N(κ)-contact Metric Manifolds"> N(κ)-contact metric manifolds" />
검색
Article Search

JMB Journal of Microbiolog and Biotechnology

OPEN ACCESS eISSN 0454-8124
pISSN 1225-6951
QR Code

Article

Kyungpook Mathematical Journal 2023; 63(2): 313-324

Published online June 30, 2023

Copyright © Kyungpook Mathematical Journal.

Generalized Ricci Solitons on N(κ)-contact Metric Manifolds

Tarak Mandal*, Urmila Biswas and Avijit Sarkar

Department of Mathematics, University of Kalyani, Kalyani-741235, West Bengal, India
e-mail : mathtarak@gmail.com, biswasurmila50@gmail.com and avjaj@yahoo.co.in

Received: November 25, 2022; Revised: May 31, 2023; Accepted: June 1, 2023

In the present paper, we study generalized Ricci solitons on N(κ)-contact metric manifolds, in particular, we consider when the potential vector field is the concircular vector field. We also consider generalized gradient Ricci solitons, and verify our results with an example.

Keywords: Nullity distribution, contact manifolds, N(κ)-contact metric manifolds, Ricci solitons, generalized Ricci solitons

The notion of a κ-nullity distribution (in brief, KND) on a Riemannian manifold (RM, in short) was coined by Tanno [12]. A KND in a RM M is described by

N(κ):qNq(κ)={V3TqM:R(V1,V2)V3      =κ[g(V2,V3)V1g(V1,V3)V2]},

for vector fields V1,V2TqM, κ being a real number, and TqM being the tangent space of M at q. A (2m+1)-dimensional contact metric manifold (CMM, in short) is called N(κ)-contact metric manifold (NCMM, in short) if the Reeb vector field 𝜃 satisfies KND. So, for a NCMM, we have

R(V1,V2)θ=κ{τ(V2)V1τ(V1)V2}.

For κ=1, the manifold is a Sasakian manifold and when κ=0 it is locally isometric to the product of an m-dimensional manifold of scalar curvature 4 with a flat (m+1)-dimensional manifold, provided m>1. If m=1 and κ=0, the manifold is flat [1]. NCMMs have been studied by many authors [1, 2, 3, 5].

Hamilton [8] introduced the famous geometric flow, named as Ricci flow, which is a kind of pseudo parabolic heat equation defined on a RM as

g(t)t=2Ric(t),

where g and Ric indicate, respectively, the Riemannian metric and the (0,2) Ricci tensor.

A Ricci soliton (RS, in short) is a fixed point of Ricci flow (RF, in short) equation (1.2). At the same time it is also a generalization of the Einstein metric. A RS on a (2m+1)-dimensional RM is given by

(LEg)(V1,V2)+2Ric(V1,V2)=2λg(V1,V2),

L being the Lie-derivative operator and λ is a constant. The nature of a RS is described by the value of λ, that means, a RS is shrinking if λ>0, it is steady if λ=0, for λ<0 it is expanding. For more about RSs, one can see the papers [13].

In [10], the authors extended the idea of an RS to a generalized Ricci soliton (GRS, in short). On a RM it is given by

(LEg)(V1,V2)+2aRic(V1,V2)+2bE(V1)E(V2)=2λg(V1,V2),

where a, b, λ and E is the canonical 1-form related with E i.e., E(V1)=g(V1,E). Similar to a RS, a GRS is shrinking or steady or expanding according as λ takes positive, zero or negative value. Here the potential vector field (PVF, in short) is termed as

Homothetic vector field if a=b=0

Killing vector field if a=b=λ=0

and the equation (1.3) is called

Ricci soliton when a=1 and b=0

Einstein-Weyl equation when a=1n1 and b=1.

GRS on different types of RS have been discussed by many authors like Ghosh and De [6, 7], Kumara, Naik and Venkatesha [9].

If the PVF be taken as the gradient of a smooth function, the GRS reduces into generalized gradient Ricci soliton (GGRS, in short). Thus the GGRS on a RM M is given by

2ψ(V1,V2)+aRic(V1,V2)+b(V1ψ)(V2ψ)=λg(V1,V2),

where 2 being the Hessian operator and ψ is a smooth function on M.

