1. Rigidity, gap theorems and maximum
principles for Ricci solitons
Manuel Fernández López
Consellería de Educación e Ordenación Universitaria
Xunta de Galicia
Galicia SPAIN
(joint work with Eduardo García Río)
Ricci Solitons Days in Pisa
4-8th April 2011
2. Outline
Rigidity of Ricci solitons
Rigidity: compact case
Rigidity: non-compact case
Locally conformally flat case
Gap theorems
Diameter bounds
Gap theorems: compact case
Gap theorems: non-compact case
Maximum principles
Introduction
Omori-Yau maximum principle
Applications
Steady solitons
Lower bound for the curvature of a steady soliton
3. Definition (Petersen and Wylie, 2007)
A Ricci soliton is said to be rigid if it is of the form N ×Γ Rk ,
where N is an Einstein manifold and Γ acts freely on N and by
orthogonal transformations on Rk .
Theorem (Petersen and Wylie, 2007)
The following conditions for a shrinking (expanding) gradient
soliton Ric + Hf = λg all imply that the metric is radially flat and
has constant scalar curvature
R is constant and sec(E, f ) ≥ 0 (sec(E, f ) ≤ 0)
R is constant and 0 ≤ Ric ≤ λg (λg ≤ Ric ≤ 0)
The curvature tensor is harmonic
Ric ≥ 0 (Ric ≤ 0) and sec(E, f) = 0
4. Definition (Petersen and Wylie, 2007)
A Ricci soliton is said to be rigid if it is of the form N ×Γ Rk ,
where N is an Einstein manifold and Γ acts freely on N and by
orthogonal transformations on Rk .
Theorem (Petersen and Wylie, 2007)
The following conditions for a shrinking (expanding) gradient
soliton Ric + Hf = λg all imply that the metric is radially flat and
has constant scalar curvature
R is constant and sec(E, f ) ≥ 0 (sec(E, f ) ≤ 0)
R is constant and 0 ≤ Ric ≤ λg (λg ≤ Ric ≤ 0)
The curvature tensor is harmonic
Ric ≥ 0 (Ric ≤ 0) and sec(E, f) = 0
5. Outline
Rigidity of Ricci solitons
Rigidity: compact case
Rigidity: non-compact case
Locally conformally flat case
Gap theorems
Diameter bounds
Gap theorems: compact case
Gap theorems: non-compact case
Maximum principles
Introduction
Omori-Yau maximum principle
Applications
Steady solitons
Lower bound for the curvature of a steady soliton
6. Theorem (Eminenti, LaNave and Mantegazza, 2008)
Let (M n , g) be an n-dimensional compact Ricci soliton. If (M, g)
is locally conformally flat then it is Einstein (in fact, a space
form).
Theorem (E. García Río and MFL, 2009)
Let (M n , g) be an n-dimensional compact Ricci soliton. Then
(M, g) is rigid if an only if it has harmonic Weyl tensor.
A gradient Ricci soliton is a Riemannian manifold such that
Ric + Hf = λg
7. Theorem (Eminenti, LaNave and Mantegazza, 2008)
Let (M n , g) be an n-dimensional compact Ricci soliton. If (M, g)
is locally conformally flat then it is Einstein (in fact, a space
form).
Theorem (E. García Río and MFL, 2009)
Let (M n , g) be an n-dimensional compact Ricci soliton. Then
(M, g) is rigid if an only if it has harmonic Weyl tensor.
A gradient Ricci soliton is a Riemannian manifold such that
Ric + Hf = λg
8. Theorem (Eminenti, LaNave and Mantegazza, 2008)
Let (M n , g) be an n-dimensional compact Ricci soliton. If (M, g)
is locally conformally flat then it is Einstein (in fact, a space
form).
Theorem (E. García Río and MFL, 2009)
Let (M n , g) be an n-dimensional compact Ricci soliton. Then
(M, g) is rigid if an only if it has harmonic Weyl tensor.
A gradient Ricci soliton is a Riemannian manifold such that
Ric + Hf = λg
9. R
The Schouten tensor S = Rc − g is a Codazzi tensor
2(n − 1)
X (R) Y (R)
( X Rc)(Y , Z ) − ( Y Rc)(X , Z ) = g(Y , Z ) − g(X , Z )
2(n − 1) 2(n − 1)
1 1
Rm(X , Y , Z , f ) = Rc(X , f )g(Y , Z ) − Rc(Y , f )g(X , Z )
n−1 n−1
f is an eigenvector of Rc
(div Rm)(X , Y , Z ) = Rm(X , Y , Z , f)
1
|div Rm|2 = | R|2
2(n − 1)
10. R
The Schouten tensor S = Rc − g is a Codazzi tensor
2(n − 1)
X (R) Y (R)
( X Rc)(Y , Z ) − ( Y Rc)(X , Z ) = g(Y , Z ) − g(X , Z )
2(n − 1) 2(n − 1)
1 1
Rm(X , Y , Z , f ) = Rc(X , f )g(Y , Z ) − Rc(Y , f )g(X , Z )
n−1 n−1
f is an eigenvector of Rc
(div Rm)(X , Y , Z ) = Rm(X , Y , Z , f)
1
|div Rm|2 = | R|2
2(n − 1)
11. R
The Schouten tensor S = Rc − g is a Codazzi tensor
2(n − 1)
X (R) Y (R)
( X Rc)(Y , Z ) − ( Y Rc)(X , Z ) = g(Y , Z ) − g(X , Z )
2(n − 1) 2(n − 1)
1 1
Rm(X , Y , Z , f ) = Rc(X , f )g(Y , Z ) − Rc(Y , f )g(X , Z )
n−1 n−1
f is an eigenvector of Rc
(div Rm)(X , Y , Z ) = Rm(X , Y , Z , f)
1
|div Rm|2 = | R|2
2(n − 1)
12. R
The Schouten tensor S = Rc − g is a Codazzi tensor
2(n − 1)
X (R) Y (R)
( X Rc)(Y , Z ) − ( Y Rc)(X , Z ) = g(Y , Z ) − g(X , Z )
2(n − 1) 2(n − 1)
1 1
Rm(X , Y , Z , f ) = Rc(X , f )g(Y , Z ) − Rc(Y , f )g(X , Z )
n−1 n−1
f is an eigenvector of Rc
(div Rm)(X , Y , Z ) = Rm(X , Y , Z , f)
1
|div Rm|2 = | R|2
2(n − 1)
13. R
The Schouten tensor S = Rc − g is a Codazzi tensor
2(n − 1)
X (R) Y (R)
( X Rc)(Y , Z ) − ( Y Rc)(X , Z ) = g(Y , Z ) − g(X , Z )
2(n − 1) 2(n − 1)
1 1
Rm(X , Y , Z , f ) = Rc(X , f )g(Y , Z ) − Rc(Y , f )g(X , Z )
n−1 n−1
f is an eigenvector of Rc
(div Rm)(X , Y , Z ) = Rm(X , Y , Z , f)
1
|div Rm|2 = | R|2
2(n − 1)
14. R
The Schouten tensor S = Rc − g is a Codazzi tensor
2(n − 1)
X (R) Y (R)
( X Rc)(Y , Z ) − ( Y Rc)(X , Z ) = g(Y , Z ) − g(X , Z )
2(n − 1) 2(n − 1)
1 1
Rm(X , Y , Z , f ) = Rc(X , f )g(Y , Z ) − Rc(Y , f )g(X , Z )
n−1 n−1
f is an eigenvector of Rc
(div Rm)(X , Y , Z ) = Rm(X , Y , Z , f)
1
|div Rm|2 = | R|2
2(n − 1)
15. |div Rm|2 e−f = | Ric|2 e−f
M M
X. Cao, B. Wang and Z. Zhang; On Locally Conformally
Flat Gradient Shrinking Ricci Solitons
1 1
| R|2 e−f ≥ | R|2 e−f
2(n − 1) M n M
For n ≤ 3 compact shrinking Ricci solitons are Einstein (n = 2
Hamilton, n = 3 Ivey)
Since n ≥ 4 one has that R is constant
(M, g) is Einstein
What about the noncompact case?
16. |div Rm|2 e−f = | Ric|2 e−f
M M
X. Cao, B. Wang and Z. Zhang; On Locally Conformally
Flat Gradient Shrinking Ricci Solitons
1 1
| R|2 e−f ≥ | R|2 e−f
2(n − 1) M n M
For n ≤ 3 compact shrinking Ricci solitons are Einstein (n = 2
Hamilton, n = 3 Ivey)
Since n ≥ 4 one has that R is constant
(M, g) is Einstein
What about the noncompact case?
