• 検索結果がありません。

with integrable screen distribution

N/A
N/A
Protected

Academic year: 2022

シェア "with integrable screen distribution"

Copied!
7
0
0

読み込み中.... (全文を見る)

全文

(1)

2

with integrable screen distribution

Makoto Sakaki

Abstract. We give the necessary and sufficient condition for a lightlike hypersurface inR42 with integrable screen distribution to be minimal. Us- ing the condition we can get many minimal lightlike hypersurfaces inR42 which are not totally geodesic.

M.S.C. 2000: 53C42, 53C50.

Key words: lightlike hypersurface, minimal, screen distribution, semi-Euclidean space.

1 Introduction

LetM be a submanifold in a semi-Riemannian manifold ( ¯M ,g). If the induced metric¯ g= ¯g|M is non-degenerate, then (M, g) becomes a semi-Riemannian manifold and it can be studied as a semi-Riemannian submanifold. Wheng is degenerate, (M, g) is called a lightlike submanifold, and many different situations appear (cf. [2]). In this case, the tangent bundleT Mand the normal bundleT Mhave a non-trivial intersec- tion, which is called the radical distribution and denoted by Rad(T M). Then we may choose a (non-unique) semi-Riemannian complementary distribution of Rad(T M) in T M, which is called the screen distribution and denoted byS(T M).

In particular, in the case of lightlike hypersurfaces, the normal bundleT M coin- cides with the radical distribution Rad(T M), and there exists a canonical transversal vector bundle tr(T M) corresponding to the screen distributionS(T M) which is called the lightlike transversal vector bundle.

Recently, Bejan and Duggal [1] introduced the notion of minimal lightlike sub- manifolds. In the proof of Theorem 3.2 of [1], they implicitly show that a lightlike hypersurfaceM in a semi-Riemannian manifold ( ¯M ,g) with integrable screen distri-¯ bution S(T M) is minimal if and only if the radical distribution Rad(T M) contains the mean curvature vector field of any leaf ofS(T M). But this statement is a general one, and not easy to use to give some examples of minimal lightlike hypersurfaces.

In this paper, we discuss minimal lightlike hypersurfaces in the 4-dimensional semi-Euclidean space R42 of index 2. We give the necessary and sufficient condition

Balkan Journal of Geometry and Its Applications, Vol.14, No.1, 2009, pp. 84-90.

c

°Balkan Society of Geometers, Geometry Balkan Press 2009.

(2)

for a lightlike hypersurface inR42 with integrable screen distribution to be minimal, as follows:

Theorem 1.1. Let (M, g, S(T M)) be a lightlike hypersurface in R42 with integrable screen distributionS(T M). ThenM is minimal if and only if the eigenvalues of the shape operator of any leaf ofS(T M)in the direction of the lightlike transversal vector bundle are both zero.

The necessary and sufficient condition in Theorem 1.1 seems stronger than that from [1, Th.3.2], but they are equivalent in our case, as we will see in Section 3. And in Section 4, using the discussion in the proof of Theorem 1.1, we give a class of minimal lightlike hypersurfaces inR42 which are not totally geodesic.

2 Preliminaries

In this section, following [2] and [1], we recall some basic facts on lightlike hypersur- faces.

Let ¯M be a semi-Riemannian manifold with metric ¯gand Levi-Civita connection

∇. Let¯ M be a lightlike hypersurface in ¯M, that is, the induce metric g = ¯g|M is degenerate. In the case of lightlike hypersurfaces, the normal bundleT M coincides with the radical distribution Rad(T M), defined by

Rad(TxM) ={ξ∈TxM|g(ξ, X) = 0, X∈TxM},

where dim(Rad(TxM)) = 1. There exists a screen distribution S(T M) which is a semi-Riemannian complementary distribution of Rad(T M) inT M, that is,

T M =S(T M)Rad(T M) =S(T M)⊥T M.

We note that if ¯M is of indexq, thenS(T M) is of indexq−1. IfS(T M) is integrable, thenM is locally a productL×d, wheredis a null geodesic in ¯M as an integral curve of Rad(T M) andLis a semi-Riemannian submanifold in ¯M as a leaf ofS(T M).