In a RM a vector field E is called concircular vector field [4] if

V1E=fV1,

for any vector field V1 on the manifold, where ∇ is the Levi-Civita connection and f is a smooth function.

The paper is embodied as follows: after a brief review of literature, we give some basic definition and curvature properties of NCMMs in the Section 2. In Section 3, we deduce certain characterizations of GRSs on NCMMs. The next section deals with generalized gradient Ricci solitons. In the last section, we give an example to support our results.

A (2m+1)-dimensional differentiable manifold endowed with a tensor field ϕ of type (1,1), a vector field 𝜃, a 1-form τ satisfying [5]

ϕ2(V1)=V1+τ(V1)θ,τ(θ)=1,

for any vector field V1χ(N), the set of all vector fields on N, is known as an almost contact manifold. An almost contact manifold is called almost contact metric manifold if it admits a Riemannian metric g such that

g(ϕV1,ϕV2)=g(V1,V2)τ(V1)τ(V2).

As a consequence of (2.2) and (2.3), we get the following:

ϕθ=0,g(V1,θ)=τ(V1),τ(ϕV1)=0,
g(ϕV1,V2)=g(V1,ϕV2),
(V1τ)(V2)=g(V1θ,V2),

for any vector fields V1,V2χ(N).

An almost contact metric manifold is called CMM whenever the almost contact metric structure (ϕ,θ,τ,g) satisfies the following condition [5]

g(V1,ϕV2)=dτ(V1,V2),

for every vector fields V1,V2χ(N). For a CMM N, we determine a symmetric (1,1)-tensor field h by h=12Lθϕ, Lθϕ indicates the Lie differentiation of ϕ in the direction 𝜃 and satisfying the following conditions

hθ=0,hϕ+ϕh=0,tr(h)=tr(hϕ)=0,
V1θ=ϕV1ϕhV1.

For a NCMM N of dimension 2m+1, m ≥ 1, we have [5]

h2=(κ1)ϕ2,
(V1ϕ)(V2)=g(V1+hV1,V2)θτ(V2)(V1+hV1),
R(V1,V2)θ=κ{τ(V2)V1τ(V1)V2},
R(θ,V1)V2=κ{g(V1,V2)θτ(V2)V1},
Ric(V1,V2)=2(m1){g(V1,V2)+g(hV1,V2)}    +{2mκ2(m1)}τ(V1)τ(V2),
Ric(V1,θ)=2mκτ(V1),
(V1τ)(V2)=g(V1+hV1,ϕV2),
(V1h)(V2)={(1κ)g(V1,ϕV2)+g(V1,hϕV2)}θ    +τ(V2){h(ϕV1+ϕhV1)},
r=2m(2m2+κ),

for every vector fields V1,V2,χ(N); R, Ric and r are the Riemannian curvature, Ricci tensor and scalar curvature, respectively.

Lemma 2.1. In a (2m+1)-dimensional NCMM N, the following holds

(V1hϕ)V2=(κ1)(g(V1,V2)θ2τ(V1)τ(V2)θ    +τ(V2)V1)g(V1,hV2)θτ(V2)hV1,

for any vector fields V1, V2 on the manifold.

Proof. By a straightforward calculation, we obtain

(V1hϕ)V2=(V1h)ϕV2+h(V1ϕ)V2.

Using (2.4) and (2.9) in the previous relation, the desired result is obtained.

Let N be a NCMM of dimension (2m+1) admitting generalized Ricci soliton. Applying covariant derivative on (1.3) in the direction V3, we obtain

(V3LEg)(V1,V2)=2a(V3Ric)(V1,V2)2b(g(V3E,V1)E(V2)      +E(V1)g(V3E,V2)).

According to Yano [14], we infer

(LEV3gV3LEg[E,V3]g)(V1,V2)=g((LE)(V3,V1),V2)            g((LE)(V3,V2),V1).