17. |div Rm|2 e−f = | Ric|2 e−f
M M
X. Cao, B. Wang and Z. Zhang; On Locally Conformally
Flat Gradient Shrinking Ricci Solitons
1 1
| R|2 e−f ≥ | R|2 e−f
2(n − 1) M n M
For n ≤ 3 compact shrinking Ricci solitons are Einstein (n = 2
Hamilton, n = 3 Ivey)
Since n ≥ 4 one has that R is constant
(M, g) is Einstein
What about the noncompact case?
18. |div Rm|2 e−f = | Ric|2 e−f
M M
X. Cao, B. Wang and Z. Zhang; On Locally Conformally
Flat Gradient Shrinking Ricci Solitons
1 1
| R|2 e−f ≥ | R|2 e−f
2(n − 1) M n M
For n ≤ 3 compact shrinking Ricci solitons are Einstein (n = 2
Hamilton, n = 3 Ivey)
Since n ≥ 4 one has that R is constant
(M, g) is Einstein
What about the noncompact case?
19. |div Rm|2 e−f = | Ric|2 e−f
M M
X. Cao, B. Wang and Z. Zhang; On Locally Conformally
Flat Gradient Shrinking Ricci Solitons
1 1
| R|2 e−f ≥ | R|2 e−f
2(n − 1) M n M
For n ≤ 3 compact shrinking Ricci solitons are Einstein (n = 2
Hamilton, n = 3 Ivey)
Since n ≥ 4 one has that R is constant
(M, g) is Einstein
What about the noncompact case?
20. |div Rm|2 e−f = | Ric|2 e−f
M M
X. Cao, B. Wang and Z. Zhang; On Locally Conformally
Flat Gradient Shrinking Ricci Solitons
1 1
| R|2 e−f ≥ | R|2 e−f
2(n − 1) M n M
For n ≤ 3 compact shrinking Ricci solitons are Einstein (n = 2
Hamilton, n = 3 Ivey)
Since n ≥ 4 one has that R is constant
(M, g) is Einstein
What about the noncompact case?
21. |div Rm|2 e−f = | Ric|2 e−f
M M
X. Cao, B. Wang and Z. Zhang; On Locally Conformally
Flat Gradient Shrinking Ricci Solitons
1 1
| R|2 e−f ≥ | R|2 e−f
2(n − 1) M n M
For n ≤ 3 compact shrinking Ricci solitons are Einstein (n = 2
Hamilton, n = 3 Ivey)
Since n ≥ 4 one has that R is constant
(M, g) is Einstein
What about the noncompact case?
22. Outline
Rigidity of Ricci solitons
Rigidity: compact case
Rigidity: non-compact case
Locally conformally flat case
Gap theorems
Diameter bounds
Gap theorems: compact case
Gap theorems: non-compact case
Maximum principles
Introduction
Omori-Yau maximum principle
Applications
Steady solitons
Lower bound for the curvature of a steady soliton
23. Theorem (E. García Río and MFL, 2009)
Let (M n , g) be a complete noncompact gradient shrinking Ricci
soliton whose curvature tensor has at most exponential growth
and having Ricci tensor bounded from below. Then (M, g) is
rigid if an only if it has harmonic Weyl tensor.
|div Rm|2 e−f = | Ric|2 e−f
M M
R is constant and Rm( f , X , X , f) = 0
P. Petersen and W. Wilye; Rigidity of gradient Ricci solitons
Theorem (Munteanu and Sesum, 2009)
Let (M, g) be a complete noncompact gradient shrinking Ricci
soliton. Then (M, g) is rigid if an only if it has harmonic Weyl
tensor.
24. Theorem (E. García Río and MFL, 2009)
Let (M n , g) be a complete noncompact gradient shrinking Ricci
soliton whose curvature tensor has at most exponential growth
and having Ricci tensor bounded from below. Then (M, g) is
rigid if an only if it has harmonic Weyl tensor.
|div Rm|2 e−f = | Ric|2 e−f
M M
R is constant and Rm( f , X , X , f) = 0
P. Petersen and W. Wilye; Rigidity of gradient Ricci solitons
Theorem (Munteanu and Sesum, 2009)
Let (M, g) be a complete noncompact gradient shrinking Ricci
soliton. Then (M, g) is rigid if an only if it has harmonic Weyl
tensor.
25. Theorem (E. García Río and MFL, 2009)
Let (M n , g) be a complete noncompact gradient shrinking Ricci
soliton whose curvature tensor has at most exponential growth
and having Ricci tensor bounded from below. Then (M, g) is
rigid if an only if it has harmonic Weyl tensor.
|div Rm|2 e−f = | Ric|2 e−f
M M
R is constant and Rm( f , X , X , f) = 0
P. Petersen and W. Wilye; Rigidity of gradient Ricci solitons
Theorem (Munteanu and Sesum, 2009)
Let (M, g) be a complete noncompact gradient shrinking Ricci
soliton. Then (M, g) is rigid if an only if it has harmonic Weyl
tensor.
26. Theorem (E. García Río and MFL, 2009)
Let (M n , g) be a complete noncompact gradient shrinking Ricci
soliton whose curvature tensor has at most exponential growth
and having Ricci tensor bounded from below. Then (M, g) is
rigid if an only if it has harmonic Weyl tensor.
|div Rm|2 e−f = | Ric|2 e−f
M M
R is constant and Rm( f , X , X , f) = 0
P. Petersen and W. Wilye; Rigidity of gradient Ricci solitons
Theorem (Munteanu and Sesum, 2009)
Let (M, g) be a complete noncompact gradient shrinking Ricci
soliton. Then (M, g) is rigid if an only if it has harmonic Weyl
tensor.
27. Theorem (E. García Río and MFL, 2009)
Let (M n , g) be a complete noncompact gradient shrinking Ricci
soliton whose curvature tensor has at most exponential growth
and having Ricci tensor bounded from below. Then (M, g) is
rigid if an only if it has harmonic Weyl tensor.
|div Rm|2 e−f = | Ric|2 e−f
M M
R is constant and Rm( f , X , X , f) = 0
P. Petersen and W. Wilye; Rigidity of gradient Ricci solitons
Theorem (Munteanu and Sesum, 2009)
Let (M, g) be a complete noncompact gradient shrinking Ricci
soliton. Then (M, g) is rigid if an only if it has harmonic Weyl
tensor.
28. Outline
Rigidity of Ricci solitons
Rigidity: compact case
Rigidity: non-compact case
Locally conformally flat case
Gap theorems
Diameter bounds
Gap theorems: compact case
Gap theorems: non-compact case
Maximum principles
Introduction
Omori-Yau maximum principle
Applications
Steady solitons
Lower bound for the curvature of a steady soliton
29. Lemma (E. García Río and MFL, 2010)
Let (M n , g) be a locally conformally flat gradient Ricci soliton.
Then it is locally (where f = 0) isometric to a warped product
(M, g) = ((a, b) × N, dt 2 + ψ(t)2 gN ),
where (N, gN ) is a space form.
Rc(V , V ) Rc(Ei , Ei ) R
W (V , Ei , Ei , V ) = − − +
(n − 1)(n − 2) n−2 (n − 1)(n − 2)
where
1
V = f
| f|
30. Lemma (E. García Río and MFL, 2010)
Let (M n , g) be a locally conformally flat gradient Ricci soliton.
Then it is locally (where f = 0) isometric to a warped product
(M, g) = ((a, b) × N, dt 2 + ψ(t)2 gN ),
where (N, gN ) is a space form.
