PARALLEL RICCI TENSOR
YOUNG JIN SUH*
Abstract. First we introduce the notion of parallel Ricci tensor for real hypersurfaces in the complex quadricQm=SOm+2/SOmSO2 . Next, according to theA-principal or theA-isotropic unit normalN, we give a complete classification of real hypersurfaces in Qm=SOm+2/SOmSO2 with parallel Ricci tensor.
1. Introduction
When we consider some Hermitian symmetric spaces of rank 2, usually we can give ex- amples of Riemannian symmetric spacesSUm+2/S(U2Um) andSU2,m/S(U2Um), which are said to be complex two-plane Grassmannians and complex hyperbolic two-plane Grass- mannians respectively (see [1], [2], [4] and [13]). Those are said to be Hermitian symmetric spaces and quaternionic K¨ahler symmetric spaces equipped with the K¨ahler structure J and the quaternionic K¨ahler structureJonSU2,m/S(U2Um). The rank ofSU2,m/S(U2Um) is 2 and there are exactly two types of singular tangent vectorsX of G2(Cm+2) which are characterized by the geometric properties J X ∈JX and J X ⊥JX respectively.
As another kind of Hermitian symmetric space with rank 2 of compact type different from the above ones, we can give an example of complex quadricQm =SOm+2/SOmSO2, which is a complex hypersurface in complex projective space CPm+1 (see Berndt and Suh [3], and Smyth [11]). The complex quadric also can be regarded as a kind of real Grassmann manifolds of compact type with rank 2 (see Kobayashi and Nomizu [7]).
Accordingly, the complex quadric admits two important geometric structures, that is, a complex conjugation structureAand a K¨ahler structureJ, which anti-commute with each other, that is, AJ =−J A. Then for m≥2 the triple (Qm, J, g) is a Hermitian symmetric space of compact type with rank 2 and its maximal sectional curvature is equal to 4 (see Klein [5] and Reckziegel [10]).
For the complex projective space CPm+1 and the quaternionic projective spaceQPm+1 some classifications related to parallel Ricci tensor were respectively investigated by Kimura [6], and P´erez and Suh [8]. For the complex 2-plane Grassmannian G2(Cm+2) = SUm+2/S(UmU2) the classification was obtained by Berndt and Suh [1]. By using such a classification P´erez and Suh [9] proved a non-existence property for Hopf hypersurfaces in
* Corresponding author. E-mail address: [email protected]
2010Mathematics Subject Classification: Primary 53C40. Secondary 53C55.
Key words: parallel Ricci tensor, A-isotropic,A-principal, K¨ahler structure, complex conjugation, com- plex quadric.
This work was supported by grant Proj. No. NRF-2012-R1A2A2A-01043023 from National Research Foundation of Korea.
YOUNG JIN SUH*
G2(Cm+2) with parallel and commuting Ricci tensor. Suh [12] extended this result to hy- persurfaces inG2(Cm+2) with parallel Ricci tensor. Moreover, Suh and Woo [15] studied another non-exitence property for Hopf hypersurfaces in complex hyperbolic two-plane Grassmannians SU2,m/S(U2Um) with parallel Ricci tensor.
It is known that the Reeb flow on a real hypersurface in G2(Cm+2) is isometric if and only if M is an open part of a tube around a totally geodesic G2(Cm+1) ⊂ G2(Cm+2) in [2]. Moreover, in [13] we have asserted that the Reeb flow on a real hypersurface in SU2,m/S(U2Um) is isometric if and only if M is an open part of a tube around a totally geodesic SU2,m−1/S(U2Um−1) ⊂ SU2,m/S(U2Um). Here, the Reeb flow on real hypersurfaces in SUm+2/S(UmU2) or SU2,m/S(U2Um) is said to be isometric if the shape operator commutes with the structure tensor. In the paper [3] due to Berndt and Suh, we investigated this problem for the complex quadric Qm =SOm+2/SOmSO2 and obtained the follwoing result:
Let M be a real hypersurface of the complex quadric Qm, m ≥3. The Reeb flow on M is isometric if and only if m is even, saym= 2k, andM is an open part of a tube around a totally geodesic CPk ⊂Q2k.
When we consider a hypersurfaceM in the complex quadricQm, the unit normal vector field N of M in Qm can be divided into two classes that N is A-isotropic or A-principal (See [3], [4], and [14]). In the first class where M has an A-isotropic unit normal N, we have asserted in [3] that M is locally congruent to a tube over a totally geodesic CPk in Q2k. In the second class forN isA-principal we have proved that a contact hypersurface M inQm is locally congruent to a tube over a totally geodesic and totally real submanifold Sm in Qm (see [4]). From such a view point, in this paper let us consider the notion of Ricci parallel for hypersurfaces in Qm , that is, ∇Ric= 0. Then motivated by a theorem due to the A-principal normal vector field [4] for contact hypersurfaces in Qm we assert the following
Theorem 1. There does not exist any Hopf hypersurfaces in the complex quadric Qm with parallel Ricci tensor and A-principal normal vector field.