As g=0 and from the above equation, we have

(V3LEg)(V1,V2)=g((LE)(V3,V1),V2)+g((LE)(V3,V2),V1).

Due to symmetry property of LE, the above equation reduces to

2g((LE)(V3,V1),V2)=(V3LEg)(V1,V2)+(V1LEg)(V3,V2)        (V2LEg)(V3,V1).

Using (3.1) in (3.2), we get

2g((LE)(V3,V1),V2)=2a[(V3Ric)(V1,V2)+(V1Ric)(V3,V2)        (V2Ric)(V3,V1)]2b[g(V3E,V1)E(V2)        +E(V1)g(V3E,V2)+g(V1E,V3)E(V2)        +E(V3)g(V1E,V2)g(V2E,V3)E(V1)        E(V3)g(V2E,V1)].

Assuming that the PVF E as concircular vector field. Then using (1.5) in (3.3), we get

2g((LE)(V3,V1),V2)=2a[(V3Ric)(V1,V2)+(V1Ric)(V3,V2)        (V2Ric)(V3,V1)]4bfg(V3,V1)E(V2).

Differentiating (2.6) covariantly with respect to V3 and utilizing (2.3), one obtains

(V3Ric)(V1,V2)=2κ(g(V3,ϕV1)τ(V2)g(V3,ϕV2)τ(V1))    +2mκ(g(V3,hϕV1)τ(V2)+g(V3,hϕV2)τ(V1)).

Applying (3.5) in (3.4), we obtain

g((LE)(V3,V1),V2)=4maκg(V3,hϕV1)τ(V2)2bfg(V3,V1)E(V2),

which implies

(LE)(V3,V1)=4maκg(V3,hϕV1)θ2bfg(V3,V1)E.

Differentiating above equation covariantly with respect to V2 and applying (1.5), we infer

(V2LE)(V3,V1)=4maκ(g(V3,(V2hϕ)V1)θ      g(V3,hϕV1)(ϕV2+ϕhV2))      2b(V2f)g(V3,V1)E2bf2g(V3,V1)V2.

Due to Yano ([14], p-23), we have

(LER)(V1,V2)V3=(V1LE)(V2,V3)(V2LE)(V1,V3).

Applying (3.8) in the above equation, we have

(LER)(V1,V2)V3=4maκ(g((V1hϕ)V2(V2hϕ)V1,V3)θ      g(hϕV2,V3)(ϕV1+ϕhV1)      +g(hϕV1,V3)(ϕV2+ϕhV2))      2b(V1f)g(V3,V2)E+2b(V2f)g(V3,V1)E      2bf2(g(V2,V3)V1g(V1,V3)V2).

Using (2.11) in (3.10), we get

(LER)(V1,V2)V3=4maκ((κ1)(g(V1,V3)τ(V2)θ      g(V2,V3)τ(V1)θ)g(hV1,V3)τ(V2)θ      +g(hV2,V3)τ(V1)θg(hϕV2,V3)(ϕV1+ϕhV1)      +g(hϕV1,V3)(ϕV2+ϕhV2))      2b(V1f)g(V3,V2)E+2b(V2f)g(V3,V1)E      2bf2(g(V2,V3)V1g(V1,V3)V2).

Setting V2=V3=θ in the above equation, we obtain

(LER)(V1,θ)θ=2b(V1f)E+2b(θf)τ(V1)E2bf2(V1τ(V1)θ).

Taking Lie derivative of R(V1,θ)θ=κ(V1τ(V1)θ) along the vector field E, we get

(LER)(V1,θ)θ=κ((LEτ)(V1)θτ(V1)LEθ)R(V1,LEθ)θR(V1,θ)LEθ.

Putting V2=θ in (1.3) and using (2.7), we infer

(LEτ)(V1)=g(V1,LEθ)2(2maκλ)τ(V1)θ2bE(V1)E(θ).

Applying (3.14) in (3.13), we have

(LER)(V1,θ)θ=κ(g(V1,LEθ)2(2maκλ)τ(V1)θ2bE(V1)E(θ)τ(V1)LEθ)R(V1,LEθ)θR(V1,θ)LEθ.