Rc(V , V ) Rc(Ei , Ei ) R
W (V , Ei , Ei , V ) = − − +
(n − 1)(n − 2) n−2 (n − 1)(n − 2)
where
1
V = f
| f|
31. 1
Rc(Ei , Ei ) = (R − Rc(V , V ))
n−1
1
Hf (Ei , Ei ) = (∆f − Hf (V , V ))
n−1
N = f −1 (c) is a totally umbilical submanifold of (M, g)
f is an eigenvector of Hf ↔ the integral curves of V are
geodesics
(M, g) is locally a warped product
N is a space form
Brozos-Vázquez, García-Río and Vázquez-Lorenzo;
Complete locally conformally flat manifolds of negative
curvature
32. 1
Rc(Ei , Ei ) = (R − Rc(V , V ))
n−1
1
Hf (Ei , Ei ) = (∆f − Hf (V , V ))
n−1
N = f −1 (c) is a totally umbilical submanifold of (M, g)
f is an eigenvector of Hf ↔ the integral curves of V are
geodesics
(M, g) is locally a warped product
N is a space form
Brozos-Vázquez, García-Río and Vázquez-Lorenzo;
Complete locally conformally flat manifolds of negative
curvature
33. 1
Rc(Ei , Ei ) = (R − Rc(V , V ))
n−1
1
Hf (Ei , Ei ) = (∆f − Hf (V , V ))
n−1
N = f −1 (c) is a totally umbilical submanifold of (M, g)
f is an eigenvector of Hf ↔ the integral curves of V are
geodesics
(M, g) is locally a warped product
N is a space form
Brozos-Vázquez, García-Río and Vázquez-Lorenzo;
Complete locally conformally flat manifolds of negative
curvature
34. 1
Rc(Ei , Ei ) = (R − Rc(V , V ))
n−1
1
Hf (Ei , Ei ) = (∆f − Hf (V , V ))
n−1
N = f −1 (c) is a totally umbilical submanifold of (M, g)
f is an eigenvector of Hf ↔ the integral curves of V are
geodesics
(M, g) is locally a warped product
N is a space form
Brozos-Vázquez, García-Río and Vázquez-Lorenzo;
Complete locally conformally flat manifolds of negative
curvature
35. 1
Rc(Ei , Ei ) = (R − Rc(V , V ))
n−1
1
Hf (Ei , Ei ) = (∆f − Hf (V , V ))
n−1
N = f −1 (c) is a totally umbilical submanifold of (M, g)
f is an eigenvector of Hf ↔ the integral curves of V are
geodesics
(M, g) is locally a warped product
N is a space form
Brozos-Vázquez, García-Río and Vázquez-Lorenzo;
Complete locally conformally flat manifolds of negative
curvature
36. 1
Rc(Ei , Ei ) = (R − Rc(V , V ))
n−1
1
Hf (Ei , Ei ) = (∆f − Hf (V , V ))
n−1
N = f −1 (c) is a totally umbilical submanifold of (M, g)
f is an eigenvector of Hf ↔ the integral curves of V are
geodesics
(M, g) is locally a warped product
N is a space form
Brozos-Vázquez, García-Río and Vázquez-Lorenzo;
Complete locally conformally flat manifolds of negative
curvature
37. Theorem (E. García Río and MFL, 2010)
Let (M n , g) be a simply connected complete locally conformally
flat gradient shrinking or steady Ricci soliton. Then it is
rotationally symmetric.
Any complete ancient solution to the Ricci flow has nonnegative
curvature operator (n = 3 Chen, n ≥ 4 Zhang)
RmN (X , Y , Y , X ) = RmM (X , Y , Y , X )+II(X , X )II(Y , Y )−II(X , Y )2
N is a standard sphere
(M n , g) is rotationally symmetric
B. Kotschwar; On rotationally invariant shrinking gradient
Ricci solitons
H.-D. Cao and Q. Chen; On Locally Conformally Flat
Gradient Steady Ricci Solitons
38. Theorem (E. García Río and MFL, 2010)
Let (M n , g) be a simply connected complete locally conformally
flat gradient shrinking or steady Ricci soliton. Then it is
rotationally symmetric.
Any complete ancient solution to the Ricci flow has nonnegative
curvature operator (n = 3 Chen, n ≥ 4 Zhang)
RmN (X , Y , Y , X ) = RmM (X , Y , Y , X )+II(X , X )II(Y , Y )−II(X , Y )2
N is a standard sphere
(M n , g) is rotationally symmetric
B. Kotschwar; On rotationally invariant shrinking gradient
Ricci solitons
H.-D. Cao and Q. Chen; On Locally Conformally Flat
Gradient Steady Ricci Solitons
39. Theorem (E. García Río and MFL, 2010)
Let (M n , g) be a simply connected complete locally conformally
flat gradient shrinking or steady Ricci soliton. Then it is
rotationally symmetric.
Any complete ancient solution to the Ricci flow has nonnegative
curvature operator (n = 3 Chen, n ≥ 4 Zhang)
RmN (X , Y , Y , X ) = RmM (X , Y , Y , X )+II(X , X )II(Y , Y )−II(X , Y )2
N is a standard sphere
(M n , g) is rotationally symmetric
B. Kotschwar; On rotationally invariant shrinking gradient
Ricci solitons
H.-D. Cao and Q. Chen; On Locally Conformally Flat
Gradient Steady Ricci Solitons
40. Theorem (E. García Río and MFL, 2010)
Let (M n , g) be a simply connected complete locally conformally
flat gradient shrinking or steady Ricci soliton. Then it is
rotationally symmetric.
Any complete ancient solution to the Ricci flow has nonnegative
curvature operator (n = 3 Chen, n ≥ 4 Zhang)
RmN (X , Y , Y , X ) = RmM (X , Y , Y , X )+II(X , X )II(Y , Y )−II(X , Y )2
N is a standard sphere
(M n , g) is rotationally symmetric
B. Kotschwar; On rotationally invariant shrinking gradient
Ricci solitons
H.-D. Cao and Q. Chen; On Locally Conformally Flat
Gradient Steady Ricci Solitons
41. Theorem (E. García Río and MFL, 2010)
Let (M n , g) be a simply connected complete locally conformally
flat gradient shrinking or steady Ricci soliton. Then it is
rotationally symmetric.
Any complete ancient solution to the Ricci flow has nonnegative
curvature operator (n = 3 Chen, n ≥ 4 Zhang)
RmN (X , Y , Y , X ) = RmM (X , Y , Y , X )+II(X , X )II(Y , Y )−II(X , Y )2
N is a standard sphere
(M n , g) is rotationally symmetric
B. Kotschwar; On rotationally invariant shrinking gradient
Ricci solitons
H.-D. Cao and Q. Chen; On Locally Conformally Flat
Gradient Steady Ricci Solitons
42. Theorem (E. García Río and MFL, 2010)
Let (M n , g) be a simply connected complete locally conformally
flat gradient shrinking or steady Ricci soliton. Then it is
rotationally symmetric.
Any complete ancient solution to the Ricci flow has nonnegative
curvature operator (n = 3 Chen, n ≥ 4 Zhang)
RmN (X , Y , Y , X ) = RmM (X , Y , Y , X )+II(X , X )II(Y , Y )−II(X , Y )2
N is a standard sphere
(M n , g) is rotationally symmetric
B. Kotschwar; On rotationally invariant shrinking gradient
Ricci solitons
H.-D. Cao and Q. Chen; On Locally Conformally Flat
Gradient Steady Ricci Solitons
43. Outline
Rigidity of Ricci solitons
Rigidity: compact case
Rigidity: non-compact case
Locally conformally flat case
Gap theorems
Diameter bounds
Gap theorems: compact case
Gap theorems: non-compact case
Maximum principles
Introduction
Omori-Yau maximum principle
Applications
Steady solitons
Lower bound for the curvature of a steady soliton
44. Theorem (E. García Río and MFL, 2008)
Let (M n , g) be a compact gradient Ricci soliton. Then
fmax − fmin fmax − fmin fmax − fmin
diam2 (M, g) ≥ 2max , ,4
λ−c C−λ C−c
where c ≤ Ric ≤ C.
Theorem (E. García Río and MFL, 2008)
Let (M n , g) be a compact gradient Ricci soliton with Ric > 0.
Then
Rmax − Rmin Rmax − Rmin Rmax − Rmin
diam2 (M, g) ≥ max , ,4
λ(λ − c) λ(C − λ) λ(C − c)
where c ≤ Ric ≤ C.
45. Theorem (E. García Río and MFL, 2008)
Let (M n , g) be a compact gradient Ricci soliton. Then
fmax − fmin fmax − fmin fmax − fmin
diam2 (M, g) ≥ 2max , ,4
λ−c C−λ C−c
where c ≤ Ric ≤ C.
Theorem (E. García Río and MFL, 2008)
Let (M n , g) be a compact gradient Ricci soliton with Ric > 0.
Then
Rmax − Rmin Rmax − Rmin Rmax − Rmin
diam2 (M, g) ≥ max , ,4
λ(λ − c) λ(C − λ) λ(C − c)
where c ≤ Ric ≤ C.
46. Outline
Rigidity of Ricci solitons
Rigidity: compact case
Rigidity: non-compact case
Locally conformally flat case
Gap theorems
Diameter bounds
Gap theorems: compact case
Gap theorems: non-compact case
Maximum principles
Introduction
Omori-Yau maximum principle
Applications
Steady solitons
Lower bound for the curvature of a steady soliton
47. Theorem (A. Futaki and Y. Sano, 2010)
Let (M, g) be an n-dimensional compact shrinking Ricci soliton.
Then
10π
diam(M, g) ≥ √ .
13 λ
Theorem (E. García Río and MFL, 2010)
Let (M, g) be an n-dimensional compact shrinking Ricci soliton.
Then (M, g) is Einstein if and only if one of the following
conditions holds:
Rmax − Rmin
(i) Ric ≥ 1 − λg,
(n − 1)λπ 2 + Rmax − Rmin
c(Rmax − Rmin )
(ii) cg ≤ Ric ≤ λ + g, for some c > 0
(n − 1)λπ 2
4(Rmax − Rmin )
(iii) cg ≤ Ric ≤ 1 + cg, for some c > 0.