In section 4 we will give this proof in detail. Now at each pointz ∈M let us consider a maximal A-invariant subspace Qz of TzM, z∈M, defined by
Qz ={X∈TzM |AX ∈TzM for all A∈Az}
ofTzM,z∈M. Thus for a case where the unit normal vector fieldN isA-isotropic it can be easily checked that the orthogonal complementQ⊥z =Cz⊖Qz,z∈M, of the distribution Q in the complex subbundleC, becomesQ⊥z = Span [Aξ, AN]. Here it can be easily checked that the vector fields Aξ and AN belong to the tangent space TzM, z∈M if the unit normal vector field N becomes A-isotropic. Then motivated by the above result, in this paper we give another theorem for real hypersurfaces in the complex quadric Qm with parallel Ricci tensor andA-isotropic unit normal as follows:
Theorem 2. Let M be a Hopf real hypersurface in the complex quadric Qm, m≥4, with parallel Ricci tensor andA-isotropic unit normal N. If the shape operator commutes with the structure tensor on the distributionQ⊥, then M has 3distinct constant principal curvatures, three of which are given by
α = 0, γ = 0, λ =−√
3, and µ=√ 3
or
α=√
2(m−3), γ = 0, λ= 0, and µ=− 2
√2(m−3) with corresponding principal curvature spaces
Tα = [ξ], Tγ = [Aξ, AN], ϕ(Tλ) = Tµ,dim Tλ =dim Tµ=m−2.
Our paper is organized as follows. In section 2 we present basic material about the complex quadric Qm, including its Riemannian curvature tensor and a description of its singular vectors ofQmlikeA-principal orA-isotropic unit normal vector filed. Apart from the complex structureJ there is another distinguished geometric structure onQm, namely a parallel rank two vector bundleAwhich contains anS1-bundle of real structures, that is, complex conjugationsAon the tangent spaces ofQm. This geometric structure determines a maximal A-invariant subbundle Qof the tangent bundle T M of a real hypersurfaceM in Qm. Accordingly, in section 3, we investigate the geometry of this subbundle Q for hypersurfaces in Qm and some equations including Codazzi and fundamental formulas related to the vector fieldsξ, N, Aξ, andAN for the complex conjugation Aof M inQm. Finally, in sections 4 and 5, we present the proof of Theorems 1 and 2, respectively.
In section 4, the first step is to derive the Ricci tensor from the equation of Gauss for real hypersurfacesM inQm and to get some formulas by the assumption of parallel Ricci tensor andA-principal normal vector field, and show that a real hypersurface inQm which is a tube over an m-dimensional unit sphere Sm does not admit a parallel Ricci tensor.
The next step is to show that there do not exist any real hypersurfaces in the complex quadric Qm with parallel Ricci tensor andA-principal normal vector field.
In section 5, we give a complete proof of Theorem 2. The first part of this proof is devoted to give some fundamental formulas from Ricci parallel and A-isotropic unit normal vector field. Then in the latter part of the proof we will use the expression of the shape operator for real hypersurfaces in Qm when the shape operator and the structure tensor commutes on the distribution Q⊥, where Q⊥ denotes an orthogonal complement of the maximalA-invariant subspace Qin the complex subbundle C of TzM,z∈M inQm.
2. The complex quadric
For more details in this section we refer to [3], [4], [5], [7], [10] and [14]. The complex quadric Qm is the complex hypersurface in CPm+1 which is defined by the equationz02+ . . .+zm+12 = 0, wherez0, . . . , zm+1 are homogeneous coordinates onCPm+1. We equipQm with the Riemannian metricgwhich is induced from the Fubini-Study metric ¯gonCPm+1 with constant holomorphic sectional curvature 4. The Fubini-Study metric ¯g is defined by
¯
g(X, Y) = Φ(J X, Y) for any vector fieldsXand Y onCPm+1 and a globally closed (1,1)- form Φ given by Φ =−4i∂∂¯logfj on an open set Uj ={[z0, z1,· · ·, zm+1]∈CPm+1|zj̸=0}, where the function fj denotes fj =∑m+1
k=0tkj¯tkj, and tkj = zzkj for j, k = 0,· · ·, m+ 1. Then naturally the K¨ahler structure onCPm+1 induces canonically a K¨ahler structure (J, g) on the complex quadric Qm.
The complex projective spaceCPm+1is a Hermitian symmetric space of the special uni- tary groupSUm+2, namelyCPm+1 =SUm+2/S(Um+1U1). We denote byo= [0, . . . ,0,1]∈
YOUNG JIN SUH*
CPm+1 the fixed point of the action of the stabilizer S(Um+1U1). The special orthogonal group SOm+2 ⊂SUm+2 acts on CPm+1 with cohomogeneity one. The orbit containing o is a totally geodesic real projective space RPm+1 ⊂ CPm+1. The second singular orbit of this action is the complex quadric Qm = SOm+2/SOmSO2. This homogeneous space model leads to the geometric interpretation of the complex quadricQm as the Grassmann manifold G+2(Rm+2) of oriented 2-planes in Rm+2. It also gives a model of Qm as a Her- mitian symmetric space of rank 2. The complex quadric Q1 is isometric to a sphere S2 with constant curvature, andQ2 is isometric to the Riemannian product of two 2-spheres with constant curvature. For this reason we will assume m ≥3 from now on.
For a nonzero vector z ∈ Cm+2 we denote by [z] the complex span of z, that is, [z] ={λz | λ ∈ C}. Note that by definition [z] is a point in CPm+1. As usual, for each [z] ∈CPm+1 we identify TzCPm+1 with the orthogonal complement Cm+2 ⊖[z] of [z] in Cm+2. For [z]∈ Qm the tangent spaceTzQm can then be identified canonically with the orthogonal complement Cm+2⊖([z]⊕[¯z]) of [z]⊕[¯z] inCm+2 (see Kobayashi and Nomizu [7]). Note that ¯z∈νzQm is a unit normal vector of Qm in CPm+1 at the point [z].
We denote byA¯zthe shape operator ofQm inCPm+1with respect to the unit normal ¯z.
