo ds and Ph ases: Lie algeb as, abelian subalgeb a, abelian ideal, sol able, supe sol able, nilpo en . 1 In oduc ion Nowadays, he e exis s an ex ensi e body o esea ch o Lie Theo y due o i s own impo ance om a heo e ical poin o iew and also due o i s 1Suppo ed by MTM2010-19336 and FEDER 1 applica ions o o he ields like Enginee ing, Physics and Applied Ma he- ma ics. Howe e , some aspec s o Lie algeb as emain unknown. Indeed, he classi ica ion o nilpo en and sol able Lie algeb as is s ill an open p ob- lem, al hough he classi ica ion o ce ain o he ypes o Lie algeb as (like semi-simple and simple ones) we e al eady ob ained in 1890, a leas o e he complex ield. In o de o make p og ess on hese and o he p oblems, he need o s udying di e en p ope ies o Lie algeb as a ises. Fo exam- ple, condi ions on he la ice o subalgeb as o a Lie algeb a o en lead o in o ma ion abou he Lie algeb a i sel . S udying abelian subalgeb as and ideals o a ini e-dimensional Lie algeb a cons i u es he main goal o his pape . Th oughou Lwill deno e a ini e-dimensional Lie algeb a o e a ield F. The assump ions on Fwill be speci ied in each esul . Algeb a di ec sums will be deno ed by ⊕, whe eas ec o space di ec sums will be deno ed by ˙ +. We conside he ollowing in a ian s o L: α(L) = max{dim(A)|Ais an abelian subalgeb a o L}, β(L) = max{dim(B)|Bis an abelian ideal o L}. Bo h in a ian s a e impo an o many easons. Fo example, hey a e e y use ul o he s udy o Lie algeb a con ac ions and degene a ions. The e is a la ge li e a u e, in pa icula o low-dimensional Lie algeb as, see [9, 6, 13, 15, 8], and he e e ences gi en he ein. The i s au ho dealing wi h he in a ian α(g) was Schu [14], who s udied in 1905 he abelian subalgeb as o maximal dimension con ained in he Lie algeb a o n×nsqua e ma ices. Schu p o ed ha he max- imum numbe o linea ly independen commu ing n×nma ices o e an algeb aically closed ield is hn2 4i+ 1, which is he maximal dimension o abelian ideals o Bo el subalgeb as in he gene al linea Lie algeb a gl(n) ( whe e [x] deno es he in ege pa o a eal numbe x). Le us no e ha his esul was ob ained only o e an algeb aically closed ield such as he complex numbe ield. Almos o y yea s la e , in 1944, Jacobson [10] ga e a simple p oo o Schu ’s esul s, ex ending hem om algeb aically closed ields o a bi a y ields. This ac allowed se e al au ho s o gain insigh in o he abelian subalgeb as o maximal dimension o many di e en ypes o Lie algeb as. Mo e speci ically, o semisimple Lie algeb as s he in a ian α(s) has been comple ely de e mined by Malce [12]. Since he e a e no abelian ideals in s, we ha e β(s) = 0. The alue o α o simple Lie algeb as is ep oduced in able 1. In his pape , we will s udy se e al p ope ies o 2 hese in a ian s and compa e hem o supe sol able, sol able and nilpo en Lie algeb as. Table 1: The in a ian α o simple Lie algeb as sdim(s)α(s) An, n ≥1n(n+ 2) ⌊(n+1 2)2⌋ B321 5 Bn, n ≥4n(2n+ 1) n(n−1) 2+ 1 Cn, n ≥2n(2n+ 1) n(n+1) 2 Dn, n ≥4n(2n−1) n(n−1) 2 G214 3 F452 9 E678 16 E7133 27 E8248 36 We shall call Lsupe