From [2, p.79], we know that for a screen distributionS(T M), there exists a unique vector bundle tr(T M) of rank 1 such that, for any non-zero local sectionξ ofT M onU there is a unique sectionN of tr(T M)|U satisfying

¯

g(ξ, N) = 1, ¯g(N, N) = ¯g(N, W) = 0

for allW Γ(S(T M)|U). This vector bundle tr(T M) is called the lightlike transversal vector bundle with respect toS(T M), and we have the decomposition

TM¯|M =T M tr(T M).

From now on,ξ denotes a non-zero local section of Rad(T M). According to the above decomposition, we have the Gauss formula

¯XY =XY +B(X, Y)N,

(3)

whereX, Y Γ(T M). Thenis a torsion-free linear connnection onM, and B is a symmetricC(M)-bilinear form on Γ(T M). This formB is called the local second fundamental form ofM, which is independent of the choice ofS(T M). WhenB= 0, M is called totally geodesic.

Following the Definition 2 of [1] in the case of lightlike hypersurfaces,M is called minimal if trace(B) = 0, where the trace is written with respect to g restricted to S(T M). This condition is independent of the choice ofS(T M) andξ.

3 Proof of Theorem 1.1

Proof of Theorem 1.1. Let (M, g, S(T M)) be a lightlike hypersurface inR42 with in- tegrable screen distributionS(T M). ThenM is locally a productL×d, where dis an open set of a lightlike line in R42 as an integral curve of Rad(T M) and L is a Lorentzian surface inR42as a leaf ofS(T M).

LetLbe an arbitrary leaf ofS(T M), andf :L→R42be the inclusion map. Along L, we choose a local frame field {e1, e2} so that {fe1, fe2} is orthonormal with signature (+,−), and a local normal orthonormal frame field{e3, e4}with signature (+,−). Then we may assume that the inclusion mapF :M R42 is given by

F(p, t) =f(p) +t(e3(p) +e4(p)), p∈L, t∈(−ε, ε).

We shall use the following ranges of indices:

1≤A, B, ...≤4, 1≤i, j, ...≤2, 3≤α, β, ...≤4.

LetωBA be the connection forms which satisfy

d(fei) = X2

j=1

ωijfej+ X4

α=3

ωαieα, deα= X2

i=1

ωαifei+ X4

β=3

ωαβeβ.

We note that, in our situation,ωBA =−ωBAif|A−B|is even, andωBA =ωAB if|A−B|

is odd. Then

de3=ω13fe1+ω32fe2+ω43e4=−ω31fe1+ω23fe2+ω34e4, de4=ω41fe1+ω42fe2+ω34e3=ω41fe1−ω24fe2+ω43e3.

Lethαij denote the components of the second fundamental formhofL, so that

ωiα= X2

j=1

hαijωj,

where1, ω2}is the coframe field dual to{e1, e2}.

Set

˜

ei(p, t) = (ei(p),0)∈T(p,t)M =TpL×Ttd.

Then{˜e1,e˜2, ∂t}is a natural frame field onM =L×d, and we obtain

(4)

Fe˜1= (1−tA11)fe1+tA12fe2+43(e1)(e3+e4), Fe˜2=−tA12fe1+ (1 +tA22)fe2+43(e2)(e3+e4),

Ft=e3+e4=:ξ, where we set

Aij =h3ij−h4ij. As the induced metricgis given by

g(X, Y) =hFX, FYi, X, Y Γ(T M), we have the components ofg by

g(˜e1,˜e1) = 12tA11+t2(A211−A212), g(˜e2,e˜2) =−1−2tA22+t2(A212−A222), g(˜e1,˜e2) =−2tA12+t2A12(A11−A22),

g(˜e1, ∂t) =g(˜e2, ∂t) =g(∂t, ∂t) = 0.

Thus, for sufficiently smallt, M is a lightlike hypersurface, andξis a non-zero local section of the radical distribution Rad(T M).

We choose the screen distribution S(T M) so that it is spanned by {Fe˜1, Fe˜2}.