It is defined byA¯zw=∇wz¯=wfor a complex Euclidean connection∇and allw∈TzQm. That is, the shape operatorA¯z is just complex conjugation restricted toTzQm. The shape operator Az¯ is an anti-commuting involution such that A2¯z = I and AJ = −J A on the complex vector spaceTzQm and
TzQm =V(A¯z)⊕J V(Az¯),
where V(Az¯) = Rm+2 ∩TzQm is the (+1)-eigenspace and J V(Az¯) = iRm+2 ∩TzQm is the (−1)-eigenspace of Az¯. That is, Az¯X =X and A¯zJ X =−J X, respectively, for any X∈V(Az¯).
Geometrically this means that the shape operator Az¯ defines a real structure on the complex vector space TzQm, or equivalently, is a complex conjugation on TzQm. Since the real codimension of Qm inCPm+1 is 2, this induces anS1-subbundle A of the endo- morphism bundle End(T Qm) consisting of complex conjugations.
There is a geometric interpretation of these conjugations. The complex quadricQm can be viewed as the complexification of the m-dimensional sphere Sm. Through each point z ∈ Qm there exists a one-parameter family of real forms of Qm which are isometric to the sphere Sm. These real forms are congruent to each other under action of the center SO2 of the isotropy subgroup of SOm+2 at z. The isometric reflection of Qm in such a real formSm is an isometry, and the differential at z of such a reflection is a conjugation on TzQm. In this way the family A of conjugations on TzQm corresponds to the family of real forms Sm of Qm containingz, and the subspaces V(A)⊂TzQm correspond to the tangent spaces TzSm of the real forms Sm of Qm.
The Gauss equation for Qm ⊂CPm+1 implies that the Riemannian curvature tensor ¯R ofQm can be described in terms of the complex structureJ and the complex conjugations A∈A:
R(X, Y¯ )Z = g(Y, Z)X−g(X, Z)Y +g(J Y, Z)J X−g(J X, Z)J Y −2g(J X, Y)J Z +g(AY, Z)AX−g(AX, Z)AY +g(J AY, Z)J AX −g(J AX, Z)J AY.
Note that J and each complex conjugationA anti-commute, that is, AJ =−J Afor each A∈A.
Recall that a nonzero tangent vector W ∈ TzQm is called singular if it is tangent to more than one maximal flat in Qm. There are two types of singular tangent vectors for the complex quadric Qm:
1. If there exists a conjugation A ∈ A such that W ∈ V(A), then W is singular.
Such a singular tangent vector is calledA-principal.
2. If there exist a conjugation A ∈ A and orthonormal vectors X, Y ∈ V(A) such thatW/||W||= (X+J Y)/√
2, thenW is singular. Such a singular tangent vector is calledA-isotropic.
For every unit tangent vector W ∈ TzQm there exist a conjugation A ∈ A and or- thonormal vectorsX, Y ∈V(A) such that
W = cos(t)X+ sin(t)J Y
for some t ∈ [0, π/4]. The singular tangent vectors correspond to the values t = 0 and t=π/4. If 0 < t < π/4 then the unique maximal flat containingW isRX⊕RJ Y. Later we will need the eigenvalues and eigenspaces of the Jacobi operator RW =R(·, W)W for a singular unit tangent vector W.
1. If W is an A-principal singular unit tangent vector with respect to A ∈ A, then the eigenvalues of RW are 0 and 2 and the corresponding eigenspaces are RW ⊕ J(V(A)⊖RW) and (V(A)⊖RW)⊕RJ W, respectively.
2. If W is an A-isotropic singular unit tangent vector with respect to A ∈ A and X, Y ∈ V(A), then the eigenvalues of RW are 0, 1 and 4 and the corresponding eigenspaces are RW ⊕C(J X +Y), TzQm⊖(CX⊕CY) and RJ W, respectively.
3. Some general equations
LetM be a real hypersurface inQmand denote by (ϕ, ξ, η, g) the induced almost contact metric structure. Note that ξ = −J N, where N is a (local) unit normal vector field of M and η the corresponding 1-form defined by η(X) = g(ξ, X) for any tangent vector field X on M. The tangent bundle T M of M splits orthogonally into T M = C ⊕Rξ, whereC = ker(η) is the maximal complex subbundle of T M. The structure tensor field ϕ restricted to C coincides with the complex structureJ restricted toC, and ϕξ= 0.
At each point z ∈ M we define a maximal A-invariant subspace of TzM, z∈M as follows:
Qz ={X ∈TzM |AX ∈TzM for all A∈Az}. Lemma 3.1. (see [14]) For each z ∈M we have
(i) If Nz isA-principal, then Qz =Cz.
(ii) IFNz is notA-principal, there exist a conjugationA∈Aand orthonormal vectors X, Y ∈V(A) such that Nz = cos(t)X+ sin(t)J Y for some t∈ (0, π/4]. Then we have Qz =Cz⊖C(J X+Y).
We now assume thatM is a Hopf hypersurface. Then we have Sξ =αξ
YOUNG JIN SUH*
with the smooth function α =g(Sξ, ξ) on M. When we consider a transform J X of the Kaehler structure J onQm for any vector field X on M inQm, we may put
J X =ϕX +η(X)N
for a unit normal N toM. Then we now consider the Codazzi equation
g((∇XS)Y −(∇YS)X, Z) = η(X)g(ϕY, Z)−η(Y)g(ϕX, Z)−2η(Z)g(ϕX, Y) +g(X, AN)g(AY, Z)−g(Y, AN)g(AX, Z) +g(X, Aξ)g(J AY, Z)−g(Y, Aξ)g(J AX, Z).
