1.
Introduction
In the seventies, Kac gave a comprehensive presentation of Lie superalgebra theory in [7]. The development of Lie superalgebra theory largely depends on Lie algebra theory. Note that every semi-simple Lie algebra L satisfies [L,L]=L, Der(L)=ad(L) and C(L)={0}. In 1970s, Flato raised the question whether three properties characterize semi-simple Lie algebras. Later, Angelopouos gave a negative answer by constructing a class of counterexamples, where the minimal dimension is 35 with a Levi subalgebra isomorphic to sl(2). After that, the concept of sympathetic Lie algebras was proposed by Benayadi. Furthermore, Benayadi gave more classical properties of semi-simple Lie algebras which are still valid for sympathetic Lie algebras in [2,3]. Subsequently, this concept attracted more and more scholars' attention.
This article aims to study what we call sympathetic Lie superalgebras, which are Lie superalgebras satisfying the three aforementioned properties.
This paper is organized as follows: In Section 2, we introduce the concepts and some simple properties needed in this article. In Section 3, we get some conclusions related to sympathetic Lie superalgebras through the graded ideals and show that some properties in [2] are still valid for sympathetic Lie superalgebras. In Section 4, we study perfect Lie superalgebras. Applying to sympathetic ones, we show that every sympathetic Lie superalgebra can be uniquely decomposed to a direct sum of irreducible sympathetic ideals. In Section 5, we focus on each sympathetic Lie superalgebra that has a maximal sympathetic ideal and a decomposition of direct sum of the maximal sympathetic ideal and a characteristic ideal, which is called sympathetic decomposition. After that, we give the concept of simply sympathetic and some properties. In Section 6, we focus on graded ideals such that L/I is a sympathetic Lie superalgebra and then research some properties. Meanwhile, we also give that the Der(L) is sympathetic, and if L is indecomposable, then Der(L) is simply sympathetic.
The field involved in this article is algebraically closed of characteristic 0.
2.
Preliminaries
Throughout this paper, L always denotes a finite-dimension Lie superalgebra over a commutative ring R with 1. A Lie superalgebra L is called perfect if the derived algebra [L,L]=L. The center of L is denoted by C(L). L is called centerless if C(L)={0}. Der(L) is the derivation algebra of L.
Definition 2.1. A Lie superalgebra L is called sympathetic if C(L)={0},[L,L]=L,Der(L)=ad(L).
Example 1. Recall the classical simple Lie superalgebras osp(n,2r)(n,r≥1), spl(n,m)(n≠m), Γ(σ1,σ2,σ3), Γ2 and Γ3. From [9], if they are perfect, then they are sympathetic.
Definition 2.2. Let L be a Lie superalgebra, I be a graded ideal of L. Then I is said to be a direct factor if there exists a graded ideal J of L such that L=I⊕J.
Definition 2.3. [4] Let L be a Lie superalgebra, I be a subspace of L. Then I is called characteristic ideal if for every D∈Der(L), D(I)⊆I.
Lemma 2.4. Let L be a Lie superalgebra, I be a graded ideal of L. If I is perfect, then I is a characteristic ideal of L.
Proof. With I a graded ideal of L, for each D∈Der(L), xi,yi∈I, we have
Since I is perfect, it then follows that D(I)⊆I. Thus I is a characteristic ideal.
Proposition 1. Suppose that L is a perfect Lie superalgebra, I is a graded ideal of L. If I is a direct factor of L, then I is perfect.
Proof. Since I is a direct factor of L, there exists a graded ideal J of L such that L=I⊕J, in particular, [I,J]={0}. It follows that [L,L]=[I,I]⊕[J,J], which implies that both I and J are perfect.
Proposition 2. Let L be a Lie superalgebra with trivial center, I be a direct factor of L. Then C(I)={0}.
3.
The sympathetic Lie superalgebras
Lemma 3.1. If I is a sympathetic graded ideal of L, then there exists a graded ideal J such that L=I⊕J.