Comparing (3.12) and (3.15), we obtain

R(V1,LEθ)θ+R(V1,θ)LEθ=2b(V1f)E2b(θf)τ(V1)E+2bf2(V1τ(V1)θ)κ(g(V1,LEθ)2(2maκλ)τ(V1)θ2bE(V1)E(θ)τ(V1)LEθ).

Contracting the above equation, we get

Ric(LEθ,θ)=b(Ef)b(θf)τ(E)+2mbf2+κ((2maκλ)+b(θ(E))2).

Using (2.7) in the above equation, we get

g(LEθ,θ)=12mκ[b(Ef)b(θf)τ(E)+2mbf2+κ(2maκλ)+b(τ(E))2].

Again, putting V1=V2=θ in (1.3), we obtain

g(LEθ,θ)=λ+2maκ+b(τ(E))2.

Comparing (3.18) and (3.19) and putting E=θ, we infer

f2=(2m1)κ2mb(2maκλ+b).

Thus we may assert the following theorem.

Theorem 3.1. If a (2m+1)-dimensional NCMM admits GRS where the PVF being the concircular vector field, then f2=(2m1)κ2mb(2maκλ+b).

Let us consider the PVF be pointwise collinear with the Reeb vector field 𝜃, i.e., E=ψθ, ψ being a smooth function on the manifold. Then, from (1.2), we have

ψg(V1θ,V2)+(V1ψ)τ(V2)+ψg(V1,V2θ)+(V2ψ)τ(V2)+2aRic(V1,V2)+2bψ2τ(V1)τ(V2)=2λg(V1,V2).

Using (2.3) in (3.21), we infer

2ψg(ϕhV1,V2)+(V1ψ)τ(V2)+(V2ψ)τ(V1)+2aRic(V1,V2)+2bψ2τ(V1)τ(V2)=2λg(V1,V2).

Setting V2=θ in the foregoing equation, we obtain

(V1ψ)+(θψ)τ(V1)=2(λ2maκbψ2)τ(V1).

Again, putting V1=θ in the above equation, we have

(θψ)=λ2maκbψ2.

Applying (3.24) in (3.23), we get

(V1ψ)=(λ2maκbψ2)τ(V1),

which implies

dψ=(λ2maκbψ2)τ.

Taking exterior derivative if (3.26) and then taking wedge product with τ, we get

(λ2maκbψ2)τdτ=0.

Since τdτ is the volume element, τdτ0. Thus, from (3.27), we infer

ψ2=λ2maκb,

which indicates that ψ is a constant. Thus we assert the following

Theorem 3.2. If a (2m+1)-dimensional NCMM admits GRS and the PVF is pointwise collinear with the Reeb vector field, then the potential vector field is a constant multiple of the Reeb vector field.

Let us suppose that the PVF be the Reeb vector field θ, then from (1.2), we infer

2g(ϕhV1,V2)+2aRic(V1,V2)+2bτ(V1)τ(V2)=2λg(V1,V2).

Assume an orthonormal frame field {ei}, i=1,2,,(2m+1) at any point on the manifold and contracting V1 and V2, yields

tr(ϕh)+ar+b=(2m+1)λ.

Since tr(ϕh)=0, we obtain from above equation

2ma(2m2+κ)+b=(2m+1)λ,

where we used (2.10). Again putting V1=V2=θ in (3.29) and using (2.7), we have

2maκ+b=λ.

Comparing (3.31) and (3.32), we get

κ=b2mam1m.

Thus we may assert the following

Theorem 3.3. If a (2m+1)-dimensional N(κ)-contact metric manifold admits generalized Ricci soliton and the potential vector is the Reeb vector field θ, then κ=b2mam1m.

In the current section we investigate the behaviour of generalized gradient Ricci solitons on N(κ)-contact metric manifolds.

Let us consider that a (2m+1)-dimensional NCMM admitting generalized gradient Ricci solitons. Then equation (1.4) can be written as

V1Dψ=λV1aQV1b(V1ψ)Dψ,

where D indicates the gradient operator. With the help of (2.6), the above equation reduces to

V1Dψ=λV12a(m1)(V1+hV1)a(2mκ2(m1))τ(V1)θb(V1ψ)Dψ.