(3.1) Putting Z =ξ in (3.1) we get
g((∇XS)Y −(∇YS)X, ξ) = −2g(ϕX, Y)
+g(X, AN)g(Y, Aξ)−g(Y, AN)g(X, Aξ)
−g(X, Aξ)g(J Y, Aξ) +g(Y, Aξ)g(J X, Aξ).
On the other hand, we have
g((∇XS)Y −(∇YS)X, ξ)
= g((∇XS)ξ, Y)−g((∇YS)ξ, X)
= (Xα)η(Y)−(Y α)η(X) +αg((Sϕ+ϕS)X, Y)−2g(SϕSX, Y).
Comparing the previous two equations and putting X =ξ yields
Y α= (ξα)η(Y)−2g(ξ, AN)g(Y, Aξ) + 2g(Y, AN)g(ξ, Aξ).
Reinserting this into the previous equation yields g((∇XS)Y −(∇YS)X, ξ)
= −2g(ξ, AN)g(X, Aξ)η(Y) + 2g(X, AN)g(ξ, Aξ)η(Y) +2g(ξ, AN)g(Y, Aξ)η(X)−2g(Y, AN)g(ξ, Aξ)η(X) +αg((ϕS +Sϕ)X, Y)−2g(SϕSX, Y).
Altogether this implies
0 =2g(SϕSX, Y)−αg((ϕS+Sϕ)X, Y)−2g(ϕX, Y) +g(X, AN)g(Y, Aξ)−g(Y, AN)g(X, Aξ)
−g(X, Aξ)g(J Y, Aξ) +g(Y, Aξ)g(J X, Aξ)
+ 2g(ξ, AN)g(X, Aξ)η(Y)−2g(X, AN)g(ξ, Aξ)η(Y)
−2g(ξ, AN)g(Y, Aξ)η(X) + 2g(Y, AN)g(ξ, Aξ)η(X).
(3.2)
At each point z ∈M we can choose A∈Az such that N = cos(t)Z1+ sin(t)J Z2
for some orthonormal vectors Z1, Z2 ∈ V(A) and 0 ≤ t ≤ π4 (see Proposition 3 in [10]).
Note thatt is a function onM. First of all, since ξ=−J N, we have N = cos(t)Z1+ sin(t)J Z2,
AN = cos(t)Z1−sin(t)J Z2, ξ= sin(t)Z2−cos(t)J Z1, Aξ = sin(t)Z2+ cos(t)J Z1.
(3.3)
This implies g(ξ, AN) = 0 and hence
0 =2g(SϕSX, Y)−αg((ϕS+Sϕ)X, Y)−2g(ϕX, Y) +g(X, AN)g(Y, Aξ)−g(Y, AN)g(X, Aξ)
−g(X, Aξ)g(J Y, Aξ) +g(Y, Aξ)g(J X, Aξ)
−2g(X, AN)g(ξ, Aξ)η(Y) + 2g(Y, AN)g(ξ, Aξ)η(X).
(3.4)
4. Proof of Theorem 1
By the equation of Gauss, the curvature tensorR(X, Y)Z for a real hypersurfaceM in Qm induced from the curvature tensor ¯R of Qm can be described in terms of the complex structure J and the complex conjugation A∈A as follows:
R(X, Y)Z = g(Y, Z)X−g(X, Z)Y +g(ϕY, Z)ϕX−g(ϕX, Z)ϕY −2g(ϕX, Y)ϕZ +g(AY, Z)AX−g(AX, Z)AY +g(J AY, Z)J AX −g(J AX, Z)J AY +g(SY, Z)SX −g(SX, Z)SY
for any X, Y, Z∈TzM, z∈M.
From this, contractingY and Z onM inQm, we have Ric(X) =(2m−1)X−X−ϕ2X−2ϕ2X
−g(AN, N)AX−X+g(AX, N)AN −g(J AN, N)J AX
−X+g(J AX, N)J AN+ (trS)SX −S2X
=(2m−1)X−3η(X)ξ−g(AN, N)AX+g(AX, N)AN
−g(J AN, N)J AX +g(J AX, N)J AN + (trS)SX −S2X,
(4.1)
where we have used the following
∑2m−1
i=1 g(Aei, ei) = T rA−g(AN, N) = −g(AN, N),
∑2m−1
i=1 g(AX, ei)Aei = ∑2m
i=1g(AX, ei)Aei−g(AX, N)AN =X−g(AX, N)AN,
∑2m−1
i=1 g(J Aei, ei)J AX = ∑2m
i=1g(J Aei.ei)J AX−g(J AN, N)J AX
= −g(J AN, N)J AX, and
∑2m−1
i=1 g(J AX, ei)J Aei = ∑2m
i=1g(J AX, ei)J Aei−g(J AX, N)J AN
= J AJ AX−g(J AX, N)J AN
= X−g(J AX, N)J AN.
Now in this section we consider only an A-principal normal vector field N, that is, AN =N, for a real hypersurfaceM inQm with parallel Ricci tensor. Then (4.1) becomes Ric(X) = (2m−1)X−2η(X)ξ−AX+hSX−S2X, (4.2)
YOUNG JIN SUH*
where h = trS denotes the mean curvature and is defined by the trace of the shape operator S of M in Qm. Then from this, by the parallel Ricci tensor, we have
0 =−2g(∇Xξ, Y)ξ−2η(Y)∇Xξ−(∇XA)Y + (Xh)SY +h(∇XS)Y −(∇XS2)Y
=−2g(ϕSX, Y)ξ−2η(Y)ϕSX−(∇XA)Y + (Xh)SY +h(∇XS)Y −(∇XS2)Y,
(4.3)
where (∇XA)Y = ∇X(AY)−A∇XY. Here, AY belongs to TzM, z∈M, from the fact that g(AY, N) = g(Y, AN) = g(Y, N) = 0 for any tangent vector Y on M. Then by puttingY =ξ in (4.3), we know that
2ϕSX =−(∇XA)ξ+ (Xh)Sξ +h(∇XS)ξ−(∇XS2)ξ
=−q(X)J Aξ−αη(X)AN +α(Xh)ξ +h(∇XS)ξ−(∇XS2)ξ.