sol able i he e is a chain 0 = L0⊂L1⊂... ⊂ Ln−1⊂Ln=L, whe e Liis an i-dimensional ideal o L. The ideals L(k)o he de i ed se ies a e de ined by L(0) =L, L(k+1) = [L(k), L(k)] o k≥0; we also w i e L2 o L(1) and L3 o [L2, L]. I is well known ha e e y supe sol able Lie algeb a is also sol able. Mo eo e , hese classes coincide o e an algeb aically closed ield o cha ac e is ic ze o (Lie’s heo em). The e a e, howe e , examples o sol able Lie algeb as o e algeb aically closed ield o non-ze o cha ac e is ic which a e no supe so able (see o ins ance [11, page 53] o [3]). The F a ini ideal o L,φ(L), is he la ges ideal o L con ained in all maximal subalgeb as o L. We will deno e he cen e o L by Z(L) = {x∈L: [x, y] = 0,∀y∈L}and he cen alize o a subalgeb a Ao Lby CL(A) = {x∈L: [x, A] = 0}. Gi en a subalgeb a Ao L, he co e o A, deno ed by AL, is he la ges ideal o Lcon ained in A. The abelian socle o L, AsocL, is he sum o he minimal abelian ideals o L. The s uc u e o his pape is as ollows. In sec ion 2 we gi e some bounds o he in a ian s αand β. In sec ion 3, we conside he classes o supe sol able, sol able and nilpo en Lie algeb as Lwi h α(L) = n−1 o n−2. In pa icula , we cha ac e ise n-dimensional sol able Lie algeb as L o which α(L) = n−2 and p o e ha e e y supe sol able Lie algeb a, L, o dimension nwi h α(L) = n−2 also sa is ies β(L) = n−2. In he inal sec ion we show he αand βin a ian s also coincide o nilpo en Lie 3 algeb as Lwi h α(L) = n−3, p o ided ha Fhas cha ac e is ic di e en om wo. We also gi e an example o show ha he es ic ion on Fis necessa y. 2 Some bounds on α(L)and β(L) We shall call a Lme abelian i L2is abelian. Fi s we ha e a bound on β(L) o ce ain me abelian Lie algeb as. P oposi ion 2.1 Le Lbe a me abelian Lie algeb a o dimension n, and suppose ha dim L2=k. Then dim(L/CL(L2)) ≤[k2/4] + 1. I , u he , L spli s o e L2 hen β(L)≥n−[k2/4] −1. P oo . Le ad : L→De L2be de ined by ad x(y) = [y, x] o all y∈L2. Then ad is a homomo phism wi h ke nel CL(L2). I ollows ha L/CL(L2)∼ = Dwhe e Dis an abelian subalgeb a o De L2∼ =gl(n, F ). I ollows om Schu ’s Theo em on commu ing ma ices (see [10]) ha dim(L/CL(L2)≤ [k2/4] + 1. Now suppose ha L=L2⊕Bwhe e Bis an abelian subalgeb a o L. Then CL(L2) = L2⊕B∩CL(L2) which is an abelian ideal o L. We call Lcomple ely sol able i L2is nilpo en . O e a ield o cha ac- e is ic ze o, e e y sol able Lie algeb a is comple ely sol able. Nex we no e ha i Lis comple ely sol able, has an abelian nil adical (so is me abelian) and he unde lying ield is pe ec hen α(L) and β(L) a e easily iden i ied. I ¯ Fis he algeb aic closu e o Fwe pu ¯ S=S⊗F¯ F o e e y subalgeb a So L. Lemma 2.2 α(¯ L)≥α(L),β(¯ L)≥β(L). Lemma 2.3 Le Lbe any sol able Lie algeb a wi h nil adical N. Then CL(N)⊆N P oo . Suppose ha CL(N)6⊆ N. Then he e is a non- i ial abelian ideal A/(N∩CL(N) o L/(N∩CL(N) inside CL(N)/(N∩CL(N). Bu now A3⊆ [A, N] = 0, so Ais a nilpo en ideal o L. I ollows ha A⊆N∩CL(N), a con adic ion.  