(n − 1)λπ 2
48. Theorem (A. Futaki and Y. Sano, 2010)
Let (M, g) be an n-dimensional compact shrinking Ricci soliton.
Then
10π
diam(M, g) ≥ √ .
13 λ
Theorem (E. García Río and MFL, 2010)
Let (M, g) be an n-dimensional compact shrinking Ricci soliton.
Then (M, g) is Einstein if and only if one of the following
conditions holds:
Rmax − Rmin
(i) Ric ≥ 1 − λg,
(n − 1)λπ 2 + Rmax − Rmin
c(Rmax − Rmin )
(ii) cg ≤ Ric ≤ λ + g, for some c > 0
(n − 1)λπ 2
4(Rmax − Rmin )
(iii) cg ≤ Ric ≤ 1 + cg, for some c > 0.
(n − 1)λπ 2
49. Assume (i) holds
(n − 1)λ2 π 2
c=
(n − 1)λπ 2 + Rmax − Rmin
Rmax − Rmin (n − 1)π 2
diam2 (M, g) ≥ ≥
λ(λ − c) c
Myers’ theorem:
n−1
Ric ≥ cg > 0 =⇒ diam(M, g) ≤ π
c
By Cheng M must be the standard sphere
CONTRADICTION!
50. Assume (i) holds
(n − 1)λ2 π 2
c=
(n − 1)λπ 2 + Rmax − Rmin
Rmax − Rmin (n − 1)π 2
diam2 (M, g) ≥ ≥
λ(λ − c) c
Myers’ theorem:
n−1
Ric ≥ cg > 0 =⇒ diam(M, g) ≤ π
c
By Cheng M must be the standard sphere
CONTRADICTION!
51. Assume (i) holds
(n − 1)λ2 π 2
c=
(n − 1)λπ 2 + Rmax − Rmin
Rmax − Rmin (n − 1)π 2
diam2 (M, g) ≥ ≥
λ(λ − c) c
Myers’ theorem:
n−1
Ric ≥ cg > 0 =⇒ diam(M, g) ≤ π
c
By Cheng M must be the standard sphere
CONTRADICTION!
52. Assume (i) holds
(n − 1)λ2 π 2
c=
(n − 1)λπ 2 + Rmax − Rmin
Rmax − Rmin (n − 1)π 2
diam2 (M, g) ≥ ≥
λ(λ − c) c
Myers’ theorem:
n−1
Ric ≥ cg > 0 =⇒ diam(M, g) ≤ π
c
By Cheng M must be the standard sphere
CONTRADICTION!
53. Assume (i) holds
(n − 1)λ2 π 2
c=
(n − 1)λπ 2 + Rmax − Rmin
Rmax − Rmin (n − 1)π 2
diam2 (M, g) ≥ ≥
λ(λ − c) c
Myers’ theorem:
n−1
Ric ≥ cg > 0 =⇒ diam(M, g) ≤ π
c
By Cheng M must be the standard sphere
CONTRADICTION!
54. Assume (i) holds
(n − 1)λ2 π 2
c=
(n − 1)λπ 2 + Rmax − Rmin
Rmax − Rmin (n − 1)π 2
diam2 (M, g) ≥ ≥
λ(λ − c) c
Myers’ theorem:
n−1
Ric ≥ cg > 0 =⇒ diam(M, g) ≤ π
c
By Cheng M must be the standard sphere
CONTRADICTION!
55. Theorem (E. García Río and MFL, 2010)
Let (M, g) be an n-dimensional compact shrinking Ricci soliton.
Then (M, g) is Einstein if and only if
2 1
Rmax − nλ ≤ 1+ | f |2 .
n vol (M, g) M
Theorem (E. García Río and MFL, 2010)
Let (M, g) be an n-dimensional compact shrinking Ricci soliton.
Then (M, g) is Einstein if and only if
−Λ + Λ2 + 8(n − 1)λΛ
|Ric − λg| ≤ c ≤ ,
4(n − 1)
1
where Λ = vol(M,g) M | f |2 denotes the average of the L2 -norm
of | f |.
56. Theorem (E. García Río and MFL, 2010)
Let (M, g) be an n-dimensional compact shrinking Ricci soliton.
Then (M, g) is Einstein if and only if
2 1
Rmax − nλ ≤ 1+ | f |2 .
n vol (M, g) M
Theorem (E. García Río and MFL, 2010)
Let (M, g) be an n-dimensional compact shrinking Ricci soliton.
Then (M, g) is Einstein if and only if
−Λ + Λ2 + 8(n − 1)λΛ
|Ric − λg| ≤ c ≤ ,
4(n − 1)
1
where Λ = vol(M,g) M | f |2 denotes the average of the L2 -norm
of | f |.
57. (i) (∆f )2 = ((n + 2)λ − R) | f |2
M M
(ii) | f |2 ≤ Rmax − R
(∆f )2 = (n + 2)λ | f |2 − R| f |2
M M M
≥ (n + 2)λ | f |2 − nλRmax vol (M, g) + R2
M M
= (n + 2)λ | f |2 − nλRmax vol (M, g)
M
+n2 λ2 vol (M, g) + M (∆f )
2
n+2 1
Rmax − nλ ≥ | f |2
n vol (M, g) M
2λf − R = | f |2 = Rmax − R ⇒ f is constant
58. (i) (∆f )2 = ((n + 2)λ − R) | f |2
M M
(ii) | f |2 ≤ Rmax − R
(∆f )2 = (n + 2)λ | f |2 − R| f |2
M M M
≥ (n + 2)λ | f |2 − nλRmax vol (M, g) + R2
M M
= (n + 2)λ | f |2 − nλRmax vol (M, g)
M
+n2 λ2 vol (M, g) + M (∆f )
2
n+2 1
Rmax − nλ ≥ | f |2
n vol (M, g) M
2λf − R = | f |2 = Rmax − R ⇒ f is constant
59. (i) (∆f )2 = ((n + 2)λ − R) | f |2
M M
(ii) | f |2 ≤ Rmax − R
(∆f )2 = (n + 2)λ | f |2 − R| f |2
M M M
≥ (n + 2)λ | f |2 − nλRmax vol (M, g) + R2
M M
= (n + 2)λ | f |2 − nλRmax vol (M, g)
M
+n2 λ2 vol (M, g) + M (∆f )
2
n+2 1
Rmax − nλ ≥ | f |2
n vol (M, g) M
2λf − R = | f |2 = Rmax − R ⇒ f is constant
60. (i) (∆f )2 = ((n + 2)λ − R) | f |2
M M
(ii) | f |2 ≤ Rmax − R
(∆f )2 = (n + 2)λ | f |2 − R| f |2
M M M
≥ (n + 2)λ | f |2 − nλRmax vol (M, g) + R2
M M
= (n + 2)λ | f |2 − nλRmax vol (M, g)
M
+n2 λ2 vol (M, g) + M (∆f )
2
n+2 1
Rmax − nλ ≥ | f |2
n vol (M, g) M
2λf − R = | f |2 = Rmax − R ⇒ f is constant
61. Outline
Rigidity of Ricci solitons
Rigidity: compact case
Rigidity: non-compact case
Locally conformally flat case
Gap theorems
Diameter bounds
Gap theorems: compact case
Gap theorems: non-compact case
Maximum principles
Introduction
Omori-Yau maximum principle
Applications
Steady solitons
Lower bound for the curvature of a steady soliton
62. Theorem (E. García Río and MFL, 2010)
Let (M, g) be a complete gradient shrinking Ricci soliton with
bounded scalar curvature. Then (M, g) is compact Einstein if
Ric( f , f ) ≥ g( f , f ),
r (x)2
for sufficiently large r (x), where > 0 and r (x) denotes the
distance from a fixed point.
Theorem (E. García Río and MFL, 2010)
Let (M, g) be an n-dimensional complete gradient steady Ricci
soliton. If
Ric( f , f ) ≥ g( f , f ),
where is any positive constant, then (M, g) is Ricci flat.
63. Theorem (E. García Río and MFL, 2010)
Let (M, g) be a complete gradient shrinking Ricci soliton with
bounded scalar curvature. Then (M, g) is compact Einstein if
Ric( f , f ) ≥ g( f , f ),
r (x)2
for sufficiently large r (x), where > 0 and r (x) denotes the
distance from a fixed point.
Theorem (E. García Río and MFL, 2010)
Let (M, g) be an n-dimensional complete gradient steady Ricci
soliton. If
Ric( f , f ) ≥ g( f , f ),
where is any positive constant, then (M, g) is Ricci flat.
64. Theorem (P. Li)
If a complete manifold has Ricci curvature bounded from below
by r (x)−2 , for some constant > 1/4 and all r (x) > 1, then M
must be compact.