(4.4)
In order to get the equation (4.4) we have used the following (∇XA)ξ = ∇X(Aξ)−A∇Xξ
= ( ¯∇X(Aξ))T −A∇Xξ
= {
( ¯∇XA)ξ+A∇¯Xξ }T
−AϕSX
= q(X)J Aξ+AϕSX+g(SX, ξ)AN −AϕSX
= q(X)J Aξ+αη(X)AN,
where (· · ·)T denotes the tangential component of the vector (· · ·) in Qm. Moreover, we get
(∇XS)ξ = ∇X(Sξ)−S∇Xξ = (Xα)ξ+αϕSX −SϕSX, and
(∇XS2)ξ = ∇X(S2ξ)−S2∇Xξ = (Xα2)ξ+α2ϕSX−S2ϕSX.
Then (4.4) can be written as follows:
2ϕSX = −q(X)J Aξ−αη(X)AN +α(Xh)ξ+h(Xα)ξ+hαϕSX−hSϕSX
−(Xα2)ξ−α2ϕSX+S2ϕSX.
Then it can be reduced as follows:
2ϕSX = −q(X)J Aξ−αη(X)AN +hαϕSX−hSϕSX
−α2ϕSX+S2ϕSX.
From this, if we take a tangential part, we have the following:
(2 +α2 −hα)ϕSX =−hSϕSX +S2ϕSX (4.5) for any tangent vectorX∈TzM,z∈M, because we have assumed that the unit vector field N isA-principal, that is, AN =N, and J Aξ =−AJ ξ =−AN.
On the other hand, by (3.4) in section 3 for hypersurface in complex quadricQm with A-principal normal vector field N, we have
2SϕSX =α(ϕS+Sϕ)X+ 2ϕX,
where we used g(X, AN) = 0 and g(J X, Aξ) = g(J X, J AN) = 0 from the facts that A-principal normalAN =N and anti-commuting AJ =−J A. From this, it follows that
2S2ϕSX =α(SϕS+S2ϕ)X+ 2SϕX
=α({α
2(Sϕ+ϕS)X+ϕX}
+S2ϕ)
X+ 2SϕX
= α2
2 (Sϕ+ϕS)X+αϕX +αS2ϕX + 2SϕX
(4.6)
Then summing up (4.5) and (4.6), we have (2 +α2−hα)ϕSX
=−h {α
2(Sϕ+ϕS)X+ϕX }
+α2
4 (Sϕ+ϕS)X +α
2ϕX+α
2S2ϕX +SϕX.
(4.7)
Remark 4.1. It is known that in Berndt and Suh [4] a real hypersurface M is a tube over Sm in Qm if and only if the shape operator S of M satisfies Sϕ+ϕS = kϕ for an non-zero constant k. Then let us check whether a tube overSm could satisfy (4.7) or not.
Then (4.7) gives
(2 +α2−hα)ϕSX = −h{αk 2 + 1}
ϕX+ α2 4 kϕX +α
2ϕX +α
2S2ϕX +SϕX.
If we consider an eigen vector such that SX = λX, then (Sϕ+ϕS)X =kϕX gives that SϕX = (k−λ)ϕX. From this, together with (4.7) usingαk =−2, the principal curvatures satisfy a quadratic equation such that
αx2 −2(α2−hα+ 1)x= 0.
Then λ = 0 or µ = √
2 tan√
2r (see [4]). Moreover, the trace h of the shape operator becomes h=α+ (m−1)k (see also [4]). But for a tube over a sphere Sm we know that
√2 tan√
2r = 2
α(α2−hα+ 1)
= 2(α−h) + 2 α
= 4(m−1) α + 2
α
= 2(2m−1) α
= −(2m−1)√
2 tan√ 2r,
YOUNG JIN SUH*
where in the third equality we used α−h =−(m−1)k = 2(m−1)α and in the fifth equality α=−√
2 cot(√
2r)respectively (see Prop. 3.4 in[4]). This gives that2m√
2 tan√
2r = 0, which gives us a contradiction. So we conclude that a real hypersurface in Qm which is a tube over a m-dimensional sphere Sm does not admit a parallel Ricci tensor. Of course, in this case the unit normal N is A-principal.
If we put SX =λX, then (4.5) gives
(2 +α2−hα)λϕX =−hλSϕX +λS2ϕX.
Moreover, (4.6) gives that
SϕX = αλ+ 2 2λ−αϕX.
From this, together with the above formula, we have (2 +α2 −hα)λϕX =−hλ
(αλ+ 2 2λ−α
)
ϕX +λ
(αλ+ 2 2λ−α
)2
ϕX. (4.8)
First we consider the case that λ̸=0 and µ = αλ+22λ−α are both non-vanishing. Then the function µ= αλ+22λ−α satisfies the following equation
µ2−hµ+hα−α2−2 = 0.
Then it is equivalent to the following for the function λ λ2−hλ+hα−α2−2 = 0.
Combining these two equations, we have
(λ−µ)(λ+µ−h) = 0.