Proof. Let J=CL(I). By [4] of Lemma 2.3, we have CL(I) is a graded ideal of L. For any x∈L, adx∈Der(I) since I◃L. By virtue of Der(I)=ad(I), there exists a derivation D in Der(I) such that adx=D. Thereby there exists y∈I such that D(z)=[x,z]=[y,z] for ∀z∈I. Then [x−y,z]=0 and x−y∈CL(I)=J. Hence x=j+y for some j∈J. I∩J=I∩CL(I)=C(I)={0} since I is sympathetic. Thus L=I⊕J.
Lemma 3.2. Let L be a sympathetic Lie superalgebra, I be a graded ideal of L. Then I is a direct factor of L if and only if I is sympathetic.
Proof. On account of I is a direct factor of L, there exists a graded ideal J of L such that L=I⊕J. Then L=[L,L]=[I,I]⊕[J,J]. Since L=I⊕J, we can get I=[I,I]. Suppose D∈Der(I). In the following we define d:L→L to be a linear map by
An easy calculation shows that D belongs to Der(L). Then there exists u∈L such that d=adu. Because of L=I⊕J, and u=uI+uJ, for uI∈I and uJ∈J, therefore D=aduI. Hence we can obtain Der(I)=ad(I). Thus, we conclude that C(I)⊂C(L)={0}. It follows that I is sympathetic.
Conversely, Let J={x∈L∣[x,I]={0}}, adx|I∈Der(I). Then there exists y of I, such that adx|I=ady. Hence [x−y,z]=0, for any z∈I, which proves that j=x−y∈J, then L=I+J. Suppose a∈I∩J,b∈L. Consider b=bI+bJ,bI∈I,bJ∈J. Therefore, we have [a,b]=[a,bI]+[a,bJ]=0. Thus, a∈C(L)={0}. As a result, L=I⊕J.
In view of Lemma 3.2, we immediately have the following consequences.
Corollary 1. Let L be a Lie superalgebra, and I be a sympathetic graded ideal of L. Then I is a direct factor of L.
Corollary 2. If Der(L)=ad(L), then Der(I)=ad(I).
Proposition 3. Let L be a Lie superalgebra, I, J be two graded ideals of L and I∩J={0}. If I and J are sympathetic, then S=I⊕J is also sympathetic.
Proof. On account of I∩J={0}, we can get
Suppose that D∈Der(S). For any D∈Der(L), xi,yi∈I, we can get
Since I is a graded ideal of L, thus we obtain D(I)⊆I. Then we have I is a characteristic ideal of S, the same for J. Therefore D|I∈Der(I), D|J∈Der(J). Hence, there exist x∈I and y∈J such that
Assume that D1=ad(x+y). In the following, we shall prove D=D1.
For any s∈S, there exist u∈I, v∈J such that s=u+v. Hence
it follows that D=ad(x,y), clearly, Der(S)=ad(S).
Set c∈C(S), c=cI+cJ with cI∈I and cJ∈J. Thus,
we can obtain cI=0 and cJ=0, hence c=0, it means that C(S)={0}, consequently, the proposition holds.
Corollary 3. Let Li(i=1,⋯,n) be sympathetic Lie superalgebras, then L1×L2×⋯×Ln is also sympathetic.
Corollary 4. Any extension of sympathetic Lie superalgebras is also sympathetic and trivial.
Proof. Set L1, L2 be two sympathetic Lie superalgebras, L be the extension. By Corollary 1, there exists a surjective morphism μ:L→L1 such that Ker(μ)=L2, therefore L2 is a direct factor of L, hence the extension L is trivial, that is to say L≅L1×L2. Following, according to Corollary 3, L is sympathetic.
Proposition 4. Let L be a Lie superalgebra over a field F and T be an extension of F. Then L⊗FT is sympathetic if and only if L is sympathetic.
Proof. By a result of [8], Der(L⊗FT)=Der(L)⊗FT.