2λf = R + | f |2
There exists c such that f (x) ≥ 1 (r (x) − c)2
4
H.-D. Cao and D. Zhou; On complete gradient shrinking
solitons
γ : [0, +∞) → M an integral curve of f (note that f is a
complete vector field)
Z.-H. Zhang; On the completeness of gradient Ricci solitons
65. Theorem (P. Li)
If a complete manifold has Ricci curvature bounded from below
by r (x)−2 , for some constant > 1/4 and all r (x) > 1, then M
must be compact.
2λf = R + | f |2
There exists c such that f (x) ≥ 1 (r (x) − c)2
4
H.-D. Cao and D. Zhou; On complete gradient shrinking
solitons
γ : [0, +∞) → M an integral curve of f (note that f is a
complete vector field)
Z.-H. Zhang; On the completeness of gradient Ricci solitons
66. Theorem (P. Li)
If a complete manifold has Ricci curvature bounded from below
by r (x)−2 , for some constant > 1/4 and all r (x) > 1, then M
must be compact.
2λf = R + | f |2
There exists c such that f (x) ≥ 1 (r (x) − c)2
4
H.-D. Cao and D. Zhou; On complete gradient shrinking
solitons
γ : [0, +∞) → M an integral curve of f (note that f is a
complete vector field)
Z.-H. Zhang; On the completeness of gradient Ricci solitons
67. Theorem (P. Li)
If a complete manifold has Ricci curvature bounded from below
by r (x)−2 , for some constant > 1/4 and all r (x) > 1, then M
must be compact.
2λf = R + | f |2
There exists c such that f (x) ≥ 1 (r (x) − c)2
4
H.-D. Cao and D. Zhou; On complete gradient shrinking
solitons
γ : [0, +∞) → M an integral curve of f (note that f is a
complete vector field)
Z.-H. Zhang; On the completeness of gradient Ricci solitons
68. For r ≥ r1
2
(R ◦ γ) (t) = 2Ric( f (γ(t)), f (γ(t)) ≥ | f |2
r (x)2
Since R is bounded, for some k1 > 0 and k2 > 0
| f |2 = 2λf − R ≥ λ2 (r (x) − c)2 − R ≥ k1 r (x)2 ≥ k2 R 2 r (x)2
for r (x) ≥ r2 ≥ r1
p ∈ Ω = {x ∈ M / 2λf (x) ≥ k3 } such that B(x0 , r2 ) ⊂ Ω
Since f is increasing along the integral curves of f , if we
suppose that γ(0) = p, then γ(t) ∈ M Ω for all t ≥ 0
69. For r ≥ r1
2
(R ◦ γ) (t) = 2Ric( f (γ(t)), f (γ(t)) ≥ | f |2
r (x)2
Since R is bounded, for some k1 > 0 and k2 > 0
| f |2 = 2λf − R ≥ λ2 (r (x) − c)2 − R ≥ k1 r (x)2 ≥ k2 R 2 r (x)2
for r (x) ≥ r2 ≥ r1
p ∈ Ω = {x ∈ M / 2λf (x) ≥ k3 } such that B(x0 , r2 ) ⊂ Ω
Since f is increasing along the integral curves of f , if we
suppose that γ(0) = p, then γ(t) ∈ M Ω for all t ≥ 0
70. For r ≥ r1
2
(R ◦ γ) (t) = 2Ric( f (γ(t)), f (γ(t)) ≥ | f |2
r (x)2
Since R is bounded, for some k1 > 0 and k2 > 0
| f |2 = 2λf − R ≥ λ2 (r (x) − c)2 − R ≥ k1 r (x)2 ≥ k2 R 2 r (x)2
for r (x) ≥ r2 ≥ r1
p ∈ Ω = {x ∈ M / 2λf (x) ≥ k3 } such that B(x0 , r2 ) ⊂ Ω
Since f is increasing along the integral curves of f , if we
suppose that γ(0) = p, then γ(t) ∈ M Ω for all t ≥ 0
71. For r ≥ r1
2
(R ◦ γ) (t) = 2Ric( f (γ(t)), f (γ(t)) ≥ | f |2
r (x)2
Since R is bounded, for some k1 > 0 and k2 > 0
| f |2 = 2λf − R ≥ λ2 (r (x) − c)2 − R ≥ k1 r (x)2 ≥ k2 R 2 r (x)2
for r (x) ≥ r2 ≥ r1
p ∈ Ω = {x ∈ M / 2λf (x) ≥ k3 } such that B(x0 , r2 ) ⊂ Ω
Since f is increasing along the integral curves of f , if we
suppose that γ(0) = p, then γ(t) ∈ M Ω for all t ≥ 0
72. We have that
(R ◦ γ) (t) = g( R(γ(t)), γ (t)) ≥ 2 k2 R 2 (γ(t)),
along γ
t t
(R ◦ γ) (t)
ds ≥ 2 k2 dt
0 (R ◦ γ)2 (t) 0
1 1
− ≥ 2 k2 t
R(γ(0)) R(γ(t))
Contradiction for t going to infinite.
73. We have that
(R ◦ γ) (t) = g( R(γ(t)), γ (t)) ≥ 2 k2 R 2 (γ(t)),
along γ
t t
(R ◦ γ) (t)
ds ≥ 2 k2 dt
0 (R ◦ γ)2 (t) 0
1 1
− ≥ 2 k2 t
R(γ(0)) R(γ(t))
Contradiction for t going to infinite.
74. We have that
(R ◦ γ) (t) = g( R(γ(t)), γ (t)) ≥ 2 k2 R 2 (γ(t)),
along γ
t t
(R ◦ γ) (t)
ds ≥ 2 k2 dt
0 (R ◦ γ)2 (t) 0
1 1
− ≥ 2 k2 t
R(γ(0)) R(γ(t))
Contradiction for t going to infinite.
75. We have that
(R ◦ γ) (t) = g( R(γ(t)), γ (t)) ≥ 2 k2 R 2 (γ(t)),
along γ
t t
(R ◦ γ) (t)
ds ≥ 2 k2 dt
0 (R ◦ γ)2 (t) 0
1 1
− ≥ 2 k2 t
R(γ(0)) R(γ(t))
Contradiction for t going to infinite.
76. Outline
Rigidity of Ricci solitons
Rigidity: compact case
Rigidity: non-compact case
Locally conformally flat case
Gap theorems
Diameter bounds
Gap theorems: compact case
Gap theorems: non-compact case
Maximum principles
Introduction
Omori-Yau maximum principle
Applications
Steady solitons
Lower bound for the curvature of a steady soliton
77. A Riemannian manifold (M, g) is said to satisfy the Omori-Yau
maximum principle if given any function u ∈ C 2 (M) with
u ∗ = supM u < +∞, there exists a sequence (xk ) of points on
M satisfying
1 1 1
i) u(xk ) > u ∗ − , ii) |( u)(xk )| < , iii) (∆u)(xk ) < ,
k k k
for each k ∈ N. If, instead of iii) we assume that
1
Hu (xk ) < g,
k
in the sense of quadratic forms, then it is said that the
Riemannian manifold satisfies the Omori-Yau maximum
principle for the Hessian.
The f -Laplacian is
∆f = ef div (e−f ) = ∆ − g( f , ·)
78. A Riemannian manifold (M, g) is said to satisfy the Omori-Yau
maximum principle if given any function u ∈ C 2 (M) with
u ∗ = supM u < +∞, there exists a sequence (xk ) of points on
M satisfying
1 1 1
i) u(xk ) > u ∗ − , ii) |( u)(xk )| < , iii) (∆u)(xk ) < ,
k k k
for each k ∈ N. If, instead of iii) we assume that
1
Hu (xk ) < g,
k
in the sense of quadratic forms, then it is said that the
Riemannian manifold satisfies the Omori-Yau maximum
principle for the Hessian.
The f -Laplacian is
∆f = ef div (e−f ) = ∆ − g( f , ·)
79. In 1967 Omori showed that the Omori-Yau maximum principle
for the Hessian is satisfied by Riemannian manifolds with
curvature bounded from below.
H. Omori; Isometric immersions of Riemannian manifolds
In 1975 Yau proved that the Omori-Yau maximum principle is
satisfied by Riemannian manifolds with Ricci curvature
bounded from below.
S. T. Yau; Harmonic functions on complete Riemannian
manifolds
From now on we will work with Ricci solitons normalized in the
sense
1
Rc + Hf = ± g
2
80. In 1967 Omori showed that the Omori-Yau maximum principle
for the Hessian is satisfied by Riemannian manifolds with
curvature bounded from below.
H. Omori; Isometric immersions of Riemannian manifolds
In 1975 Yau proved that the Omori-Yau maximum principle is
satisfied by Riemannian manifolds with Ricci curvature
bounded from below.
