Monatsh Math (2012) 166:525–542 DOI 10.1007/s00605-011-0332-2

Isoperimetric inequalities for submanifolds with bounded mean curvature Keomkyo Seo

Received: 18 October 2010 / Accepted: 8 July 2011 / Published online: 23 July 2011 © Springer-Verlag 2011

Abstract In this paper, we provide various Sobolev-type inequalities for smooth nonnegative functions with compact support on a submanifold with variable mean curvature in a Riemannian manifold whose sectional curvature is bounded above by a constant. We further obtain the corresponding linear isoperimetric inequalities involving mean curvature. We also provide various first Dirichlet eigenvalue estimates for submanifolds with bounded mean curvature. Keywords Isoperimetric inequality · Sobolev-type inequality · First eigenvalue · Mean curvature Mathematics Subject Classification (2000)

58E35 · 58C40

1 Introduction The classical isoperimetric inequality says that for a given bounded domain D in Rn , if ωn is the volume of a unit ball in Rn , then n n ωn Vol(D)n−1 ≤ Vol(∂ D)n and equality holds if and only if D is a ball. As an extension of the above classical isoperimetric inequality, it is conjectured that any n-dimensional compact minimal submanifold M of Rm still satisfies the above inequality. Moreover, equality holds Communicated by Andreas Cap. This research was supported by the Sookmyung Women’s University Research Grants 2011. K. Seo (B) Department of Mathematics, Sookmyung Women’s University, Hyochangwongil 52, Yongsan-ku, Seoul 140-742, Korea e-mail: [email protected]

123

526

K. Seo

if and only if M is an n-dimensional ball. When n = 2, there are several partial results ([9,16,18], see also [15] for more references). In particular, for any minimal surface  in Rn with at most two boundary components,  satisfies the following sharp isoperimetric inequality (see [6,12]): 4π Area() ≤ Length(∂)2 , where equality holds if and only if  is a disk. When n ≥ 3, the only known result was due to Almgren [2] who proved that the above conjecture is true for n-dimensional area-minimizing submanifolds in Rm . Even though the conjecture is still open, one may try to extend the classical isoperimetric inequality to more general submanifold with variable mean curvature vector in a Riemannian manifold. In this direction, Allard [1] proved an isoperimetric inequality which involves mean curvature term. More precisely, he showed that given n-dimensional compact submanifold M in Rm , ⎛ ⎞n  (1.1) c(n) Vol(M)n−1 ≤ ⎝Vol(∂ M) + |H |⎠ , M

where c(n) is a constant which depends only on n and H is the mean curvature vector of M in Rm . In [14], Michael and Simon also obtained the similar isoperimetric inequality by proving the corresponding Sobolev inequality on a submanifold in a Euclidean space. One year later, Hoffman and Spruck [10] generalized the inequality on a submanifold in a Riemannian manifold. In this paper, we study linear isoperimetric inequalities and the first (Dirichlet) eigenvalue estimates for submanifolds with variable mean curvature in a Riemannian manifold whose sectional curvature is bounded above. In Sect. 2, we derive some useful inequalities on the Laplacian of functions of distance by virtue of Hessian comparison theorem. For minimal surface cases, such inequalities were obtained by Choe [7]. We modify his arguments to obtain more general inequalities (Lemmas 2.3–2.5) for submanifolds (not necessarily minimal) with variable mean curvature. Let  be a domain in a hyperbolic space Hn or an n-dimensional minimal submanifold of Hm . Then the following linear isoperimetric inequality holds [8,19]: (n − 1)Vol() ≤ Vol(∂). In Sect. 3, we extend the above result and obtain more general linear isoperimetric inequalities for submanifolds in a complete simply connected Riemannian manifold with sectional curvature bounded above by a constant. We further obtain the corresponding Sobolev-type inequalities involving mean curvature for smooth nonnegative functions with compact support on submanifolds. In Sect. 4, we provide various first (Dirichlet) eigenvalue estimates for submanifolds with bounded mean curvature. The first (Dirichlet) eigenvalue of a Riemannian manifold  with boundary is characterized as  |∇ f |2 , λ1 = inf  2 f  f

123

Isoperimetric inequalities for submanifolds with bounded mean curvature

527

where the infimum is taken over all compactly supported smooth functions on . In [5], Cheung and Leung obtained the first eigenvalue estimate for a complete noncompact submanifold with bounded mean curvature in hyperbolic space. More precisely, for a complete minimal submanifold M n in Hm , they proved 1 (n − 1)2 ≤ λ1 (M). 4 Here this inequality is sharp because equality holds when M is totally geodesic [13]. We improve their estimate for a compact submanifold with bounded mean curvature in a complete simply connected Riemannian manifold with sectional curvature bounded above by a constant. 2 Laplacian comparisons of functions of distance Consider an n-dimensional submanifold  in an m-dimensional Riemannian manifold M. Let e1 , . . . , en be orthonormal vector fields on a neighborhood of a point of . These vector fields can be smoothly extended to orthonormal vector fields e¯1 , . . . , e¯m on M satisfying that e¯i = ei for 1 ≤ i ≤ n on . If H is the mean curvature vector of  in M, then H=

n 

∇¯ e¯i e¯i −

i=1

n 

∇ei ei ,

i=1

where ∇¯ and ∇ are the connections of M and , respectively. Thus for f ∈ C ∞ (M) we have f = =

n  (ei ei f − ∇ei ei f ) i=1 m  i=1

m 

e¯i e¯i f −

e¯α e¯α f −

α=n+1 m 

¯ f + Hf − =

m 

∇¯ e¯i e¯i f + H f +

m 

∇¯ e¯α e¯α f

α=n+1

i=1

∇¯ 2 f (e¯α , e¯α ),

α=n+1

¯ are the Laplacians on  and M, respectively. Hence we get the where  and  following lemma, which was proved by Choe and Gulliver [8] (see also [11]). Lemma 2.1 [8] Let f be a smooth function on an m-dimensional Riemannian manifold M and  an n-dimensional submanifold of M. Let e¯n+1 , . . . , e¯m be orthonormal vectors perpendicular to . Then for the mean curvature vector H of  in M, we have m  2 ¯ f )| + H, ∇¯ f | − ( f | ) = ( ∇¯ f (e¯α , e¯α ) , α=n+1