Lemma 3.3. [9] Let L be a Lie superalgebra over a field F. Each of the following statements is strictly stronger than the foregoing one.
(1) L does not contain non-zero solvable graded ideals.
(2) L is the direct product of finitely many simple Lie superalgebras.
(3) The killing form of L is nondegenerate.
(4) All finite-dimensional graded representations of L are completely reducible.
Theorem 3.4. Let L be a perfect Lie superalgebra. If L=L0⊕L1 has the nondegenerate Killing form over a field F, then L0 and L are sympathetic.
Proof. By virtue of L has the nondegenerate Killing form, Der(L)=ad(L) by [9]. Since C(L) is abelian, C(L) is solvable. Then C(L)={0} by Lemma 3.3. Thus L is sympathetic. By the hypothesis of this theorem, we have L is sympathetic. Set K be the nondegenerate Killing form of L. Then it restriction to L0, K∣L0×L0 is the Killing form of L0, which is nondegenerate. Thereby L0 is a semisimple Lie algebra and Der(L)=ad(L) by [5]. By virtue of L0 has no nonzero solvable ideal and C(L0)={0}, L0 is sympathetic.
Definition 3.5. [4] Let L be a Lie superalgebra and h(L)=L⊕Der(L). Define the bracket in h(L) by [x+D,y+E]=[x,y]+Dy(−1)αβEx+[D,E], where E∈(Der(L))α, x∈Lβ. Then h(L) is a Lie superalgebra. We call h(L) a holomorph Lie superalgebra.
Remark 1. [4]
(1) If C(L)={0}, then C(Der(L))={0}, since C(Der(L))⊆Der(L) and C(Der(L))⊆CDer(L)(ad(L))={0}.
(2) L◃h(L) and h(L)/g≃Der(L).
(3) L∩Ch(L)(L)=C(L).
Theorem 3.6. Let L be a perfect Lie superalgebra. The following conditions are equivalent:
(1) L is a sympathetic Lie superalgebra.
(2) Any splitting extension e by L is a trivial extension and e=L⊕Ce(L).
(3) h(L)=L⊕Ch(L)(L).
Proof. (1) ⇒ (2) Let e be a splitting extension by L. Then L◃e and Ce(L)◃e. By virtue of C(L)={0} by (1), L∩Ce(L)={0}. Let x∈e. Since L◃e, adx∈L. Therefore the restriction adx|L on L is a derivation of L. But L is sympathetic, thus adx|L is an inner derivation of L. We define θ by θ(x)=adx|L for ∀x∈e. Since Der(L)=ad(L)≃L, the map θ is a homomorphism from e onto Der(L) with kernal Ce(L). Hence dime=dimL+dimCe(L).
(2) ⇒ (3) It is clear by setting e=h(L).
(3) ⇒ (1) According to Remark 1 C(L)=L∩Ch(L)(L)={0}. Ch(L)(L)≃h(L)/L by (3) and h(L)/L≃Der(L)≃Ch(L)(L)≃L from Remark 1. By virtue of C(L)={0}, L≃ad(L). Hence Der(L)≃ad(L).
4.
Decomposition of perfect Lie superalgebras
Definition 4.1. A Lie superalgebra L is called irreducible if it has no non-zero direct factor.
Lemma 4.2. Let L be a Lie superalgebra. Then L=I1⊕⋯⊕In, with every Ii(i=1,…,n) is an irreducible ideal of L.
Proof. Induction on the dimension of L. If dimL=0 or 1, then the lemma is true. It follows that we suppose that the lemma is true if dimL<n. Set L be a Lie superalgebra with dimL=n. If L is reducible, then there exist two proper ideals I,J of L such that L=I⊕J.
According to inductive hypothesis, we can get I=I1⊕⋯⊕In with every Ii is an irreducible graded ideal of I and J=J1⊕⋯⊕Jm with every Ji is an irreducible graded ideal of J, consequently, it's sufficient to see that every graded ideal of I(resp. J) is a graded ideal of L.