S. T. Yau; Harmonic functions on complete Riemannian
manifolds
From now on we will work with Ricci solitons normalized in the
sense
1
Rc + Hf = ± g
2
81. In 1967 Omori showed that the Omori-Yau maximum principle
for the Hessian is satisfied by Riemannian manifolds with
curvature bounded from below.
H. Omori; Isometric immersions of Riemannian manifolds
In 1975 Yau proved that the Omori-Yau maximum principle is
satisfied by Riemannian manifolds with Ricci curvature
bounded from below.
S. T. Yau; Harmonic functions on complete Riemannian
manifolds
From now on we will work with Ricci solitons normalized in the
sense
1
Rc + Hf = ± g
2
82. Outline
Rigidity of Ricci solitons
Rigidity: compact case
Rigidity: non-compact case
Locally conformally flat case
Gap theorems
Diameter bounds
Gap theorems: compact case
Gap theorems: non-compact case
Maximum principles
Introduction
Omori-Yau maximum principle
Applications
Steady solitons
Lower bound for the curvature of a steady soliton
83. Theorem (E. García Río and MFL, 2010)
Let (M n , g) be an n-dimensional complete noncompact gradient
shrinking Ricci soliton. Then (M, g) satisfies the Omori-Yau
maximum principle.
Moreover, if there exists C > 0 such that Ric ≥ −Cr (x)2 , where
r (x) denotes the distance to a fixed point, then the Omori-Yau
maximum principle for the Hessian holds on (M, g).
Theorem (E. García Río and MFL, 2010)
Let (M n , g) be an n-dimensional complete noncompact gradient
shrinking Ricci soliton. Then (M, g) satisfies the Omori-Yau
maximum principle for the f -Laplacian.
84. Theorem (E. García Río and MFL, 2010)
Let (M n , g) be an n-dimensional complete noncompact gradient
shrinking Ricci soliton. Then (M, g) satisfies the Omori-Yau
maximum principle.
Moreover, if there exists C > 0 such that Ric ≥ −Cr (x)2 , where
r (x) denotes the distance to a fixed point, then the Omori-Yau
maximum principle for the Hessian holds on (M, g).
Theorem (E. García Río and MFL, 2010)
Let (M n , g) be an n-dimensional complete noncompact gradient
shrinking Ricci soliton. Then (M, g) satisfies the Omori-Yau
maximum principle for the f -Laplacian.
85. S. Pigola, M. Rigoli and A. Setti; Maximum principles on
Riemannian manifolds and applications
(M, g) satisfies the Omori-Yau maximum principle if ∃0 ≤ ϕ ∈ C 2 , s. t.
ϕ(x) −→ +∞ as x −→ ∞, (1)
√
∃A < 0 such that | ϕ| ≤ A ϕ off a compact set, and (2)
√ √
∃B > 0 s. t. ∆ϕ ≤ B ϕ G( ϕ), off a compact set, (3)
where G is a smooth function on [0, +∞) satisfying
i) G(0) > 0, ii) G (t) ≥ 0, on [0, +∞),
√
∞
dt tG( t) (4)
iii) = ∞, iv ) lim sup < ∞.
0 G(t) t→∞ G(t)
√ √
∃B > 0 s. t. Hϕ ≤ B ϕ G( ϕ), off a compact set (5)
(M, g) satisfies the Omori-Yau maximum principle for Hessian.
86. S. Pigola, M. Rigoli and A. Setti; Maximum principles on
Riemannian manifolds and applications
(M, g) satisfies the Omori-Yau maximum principle if ∃0 ≤ ϕ ∈ C 2 , s. t.
ϕ(x) −→ +∞ as x −→ ∞, (1)
√
∃A < 0 such that | ϕ| ≤ A ϕ off a compact set, and (2)
√ √
∃B > 0 s. t. ∆ϕ ≤ B ϕ G( ϕ), off a compact set, (3)
where G is a smooth function on [0, +∞) satisfying
i) G(0) > 0, ii) G (t) ≥ 0, on [0, +∞),
√
∞
dt tG( t) (4)
iii) = ∞, iv ) lim sup < ∞.
0 G(t) t→∞ G(t)
√ √
∃B > 0 s. t. Hϕ ≤ B ϕ G( ϕ), off a compact set (5)
(M, g) satisfies the Omori-Yau maximum principle for Hessian.
87. Theorem (S. Pigola, M. Rimoldi and A. Setti, 2009)
Let (M, g) be a complete gradient (shrinking, steady or
expanding) Ricci solitons. Then, the weak maximum principle
at infinity for the f -Laplacian holds.
Given any function u ∈ C 2 (M) with u ∗ = supM u < +∞, there
exists a sequence (xk ) of points on M satisfying
1 1
i) u(xk ) > u ∗ − , ii) (∆f u)(xk ) < ,
k k
for each k ∈ N.
Theorem (S. Pigola, M. Rimoldi and A. Setti, 2009)
Let (M, g) be a complete gradient (shrinking, steady or
expanding) Ricci solitons. Then it is stochastically complete.
88. Theorem (S. Pigola, M. Rimoldi and A. Setti, 2009)
Let (M, g) be a complete gradient (shrinking, steady or
expanding) Ricci solitons. Then, the weak maximum principle
at infinity for the f -Laplacian holds.
Given any function u ∈ C 2 (M) with u ∗ = supM u < +∞, there
exists a sequence (xk ) of points on M satisfying
1 1
i) u(xk ) > u ∗ − , ii) (∆f u)(xk ) < ,
k k
for each k ∈ N.
Theorem (S. Pigola, M. Rimoldi and A. Setti, 2009)
Let (M, g) be a complete gradient (shrinking, steady or
expanding) Ricci solitons. Then it is stochastically complete.
89. Theorem (S. Pigola, M. Rimoldi and A. Setti, 2009)
Let (M, g) be a complete gradient (shrinking, steady or
expanding) Ricci solitons. Then, the weak maximum principle
at infinity for the f -Laplacian holds.
Given any function u ∈ C 2 (M) with u ∗ = supM u < +∞, there
exists a sequence (xk ) of points on M satisfying
1 1
i) u(xk ) > u ∗ − , ii) (∆f u)(xk ) < ,
k k
for each k ∈ N.
Theorem (S. Pigola, M. Rimoldi and A. Setti, 2009)
Let (M, g) be a complete gradient (shrinking, steady or
expanding) Ricci solitons. Then it is stochastically complete.
90. Outline
Rigidity of Ricci solitons
Rigidity: compact case
Rigidity: non-compact case
Locally conformally flat case
Gap theorems
Diameter bounds
Gap theorems: compact case
Gap theorems: non-compact case
Maximum principles
Introduction
Omori-Yau maximum principle
Applications
Steady solitons
Lower bound for the curvature of a steady soliton
91. Theorem (E. García Río and MFL, 2010)
Let (M n , g) be an n-dimensional complete gradient shrinking
Ricci soliton. Then:
(i) (M, g) has constant scalar curvature if and only if
| R|2
2|Ric|2 ≤ R + c , for some c ≥ 0.
R+1
(ii) (M, g) is isometric to (Rn , geuc ) if and only if
| R|2
2|Ric|2 ≤ (1 − )R + c , for some c ≥ 0 and > 0.
R+1
92. Theorem (E. García Río and MFL, 2010)
Let (M, g) be a complete gradient shrinking Ricci soliton with
bounded nonnegative Ricci tensor. Then (M, g) is rigid if and
only if the sectional curvature is bounded from above by
|Ric|2
2(R 2 −|Ric|2 )
.
We consider an orthonormal frame {E1 , . . . , En } formed by
eigenvectors of the Ricci operator.
∆Rii = g( Rii , f ) + Rii − 2Rijji R jj ,
where Rii = Ric(Ei , Ei ), Rijji = R(Ei , Ej , Ej , Ei )
∆f |Ric|2 = 2|Ric|2 −4R ii Rijji R jj +2 Rii R ii ≥ 2|Ric|2 −4R ii Rijji R jj .
93. Theorem (E. García Río and MFL, 2010)
Let (M, g) be a complete gradient shrinking Ricci soliton with
bounded nonnegative Ricci tensor. Then (M, g) is rigid if and
only if the sectional curvature is bounded from above by
|Ric|2
2(R 2 −|Ric|2 )
.
We consider an orthonormal frame {E1 , . . . , En } formed by
eigenvectors of the Ricci operator.
∆Rii = g( Rii , f ) + Rii − 2Rijji R jj ,
where Rii = Ric(Ei , Ei ), Rijji = R(Ei , Ej , Ej , Ei )
∆f |Ric|2 = 2|Ric|2 −4R ii Rijji R jj +2 Rii R ii ≥ 2|Ric|2 −4R ii Rijji R jj .