123

528

K. Seo

¯ are the connection and Laplacian on M respectively, and  is the where ∇¯ and  Laplacian on . In order to study a Riemannian manifold of variable curvature, we need the following useful lemma, which is so-called the H essian comparison theor em. Lemma 2.2 [17] Let M be a Riemannian manifold with the sectional curvature

be a space form of constant sectional curvature K sec ≤ K for some constant K and M

for K . Let r (·) = dist( p, ·) and rˆ (·) = dist( p, ˆ ·) be distance functions on M and M

respectively. Assume that γ is a geodesic from p to fixed points p ∈ M and pˆ ∈ M, q and v ∈ Tq M is a vector at q perpendicular to γ . Then we have ∇ 2 r (v, v) ≥ ∇ˆ 2 rˆ (u, u),

is a vector at qˆ with |u| = |v| and rˆ (q) ˆ = r (q), which is perpendicular where u ∈ Tqˆ M

from pˆ to q. to the geodesic

γ ⊂M ˆ As a consequence of this comparison, we have the following lemmas on the Laplacian of some functions of distance. Lemma 2.3 Let  be an n-dimensional submanifold in a complete simply connected Riemannian manifold M of nonpositive sectional curvature. Let ∇ and  be the connection and Laplacian on  respectively, and ∇¯ the connection on M. Define r (·) = dist( p, ·) for fixed p ∈ M. Then we have (a) (b)

¯ | r 2 ≥ 2n + 2r H, ∇r n−|∇r |2 ¯ | r ≥ + H, ∇r r

(c)

 log r ≥

n−2|∇r |2 r2

¯ | + r1 H, ∇r

of constant secProof Denote the metrics of M and an m-dimensional space form M

tional curvature K by g and g, ˆ respectively. We may assume that M = Rm . It is easy 2 2 ˆ ˆ Since to see that ∇ rˆ = 2g. ∇ 2 r = 2r ∇ 2 r + 2∇r ⊗ ∇r and ∇ˆ 2 rˆ 2 = 2ˆr ∇ˆ 2 rˆ + 2∇ˆ rˆ ⊗ ∇ˆ rˆ , Lemmas 2.1 and 2.2 imply ¯ 2 | = 2n + 2r H, ∇r ¯ | , ∇ 2 r ≥ 2n + H, ∇r which gives (a). For (b), we compute 1

r = div∇(r 2 ) 2 = div

123

r 2 1 ∇r 2 = − 2 ∇r, 2r ∇r  2r 2r 2r n − |∇r |2 ¯ | ≥ + H, ∇r r

Isoperimetric inequalities for submanifolds with bounded mean curvature

529

Similarly,  log r = div

∇r n − 2|∇r |2 1 r |∇r |2 ¯ | , + H, ∇r = − 2 ≥ r r r r2 r

which gives (c).

Lemma 2.4 Let  be an n-dimensional submanifold in a complete simply connected Riemannian manifold M with sectional curvature bounded above by a positive constant K = k 2 for k > 0. Let ∇ and  be the connection and Laplacian on  respectively, and ∇¯ the connection on M. Define r (·) = dist( p, ·) for fixed p ∈ M. π . Then we have Assume that r < 2k ¯ | (a) r ≥ k(n − |∇r |2 ) cot kr + H, ∇r 2 ¯ | (b)  sin kr ≥ sink kr (n cos2 kr − |∇r |2 ) + k cos kr H, ∇r 2 ¯ (c)  cos kr ≤ −nk cos kr − k sin kr H, ∇r | 2 |2 2 ¯ (d)  log sin kr ≥ nk 2 cot 2 kr − ksin|∇r 2 kr (1 + cos kr ) + k cot kr H, ∇r | 2 2 2 2 ¯ (e)  log cos kr ≤ −nk − k |∇r | tan kr − k tan kr H, ∇r |

of constant secProof Denote the metrics of M and an m-dimensional space form M tional curvature K by g and g, ˆ respectively. Assume finally that K = k 2 > 0 for some constant k > 0. As in the case of K < 0 above, we have ∇ˆ 2 rˆ = k cot k rˆ (gˆ − ∇ˆ rˆ ⊗ ∇ˆ rˆ )

of constant sectional curvature K . Using Lemmas 2.1 and 2.2 again, we get (a). in M For (b), we compute  sin kr = div(k cos kr ∇r ) = −k 2 sin kr |∇r |2 + k cos kr r ¯ | ] ≥ −k 2 sin kr |∇r |2 + k cos kr [k(n − |∇r |2 ) cot kr + H, ∇r 2 k ¯ | (n cos2 kr − |∇r |2 ) + k cos kr H, ∇r = sin kr Similarly,  cos kr = div(−k sin kr ∇r ) = −k 2 cos kr |∇r |2 − k sin kr r

¯ | ], ≤ −k 2 cos kr |∇r |2 − k sin kr [k(n − |∇r |2 ) cot kr + H, ∇r

which gives (c). For (d),

k cos kr ∇r  log sin kr = div sin kr

=−

k 2 |∇r |2 k cos kr r + 2 sin kr sin kr

k 2 |∇r |2 k cos kr ¯ | ) (k(n − |∇r |2 ) cot kr + H, ∇r + 2 sin kr sin kr k 2 |∇r |2 ¯ | . = nk 2 cot 2 kr − (1 + cos2 kr ) + k cot kr H, ∇r sin2 kr ≥−

123

530

K. Seo

For (e),



−k sin kr ∇r  log cos kr = div cos kr

=−

k 2 |∇r |2 k sin kr − r 2 cos kr cos kr

k 2 |∇r |2 k sin kr ¯ | ) − (k(n − |∇r |2 ) cot kr + H, ∇r 2 cos kr cos kr ¯ | , = −nk 2 − k 2 |∇r |2 tan2 kr − k tan kr H, ∇r ≤−

which completes the proof.