Differentiating covariantly of the equation (4.2), we infer

V2V1Dψ=(λ2a(m1))V2V12a(m1)V2hV1a(2mκ2(m1))(V2τ(V1)θ+τ(V1)V2θ)b(V2(V1ψ))Dψb(V1ψ)V2Dψ.

Altering V1 and V2 in the foregoing equation, we have

V1V2Dψ=(λ2a(m1))V1V22a(m1)V1hV2a(2mκ2(m1))(V1τ(V2)θ+τ(V2)V1θ)b(V1(V2ψ))Dψb(V2ψ)V1Dψ.

Also, from (4.2), we get

[V1,V2]Dψ=(λ2a(m1))[V1,V2]2a(m1)h[V1,V2]a(2mκ2(m1))τ([V1,V2])θb([V1,V2]ψ)Dψ,

Using equations (4.3)-(4.5), we have

R(V1,V2)Dψ=2a(m1){(V1h)V2(V2h)V1}a(2mκ2(m1)){(V1τ)(V2)θ(V2τ)(V1)θ+τ(V2)V1θτ(V1)V2θ}b{(V2ψ)V1Dψ(V1ψ)V2Dψ}.

Remembering (2.3), (2.8), (2.9) and (4.2), the above equation reduces to

R(V1,V2)Dψ=2a(m1){2(1κ)g(V1,ϕV2)θ+τ(V2)(h(ϕV1+ϕhV1))τ(V1)(h(ϕV2+ϕhV2))}a(2mκ2(m1)){2g(V1,ϕV2+hϕV2)θ+τ(V1)(ϕV2+ϕhV2)τ(V2)(ϕV1+ϕhV1)}b{2a(m1)((V2ψ)(V1+hV1)(V1ψ)(V2+hV2))a(2mκ2(m1))((V2ψ)τ(V1)θ(V1ψ)τ(V2)θ)+λ((V2ψ)V1(V1ψ)V2)}.

Taking inner product with θ, we have

g(R(V1,V2)Dψ,θ)=4a(m1)g(V1,ϕV2)2a(2mκ2(m1))g(V1,ϕV2+hϕV2)b(λ2amκ)((V2ψ)τ(V1)(V1ψ)τ(V2)).

Again, taking inner product of (2.5) with Dψ, we get

g(R(V1,V2)θ,Dψ)=κ((V1ψ)τ(V2)(V2ψ)τ(V1)).

Equations (4.8) and (4.9) together give

(κ+b(λ2amκ))((V2ψ)τ(V1)(V1ψ)τ(V2))=4a(m1)(1κ)g(V1,ϕV2)2a(2mκ2(m1))g(V1,ϕV2+hϕV2).

Setting V2=θ in (4.10), we infer

(κ+b(λ2amκ))((θψ)τ(V1)(V1ψ))=0,

which indicates that either λ=2amκκb or Dψ=(θψ)θ. Thus we conclude that

Theorem 4.1. If a (2m+1)-dimensional NCMM admits GGRS, then either λ=2amκκb or the potential vector field is pointwise collinear with the characteristic vector field θ.

In [5], De at el. initiated an example of a N(κ)-contact metric manifold. Following that example we construct the following.

Let us consider the manifold M={x1,x2,x33:x30} of dimension 3, where (x1,x2,x3) are standard co-ordinates in 3. We choose the vector fields W1, W2 and W3 which satisfy

[W1,W2]=3W3,[W1,W3]=W2,[W2,W3]=2W1.

The Riemannian metric tensor g is considered as

g(Wi,Wi)=1,i=1,2,3

and

g(Wi,Wj)=0,ij.

The 1-form τ is defined by

τ(V)=g(V,W1),

for every vector field V on M. The (1,1) type tensor field ϕ is given by

ϕ(W1)=0,ϕ(W2)=W3,ϕ(W3)=W2.