Theo em 2.4 I Fis a pe ec ield and Lis a comple ely sol able Lie algeb a wi h abelian nil adical N hen α(L) = β(L) = dim N. 4 P oo . I is clea ha Nis he unique maximal abelian ideal o L. Le Abe an abelian subalgeb a o Lo maximal dimension. I N⊆A, hen A⊆CL(N) = N, by Lemma 2.3, so N=Aand he esul is clea , so suppose ha N6⊆ Aand pu U=N+A. Conside i s he case whe e Fis algeb aically closed and φ(L) = 0. Pick any a∈Aand pu C=N+Fa. Then φ(C) = 0, by [18, Theo em 2.5], so N⊆N(C) = Asoc(C) by [17, Theo em 7.4], and Nis comple ely educible as an Fa-module. W i e N=⊕k i=1Ni, whe e Niis an i educible Fa-module o 1 ≤i≤k. Then he minimal polynomial o he es ic ion o ad a o Niis i educible o each i, and so {(ada)|N:a∈A}is a se o commu ing diagonalizable ope a o s. Thus N= AsocL=Fn1+. . . Fn , whe e Fniis a minimal ideal o L o 1 ≤i≤ . I ni∈A, hen CU(ni) = U; i ni/∈A hen dim CU(ni)≥dim U−1, since U/CU(ni)∼ =D, whe e Dis a subalgeb a o De Fni. Bu N=CL(N)⊇CU(N) = i=1 CU(Fni), so dim N≥dim U−( −dim(N∩A)) = dim A, and he esul holds in his case. So suppose now ha φ(L) is no necessa ily i ial. Then N/φ(L) is he nil adical o L/φ(L), by [17, Theo em 6.1], and so L/φ(L) has abelian nil adical. Also (A+φ(L))/φ(L) is an abelian subalgeb a o L/φ(L). I ollows om he abo e ha dim A+φ(L) φ(L)≤dim N φ(L),whence dim A≤dim N. Finally conside he case whe e Fis no necessa ily algeb aically closed. Since Fis pe ec , N(L) = N(¯ L), by [4, page 42]. Hence β(L)≤α(L)≤α(¯ L) = β(¯ L) = dim N(¯ L) = dim N(L) = dim N(L) = β(L) by Lemma 2.2 and he abo e.  We ob ain bounds o supe sol able Lie algeb as by ollowing a de elop- men simila o [16, Lemma 2]. Lemma 2.5 Le Lbe a supe sol able Lie algeb a and le Abe a maximal abelian ideal o L. Then CL(A) = A. 5 P oo . We ha e ha CL(A) is an ideal o L. Suppose ha CL(A)6=A. Le B/A be a minimal ideal o L/A wi h B⊂CL(A). Then, o some b∈B, B=A+Fb, which is an abelian ideal o L, con adic ing he maximali y o A. The esul ollows.  P oposi ion 2.6 Le Lbe a supe sol able Lie algeb a, Aany maximal abelian ideal o L. Suppose dim A=k. Then L/A is isomo phic o a Lie algeb a o k×klowe iangula ma ices. P oo . Le ad : L→De Abe de ined by ad x(y) = [y, x]. Then ad is a homomo phism wi h ke nel CL(A) = A, by Lemma 2.5. Since Lis supe sol able he e is a lag o ideals 0 = A0⊂A1⊂... ⊂Ak=Ao L. Choose a basis e1,...,ek o Awi h ei∈Ai. Wi h espec o his basis he ac ion o Lon Ais ep esen ed by k×klowe iangula ma ices, since [Ai, L]⊆Ai o each 0 ≤i≤k. Co olla y 2.7 Le Lbe a supe sol able Lie algeb a wi h a maximal abelian ideal Ao dimension k. Then dim L≤k(k+3) 2and Lhas de i ed leng h a mos k+ 1. P oo . The Lie algeb a o k×klowe iangula ma ices has dimension k(k+1) 2and de i ed leng h k. Co olla y 2.8 Le Lbe a supe sol able Lie algeb a o dimension n. Then β(L)≥√8n+ 9 −3 2. Co olla y 2.9 Le Lbe a non-abelian sol able Lie algeb a o dimension n o e an algeb aically closed ield o cha ac e is ic ze o. Then √8n+ 9 −3 2≤α(L)≤n−1. P oo . Simply use Co olla y 2.8 and [7, P oposi ion 2.5].  