Lemma 2.5 Let  be an n-dimensional submanifold in a complete simply connected Riemannian manifold M with sectional curvature bounded above by a negative constant K = −k 2 for k > 0. Let ∇ and  be the connection and Laplacian on  respectively, and ∇¯ the connection on M. Define r (·) = dist( p, ·) for fixed p ∈ M. Then we have (a) (b) (c) (d) (e)

¯ | r ≥ k(n − |∇r |2 ) coth kr + H, ∇r k2 2 ¯ |  sinh kr ≥ sinh kr (n cosh kr − |∇r |2 ) + k cosh kr H, ∇r 2 ¯ |  cosh kr ≥ nk cosh kr + k sinh kr H, ∇r k 2 |∇r |2 2 2 ¯ |  log sinh kr ≥ nk coth kr − sinh2 kr (1 + cosh2 kr ) + k coth kr H, ∇r ¯ |  log cosh kr ≥ nk 2 − k 2 |∇r |2 tanh2 kr + k coth kr H, ∇r

of constant secProof Denote the metrics of M and an m-dimensional space form M tional curvature K by g and g, ˆ respectively. Assume now that K = −k 2 < 0 for some constant k > 0. It is well known that ∇ˆ 2 rˆ = k coth k rˆ (gˆ − ∇ˆ rˆ ⊗ ∇ˆ rˆ ). Thus (a) follows from Lemmas 2.1 and 2.2. For (b), we compute  sinh kr = div(k cosh kr ∇r ) = k 2 sinh kr |∇r |2 + k cosh kr r

¯ | ] ≥ k 2 sinh kr |∇r |2 + k cosh kr [k(n − |∇r |2 ) coth kr + H, ∇r k2 ¯ | (n cosh2 kr − |∇r |2 ) + k cosh kr H, ∇r = sinh kr

Similarly,  cosh kr = div(−k sinh kr ∇r ) = k 2 cosh kr |∇r |2 + k sinh kr r

¯ | ], ≥ k 2 cosh kr |∇r |2 + k sinh kr [k(n − |∇r |2 ) coth kr + H, ∇r

which gives (c). For (d),

k cosh kr (k 2 sinh2 kr −k 2 cosh2 kr )|∇r |2 k cosh kr  log sinh kr = div ∇r = r + sinh kr sinh kr sinh2 kr k 2 |∇r |2 k cosh kr ¯ | ) (k(n − |∇r |2 ) coth kr + H, ∇r + 2 sinh kr sinh kr k 2 |∇r |2 ¯ | , = nk 2 coth2 kr − (1 + cosh2 kr ) + k coth kr H, ∇r sinh2 kr ≥−

123

Isoperimetric inequalities for submanifolds with bounded mean curvature

531

where we used (a) in the above inequality. Similarly, (k 2 cosh2 kr −k 2 sinh2 kr )|∇r |2 k sinh kr k sinh kr ∇r = r  log cosh kr = div + cosh kr cosh kr cosh2 kr k 2 |∇r |2 k sinh kr ¯ | ) (k(n − |∇r |2 ) coth kr + H, ∇r + 2 cosh kr cosh kr ¯ | , = nk 2 − k 2 |∇r |2 tanh2 kr + k coth kr H, ∇r ≥



which proves (e).

Remark If  is a minimal surface in a complete simply connected Riemannian manifold with sectional curvature bounded above by a constant, Lemmas 2.3–2.5 were proved by Choe [7] as mentioned in the introduction. Therefore the above lemmas can be regarded as an extension of his results. 3 Isoperimetric inequalities and Sobolev-type inequalities In this section, we investigate Sobolev-type inequalities for submanifolds with variable mean curvature vector in a Riemannian manifold whose sectional curvature is bounded above by a constant. We also derive linear isoperimetric inequalities via these Sobolev-type inequalities. When the ambient space has nonpositive sectional curvature (i.e., the sectional curvature is bounded above by 0), the following theorem holds: Theorem 3.1 Let f be a compactly supported smooth nonnegative function on an n-dimensional compact submanifold  with boundary in a complete simply connected Riemannian manifold M of nonpositive sectional curvature. Let H denote the mean curvature vector of  in M. Assume that  is contained in a geodesic ball B p (R) ⊂ M of radius R centered at a fixed point p ∈ M. Then we have    n−1 f ≤ |∇ f | + |H | f. R 





Proof Let r (·) = dist( p, ·) be the distance function on M for p ∈ M. From Lemma 2.3 (b), we have div( f ∇r ) = ∇ f, ∇r  + f r f ¯ | ≥ ∇ f, ∇r  + (n − |∇r |2 ) + f H, ∇r r f ≥ ∇ f, ∇r  + (n − 1) − |H | f. r Integrating both sides over  gives    f (n − 1) − |H | f ≤ |∇ f |. r 





The conclusion follows from the assumption that  is contained in B p (R).