Lemma 4.3. Suppose that L is a Lie superalgebra, I is a perfect graded ideal of L. If L=I1⊕⋯⊕In with every Ii is a graded ideal of L, then
Proof. By the assumption we have
then we can obtain
hence the lemma holds.
Theorem 4.4. Suppose that L is a perfect Lie superalgebra. Then L=s1⊕⋯⊕sn where every si is an irreducible perfect graded ideal of L. Moreover, this decomposition is unique.
Proof. According to Lemmas 3.2 and 4.2, we know that the decomposition exists. Let
be two decompositions of L into irreducible perfect graded ideals. By Lemma 4.3, we have
where i∈{1,⋯,n}. Hence there exists α(i)∈{1,⋯,m} such that si=si∩s′α(i) since si is irreducible, we can obtain si⊆s′α(i). Again according to Lemma 4.3, we have
there exists a j∈{1,⋯,n} such that
thus s′α(i)⊆sj, which implies that
Therefore i=j and si=s′α(i). It follows that n=m and there exists a permutation α of {1,⋯,n} such that si=s′α(i) for any i∈{1,⋯,n}, hence the decomposition is unique.
Corollary 5. Suppose that L is a perfect Lie superalgebra and I is a proper ideal of L. Set L=s1⊕⋯⊕sn be the unique decomposition of L into irreducible perfect graded ideals, then the following statements hold
(1) I is an irreducible direct factor of L if and only if there exists an i∈{1,⋯,n} such that I=si.
(2) I is a direct factor of L if and only if there exist {i1,⋯,im}⊆{1,⋯,n} such that I=si1⊕⋯⊕sim.
Proof. (1) Set I be an irreducible direct factor of L, there exists an i∈{1,⋯,n} such that I=I∩si by Lemmas 3.2 and 4.3, so that I⊆si. Since I is a direct factor of L, there exists a graded ideal J of L such that L=I⊕J and by Lemma 4.3, si=si∩I. According to Lemma 4.3, we have
On account of si is irreducible, then si=si∩I or si=si∩J. If si=si∩J, then I⊆si⊆J, contradiction. Consequently, si⊆I and I=si.
(2) Assume that I is a direct factor of L. Then according to Lemma 3.2, we have I is perfect. Therefore, there exist I1,⋯,Im irreducible prefect graded ideals of I such that I=I1⊕⋯⊕Im by Theorem 4.4. Since I is a direct factor of L, it is easy to see that every Ij is an irreducible direct factor of L. Apply (1), there exists an ij∈{1,⋯,n} such that Ij=sij for every Ij. Hence, I=si1⊕⋯⊕sim.
Proposition 5. Let L be a perfect Lie superalgebra such that Der(L)=ad(L) (resp. C(L)={0}). Then
where each si is an irreducible perfect graded ideal of L such that
Proposition 6. Let L be a sympathetic Lie superalgebra. Then
where each si is an irreducible sympathetic graded ideal of L. Moreover, the decomposition is unique.
According to Lemma 3.2, Corollary 5 and Proposition 5, it is straightforward to show the following corollary.
Corollary 6. Let L be a sympathetic Lie superalgebra.
(1) L has a finite number of irreducible sympathetic graded ideals s1,⋯,sn and L=s1⊕⋯⊕sn.
(2) A graded ideal I of L is sympathetic if and only if I=si1⊕⋯⊕sim with {i1,⋯,im}⊆{1,⋯,n}.
5.
The sympathetic decomposition of Lie superalgebras
Lemma 5.1. Let L be a Lie superalgebra. If I and J are two graded ideals satisfying the following conditions
(1) I is a sympathetic graded ideal of L,
(2) there exists a vector subspace W of L such that L=J⊕W and [I,W]⊆W, then I∩J is a sympathetic graded ideal of L.