3 Supe sol able Lie algeb as wi h α(L) = n−1o n−2 P oposi ion 3.1 Le Abe an abelian subalgeb a o a Lie algeb a L. Suppose ha K=A+Fe1is a subalgeb a o L, and ha he e is an x∈Lsuch ha [x, K]⊆K, bu [x, A]6⊆ A. Then ei he Kis abelian o K2is one dimensional and Z(K)has codimension a mos one in A. 6 P oo . Le e2,...,ekbe a basis o Asuch ha e1= [x, e2], say. Le [x, ej] = Pk i=1 αjiei o 1 ≤j≤k. Then [e2,[x, ej]] = αj1[e2, e1], so, o 2≤j≤k, 0 = [x, [e2, ej]] = −[e2,[ej, x]] −[ej,[x, e2]] = αj1[e2, e1]−[ej, e1]. Hence [e1, ej] = αj1[e1, e2]. I ollows ha K2=F[e1, e2]. Pu j=αj1e2− ej o 3 ≤j≤k. Then 3,..., k∈Z(K)∩A. The abo e esul deals wi h he case whe e an abelian subalgeb a o maximal dimension has codimension one in an ideal o L. Co olla y 3.2 Le Lbe a supe sol able Lie algeb a and le Abe an abelian subalgeb a o maximal dimension in L. I A⊂Kwhe e Kis an ideal o L and Ahas codimension 1in K, hen α(L) = β(L). P oo . I Ais an ideal o L hen he esul is clea , so suppose ha i is no an ideal o L. Wi h he same no a ion as in P oposi ion 3.1 he hypo heses o ha esul a e sa is ied. Then 3,..., k∈Z(K); in ac , he maximali y o Agi es Z(K) = F 3+···+F k. Le B/(F 3+···+F k) be a chie ac o o Lwi h B⊂K. Then Bis an abelian ideal o Lwi h he same dimension as A. The esul ollows.  Nex we conside he si ua ion whe e Lhas a maximal subalgeb a ha is abelian: i s when Lis any non-abelian Lie algeb a and Fis algeb aically closed, and hen when Lis sol able bu Fis a bi a y. P oposi ion 3.3 Le Lbe a non-abelian Lie algeb a o dimension no e an algeb aically closed ield Fo any cha ac e is ic. Then Lhas a maximal subalgeb a M ha is abelian i and only i L=A˙ +F o some ∈gl(V), whe e 6≡ 0,Ais abelian and [ , ] = −[ , ] = ( ). In pa icula , α(L) = β(L) = n−1. P oo . Suppose i s ha Lhas a maximal subalgeb a M ha is abelian. I Mis an ideal o Lwe ha e inished. So suppose ha Mis sel -idealising, in which case L2is one-dimensional and φ(L) = 0, by [19, P oposi ion 3.2]. W i e L2=Fb and no e ha AsocL=L2⊕Z(L). Now L= AsocL˙ +C, whe e Cis abelian, by [17, Theo em 7.4]. Fo each c∈Cwe ha e ha [c, b] = λ(c)b, o some λ(c)∈F. Since C∩Z(L) = 0, Cmus be one- dimensional and he esul ollows. The con e se is clea .  The ollowing is a gene alisa ion o [19, P oposi ion 3.1] 7 P oposi ion 3.4 Le Lbe a sol able Lie algeb a. Then Lhas a maximal subalgeb a M ha is abelian i and only i ei he (i) Lhas an abelian ideal o codimension one in L; o (ii) L(2) =φ(L) = Z(L),L2/L(2) is a chie ac o o L, and Lspli s o e L2. P oo . By [19, P oposi ion 3.1] i su ices o show ha i Lhas a maximal subalgeb a M ha is no an ideal hen L2is nilpo en . By maximali y, M is sel -idealising, and om sol abili y he e is a k≥1 such ha L(k)6⊆ M bu L(k+1) ⊆M. Then L=M+L(k), whence L2⊆L(k). I ollows ha L(2) ⊆Mand we ha e M⊆CL(L(2)). By maximali y and he ac ha M is sel -idealising, we ge CL(L(2)) = L, which yields ha L2is nilpo en .  