123

532

K. Seo

It is well-known that the isoperimetric inequalities can be recovered from Sobolevtype inequalities using smooth characteristic functions as test functions. Therefore, as a consequence of the above theorem, we get the following linear isoperimetric inequality involving mean curvature. Corollary 3.2 Under the same hypothesis as in Theorem 3.1, we have (n − 1) Vol() ≤ Vol(∂) + R

 |H |. 

Remark It should be mentioned that Cheung and Leung [4] proved that if M is an n-dimensional complete noncompact submanifold with bounded mean curvature in the Euclidean space or hyperbolic space, then the rate of volume growth of M is at least linear, that is, for any p ∈ M and for a sufficiently large R > 0, Vol(B p (R)) ≥ α R

(3.1)

for some constant α > 0, where B p (R) denotes the geodesic ball centered at p with radius R. In this case, if one applies the general isoperimetric inequality (1.1) in the introduction, then one immediately gets 1

c(n) n (α R)

n−1 n

1

≤ c(n) n Vol(B p (R))

n−1 n



≤ Vol(∂(B p (R))) +

|H |,

(3.2)

B p (R)

which implies that the term Vol(∂(B p (R))) + 1 n



B p (R) |H |

has a lower bound

n−1 n

c(n) (α R) for a sufficiently large R > 0. Note that the volume growth condition (3.1) is not satisfied for small R > 0. Therefore one cannot expect the inequality (3.2) for small R > 0. However Corollary 3.2 holds for any R > 0. Hence one sees that even though a complete noncompact submanifold with bounded mean curvature has a linear volume growth, the linear isoperimetric inequality in Corollary 3.2 is different from the well-known general isoperimetric inequality (1.1). When the sectional curvature of the ambient space is bounded above by a positive constant, the following theorem holds: Theorem 3.3 Let f be a compactly supported smooth nonnegative function on an n-dimensional compact submanifold  with boundary in a complete simply connected Riemannian manifold M with sectional curvature bounded above by a positive constant K = k 2 for k > 0. Let H denote the mean curvature vector of  in M. Assume π . Then we have that the diameter diam() of  satisfies diam() ≤ d < 2k nk tan kd

123



 f ≤ 

 |∇ f | +



|H | f. 

Isoperimetric inequalities for submanifolds with bounded mean curvature

533

Proof From Lemma 2.4 (e), we have div( f ∇ log cos kr ) = ∇ f, −k tan kr ∇r  + f  log cos kr ≤ −k tan kr ∇ f, ∇r  − f (nk 2 + k 2 |∇r |2 tan2 kr ¯ | ) + k tan kr H, ∇r ≤ −k tan kr ∇ f, ∇r  − f nk 2 + f |H |k tan kd. Integrating both sides over  gives 

 nk

f ≤ k tan kd

2 

 |∇ f | + k tan kd



|H | f, 



which completes the proof. Corollary 3.4 Under the same hypothesis as in Theorem 3.3, we have nk Vol() ≤ Vol(∂) + tan kd

 |H |. 

If the sectional curvature is bounded above by a negative constant, we have the following theorem. Theorem 3.5 Let f be a compactly supported smooth nonnegative function on an n-dimensional compact submanifold  with boundary in a complete simply connected Riemannian manifold M with sectional curvature bounded above by a negative constant K = −k 2 for k > 0. Let H denote the mean curvature vector of  in M. Then we have    f ≤ |∇ f | + |H | f. k(n − 1) 





Proof Since ¯ | ] div( f ∇r ) ≥ ∇ f, ∇r  + f [k(n − |∇r |2 ) coth kr + H, ∇r ≥ ∇ f, ∇r  + (k(n − 1) − |H |) f,

we get the conclusion. Corollary 3.6 Under the same hypothesis as in Theorem 3.5, we have  |H |.

k(n − 1)Vol() ≤ Vol(∂) + 

123

534

K. Seo

Remark For a domain in hyperbolic space Hn or an n-dimensional minimal submanifold of hyperbolic space, Corollary 3.6 was obtained by Yau [19] and Choe-Gulliver [8]. Hence our results can be regarded as an extension of their results. When the ambient space has sectional curvature bounded by a negative constant, we have more Sobolev-type inequalities and the corresponding linear isoperimetric inequalities. Theorem 3.7 Let f be a compactly supported smooth nonnegative function on an n-dimensional compact submanifold  with boundary in a complete simply connected Riemannian manifold M with sectional curvature bounded above by a negative constant K = −k 2 for k > 0. Let H denote the mean curvature vector of  in M. Assume that  is contained in a geodesic ball B p (R) ⊂ M of radius R centered at a fixed point p ∈ M. Then we have nk sinh k R



 f ≤ 

 |∇ f | +



|H | f. 

Proof Using ¯ | ] div( f ∇ cosh kr ) ≥ ∇ f, k sinh kr ∇r  + f [nk 2 cosh kr + k sinh kr H, ∇r 2 ≥ k sinh kr ∇ f, ∇r  + nk f − |H |k f sinh k R and integrating both sides over  give the proof.



Corollary 3.8 Under the same hypothesis as in Theorem 3.7, we have nk Vol() ≤ Vol(∂) + sinh k R

 |H |. 

In particular, if we restrict the dimension of submanifold to 2, then we have the following additional Sobolev-type inequalities for a surface with bounded mean curvature and the corresponding isoperimetric inequalities. Theorem 3.9 Let f be a compactly supported smooth nonnegative function on a surface  in a complete simply connected Riemannian manifold M with sectional curvature bounded above by a positive constant K = k 2 for k > 0. Let H denote the π , then mean curvature vector of  in M with |H | ≤ α. If diam() < 2k  2π

⎛ f2 ≤ ⎝







⎞2



|∇ f |⎠ + 2(k + α) ⎝





⎞⎛ ⎞ ⎛ ⎞2   |∇ f |⎠ ⎝ f ⎠ + α(2k + α) ⎝ f ⎠ . 