Then, for eacht, the mapFt:L→R42 defined byFt(p) =F(p, t) becomes an inclu- sion map of a leaf ofS(T M). Let tr(T M) be the lightlike transversal vector bundle corresponding toS(T M), and N be the local section of tr(T M) which corresponds toξas in Section 2. ThenAij are the components of the second fundamental formh ofLin the directionN.

With respect to the local second fundamental formB, we may obtain B(˜e1,e˜1) =hD˜e1Fe˜1, ξi=A11−t(A211−A212), B(˜e1,e˜2) =hD˜e2Fe˜1, ξi=A12−tA12(A11−A22),

B(˜e2,e˜2) =hD˜e2Fe˜2, ξi=A22−t(A212−A222),

whereD is the induced connection from the flat connection onR42. By the definition, M is minimal if and only if trace(B) = 0 where the trace is written with respect tog restricted toS(T M), which is now equivalent to that

(3.1) g(˜e2,e˜2)B(˜e1,˜e1)2g(˜e1,˜e2)B(˜e1,e˜2) +g(˜e1,˜e1)B(˜e2,˜e2) = 0.

It is a cubic identity fort, and is equivalent to that (3.2) A11=A22, A211=A212.

Let us consider

Aij = (h3)ij(h4)ij,

which are the components of the shape operator ofLin the directionN. Noting that A11=A11, A21=A21=A12, A12=−A12, A22=−A22,

(5)

we can see that the condition (3.2) is equivalent to that the trace and the determinant of (Aij) are both zero, which is also equivalent to that the eigenvalues of (Aij) are both zero. Thus we have proved the theorem. 2

Let M be a lightlike hypersurface in a semi-Riemannian manifold ( ¯M ,¯g) with integrableS(T M). As noted in the introduction, the proof of [1, Th.3.2] implies that M is minimal if and only if (∗) ”Rad(T M) contains the mean curvature vector field of any leaf ofS(T M)”. So the necessary and sufficient condition in Theorem 1.1 seems stronger than the above condition (∗), but they are equivalent in our case, as we have shown. We note that the condition (∗) corresponds to the equation (3.1), which is equivalent to (3.2) for an arbitrary leaf. Namely, for a lightlike hypersurfaceM inR42 with integrableS(T M), if Rad(T M) contains the mean curvature vector field of any leaf ofS(T M), then the leaves must satisfy another condition.

4 A class of minimal lightlike hypersurfaces

In this section, using the discussion in Section 3, we give a class of minimal lightlike hypersurfaces inR42which are not totally geodesic. In fact, we give a class of Lorentzian surfaces inR42 which satisfy the condition (3.2).

Let{x1, x2, x3, x4} be the standard coordinate system forR42with metric ds2=dx21−dx22+dx23−dx24.

Proposition 4.1. LetQ1(z), Q2(z), Q3(z)andQ4(z)be smooth functions. Set

f(u, v) =



Q1(u+v) +Q2(u−v) Q1(u+v)−Q2(u−v) Q3(u+v) +Q4(u−v) Q3(u+v)−Q4(u−v)



,

and assume that

Q01(u+v)Q02(u−v) +Q03(u+v)Q04(u−v)>0.

Thenf gives a Lorentzian surface inR42 which satisfies the condition (3.2).

Proof. First we have

fu=



Q01(u+v) +Q02(u−v) Q01(u+v)−Q02(u−v) Q03(u+v) +Q04(u−v) Q03(u+v)−Q04(u−v)



, fv=



Q01(u+v)−Q02(u−v) Q01(u+v) +Q02(u−v) Q03(u+v)−Q04(u−v) Q03(u+v) +Q04(u−v)



,

and

(6)

hfu, fui= 4{Q01(u+v)Q02(u−v) +Q03(u+v)Q04(u−v)}=:E >0, hfu, fvi= 0, hfv, fvi=−E.

Sof gives a Lorentzian surface in R42. Set

e1= 1

√E

∂u, e2= 1

√E

∂v.

Then{e1, e2}is an orthonormal frame field with signature (+,−). Set

e3= 1

√E



−Q03(u+v)−Q04(u−v) Q03(u+v)−Q04(u−v) Q01(u+v) +Q02(u−v)

−Q01(u+v) +Q02(u−v)



, e4= 1

√E



−Q03(u+v) +Q04(u−v) Q03(u+v) +Q04(u−v) Q01(u+v)−Q02(u−v)

−Q01(u+v)−Q02(u−v)



.