Nex we cha ac e ise sol able Lie algeb as Lwhose bigges abelian subal- geb as ha e codimension wo in L. The ollowing p oo elies on [7, P opo- si ions 3.1 and 5.1] which a e only s a ed o Lie algeb as o e ields o cha ac e is ic ze o. Howe e , i is easy o see ha his assunp ion is no used in hei p oo s, and ha he esul s a e, in ac , alid o e an a bi a y ield. Theo em 3.5 Le Lbe a sol able Lie algeb a o dimension nwi h α(L) = n−2, and le Abe an abelian subalgeb a o dimension n−2. Then one o he ollowing occu s: (i) β(L) = n−2; (ii) L=L2˙ +B, whe e Bis an abelian subalgeb a o L,L2is he h ee- dimensional Heisenbe g algeb a, L(2) =φ(L) = Z(L)and L2/Z(L)is a wo-dimensional chie ac o o L(in which case β(L)≤n−3); (iii) Ahas codimension one in he nil adical, N, o L, which i sel has codimension one in L. Mo eo e , N2is one dimensional, Z(N)is an abelian ideal o maximal dimension and β(L) = n−3. P oo . Le Abe a maximal abelian subalgeb a o Lo dimension n−2 and suppose ha (i) doesn’ hold. (a) Suppose i s ha Ais a maximal subalgeb a o L. Then Lis as in P oposi ion 3.4(ii) and Ais a Ca an subalgeb a o L. Le L=A˙ +L1 be he Fi ing decomposi ion o L ela i e o A. Then L1⊆L2and dim L1= 2. Le L1=Fx +Fy. 8 I [x, y] = 0 hen L1is an ideal o Land L/L1is abelian, so L2⊆ L1⊆L2. This yields ha Lis me abelian and L2is a wo dimensional minimal ideal o e which Lspli s. I ollows om P oposi ion 2.1 ha β(L) = n−2, a con adic ion. I [x, y]6= 0, hen L2=F[x, y] + L1and F[x, y]⊆L(2) =Z(L), so F[x, y] = Z(L) and we ha e case (ii). Mo eo e , i Cis a maximal abelian ideal o L, hen Z(L)⊆Cand L26⊆ C. I ollows ha C∩L2=Z(L). I dim C=n−2, hen dim(L2+C) = dim L2+ dim C−dim C∩L2= 3 + n−2−1 = n, so L=L2+C. Bu hen L2=Z(L), a con adic ion. Hence, β(L)≤n−3. (b) So suppose ha Ais no a maximal subalgeb a o L. Then A⊂M⊂L, whe e dim M=n−1. Mo eo e , he e is such a subalgeb a Ao M which is an ideal o M, by [7, P oposi ion 3.1]. Suppose i s ha A does no ac nilpo en ly on L. Then he Fi ing decomposi ion o L ela i e o Ais L=M˙ +L1, and L1is a one-dimensional ideal o L. Pu B=A˙ +L1, which is an ideal o L. Then CB(L1) has codimension one in Band so is an abelian ideal o codimension wo in L. I ollows ha β(L) = n−2, a con adic ion. Finally, suppose ha Ais an ideal o Mand ha Aac s nilpo en ly on L. Then he e is a k≥0 such ha L(ad A)k6⊆ Mbu L(ad A)k+1 ⊆ M. Le x∈L(ad A)k M, so L=M˙ +Fx. Suppose i s ha Mis no an ideal o L. Then he co e o M,MLhas codimension one in M, by [1, Theo em 3.1 and 3.2]. I A=ML hen we ha e case (i), so suppose ha A6=MLand M=A+ML. Then [A, x]⊆Mwhich implies ha [L, A]⊆Mand [L, M] = [L, ML] + [L, A]⊆M; ha is, Mis an ideal o L. Le Nbe he nil adical o L. I N⊆A hen A⊆CL(N)⊆N, so N=Aand we ha e case (i) again. I A⊂N hen N=Lo we can assume ha N=M. I A6⊆ Nand N6⊆ A hen ei he A+N=L, in which case Lis nilpo en , o we can assume ha A+N=M, in which case Mis a nilpo en ideal o Land so M=N. I Lis nilpo en , hen we ha e case (i), by [7, P oposi ion 5.1].I no , hen we ha e case (iii) by P oposi ion 3.1.  Co olla y 3.6 Le Lbe a supe sol able Lie algeb a o dimension nwi h α(L) = n−2. Then β(L) = n−2. 9 [18] D. A. Towe s and V. R. Va ea, ‘Elemen a y Lie Algeb as and Lie A-algeb as’, J. Algeb a 312 (2007), 891–901. [19] D. A. Towe s, ‘The index complex o a maximal subalgeb a o a Lie algeb a’, P oc. Edin. Ma h. Soc. 54 (2011), 531–542. 16