Proof Consider the distance function r (·) = dist( p, ·) on M for p ∈ M. From Lemma 2.4 (a), we have

123

Isoperimetric inequalities for submanifolds with bounded mean curvature

div

f k∇r sin kr



535



 sin kr r − k cos kr |∇r |2 k∇r = ∇ f, + kf sin kr sin2 kr ¯ | ] k f [k(2 − |∇r |2 ) cot kr + H, ∇r k ∇ f, ∇r  + sin kr sin kr 2 2 f k cos kr |∇r | − sin2 kr  k kf  ¯ | = k(2−2|∇r |2 ) cos kr +sin kr H, ∇r ∇ f, ∇r + 2 sin kr sin kr kα f k ∇ f, ∇r  − . (3.3) ≥ sin kr sin kr ≥

Take y = p ∈ . Let B y (t) ⊂ M be the geodesic ball of radius t centered at y ∈ , t < dist(y, ∂), t =  \ B y (t), and Ct =  ∩ ∂ B y (t). Then

 div 

f k∇r sin kr



 = ∂t

k f ∂r = sin kr ∂ν

 ∂

k f ∂r − sin kr ∂ν

 Ct

k f ∂r =− sin kr ∂ν

 Ct

k f ∂r , sin kr ∂ν

where ν on ∂t is the outward unit conormal to ∂t . Here we used the assumption ∂r → 1 on Ct as t → 0, we have that f ∈ C0∞ (). Since ∂ν  2π f (y) = lim

t→0 Ct

k f ∂r =− sin kr ∂ν



 div 

f k∇r sin kr

.

Thus integrating both sides of (3.3) over  gives  2π f (y) ≤ 

k|∇ f (x)| d Vx + α sin kr y (x)



k f (x) d Vx , sin kr y (x)



(3.4)

where r y (x) = dist(x, y). Moreover div( f ∇ sin kr ) ≥ ∇ f, k cos kr ∇r  + f ¯ | + k cos kr H, ∇r



≥ k cos kr ∇ f, ∇r  + f

k2 (2 cos2 kr − |∇r |2 ) sin kr

k2 2 − 2k − αk . sin kr

So  k 

f ≤ (2k + α) sin kr



 f +



|∇ f |.

(3.5)



123

536

K. Seo

Thus applying the inequalities (3.4) and (3.5), we obtain 







k|∇ f (x)| d Vx + α sin kr y (x)





k f (x) d Vx ⎠ d Vy sin kr y (x)        k f (y) k f (y) d Vy d Vx + α d Vy d Vx = |∇ f (x)| f (x) sin kr x (y) sin kr x (y)     ⎛ ⎞⎛ ⎞     ≤ ⎝ |∇ f | + α f ⎠ ⎝(2k + α) f + |∇ f |⎠ f (y) ⎝

f2 ≤













⎛ ⎞2 ⎛ ⎞⎛ ⎞ ⎛ ⎞2     ≤ ⎝ |∇ f |⎠ + 2(k + α) ⎝ |∇ f |⎠ ⎝ f ⎠ + α(2k + α) ⎝ f ⎠ , 









which completes the proof. Corollary 3.10 Under the same hypothesis as in Theorem 3.9, we have 2π Area() ≤ Length(∂)2 +2(k +α)Length(∂)Area()+α(2k +α)Area()2 .

Remark For a minimal surface in a complete simply connected Riemannian manifold with sectional curvature bounded above, Choe [7] obtained this result. Theorem 3.11 Let f be a compactly supported smooth nonnegative function on a surface  in a complete simply connected Riemannian manifold M with sectional curvature bounded above by a positive constant K = k 2 for k > 0. Let H denote the π , then mean curvature vector of  in M. If diam() ≤ d < 2k  2π 

⎞2 ⎛   1 ⎝ |∇ f | + |H | f ⎠ . f2 ≤ cos kd 



Proof Since ¯ | ] div( f ∇r ) ≥ ∇ f, ∇r  + f [k(2 − |∇r |2 ) cot kr + H, ∇r cos kd f k − |H | f, ≥ ∇ f, ∇r  + sin kr it follows  

123

⎞ ⎛   1 ⎝ kf ≤ |∇ f | + |H | f ⎠ . sin kr cos kd 



(3.6)

Isoperimetric inequalities for submanifolds with bounded mean curvature

537

Therefore applying the inequalities (3.4) and (3.6) gives  2π 



⎞ k|H (x)| f (x) d Vx ⎠ d Vy f2 ≤ f (y) ⎝ sin kr y (x)        k f (y) k|H (x)| f (y) d Vy d Vx + d Vy d Vx = |∇ f (x)| f (x) sin kr x (y) sin kr x (y) 







k|∇ f (x)| d Vx + sin kr y (x)



1 cos kd



⎞2 ⎛   ⎝ |∇ f | + |H | f ⎠ . 







Corollary 3.12 Under the same hypothesis as in Theorem 3.11, we have ⎛ 2π Area() ≤

1 ⎝ Length(∂) + cos kd



⎞2 |H |)⎠ .