94. Theorem (E. García Río and MFL, 2010)
Let (M, g) be a complete gradient shrinking Ricci soliton with
bounded nonnegative Ricci tensor. Then (M, g) is rigid if and
only if the sectional curvature is bounded from above by
|Ric|2
2(R 2 −|Ric|2 )
.
We consider an orthonormal frame {E1 , . . . , En } formed by
eigenvectors of the Ricci operator.
∆Rii = g( Rii , f ) + Rii − 2Rijji R jj ,
where Rii = Ric(Ei , Ei ), Rijji = R(Ei , Ej , Ej , Ei )
∆f |Ric|2 = 2|Ric|2 −4R ii Rijji R jj +2 Rii R ii ≥ 2|Ric|2 −4R ii Rijji R jj .
95. Theorem (E. García Río and MFL, 2010)
Let (M, g) be a complete gradient shrinking Ricci soliton with
bounded nonnegative Ricci tensor. Then (M, g) is rigid if and
only if the sectional curvature is bounded from above by
|Ric|2
2(R 2 −|Ric|2 )
.
We consider an orthonormal frame {E1 , . . . , En } formed by
eigenvectors of the Ricci operator.
∆Rii = g( Rii , f ) + Rii − 2Rijji R jj ,
where Rii = Ric(Ei , Ei ), Rijji = R(Ei , Ej , Ej , Ei )
∆f |Ric|2 = 2|Ric|2 −4R ii Rijji R jj +2 Rii R ii ≥ 2|Ric|2 −4R ii Rijji R jj .
96. Under our assumption one has
4|Ric|2
4 R ii Rijji R jj ≤ (R 2 − |Ric|2 ) = 2|Ric|2 .
2(R 2 − |Ric|2 )
Then ∆f |Ric|2 ≥ 0, and by f -parabolicity it follows that |Ric|2 is
constant.
0 = ∆f |Ric|2 = 2|Ric|2 − 4R ii Rijji R jj + 2 Rii R ii ≥ 0
n
⇒ | Rii |2 = 0
i=1
The Ricci soliton is rigid.
97. Under our assumption one has
4|Ric|2
4 R ii Rijji R jj ≤ (R 2 − |Ric|2 ) = 2|Ric|2 .
2(R 2 − |Ric|2 )
Then ∆f |Ric|2 ≥ 0, and by f -parabolicity it follows that |Ric|2 is
constant.
0 = ∆f |Ric|2 = 2|Ric|2 − 4R ii Rijji R jj + 2 Rii R ii ≥ 0
n
⇒ | Rii |2 = 0
i=1
The Ricci soliton is rigid.
98. Under our assumption one has
4|Ric|2
4 R ii Rijji R jj ≤ (R 2 − |Ric|2 ) = 2|Ric|2 .
2(R 2 − |Ric|2 )
Then ∆f |Ric|2 ≥ 0, and by f -parabolicity it follows that |Ric|2 is
constant.
0 = ∆f |Ric|2 = 2|Ric|2 − 4R ii Rijji R jj + 2 Rii R ii ≥ 0
n
⇒ | Rii |2 = 0
i=1
The Ricci soliton is rigid.
99. Under our assumption one has
4|Ric|2
4 R ii Rijji R jj ≤ (R 2 − |Ric|2 ) = 2|Ric|2 .
2(R 2 − |Ric|2 )
Then ∆f |Ric|2 ≥ 0, and by f -parabolicity it follows that |Ric|2 is
constant.
0 = ∆f |Ric|2 = 2|Ric|2 − 4R ii Rijji R jj + 2 Rii R ii ≥ 0
n
⇒ | Rii |2 = 0
i=1
The Ricci soliton is rigid.
100. Theorem (E. García Río and MFL, 2010)
Let (M, g) be a complete gradient steady Ricci soliton. Then
R∗ = infM R = 0.
∆f R = −2|Ric|2 .
There exists a sequence (xk ) of points of M such that
1 1
R(xk ) < R∗ + k and (∆f R)(xk ) > − k .
1 2R(xk )2
≥ 2|Ric(xk )|2 ≥ .
k n
Taking the limit when k goes to infinity we get that R∗ = 0.
101. Theorem (E. García Río and MFL, 2010)
Let (M, g) be a complete gradient steady Ricci soliton. Then
R∗ = infM R = 0.
∆f R = −2|Ric|2 .
There exists a sequence (xk ) of points of M such that
1 1
R(xk ) < R∗ + k and (∆f R)(xk ) > − k .
1 2R(xk )2
≥ 2|Ric(xk )|2 ≥ .
k n
Taking the limit when k goes to infinity we get that R∗ = 0.
102. Theorem (E. García Río and MFL, 2010)
Let (M, g) be a complete gradient steady Ricci soliton. Then
R∗ = infM R = 0.
∆f R = −2|Ric|2 .
There exists a sequence (xk ) of points of M such that
1 1
R(xk ) < R∗ + k and (∆f R)(xk ) > − k .
1 2R(xk )2
≥ 2|Ric(xk )|2 ≥ .
k n
Taking the limit when k goes to infinity we get that R∗ = 0.
103. Theorem (E. García Río and MFL, 2010)
Let (M, g) be a complete gradient steady Ricci soliton. Then
R∗ = infM R = 0.
∆f R = −2|Ric|2 .
There exists a sequence (xk ) of points of M such that
1 1
R(xk ) < R∗ + k and (∆f R)(xk ) > − k .
1 2R(xk )2
≥ 2|Ric(xk )|2 ≥ .
k n
Taking the limit when k goes to infinity we get that R∗ = 0.
104. Theorem (E. García Río and MFL, 2010)
Let (M, g) be a complete gradient steady Ricci soliton. Then
R∗ = infM R = 0.
∆f R = −2|Ric|2 .
There exists a sequence (xk ) of points of M such that
1 1
R(xk ) < R∗ + k and (∆f R)(xk ) > − k .
1 2R(xk )2
≥ 2|Ric(xk )|2 ≥ .
k n
Taking the limit when k goes to infinity we get that R∗ = 0.
105. Theorem (E. García Río and MFL, 2010)
Let (M, g) be a complete gradient expanding Ricci soliton with
Ric ≤ 0. If R ∗ = supM R < 0 then − n ≤ R ≤ − 1 .
2 2
∆f R = −R − 2|Ric|2 ≥ −R − 2R 2 = −R(1 + 2R)
1
R ∗ (1 + 2R ∗ ) ≥ 0 ⇒ R ∗ ≤ −
2
106. Theorem (E. García Río and MFL, 2010)
Let (M, g) be a complete gradient expanding Ricci soliton with
Ric ≤ 0. If R ∗ = supM R < 0 then − n ≤ R ≤ − 1 .
2 2
∆f R = −R − 2|Ric|2 ≥ −R − 2R 2 = −R(1 + 2R)
1
R ∗ (1 + 2R ∗ ) ≥ 0 ⇒ R ∗ ≤ −
2
107. Theorem (E. García Río and MFL, 2010)
Let (M, g) be a complete gradient expanding Ricci soliton with
Ric ≤ 0. If R ∗ = supM R < 0 then − n ≤ R ≤ − 1 .
2 2
∆f R = −R − 2|Ric|2 ≥ −R − 2R 2 = −R(1 + 2R)
1
R ∗ (1 + 2R ∗ ) ≥ 0 ⇒ R ∗ ≤ −
2
108. Outline
Rigidity of Ricci solitons
Rigidity: compact case
Rigidity: non-compact case
Locally conformally flat case
Gap theorems
Diameter bounds
Gap theorems: compact case
Gap theorems: non-compact case
Maximum principles
Introduction
Omori-Yau maximum principle
Applications
Steady solitons
Lower bound for the curvature of a steady soliton
109. Theorem
Let (M n , g, f ) be a complete noncompact nonflat shrinking
gradient Ricci soliton. Then for any given point O ∈ M there
−1
exists a constant CO > 0 such that R(x)d(x, O)2 ≥ CO
wherever d(x, O) ≥ CO .
B. Chow, P. Lu and B. Yang; A lower bound for the scalar
curvature of noncompact nonflat Ricci shrinkers
Theorem
Let (M n , g, f ) be a complete steady gradient Ricci solitons with
Rc = −Hf and R + | f |2 = 1. If limx→∞ f (x) = −∞ and f ≤ 0,
1
then R ≥ √ n ef .
2
+2
B. Chow, P. Lu and B. Yang; A lower bound for the scalar
curvature of certain steady gradient Ricci solitons
110. Theorem
Let (M n , g, f ) be a complete noncompact nonflat shrinking
gradient Ricci soliton. Then for any given point O ∈ M there
−1
exists a constant CO > 0 such that R(x)d(x, O)2 ≥ CO
wherever d(x, O) ≥ CO .