Then we know that the functionsλ andµare distinct. So it implies thath=λ+µ. Then it follows that
h=λ+µ=α+ (m−1)(λ+µ) =α+ (m−1)h, this implies that h=−mα−2, which gives us a contradiction.
In fact, in (4.7) with h=−mα−2 it becomes {
2 + (m−1)α2 m−2
}
ϕSX = α2
2(m−2)(ϕS+Sϕ)X−hϕX+ α2
4 (Sϕ+ϕS)X + α
2ϕX+α
2S2ϕX+SϕX.
From this, by taking a symmetric part, we get the following {
3 + (m−1)α2 m−2
}
g((ϕS −Sϕ)X, Y) = α
2g((S2ϕ−ϕS2)X, Y).
Then for any principal vector X such that SX = λX and SϕX = µϕX with distinct principal curvatures λ and µwe know
(λ−µ) {
3 + (m−1)α2 m−2 − α
2(λ+µ) }
ϕX = 0.
This implies 3 + (2(m2(m−2)−1)+1)α2 = 0, which gives a contradiction.
Next we consider for the case that λ = 0 and µ= −α2̸=0. In this case, we can easily verify that h=−α2. Then the equation (4.7), together with hα=−2, gives
(α2+ 4)ϕSX =(Sϕ+ϕS)X−hϕX+ α2
4 (Sϕ+ϕS)X +α
2ϕX+α
2S2ϕX +SϕX.
(4.9) Then by taking a skew-symmetric part of (4.9)we have
(α2+ 5)(ϕS −Sϕ) = α
2(S2ϕ−ϕS2). (4.10)
From this, using SX =λX and SϕX =µϕX, whereµ= αλ+22λ−α, then we have (λ−µ){
(α2+ 5)−α
2(λ+µ)}
ϕX = 0 (4.11)
Since λ̸=µand hα =−2, (4.10) implies the following 0 =α2+ 5− α
2(λ+µ) = α2+ 5− α
2h=α2+ 6,
which gives also a contradiction even in this case. So we conclude that there do not exist any real hypersurfaces in complex quadric Qm with parallel Ricci tensor andA-principal normal vector field.
5. Proof of Theorem 2
In this section we want to prove Theorem 2 for real hypersurfaces with parallel Ricci tensor and A-isotropic unit normal vector field. Then in this case we want to make the derivative of the Ricci tensor as follows:
(∇YRic)X =∇Y(Ric(X))−Ric(∇YX)
=−3(∇Yη)(X)ξ−3η(X)∇Yξ
+g(X,∇Y(AN))AN −g(AX, N)∇Y(AN) +g((∇Y(Aξ), X)Aξ+η(AX)∇Y(Aξ) + (Y h)SX +h(∇YS)X−(∇YS2)X.
(5.1)
Since we assumed that the unit normal N isA-isotropic, by the defintion in section 3 we know that t= π4. Then from the expression of the A-isotropic unit normal vector field it becomes N = √1
2Z1+√1
2Z2 . This implies that
g(ξ, Aξ) = 0, g(ξ, AN) = 0, g(AN, N) = 0, g(Aξ, N) = 0, and
g(J AN, ξ) =−g(AN, N) = 0.
Then it follows that
∇Y(BN) =∇Y(AN) ={( ¯∇YA)N +A∇¯YN}T ={q(Y)J AN −ASY}T, and
∇Y(Aξ) ={( ¯∇YA)ξ+A∇¯Yξ}T
={q(Y)J Aξ+AϕSY}T,
YOUNG JIN SUH*
where (· · ·)T denotes the tangential component of the vector (· · ·) inQm. By our assump- tion of Ricci parallel, the above formula becomes
0 =−3g(ϕSY, X)ξ−3η(X)ϕSY +{q(Y)g(J AN, X)−g(ASY, X)}AN
−g(AX, N){q(Y)J AN −ASY}T +{q(Y)g(J Aξ, X) +g(AϕSY, X)}Aξ+η(AX){q(Y)J Aξ+AϕSY}T + (Y h)SX +h(∇YS)X−(∇YS2)X.
(5.2)
By taking an inner product (5.2) with the Reeb vector field ξ, we have 0 =−3g(ϕSY, X) +g(AX, N)η(ASY) +η(AX)η(AϕSY)
+ (Y h)αη(X) +hg((∇YS)X, ξ)
−g((∇YS2)X, ξ).
(5.3) On the other hand, let us use the following calculation
(∇XS)ξ = (Xα)ξ+αϕSX −SϕSX, (∇XS2)ξ = (Xα2)ξ+α2ϕSX−S2ϕSX.
From this, together with puttingX =ξ in (5.2) and usingg(ξ, AN) = g(Aξ, ξ) = 0, we have
3ϕSY =(Y h)Sξ+h(∇YS)ξ−(∇YS2)ξ
−g(ASY, ξ)AN +g(AϕSY, ξ)Aξ
=(Y h)αξ+h{(Y α)ξ+αϕSY −SϕSY}
− {(Y α2)ξ+α2ϕSY −S2ϕSY}
=αhϕSY −hSϕSY −α2ϕSY +S2ϕSY.