For a surface with variable mean curvature in a negatively curved manifold, we have the following Sobolev-type inequality which contains a mean curvature term. Theorem 3.13 Let f be a compactly supported smooth nonnegative function on a compact surface  with boundary in a complete simply connected Riemannian manifold M with sectional curvature bounded above by a negative constant K = −k 2 for k > 0. Let H denote the mean curvature vector of  in M. Assume that  is contained in a geodesic ball B p (R) ⊂ M of radius R centered at a fixed point p ∈ M. Then we have ⎛ ⎞2    f 2 ≤ cosh k R ⎝ |∇ f | + |H | f ⎠ . 2π 





Proof From Lemma 2.5 (a), we have

f ∇r div sinh kr



 ∇r sinh kr r − k cosh kr |∇r |2 = ∇ f, + f sinh kr sinh2 kr

1 ∇ f, ∇r  + sinh kr f k cosh kr |∇r |2 − sinh2 kr 1 = ∇ f, ∇r  + sinh kr 1 ≥ ∇ f, ∇r  − sinh kr ≥

¯ | ] f [k(2 − |∇r |2 ) coth kr + H, ∇r sinh kr

¯ | f H, ∇r f [k(2 − 2|∇r |2 ) cosh kr ] + sinh kr sinh2 kr |H | f . sinh kr

123

538

K. Seo

As in the proof of Theorem 3.9, integrating both sides over  yields  2π f (y) ≤ 

k|∇ f (x)| d Vx + sinh kr y (x)



k|H (x)| f (x) d Vx , sinh kr y (x)



where r y (x) = dist(x, y). Moreover div( f ∇ sinh kr ) ≥ ∇ f, k cosh kr ∇r  + f ¯ | + k cosh kr H, ∇r



≥ k cosh kr ∇ f, ∇r  + f

k2 (2 cosh2 kr − |∇r |2 ) sinh kr

k2 − k|H | f cosh kr. sinh kr

Thus  

kf ≤ sinh kr



 |∇ f | cosh kr +



|H | f cosh kr. 

Hence using the inequality (3.4), we get  2π 



⎞ k|H (x)| f (x) d Vx ⎠ d Vy f2 ≤ f (y) ⎝ sinh kr y (x)        k f (y) k|H (x)| f (y) d Vy d Vx + d Vy d Vx = |∇ f (x)| f (x) sinh kr x (y) sinh kr x (y)     ⎛ ⎞⎛ ⎞     ≤ ⎝ |∇ f | + |H | f ⎠ ⎝ |∇ f | cosh kr + |H | f cosh kr ⎠ 





k|∇ f (x)| d Vx + sinh kr y (x)







⎛ ⎞2   ≤ cosh k R ⎝ |∇ f | + |H | f ⎠ , 







which completes the proof. Corollary 3.14 Under the same hypothesis as in Theorem 3.13, we have ⎛ 2π Area() ≤ cosh k R ⎝Length(∂) +

 

123

⎞2 |H |⎠ .

Isoperimetric inequalities for submanifolds with bounded mean curvature

539

4 First eigenvalue estimates In this section, we obtain various first Dirichlet eigenvalue estimates for a compact submanifold with bounded mean curvature in a complete simply connected Riemannian manifold with sectional curvature bounded above by a constant. Theorem 4.1 Let  be an n-dimensional compact submanifold with boundary in a complete simply connected Riemannian manifold M of nonpositive sectional curvature. Let H denote the mean curvature vector of  in M. Assume that  is contained in a geodesic ball B p (R) ⊂ M of radius R centered at a fixed point p ∈ M. If |H | ≤ α for some constant α < n−1 R , then we have 1 4



n−1 −α R

2 ≤ λ1 ().

Proof By Lemma 2.3 (b), we have for any f ∈ C0∞ (M) div( f 2 ∇r ) = ∇ f 2 , ∇r  + f 2 r f2 ¯ | (n − |∇r |2 ) + f 2 H, ∇r ≥ ∇ f 2 , ∇r  + r f2 − α f 2. ≥ −2| f ||∇ f | + (n − 1) r Integrating both sides over  and applying Schwarz inequality, we get n−1 R



 f2 −α 

 f2 ≤ 2



 | f ||∇ f | ≤ ε



f2 + 

1 ε

 |∇ f |2 . 

Therefore ε

n−1 −α−ε R



 f2 ≤



|∇ f |2 . 

Since the maximum value of the quadratic function ε( n−1 R − α − ε) is attained at n−1 − α ε= R , we obtain 2 2  |∇ f |2 1 n−1 − α ≤  , 2 4 R  f which completes the proof.



Theorem 4.2 Let  be an n-dimensional compact submanifold with boundary in a complete simply connected Riemannian manifold M with sectional curvature bounded

123

540

K. Seo

above by a positive constant K = k 2 for k > 0. Let H denote the mean curvature π . vector of  in M. Assume that the diameter diam() of  satisfies diam() ≤ d < 2k nk If |H | ≤ α for some constant α < tan kd , then we have 1 4



nk −α tan kd

2 ≤ λ1 ().

Proof Using Lemma 2.4 (e), we see that for any f ∈ C0∞ (M) div( f 2 ∇ log cos kr ) = ∇ f 2 , −k tan kr ∇r  + f 2  log cos kr ≤ −k tan kr ∇ f 2 , ∇r  − f 2 (nk 2 + k 2 |∇r |2 tan2 kr ¯ | ) + k tan kr H, ∇r ≤ −k tan kr ∇ f 2 , ∇r  − nk 2 f 2 + f 2 αk tan kd. Integrating both sides over  and applying Schwarz inequality for ε > 0 give  (nk 2 −αk tan kd)

 f 2 ≤ 2k tan kd



⎛ | f ||∇ f | ≤ k tan kd ⎝ε



 f 2+ 

1 ε



⎞ |∇ f |2 ⎠ ,



which implies  ε (nk − α tan kd − ε tan kd) ≤

 |∇ 

f |2

f2

.

Since the minimum of the left hand side is 14 ( tannkkd − α)2 at ε = 1 4



nk −α tan kd

nk−α tan kd 2 tan kd , we obtain

2 ≤ λ1 ().