B. Chow, P. Lu and B. Yang; A lower bound for the scalar
curvature of noncompact nonflat Ricci shrinkers
Theorem
Let (M n , g, f ) be a complete steady gradient Ricci solitons with
Rc = −Hf and R + | f |2 = 1. If limx→∞ f (x) = −∞ and f ≤ 0,
1
then R ≥ √ n ef .
2
+2
B. Chow, P. Lu and B. Yang; A lower bound for the scalar
curvature of certain steady gradient Ricci solitons
111. Theorem (E. García Río and MFL, 2011)
Let (M, g) be a complete gradient steady Ricci soliton satisfying
2
|Ric|2 ≤ R . Then
2
r (x)
R(x) ≥ k sech2 ,
2
where r (x) is the distance from a fixed point O ∈ M and k ≤ 1
is a constant that only depends on O and R(O).
√ | R|2
|Hf |2 = | f |2 ≥ | | f ||2 = | 1 − R|2 =
4| f |2
| R|2 ≤ 4|Hf |2 | f |2
R2 | R|
|Hf |2 = |Rc|2 ≤ ⇒ √ ≤1
2 R 1−R
112. Theorem (E. García Río and MFL, 2011)
Let (M, g) be a complete gradient steady Ricci soliton satisfying
2
|Ric|2 ≤ R . Then
2
r (x)
R(x) ≥ k sech2 ,
2
where r (x) is the distance from a fixed point O ∈ M and k ≤ 1
is a constant that only depends on O and R(O).
√ | R|2
|Hf |2 = | f |2 ≥ | | f ||2 = | 1 − R|2 =
4| f |2
| R|2 ≤ 4|Hf |2 | f |2
R2 | R|
|Hf |2 = |Rc|2 ≤ ⇒ √ ≤1
2 R 1−R
113. Theorem (E. García Río and MFL, 2011)
Let (M, g) be a complete gradient steady Ricci soliton satisfying
2
|Ric|2 ≤ R . Then
2
r (x)
R(x) ≥ k sech2 ,
2
where r (x) is the distance from a fixed point O ∈ M and k ≤ 1
is a constant that only depends on O and R(O).
√ | R|2
|Hf |2 = | f |2 ≥ | | f ||2 = | 1 − R|2 =
4| f |2
| R|2 ≤ 4|Hf |2 | f |2
R2 | R|
|Hf |2 = |Rc|2 ≤ ⇒ √ ≤1
2 R 1−R
114. Theorem (E. García Río and MFL, 2011)
Let (M, g) be a complete gradient steady Ricci soliton satisfying
2
|Ric|2 ≤ R . Then
2
r (x)
R(x) ≥ k sech2 ,
2
where r (x) is the distance from a fixed point O ∈ M and k ≤ 1
is a constant that only depends on O and R(O).
√ | R|2
|Hf |2 = | f |2 ≥ | | f ||2 = | 1 − R|2 =
4| f |2
| R|2 ≤ 4|Hf |2 | f |2
R2 | R|
|Hf |2 = |Rc|2 ≤ ⇒ √ ≤1
2 R 1−R
115. −(R◦γ)
Integrating √
R 1−R
along a minimizing geodesic γ(s)
√ t l t
1+ 1−R (R ◦ γ) | R|
ln √ =− √ ds ≤ √ ds ≤ t
1− 1−R 0 0 R 1−R 0 R 1−R
√
1+ 1−R(O)
Writing c = √ we get that
1− 1−R(O)
1+ 1 − R(γ(t)) ≤ cet (1 − 1 − R(γ(t)))
4c
R(γ(t)) ≥
c 2 et + 2c + e−t
Since c ≥ 1 we have that
4c 4c 1 t
R(γ(t)) ≥ −t
≥ 2 t 2 + c 2 e−t
= sech2
c 2 et + 2c + e c e + 2c c 2
116. −(R◦γ)
Integrating √
R 1−R
along a minimizing geodesic γ(s)
√ t l t
1+ 1−R (R ◦ γ) | R|
ln √ =− √ ds ≤ √ ds ≤ t
1− 1−R 0 0 R 1−R 0 R 1−R
√
1+ 1−R(O)
Writing c = √ we get that
1− 1−R(O)
1+ 1 − R(γ(t)) ≤ cet (1 − 1 − R(γ(t)))
4c
R(γ(t)) ≥
c 2 et + 2c + e−t
Since c ≥ 1 we have that
4c 4c 1 t
R(γ(t)) ≥ −t
≥ 2 t 2 + c 2 e−t
= sech2
c 2 et + 2c + e c e + 2c c 2
117. −(R◦γ)
Integrating √
R 1−R
along a minimizing geodesic γ(s)
√ t l t
1+ 1−R (R ◦ γ) | R|
ln √ =− √ ds ≤ √ ds ≤ t
1− 1−R 0 0 R 1−R 0 R 1−R
√
1+ 1−R(O)
Writing c = √ we get that
1− 1−R(O)
1+ 1 − R(γ(t)) ≤ cet (1 − 1 − R(γ(t)))
4c
R(γ(t)) ≥
c 2 et + 2c + e−t
Since c ≥ 1 we have that
4c 4c 1 t
R(γ(t)) ≥ −t
≥ 2 t 2 + c 2 e−t
= sech2
c 2 et + 2c + e c e + 2c c 2
118. −(R◦γ)
Integrating √
R 1−R
along a minimizing geodesic γ(s)
√ t l t
1+ 1−R (R ◦ γ) | R|
ln √ =− √ ds ≤ √ ds ≤ t
1− 1−R 0 0 R 1−R 0 R 1−R
√
1+ 1−R(O)
Writing c = √ we get that
1− 1−R(O)
1+ 1 − R(γ(t)) ≤ cet (1 − 1 − R(γ(t)))
4c
R(γ(t)) ≥
c 2 et + 2c + e−t
Since c ≥ 1 we have that
4c 4c 1 t
R(γ(t)) ≥ −t
≥ 2 t 2 + c 2 e−t
= sech2
c 2 et + 2c + e c e + 2c c 2
119. The scalar curvature of Hamilton’s cigar soliton
dx 2 + dy 2
R2 ,
1 + x2 + y2
satisfies
R(x) = 4sech2 r (x)
The scalar curvature of normalized Hamilton’s cigar soliton
4(dx 2 + dy 2 )
R2 ,
1 + x2 + y2
satisfies
r (x)
R(x) = sech2
2
Our inequality is SHARP
120. The scalar curvature of Hamilton’s cigar soliton
dx 2 + dy 2
R2 ,
1 + x2 + y2
satisfies
R(x) = 4sech2 r (x)
The scalar curvature of normalized Hamilton’s cigar soliton
4(dx 2 + dy 2 )
R2 ,
1 + x2 + y2
satisfies
r (x)
R(x) = sech2
2
Our inequality is SHARP
121. The scalar curvature of Hamilton’s cigar soliton
dx 2 + dy 2
R2 ,
1 + x2 + y2
satisfies
R(x) = 4sech2 r (x)
The scalar curvature of normalized Hamilton’s cigar soliton
4(dx 2 + dy 2 )
R2 ,
1 + x2 + y2
satisfies
r (x)
R(x) = sech2
2
Our inequality is SHARP
122. Theorem (E. García Río and MFL, 2011)
Let (M, g) be a complete gradient steady Ricci soliton with
nonnegative Ricci curvature normalized as before. Then
R(x) ≥ k sech2 r (x),
where r (x) is the distance from a fixed point O ∈ M and k ≤ 1
is a constant that only depends on O and R(O).
| R|2 ≤ 4|Hf |2 | f |2
Since |Hf |2 = |Rc|2 ≤ R 2 one has
| R|
√ ≤2
R 1−R
123. Theorem (E. García Río and MFL, 2011)
Let (M, g) be a complete gradient steady Ricci soliton with
nonnegative Ricci curvature normalized as before. Then
R(x) ≥ k sech2 r (x),
where r (x) is the distance from a fixed point O ∈ M and k ≤ 1
is a constant that only depends on O and R(O).
| R|2 ≤ 4|Hf |2 | f |2
Since |Hf |2 = |Rc|2 ≤ R 2 one has
| R|
√ ≤2
R 1−R
124. Theorem (E. García Río and MFL, 2011)
Let (M, g) be a complete gradient steady Ricci soliton with
nonnegative Ricci curvature normalized as before. Then
R(x) ≥ k sech2 r (x),
where r (x) is the distance from a fixed point O ∈ M and k ≤ 1
is a constant that only depends on O and R(O).
| R|2 ≤ 4|Hf |2 | f |2
Since |Hf |2 = |Rc|2 ≤ R 2 one has
| R|
√ ≤2
R 1−R