(5.4)
From this, it follows that the functionhα−α2is constant onM. Then it can be rearranged as follows:
(3 +α2−αh)ϕSY =−hSϕSY +S2ϕSY −g(ASY, ξ)AN +g(AϕSY, ξ)Aξ. (5.5) Since the unit normal N isA-isotropic, we know thatg(ξ, Aξ) = 0. Moreover, by Lemma 4.2 due to [14], we have the following
2SϕSX =α(Sϕ+ϕS)X+ 2ϕX−2g(X, AN)Aξ+ 2g(X, Aξ)AN. (5.6) Now let us consider the distribution Q⊥, which is an orthogonal complement of the maximalA-invariant subspaceQin the complex subbundleC of TzM,z∈M inQm. Then by Lemma 3.1 in section 3, the orthogonal complement Q⊥ = C⊖Q becomes C⊖Q = Span [AN, Aξ]. From the assumption ofSϕ =ϕS on the distributionQ⊥ it can be easily checked that the distribution Q⊥ is invariant by the shape operator S. Then (5.6) gives the following for SAN =λAN
(2λ−α)SϕAN = (αλ+ 2)ϕAN −2Aξ
= (αλ+ 2)ϕAN −2ϕAN
= αλϕAN.
Then Aξ =ϕAN gives the following
SAξ = αλ
2λ−αAξ. (5.7)
Then from the assumptionSϕ =ϕS on Q⊥ =C⊖Q it follows that λ= 2λαλ−α gives
λ= 0 or λ=α. (5.8)
On the other hand, on a distributionQwe know thatAX∈TzM,z∈M, becauseAN∈Q.
So (5.6), together with the fact thatg(X, Aξ) = 0 andg(X, AN) = 0 for anyX∈Q, imply that
2SϕSX =α(Sϕ+ϕS)X+ 2ϕX. (5.9)
Then we can take an orthonormal basis X1,· · ·, X2(m−2)∈Q such that AXi = λiXi for i= 1,· · ·, m−2. Then by (5.6) we know that
SϕXi = αλi + 2 2λi−αϕXi.
Accordingly, by (5.8) the shape operator S can be expressed in such a way that
S =
α 0 0 0 · · · 0 0 · · · 0 0 0(α) 0 0 · · · 0 0 · · · 0 0 0 0(α) 0 · · · 0 0 · · · 0 0 0 0 λ1 · · · 0 0 · · · 0 ... ... ... ... . .. ... ... · · · ... 0 0 0 0 · · · λm−2 0 · · · 0 0 0 0 0 · · · 0 µ1 · · · 0 ... ... ... ... ... ... ... . .. ... 0 0 0 0 · · · 0 0 · · · µm−2
So on the distribution Q, for any X, ϕX∈Q such that SX = λX and SϕX = µϕX, µ= αλ+22λ−α, (5.5) becomes the following
(3 +α2 −αh)λϕX =−hλSϕX+λS2ϕX
+λg(X, Aξ)AN +λg(AϕX, ξ)Aξ
=−hλSϕX+λµ2ϕX
=−hλµϕX+λµ2ϕX,
(5.10)
where we used C⊖Q= Span [Aξ, AN] in the second equality.
From this, if there exists a non-vanishing principal curvature λ̸=0, then any principal curvature of the shape operator on the distribution Q satisfies the following quadratic equation
x2−hx+ (αh−α2−3) = 0. (5.11)
Then the discreminant always D = (h−2α)2 + 12 > 0, we conclude that there exist two principal curvatures λ and µ satisfy h = λ+µ. Moreover, by the assumption of parallel Ricci tensor ∇Ric = 0, and (5.4) the function hα−α2 should be constant. Then summing up these properties and the expressions above, let us consider the first case for non-vanishing λ̸=µas follows:
YOUNG JIN SUH*
h=λ+µ=α+ (m−2)(λ+µ)
=α+ (m−2)h. (5.12)
From this, the constant hα−α2 becomes
hα−α2 =α{α+ (m−2)(λ+µ} −α2
=(m−2)(λ+µ)α
=(m−2)hα.
Then the constanthαimplies thatα must be constant. Thus the trace of the shape oper- atorhshould be constant. So the above quadratic equation has with constant coefficients.
On the other hand, for SAξ=αAξ and SAN =αAN the constant h becomes h=λ+µ
=3α+ (m−2)(λ+µ)
=3α+ (m−2)h.
(5.13) So if we compare the above two equations (5.12) and (5.13) for the constant mean curvature h, we know that the Reeb functionα should vanish, that is, α= 0. From this, together with the assumptionm≥4, it follows that the trace h=λ+µ= 0, that is,
λ=−µ. (5.14)
Moreover, from the equation (5.11) we may put the principal curvature λ=−√
3 tanr and µ=√
3 cotr. (5.15)
Since the equation (5.11) has only distinct two principal curvatures λ and µ and the princicipal curvature µrelated to the almost complex structure ϕ such that SϕX =µϕX for SX =λX on the distribution Q, that is, Q=V(λ)⊕V(µ), where dimQ= 2(m−2), V(λ) = {X∈Q|SX = λX} and V(µ) = {X∈Q|SX = µX}. Then, we know that dimV(λ) = dimV(µ) =m−2, because λ and µ are different roots of the equation (5.11) and ϕV(λ) = V(µ). Moreover, we know that the functionρ= 0 or ρ=α by the property of the vector AN and Aξ.
Summing up these results, together with (5.14) and (5.15), we may put α = 0, ρ= 0, λ=−√
3 and µ=√
3 respectively with multiplicities 1, 2, (m−2) and (m−2).
Next for the case whereλ= 0 we conclude that on the distributionξ⊕[C⊖Q]⊕Q, where C⊖Q= Span[AN, Aξ], the shape operator becomes
S =
α 0 0 0 · · · 0 0 · · · 0 0 0(α) 0 0 · · · 0 0 · · · 0 0 0 0(α) 0 · · · 0 0 · · · 0 0 0 0 −α2 · · · 0 0 · · · 0 ... ... ... ... . .. ... ... · · · 0 0 0 0 0 · · · −α2 0 · · · 0 0 0 0 0 · · · 0 0 · · · 0 ... ... ... ... ... ... ... . .. ...