Proof. On account of I is sympathetic, we have I=[I,L], then
Since [[I,W],I]⊆W, therefore [I,W] is an ideal of I, thus [I,J] is a direct factor of I. According to Lemma 3.2, it is easy to see that [I,J] and [I,W] are sympathetic. Thus
Considering the canonical surjection ϕ:I→I/[I,J]. Set x∈I∩J, y∈I,
it follows that
we have x∈[I,J]. Hence, I∩J=[I,J], i.e., I∩J is a sympathetic graded ideal of L.
Corollary 7. Suppose that L is a Lie superalgebra, I and J are two graded ideals of L.
(1) If I is a sympathetic graded ideal of L and J is a direct factor of L, then I∩J is a sympathetic graded ideal of L.
(2) If I and J are sympathetic graded ideals of L, then I∩J and I+J are sympathetic graded ideals of L.
Proof. (1) The assertion is obvious by Lemma 5.1.
(2) According to Corollary 1, we can get J is a direct factor of L, thus I∩J is a sympathetic graded ideal of L. The following sequence is exact:
defined by φ(x)=x for any x∈I and φ(xI+xJ)=T(xJ) for any xI∈I, xJ∈J where T:J→J/I∩J is the canonical surjection. Thus I+J is an extension of sympathetic Lie superalgebras. According to Corollary 4, we obtain I+J is a sympathetic graded ideal of L.
Proposition 7. Each Lie superalgebra contains a maximal sympathetic graded ideal.
Proof. Using dimensions to prove. Suppose that L is a Lie superalgebra.
is a non-empty subset of N bounded above by dimL, then it has a maximal element m, consequently there exists a sympathetic graded ideal M of L such that dimM=m. Set I be a sympathetic graded ideal of L, thus M+I is a sympathetic graded ideal of L by Corollary 7. Hence M=M+I, thereby I⊆M, which implies that M is the maximal graded ideal of L, the proposition is certified.
Lemma 5.2. Suppose that L is a Lie superalgebra, set I be a graded ideal of L. If there exists a graded ideal J of L satisfies the following conditions
(1) L=I⊕J,
(2) C(J)={0} and [J,J]=J,
then I is a characteristic ideal of L.
Proof. Set x∈I and d∈Der(L). Thus
where y∈I, c∈J. Set c1∈J. Then
since J is a characteristic ideal by Lemma 2.4, it follows that c∈C(J)=0, hence d(x)∈I, which implies that d(I)⊆I.
Theorem 5.3. Let L be a Lie superalgebra and M be a maximal sympathetic graded ideal of L. Then there exists a characteristic ideal I of L satisfying
(1) L=M⊕I,
(2) I is the greatest among the ideals of L which are direct factors of L and do not contain any non-zero sympathetic graded ideal.
Proof. (1) According to Corollary 1, M is a direct factor of L. So there exists an ideal I of L such that L=M⊕I. Set x∈I and d∈Der(L). Thus
where y∈I, z∈M. Let z1∈M. Then
therefore z∈C(M)=0, hence d(x)∈I, then we obtain d(I)⊆I. It follows that I is a characteristic ideal of L.
Let S be a sympathetic graded ideal of I, then S is also a sympathetic graded ideal of L, consequently S⊆I∩M={0}.
(2) Let J be a graded ideal of L which is a direct factor of L and doesn't contain any non-zero sympathetic graded ideal, then there exists a graded ideal Q of L such that L=J⊕Q.
Set M′ be the maximal sympathetic graded ideal of Q, then there exists a graded ideal I′ of Q such that Q=M′⊕I′ and I′ is a characteristic ideal of Q which doesn't have any non-zero sympathetic graded ideal.
Therefore, L=M′⊕I′⊕J where M′ is a sympathetic graded ideal of L. I′ and J are graded ideals of L which don't contain any non-zero sympathetic graded ideal.
According to Lemma 4.3, we get
Claim that M∩I′={0}(resp. M∩J={0}). Otherwise, M∩I′(resp. M∩J) is a non-zero sympathetic graded ideal of L by Corollary 7(1). Note that [M∩I′,I′]⊆M∩I′(resp. [M∩J,J]⊆M∩J), we have M∩I′(resp. M∩J) is a non-zero sympathetic graded ideal of I′(resp. J), contradiction. So M∩I′={0}(resp. M∩J={0}), consequently, M=M∩M′, which implies that M=M′.