When the ambient space is a complete simply connected Riemannian manifold with sectional curvature bounded above by a negative constant, we have the following first eigenvalue estimate which was proved by Bessa and Montenegro [3]. Here we give a simple proof of their theorem. Theorem 4.3 [3] Let  be an n-dimensional submanifold in a complete simply connected Riemannian manifold M with sectional curvature bounded above by a negative constant K = −k 2 for k > 0. Let H denote the mean curvature vector of  in M. If |H | ≤ α for some constant α < (n − 1)k, then we have [(n − 1)k − α]2 ≤ λ1 () 4

123

Isoperimetric inequalities for submanifolds with bounded mean curvature

541

Proof Since ¯ | ] div( f 2 ∇r ) ≥ ∇ f 2 , ∇r  + f 2 [k(n − |∇r |2 ) coth kr + H, ∇r for any f ∈ C0∞ (M), integrating both sides over  and applying Schwarz inequality for ε > 0 as before give     1 [(n − 1)k − α] f 2 ≤ 2| f ||∇ f | ≤ ε f2 + |∇ f |2 , ε 







which implies  ε ((n − 1)k − α − ε)

 f2 ≤



As before, taking ε =

(n−1)k−α 2

|∇ f |2 . 



gives the conclusion.

Theorem 4.4 Let  be an n-dimensional compact submanifold with boundary in a complete simply connected Riemannian manifold M with sectional curvature bounded above by a negative constant K = −k 2 for k > 0. Let H denote the mean curvature vector of  in M. Assume that  is contained in a geodesic ball B p (R) ⊂ M of radius R centered at a fixed point p ∈ M. If |H | ≤ α for some constant α < sinhnkk R , then we have 1 4



nk −α sinh k R

2 ≤ λ1 ().

Proof Using ¯ | ] div( f 2 ∇ cosh kr ) ≥ ∇ f 2 , k sinh kr ∇r  + f 2 [nk 2 cosh kr + k sinh kr H, ∇r 2 2 ≥ k sinh kr ∇ f , ∇r  + nk f − αk f sinh k R for any f ∈ C0∞ (M), integrating both sides over  and applying Schwarz inequality for ε > 0, we see that ⎛ ⎞     1 (nk − α sinh k R) f 2 ≤ sinh k R 2| f ||∇ f | ≤ sinh k R ⎝ε f2 + |∇ f |2 ⎠ ε 







which implies

nk ε −α−ε sinh k R Therefore we obtain the conclusion.



 f ≤

|∇ f |2 .

2







123

542 Acknowledgements suggestions.

K. Seo The author would like to thank the referees for the valuable comments and

References 1. Allard, W.: On the first variation of a varifold. Ann. Math. 95(2), 417–491 (1972) 2. Almgren, F.: Optimal isoperimetric inequalities. Indiana Univ. Math. J. 35(3), 451–547 (1986) 3. Bessa, G., Montenegro, J.: Eigenvalue estimates for submanifolds with locally bounded mean curvature. Ann. Global Anal. Geom. 24(3), 279–290 (2003) 4. Cheung, L.F., Leung, P.F.: The mean curvature and volume growth of complete noncompact submanifolds. Differ. Geom. Appl. 8(3), 251–256 (1998) 5. Cheung, L.F., Leung, P.F.: Eigenvalue estimates for submanifolds with bounded mean curvature in the hyperbolic space. Math. Z. 236, 525–530 (2001) 6. Choe, J.: The isoperimetric inequality for a minimal surface with radially connected boundary. Ann. Scuola Norm. Sup. Pisa Cl. Sci. 17(4), 583–593 (1990) 7. Choe, J.: The isoperimetric inequality for minimal surfaces in a Riemannian manifold. J. Reine Angew. Math. 506, 205–214 (1999) 8. Choe, J., Gulliver, R.: Isoperimetric inequalities on minimal submanifolds of space forms. Manuscripta. Math. 77, 169–189 (1992) 9. Feinberg, J.: The isoperimetric inequality for doubly-connected minimal surfaces in R n . J. Analyse Math. 32, 249–278 (1977) 10. Hoffman, D., Spruckc, J.: Sobolev and isoperimetric inequalities for Riemannian submanifolds. Comm. Pure Appl. Math. 27, 715–727 (1974) 11. Karp, L.: Differential inequalities on complete Riemannian manifolds and applications. Math. Ann. 272(4), 449–459 (1985) 12. Li, P., Schoen, R., Yau, S.T.: On the isoperimetric inequality for minimal surfaces. Ann. Scuola Norm. Sup. Pisa Cl. Sci. 11(2), 237–244 (1984) 13. McKean, H.P.: An upper bound to the spectrum of  on a manifold of negative curvature. J. Diff. Geom. 4, 359–366 (1970) 14. Michael, J., Simon, L.M.: Sobolev and mean-value inequalities on generalized submanifolds of Rn . Comm. Pure. Appl. Math. 26, 361–379 (1973) 15. Osserman, R.: The isoperimetric inequality. Bull. Amer. Math. Soc. 84, 1182–1238 (1978) 16. Osserman, R., Schiffer, M.: Doubly-connected minimal surfaces. Arch. Rational Mech. Anal. 58, 285– 307 (1975) 17. Schoen, R., Yau, S.-T.: Lectures on differential geometry. Conference Proceedings and Lecture Notes in Geometry and Topology. I. International Press (1994) 18. Stone, A.: On the isoperimetric inequality on a minimal surface. Calc. Var. Partial Differ. Equ. 17, 369–391 (2003) 19. Yau, S.-T.: Isoperimetric constants and the first eigenvalue of a compact manifold. Ann. Sci. Ecole Norm. Sup. 8, 487–507 (1975)

123

Isoperimetric inequalities for submanifolds with ... - Springer Link

Jul 23, 2011 - if ωn is the volume of a unit ball in Rn, then. nnωnVol(D)n−1 ≤ Vol(∂D)n and equality holds if and only if D is a ball. As an extension of the above classical isoperimetric inequality, it is conjectured that any n-dimensional compact minimal submanifold M of Rm still satisfies the above inequality. Moreover ...