0 0 0 0 · · · 0 0 · · · 0
When SAN = 0 and SAξ= 0, the expression for the constant h becomes h =λ+µ=α+ (m−2)(λ+µ).
This gives that α+ (m−3)h= 0.
On the other hand, for λ = 0 we know µ = −α2. So it follows that h =λ+µ = −α2. Then these two equations give α2 = 2(m−3). So let us putα=√
2(m−3). Then from the expressions above the trace of the shape operator becomes respectively
h=√
2(m−3)− 2(m−2)
√2(m−3) =− 2
√2(m−3) or
h=√
2(m−3) + 2√
2(m−3)− 2(m−2)
√2(m−3) =− 2
√2(m−3)
according to the eigenvalue 0 or α on the distribution C⊖Q = Span[Aξ, AN]. The first equation is satisfied for all of m. So in this case we may put
α=√
2(m−3), γ = 0, λ= 0, and µ=− 2
√2(m−3) with corresponding principal curvature spaces
Tα = [ξ], Tγ = [Aξ, AN], ϕ(Tλ) = Tµ,dim Tλ = dim Tµ=m−2.
But if we compare the latter equation, we have m= 3. This gives us a contradiction. So the latter case can not be appeared.
Summing up the above discussions, we give a complete proof of our Main Theorem 2 in the introduction.
Remark 5.1. In this remark let us check that whether the Ricci tensor of the tube M over a totally geodesic PkC in Qm, m = 2k, mentioned in Berndt and Suh [3] is parallel or not. Then by a theorem due to [3], the shape operator S commutes with the structure tensor ϕ, that is, Sϕ=ϕS. In this case we know that the normal vector field N of M in Q2k isA-isotropic. So let us suppose that the Ricci tensor of M is parallel. Then for any Y∈Q such that SY =λY the equation (5.5) gives that
(3 +α2−αh)λϕY =−hλ2ϕY +λ3ϕY. (5.16) On the other hand, by (5.6), together with the commuting property Sϕ = ϕS, we know that
(2λ−α)λϕY = (αλ+ 2)ϕY (5.17) From this, we see that the function λ̸=0. Then by (5.16) with λ̸=0, we get
λ2−hλ+ (αh−α2−3) = 0.
Moreover, (5.17) gives that
λ2−αλ−1 = 0.
From these two equations we know that h =α and finally gives us a contradiction. So a tube over PkC never has a parallel Ricci tensor.
YOUNG JIN SUH*
Acknowledgements The present author would like to express his deep gratitude to Professor J¨urgen Berndt for his great contributions to our works due to [3] and [4].
Motivated by these works, the author has considered such a notion of Ricci parallel for real hypersurfaces in the complex quadricQm. The author also wish his thanks to the referee for many kinds of valuable comments to develop the first version of this manuscript.
References
[1] J. Berndt and Y.J. Suh, Real hypersurfaces in complex two-plane Grassmannians, Monatsh. Math.
127(1999), 1–14.
[2] J. Berndt and Y.J. Suh, Real hypersurfaces with isometric Reeb flow in complex two-plane Grass- mannians, Monatsh. Math.137(2002), 87–98.
[3] J. Berndt and Y. J. Suh, Real hypersurfaces with isometric Reeb flow in complex quadrics, Interna- tional J. Math.24(2013), 1350050, 18pp.
[4] J. Berndt and Y. J. Suh,Contact hypersurfaces in Kaehler manifold, Proceedings of AMS.23(2015), (in press).
[5] S. Klein,Totally geodesic submanifolds in the complex quadric, Diff. Geom. and Its Appl.26(2008), 79–96.
[6] M. Kimura, Real hypersurfaces of a complex projective space, Bull. Austral. Math. Soc. 33 (1986), 383–387.
[7] S. Kobayashi and K. Nomizu, Foundations of Differential Geometry, Vol. II, A Wiley-Interscience Publ., Wiley Classics Library Ed., 1996.
[8] J.D. P´erez and Y.J. Suh, Real hypersurfaces of quaternionic projective space satisfying ∇UiR = 0, Diff. Geom. and Its Appl.7(1997), 211–217.
[9] J.D. P´erez and Y.J. Suh,The Ricci tensor of real hypersurfaces in complex two-plane Grassmannians, J. of Korean Math. Soc.44(2007), 211–235.
[10] H. Reckziegel,On the geometry of the complex quadric, in: Geometry and Topology of Submanifolds VIII(Brussels/Nordfjordeid 1995), World Sci. Publ., River Edge, NJ, 1995, pp. 302–315.
[11] B. Smyth,Differential geometry of complex hypersurfaces, Ann. Math.85(1967), 246-266.
[12] Y.J. Suh, Real hypersurfaces in complex two-plane Grassmannians with parallel Ricci tensor, Proc.
Royal Soc. Edinburgh,142 A(2012), 1309-1324.
[13] Y.J. Suh,Hypersurfaces with isometric Reeb flow in complex hyperbolic two-plane Grassmannians, Advances in Applied Mathematics,50(2013), 645-659.
[14] Y.J. Suh,Real hypersurfaces in the complex quadric with Reeb parallel shape operator, International J. Math.25(2014), 1450059, 17pp.
[15] Y.J. Suh and C. Woo,Real hypersurfaces in complex hyperbolic two-plane Grassmannians with par- allel Ricci tensor, Math. Nachr.287(2014), 1524–1529.
Kyungpook National University, College of Natural Sciences, Department of Math- ematics, Daegu 702-701, Republic of Korea