Let x∈J, then x=y+z with y∈M and z∈I. Note that [M,J−I]={0}, we get
which implies that y∈C(M)=0, consequently x=z, i.e., J⊆I.
Definition 5.4. Let L be a Lie superalgebra. We call the decomposition of Theorem 5.3 is a sympathetic decomposition, i.e., the decomposition L=M⊕I where M is the maximal sympathetic graded ideal of L and I is the greatest among the graded ideals of L. The I is a direct factor of L and it doesn't contain any non-zero sympathetic graded ideal.
Proposition 8. Suppose that L is a Lie superalgebra, M is a maximal sympathetic graded ideal of L, L=M⊕I is a sympathetic decomposition. If I={0}, then Rad(I)={0}.
Proposition 9. Let L be a Lie superalgebra, M be a maximal sympathetic graded ideal of L, L=M⊕I is a sympathetic decomposition. For I={0} if and only if L is a sympathetic Lie superalgebra.
Proof. The necessity is obvious. Then we prove the sufficiency. By Lemma 3.2, I is a sympathetic Lie superalgebra, thus I={0}.
Lemma 5.5. Let the Lie superalgebra L be decomposed into the direct sum of two graded ideals. i.e., L=I⊕J. Then we have the following conclusions
(1) C(L) has the decomposition C(L)=C(I)⊕C(J).
(2) If C(L)={0}, then ad(L)=ad(I)⊕ad(J) and Der(L)=Der(I)⊕Der(J).
(3) If L, I and J are perfect, then L is sympathetic if and only if I and J are sympathetic.
Proof. (1) According to Proposition 4, C(I) and C(J) are ideals of L. C(I)∩C(J)={0}, since I∩J={0}. Set
with a∈C(I) and b∈C(J). Then [a,I]=0 and [b,J]=0. Set
with c∈I and d∈J. Then
by virtue of [b,c] and [a,d] belong to I∩J. Thus a+b∈C(L) and C(I)⊕C(J)⊆C(L). Set m=x+y∈C(L) with x∈I and y∈J. Then
and
by virtue of m∈C(L) and [y,I]⊆[J,I]=0. Thereby x∈C(I). Similarly, y∈C(J). Hence C(L)⊆C(I)⊕C(J).
(2) For D∈Der(I), extend it to a linear transformation on L by setting D(a+b)=D(a), for a∈I and b∈J. It is easy to see that D∈Der(L) and Der(I)⊆Der(L). Similarly, Der(J)⊆Der(L). Let a∈I, b∈J and D∈Der(L). Then
By virtue of I∩J={0}, thus [D(a),b]=[a,D(b)]=0. Assume D(a)=a′+b′, where a′∈I, then b′∈J. Thus [D(a),b]=[a′,b]+[b′,b]=0 for every b∈J and b′∈C(J). By the hypothesis C(L)=C(I)⊕C(J)={0}, we have b′=0. Hence D(a)=a′∈I. Consequently, D(I)⊆I. Similarly, D(J)⊆J.
Set D∈Der(L) and a+b∈I⊕J with a∈I and b∈J. Define W and V by
Thus W∈Der(I) and V∈Der(J). Hence D=W+V∈Der(I)+Der(J). On occount of Der(I)∩Der(J)={0}, Der(L)=Der(I)⊕Der(J) as a vector space. Set D∈(Der(I)), W∈(Der(L)) and b∈J. Then
Which implies that Der(I)◃Der(L). Similarly, we can obtain Der(J)◃Der(L).
(3) First prove the necessity, by virtue of L is sympathetic, C(L)={0}. And from (1), we have C(I)=C(J)={0}. Since ad(L)=Der(L), ad(L)=ad(L)⊕ad(J) and Der(L)=Der(I)⊕Der(J) by (2). Since ad(I)⊂Der(I) and ad(J)⊂Der(J), then ad(I)=Der(I) and ad(J)=Der(J). Hence I and J are sympathetic Lie superalgebras.