192KB Sizes 0 Downloads 406 Views

Recommend Documents

Lp AFFINE ISOPERIMETRIC INEQUALITIES Erwin ...
If ρK is positive and continuous, call K a star body (about the origin). Two star bodies K and L are said to be ..... bridge, 1993. [SW]. R. Schneider and W. Weil, ...

Relative isoperimetric inequalities for minimal ... - Wiley Online Library
Mar 6, 2012 - 10, 1264 – 1273 (2012) / DOI 10.1002/mana.201100078. Relative isoperimetric inequalities for minimal submanifolds outside a convex set. Keomkyo Seo. ∗. Department of Mathematics, Sookmyung Women's University, Hyochangwongil 52, Yong

Data integration with uncertainty - Springer Link
Nov 14, 2008 - sources by automatic methods (e.g., HTML pages, emails, blogs). ..... If a tuple is an answer to Q under multiple mappings in m, then we add.

Plant location with minimum inventory - Springer Link
fractional solution we were able to derive integer solutions within 4% of optimality. ... F. Barahona, D. Jensen / Mathematical Programming 83 (1998) 101-111.

Explicit formulas for repeated games with absorbing ... - Springer Link
Dec 1, 2009 - mal stationary strategy (that is, he plays the same mixed action x at each period). This implies in particular that the lemma holds even if the players have no memory or do not observe past actions. Note that those properties are valid

Energy efficient routing with delay guarantee for sensor ... - Springer Link
Jun 15, 2006 - shown in [2], most of the battery energy is consumed by the radio. A Time ..... can then prove that the head of each arc in G v is closer to the.

Indirect Pharmacodynamic Models for Responses with ... - Springer Link
ûariable and simple first-order loss (kout) of R. These models were extended using ... of two-compartment distribution of R and or polyexponential loss of R.

Adaptive Finite Elements with High Aspect Ratio for ... - Springer Link
An adaptive phase field model for the solidification of binary alloys in two space dimensions is .... c kρsφ + ρl(1 − φ). ( ρv + (k − 1)ρsφvs. )) − div. (. D(φ)∇c + ˜D(c, φ)∇φ. ) = 0, (8) where we have set .... ena during solidif

A link between complete models with stochastic ... - Springer Link
classical ARCH models, a stationary solution with infinite variance may exists. In ..... must compute the required conditional expectations and variances. Setting ...

MINIMAL SUBMANIFOLDS WITH SMALL TOTAL ...
Keomkyo Seo. Abstract. Let M be an n-dimensional complete minimal submanifold in Rnюp. Lei Ni proved that if M has sufficiently small total scalar curvature, ...

Exploiting Graphics Processing Units for ... - Springer Link
Then we call the CUDA function. cudaMemcpy to ..... Processing Studies (AFIPS) Conference 30, 483–485. ... download.nvidia.com/compute/cuda/1 1/Website/.

Evidence for Cyclic Spell-Out - Springer Link
Jul 23, 2001 - embedding C0 as argued in Section 2.1, this allows one to test whether object ... descriptively head-final languages but also dominantly head-initial lan- ..... The Phonology-Syntax Connection, University of Chicago Press,.

MAJORIZATION AND ADDITIVITY FOR MULTIMODE ... - Springer Link
where 〈z|ρ|z〉 is the Husimi function, |z〉 are the Glauber coherent vectors, .... Let Φ be a Gaussian gauge-covariant channel and f be a concave function on [0, 1].

Tinospora crispa - Springer Link
naturally free from side effects are still in use by diabetic patients, especially in Third .... For the perifusion studies, data from rat islets are presented as mean absolute .... treated animals showed signs of recovery in body weight gains, reach

Chloraea alpina - Springer Link
Many floral characters influence not only pollen receipt and seed set but also pollen export and the number of seeds sired in the .... inserted by natural agents were not included in the final data set. Data were analysed with a ..... Ashman, T.L. an

GOODMAN'S - Springer Link
relation (evidential support) in “grue” contexts, not a logical relation (the ...... Fitelson, B.: The paradox of confirmation, Philosophy Compass, in B. Weatherson.

Bubo bubo - Springer Link
a local spatial-scale analysis. Joaquın Ortego Æ Pedro J. Cordero. Received: 16 March 2009 / Accepted: 17 August 2009 / Published online: 4 September 2009. Ó Springer Science+Business Media B.V. 2009. Abstract Knowledge of the factors influencing

Quantum Programming - Springer Link
Abstract. In this paper a programming language, qGCL, is presented for the expression of quantum algorithms. It contains the features re- quired to program a 'universal' quantum computer (including initiali- sation and observation), has a formal sema

BMC Bioinformatics - Springer Link
Apr 11, 2008 - Abstract. Background: This paper describes the design of an event ontology being developed for application in the machine understanding of infectious disease-related events reported in natural language text. This event ontology is desi

Probabilities for new theories - Springer Link
where between 0 and r, where r is the prior probability that none of the existing theories is ..... theorist's internal programming language"(Dorling 1991, p. 199).

A Process Semantics for BPMN - Springer Link
Business Process Modelling Notation (BPMN), developed by the Business ..... In this paper we call both sequence flows and exception flows 'transitions'; states are linked ...... International Conference on Integrated Formal Methods, pp. 77–96 ...

Unsupervised Learning for Graph Matching - Springer Link
Apr 14, 2011 - Springer Science+Business Media, LLC 2011. Abstract Graph .... tion as an integer quadratic program (Leordeanu and Hebert. 2006; Cour and Shi ... computer vision applications such as: discovering texture regularity (Hays et al. .... fo