Then{e3, e4}is a normal orthonormal frame field with signature (+,−).

Lethαij denote the components of the second fundamental form off with respect to these frames. Then we can get

h311= 1

Ehfuu, e3i

= 2E−3/2{Q01(u+v)Q003(u+v)−Q001(u+v)Q03(u+v)+Q02(u−v)Q004(u−v)−Q002(u−v)Q04(u−v)}

=h322=h412, and

h312=h411=h422

= 2E−3/2{Q01(u+v)Q003(u+v)−Q001(u+v)Q03(u+v)−Q02(u−v)Q004(u−v)+Q002(u−v)Q04(u−v)}.

Thus the condition (3.2) is satisfied, and we have proved the proposition. 2 Remark 1 This construction is inspired by the previous paper [5] and the struc- ture of complex curves inR4=C2.

By the discussion in the proof of Theorem 1.1, we get the following result Theorem 4.2. Let f, e3, e4 be as in Proposition 4.1. Then the map

F(u, v, t) =f(u, v) +t(e3+e4)

gives a minimal lightlike hypersurface inR42, which is not totally geodesic if Q02(u−v)Q004(u−v)−Q002(u−v)Q04(u−v)6= 0.

By Proposition 4.1 and Theorem 4.2, we can get many minimal lightlike hyper- surfaces inR42 which are not totally geodesic. For example, when

Q1(z) =Q2(z) =z, Q3(z) =Q4(z) =ez, we have

E= 4{Q01(u+v)Q02(u−v) +Q03(u+v)Q04(u−v)}= 4(1 +e2u)>0, and

Q02(u−v)Q004(u−v)−Q002(u−v)Q04(u−v) =eu−v6= 0.

We point out, that related information to this subject can be found in [3] and [4].

(7)

References

[1] C. L. Bejan and K. L. Duggal,Global lightlike manifolds and harmonicity, Kodai Math. J. 28 (2005), 131-145.

[2] K. L. Duggal and A. Bejancu,Lightlike Submanifolds of Semi-Riemannian Man- ifolds and Applications, Kluwer Academic Publishers, 1996.

[3] F. Massamba,Lightlike hypersurfaces of indefinite Sasakian manifolds with par- allel symmetric bilinear forms, Differ. Geom. Dyn. Syst. 10 (2008), 226-234.

[4] B. Sahin,Slant lightlike submanifolds of indefinite Hermitian manifolds, Balkan Jour. Geom. Appl. 13, 1 (2008), 107-119.

[5] M. Sakaki, Two classes of Lorentzian stationary surfaces in semi-Riemannian space forms, Nihonkai Math. Jour. 15 (2004), 15-22.

Author’s address:

Makoto Sakaki

Graduate School of Science and Technology, Hirosaki University, Hirosaki 036-8561, Japan.

E-mail address: [email protected]

参照

関連したドキュメント

In section 2, we firstly prove a theorem for this factor groupoid (Theorem 1). Then, our intended theorem is obtained as a speeial case of Theorem 1. In seetion 3, we g,ive a proof

We give an elementary proof of a theorem, which is originally due to Mattila on the vanishing of the analytic capacity of a certain natural subclass of the totally unrectifiable

Weyl’s theorem holds also for paranormal operators on Hilbert spaces, since these operators satisfies property (3) and have SVEP, see [2].. Note that the class of totally

The proof of Lemma 3 is based on the following trivial, but useful statement which was also applied in the proof of [2, Theorem V1 A]..

In particular, the [proof of the] main result does not give an alternative proof of the Neukirch-Uchida theorem.... Mono-anabelian Reconstruction Algorithm (2) (MRA 1 ) What is

Then $\phi$ interchanges the twofixed points if and only $\iota f$ it rotates the period 2 orbit,.. Proof of

The proof of the following lemma is obtained from the proof of [OKY,Lemma 4] and that of [KY,Theorem 1] without the assumption that a is anti-integral over R?.

The proof of Corollary 2.2 is bit misleading, but we apply Theorem 2.1 in the proof with q = 4, and use “higher” q to obtain the