On the contrary, according to C(L)=C(I)⊕C(J)={0}. Then Der(L)=Der(I)⊕Der(J)=ad(I)⊕ad(J)=ad(L) by the (2) of this Lemma.
Definition 5.6. A sympathetic Lie superalgebra L is called a simply sympathetic Lie superalgebra if any non-trivial graded ideal of L is not sympathetic.
Example 2. A simple and sympathetic Lie superalgebra is a simply sympathetic Lie superalgebra.
Lemma 5.7. (1) Any sympathetic Lie superalgebra can be decomposed into the direct sum of simply sympathetic ideals.
(2) A sympathetic Lie superalgebra is simply sympathetic if and only if it is indecomposable.
Proof. (1) If L is simply sympathetic, the result is true. If L is not simply sympathetic, there exists a nonzero minimal sympathetic graded ideal I of L such that L=I⊕CL(I) by Lemma 3.1. Continuing this process to CL(I), we get the decomposition of L into simply sympathetic graded ideals since a graded ideal of CL(I) is also a graded ideal of L. (2) The conclusion follows from (1).
6.
Sympathetic Lie superalgebra L/I and Der(L)
Theorem 6.1. Let L be a Lie superalgebra and I be a graded ideal of L such that C(L/I)={0}. The following conclusions hold
(1) C(L)⊆I.
(2) If L=s1⊕⋯⊕sn where every si is a graded ideal of L, then I=I∩s1⊕⋯I∩sn.
Proof. Considering the canonical surjection ψ:L→L/I.
(1) Let x∈C(L) and y∈L,
then ψ(x)∈C(L/I)={0}, it follows that x∈I. Consequently, C(L)⊆I.
(2) Let x∈I and y∈L,
where xi,yi are elements of si for all i∈{1,⋯,n}, then [ψ(xi),ψ(y)]=[ψ(x),ψ(yi)]=0, conclude that ψ(xi)∈C(L/I)={0}, thus, xi∈si∩I, consequently,
Lemma 6.2. Let L be a Lie superalgebra. Suppose that I and J are graded ideals of L such that L=I⊕J and I′(resp. J′) is a graded ideal of I(resp. J). Then the following conditions are equivalent
(1) L/(I′⊕J′) is a sympathetic Lie superalgebra;
(2) I/I′ and J/J′ are sympathetic Lie superalgebras.
Proof. By virtue of I and J are direct factors of L, I′ and J′ are graded ideals of L, we have I′⊕J′ is a graded ideal of L and L/(I′⊕J′) is isomorphic to I/I′⊕J/J′. According to Lemma 3.2 and Proposition 3, we can obtain the conditions (1) and (2) are equivalent.
Proposition 10. Let L be a Lie superalgebra. Suppose that I and J are graded ideals of L such that L=I⊕J and S is a graded ideal of L. Then the following conditions are equivalent
(1) L/S is a sympathetic Lie superalgebra;
(2) I/I∩S and J/J∩S are sympathetic Lie superalgebras.
Proof. (1) ⇒ (2) By Theorem 6.1, S=S∩I⊕S∩J, and by Lemma 6.2(2), we have I/I∩S and J/J∩S are sympathetic Lie superalgebras.
(2) ⇒ (1) We claim that S=S∩I⊕S∩J. Set x∈S, then x=y+z with y∈I and z∈J. Considering the canonical surjections
Let y′∈I, then
we get that [y,y′]∈S∩I, therefore
Therefore
then y∈I∩S. Similarly, we have z∈J∩S. It implies that S=S∩I⊕S∩J. Consequently, L/S is a sympathetic Lie superalgebra by Lemma 6.2.