Then we find that

τ(W1)=1,ϕ2V1=V1+τ(V1)W1,
g(ϕV1,ϕV2)=g(V1,V2)τ(V1)τ(V2),dτ(V1,V2)=g(V1,ϕV2),

for every vector fields V1, V2 on M. Thus (ϕ,W1,τ,g) defines a contact structure.

Due to Koszul's famous formula, we obtain the following

W1W1=0,W1W2=0,W1W3=0,
W2W2=0,W2W1=3W3,W2W3=3W1,
W3W3=0,W3W1=W2,W3W2=W1.

From the above expressions of ∇, we obtain

hW1=0,hW2=2W2,hW3=2W3.

We also have

R(W1,W2)W2=3W1,R(W2,W1)W1=3W2,R(W2,W3)W3=3W2,
R(W3,W2)W2=3W3,R(W1,W3)W3=3W1,R(W3,W1)W1=3W3,
R(W1,W2)W3=0,R(W2,W3)W1=0,R(W1,W3)W2=0.

Thus the manifold is an N(κ)-contact metric manifold with κ=3.

On contraction of curvature tensor, we infer

Ric(W1,W1)=6,Ric(W2,W2)=0,Ric(W3,W3)=0.

The scalar curvature r of the manifold is given by

r=Ric(W1,W1)+Ric(W2,W2)+Ric(W3,W3)=6.

Let the potential vector field E=W1, then

(LW1g)(W1,W1)=0,(LW1g)(W2,W2)=0,(LW1g)(W3,W3)=0.

The above data indicates that the given manifold admits generalized Ricci soliton with λ=0 and hence the soliton is steady type. Also, from (1.3), we see that 6a=b and these data satisfy equation (3.33). Hence Theorem 3.3 is verified.

  1. D. E. Blair. Contact Manifolds in Riemannian Geometry. Lecture Notes in Math. Berlin-New York: Springer-Verlag; 1976.
    CrossRef
  2. D. E. Blair, Two remarks on contact metric structure, Tohoku Math., J., 29(1977), 319-324.
    CrossRef
  3. D. E. Blair, T. Koufogiorgos and B. J. Papantoniou, Contact metric manifolds satisfying a nullity condition, Israel J. Math., 91(1995), 189-214.
    CrossRef
  4. F. Brickell and K. Yano, Concurrent vector fields and Minkowski structure, Kodai Math. Sem. Rep., 26(1974), 22-28.
    CrossRef
  5. U. C. De, A. Yildiz and S. Ghosh, On a class of N(κ)-contact metric manifolds, Math. Reports, 16(66)(2014), 207-217.
  6. G. Ghosh and U. C. De, Generalized Ricci solitons on K-contact manifolds, Math. Sci. Appl. E-Notes, 8(2)(2020), 165-169.
    CrossRef
  7. G. Ghosh and U. C. De, Generalized Ricci solitons on contact metric manifolds, Afr. Mat., 33(2022), 1-6.
    CrossRef
  8. R. S. Hamilton, The Ricci flow on Surfaces, Contemp. Math., 71(1988), 237-262.
    CrossRef
  9. H. A. Kumara, D. M. Naik and V. Venkatesha, Geometry of generalized Ricci-type solitons on a class of Riemannian manifolds, J. Geom. Phys., 176(2022).
    CrossRef
  10. P. Nurowski and M. Randall, Generalized Ricci solitons, J. Geom. Anal., 26(2016), 1280-1345.
    CrossRef
  11. A. Sarkar and A. Sardar, η-Ricci solitons on N(κ)-contact metric manifolds, Filomat, 35(11)(2021), 3879-3889.
    CrossRef
  12. S. Tanno, The topology of contact Riemannian manifolds, Illinois J. Math., 12(1968), 700-717.
    CrossRef
  13. M. Turan, U. C. De and A. Yildiz, Ricci solitons and gradient Ricci solitons in three-dimensional trans-Sasakian manifolds, Filomat, 26(2)(2012), 363-370.
    CrossRef
  14. K. Yano. Integral formulas in Riemannian Geometry. New York: Marcel Dekker; 1970.