Proposition 11. Let L be a Lie superalgebra. Suppose that I and J are graded ideals of L such that L=I⊕J and S a graded ideal of L. Then the following conditions are equivalent
(1) S is the smallest among the graded ideals A of L for which L/A is a sympathetic Lie superalgebra;
(2) S∩I(resp. S∩J) is the smallest among the graded ideals A of I(resp. J) for which I/A(resp. J/A) is a sympathetic Lie superalgebra.
Proof. (1) ⇒ (2) According to Proposition 10, I/I∩S and J/J∩S are sympathetic Lie superalgebras. Set I′(resp. J′) be a graded ideal of I(resp. J) such that I′⊆I∩S(resp. J′⊆J∩S) and I/I′(resp. J/J′) is a sympathetic Lie superalgebra.
According to Lemma 6.2, we get that L/(I′⊕J′) is a sympathetic Lie superalgebra, consequently, S=I′⊕J′. Then
Similarly, J′=J∩H.
(2) ⇒ (1) Set S′ be a graded ideal of L such that S′⊆S and L/S′ is a sympathetic Lie superalgebra. According to Proposition 10, I/I∩S′ and J/J∩S′ are sympathetic Lie superalgebras. By the assumption of this proposition we have I∩S′=I∩S and J∩S′=J∩S. Then we obtain S=S∩I⊕S∩J=S′∩I⊕S′∩J=S′.
Definition 6.3. Let L be a Lie superalgebra. An ascending central series of L is a family graded ideals {Ai(L)}i≥0 of L satisfying
Lemma 6.4. Let L be a Lie superalgebra, {Ai(L)}i≥0 the ascending central series of L and A∞(L) the union of the ascending central series of L. Then A∞(L) is the smallest among the graded ideals I of L such that C(L/I)={0}.
Proof. There exists a positive integer n such that A∞(L)=An(L). Then we have
Set J be a graded ideal of L such that C(L/J)={0}. By Lemma 6.1, C(L)⊆J, i.e., A1(L)⊆J. Suppose that Ai(L)⊆J, and show that Ai+1(L)⊆J.
Set H=J/Ai(L) be a graded ideal of G=L/Ai(L), and G/H is isomorphic to L/J, which implies that C(G/H)={0}. By Lemma 6.1, we have that C(G)⊆H.
Considering the canonical surjection f:L→G, then J=f−1(H) contains Ai+1(L)=f−1(C(G)). It follows that Ai(L)⊆J for all positive integers i, hence A∞(L)⊆J.
Theorem 6.5. Let L be a perfect Lie superalgebra with trivial centre and ad(L) be a characteristic ideal of Der(L). Then Der(L) is sympathetic. If, in addition, L is indecomposable, then Der(L) is simply sympathetic.
Proof. By virtue of C(L)={0}, we have L≃ad(L). Let f=Der(L), then L◃f. Set e be a splitting extension by f, we have f◃e. Therefore we can get ade∈Der(f) with ∀e∈e. On account of L is a characteristic ideal of f, there exists an element f∈f such that adf|L=ade|L. Then ad(f−e)|L=0. Thus f−e∈Cf(L). It follows that we obtain
But f∩Ce(L)=Cf(L)=0 and f◃e. Thereby
Immediately, we have
which implies that e=f⊕Ce(f). Consequently, f is a sympathetic Lie superalgebra by Theorem 3.6.
Assume that Der(L) is not simply sympathetic. Then there exists a simply sympathetic ideal I. And there exists a graded ideal J such that f=I⊕J by Lemma 3.1. For x,y∈L, there exist x1,y1∈I,x2,y2∈J such that x=xl+x2 and y=y1+y2. Hence
where [x1,y]∈I∩L and [x2,y]∈J∩L. Thus
Since L is indecomposable, then I∩L={0} (or J∩L={0}). Therefore L⊂J∩L and L⊂J. It follows that I⊂Cf(L)={0}. According to Lemma 5.7, we obtain f is indecomposable and Der(L) is simply sympathetic. The theorem holds.
Acknowledgments
The authors would like to thank the referee for valuable comments and suggestions on this article.