Skip to main content
Springer logoLink to Springer
. 2017 Aug 15;2017(1):190. doi: 10.1186/s13660-017-1462-4

Bonnesen-style inequality for the first eigenvalue on a complete surface of constant curvature

Niufa Fang 1, Jiazu Zhou 1,2,
PMCID: PMC5557883  PMID: 28860691

Abstract

By Cheeger’s isoperimetric constants, some lower bounds and upper bounds of λ1, the first eigenvalue on a complete surface of constant curvature, are given. Some Bonnesen-style inequalities and reverse Bonnesen-style inequalities for the first eigenvalue are obtained. Those Bonnesen-style inequalities obtained are stronger than the well-known Osserman’s results and the upper bound is stronger than Osserman’s results (Osserman in Proceedings of the International Congress of Mathematicians, Helsinki, 1978).

Keywords: the first eigenvalue, Cheeger’s isoperimetric constants, Bonnesen-style inequality

Introduction

The classical isoperimetric problem is to determine a plane figure of largest possible area whose boundary has a specific length and it was known in Ancient Greece. However, the first mathematically rigorous proof was obtained only in the nineteenth century and it was well recognized by Weierstrass though Bernoulli, Euler and Lagrange once claimed the proof that was found flawed later. Hurwitz published a short proof using the Fourier series that applies to arbitrary plane domain D whose boundary ∂D was not assumed to be smooth. An elegant direct proof, based on the comparison of a smooth simple closed curve with a circle, was given by Schmidt in 1938 by using only the arc length formula, expression for the area of a plane region from Green’s theorem, and the Cauchy-Schwarz inequality [2]. Many other proofs have been found and some of them were stunningly simple. The isoperimetric problem has been extended in multiple ways, for example, to domains on surfaces and in higher dimensional spaces, or more generally to integral currents and analytic manifolds, but the proof is too difficult.

Let D be a domain (subset with nonempty interiors) in the Euclidean plane R2 with the boundary composing of the simple curve of length L and area A. Then

L24πA0, 1.1

the equality holds when and only when D is a disc.

It is known that the isoperimetric inequality (1.1) is equivalent to the following Sobolev inequality (see [3]):

If f has compact support in DR2, then

(D|f|)24πDf20. 1.2

Here ∇ denotes the gradient operator. The equality holds in (1.2) if and only if f is the characteristic function of balls.

During 1920s, Bonnesen proved a series of inequalities of the form

L24πABD, 1.3

where the quantity BD on the right-hand side is a non-negative geometric invariant of significance and vanishes only when D is a disc.

An inequality of the form (1.3) is called the Bonnesen-style inequality, and it is stronger than the classical isoperimetric inequality. The Bonnesen-style inequality has been extended to surfaces of constant curvature and higher dimensions and many Bonnesen-style inequalities have been found during the past. Mathematicians are still working on unknown Bonnesen-style inequalities of geometric significance [36]. The isoperimetric inequality for domains on surfaces M of constant curvature can be stated as follows.

Let D be a compact domain on the surface M of constant curvature. Let A and L denote the area and the boundary length of D, respectively. Then

L24πA+MA20,

the equality holds if and only if D is a geodesic disc.

The Bonnesen-style inequality for domains on surfaces of constant curvature was first investigated by Santaló [7, 8]. Klain obtained some new Bonnesen-style inequalities for domains on surfaces of constant curvature [9]. By the kinematic formulas in integral geometry, Xu, Zhou et al. also obtained Bonnesen-style inequalities on a complete surface of constant curvature (see [10, 11]). Osserman [5] studied the Bonnesen-style inequality for the domains on surfaces with the bounded Gauss curvature. More Bonnesen-style homothetic (Wulff) inequalities were obtained in [4, 12, 13]. Another important extension of the isoperimetric problem in analysis is eigenvalues of the Laplacian.

Eigenvalues of the Laplacian. Let D be a domain with smooth boundary ∂D on a compact Riemannian surface M. The eigenvalue problem

Δu+λu=0in D;u|D=0,

is known to have a complete system of eigenfunctions u=ϕn, with corresponding eigenvalues λn, where

0=λ0<λ1<λ2<<λp<.

One would ask naturally a basic question: how are the properties of domain D on a compact Riemannian surface M, that is, area of DM, length and integrals of curvature of ∂D, reflected in the set of eigenvalues {λn}?

In this paper, we will investigate the Bonnesen-style inequalities for the first eigenvalue λ1 of Laplacian on the complete surface. Let M be a compact Riemannian surface and Δ be the Laplacian-Beltrami operator acting on differential functions C(M). It is known that Δ is an elliptic operator. The first eigenvalue λ1 on domain DM can be also characterized by [1]

λ1(D)=inffFD|f|2D|f|2, 1.4

where F is the set of smooth functions in D vanishing on the boundary.

The Laplace operator on a Riemannian manifold, its spectral theory and the relations between its first eigenvalue and the geometrical data of the manifold, such as curvatures, diameter, injectivity radius and volume, have been extensively studied in the recent mathematical literature. Amazing connections between the isoperimetric inequality and the first eigenvalue of Laplacian operator have been found during the past decades. One of the basic results is that Cheeger connected the first eigenvalue λ1 of the Laplacian on a manifold to certain isoperimetric constants. For a domain D on a two-dimensional surface, Cheeger considers the quantity

h=infDFLA, 1.5

where F is the family of relatively compact subdomains of D, A and L are the area and the boundary length of subdomain DD, respectively. Cheeger proved that

λ1(D)14h2. 1.6

The upper estimate of the first eigenvalue of Laplacian has been discussed by geometers and analysts. Hersch [14] obtained an upper bound for manifolds homeomorphic to the two sphere. Cheeger [15], Chavel and Feldman [16] obtained upper bound for manifolds with non-negative Ricci curvature. The comparison theorem of Cheng [17] gives a sharp upper bound for general Riemannian manifold in terms of the Ricci curvature and the diameter of domain.

While the progress has been made on the upper bound, not too much is known about the lower bound of the first eigenvalue. The best result is due to Lichnerowicz [18] who gives a computable sharp lower bound for manifolds whose Ricci curvature is bounded from below by a positive constant. Cheeger [19] also gives a lower estimate for general manifolds in terms of some isoperimetric constants. These constants of Cheeger, however, are not computable. Cheng [20] observed that if the manifold is a two-dimensional convex surface, then the isoperimetric constant has a lower bound in terms of the diameter. Since 1979, Li and Yau have been trying to obtain the lower bound of the first eigenvalue [21, 22]. Chen introduced the method in probability theory to improve almost all results proved by others in [23]. For more detailed isoperimetric properties and the first eigenvalue, one can refer to [1, 18, 2431].

In [5], Osserman considered the first eigenvalue λ1 on the two-dimensional manifolds with bounded Gauss curvature and obtained some lower and upper bounds of the first eigenvalue by using Cheeger’s isoperimetric constant as follows:

Let S be a compact simply connected surface with Gauss curvature K, Kα2, α>0. If DS is simply connected and ρ is inradius of D, then

λ1(D)α24α24csch2αρ. 1.7

In this paper, we obtain the following lower bound of the first eigenvalue that is stronger than Osserman’s result (1.7):

Let S be a simply connected complete surface with Gauss curvature K0 everywhere. For any simply connected domain DS , let ρ denote its inradius and R denote its circumradius. Let

α2=infD(K),β2=supD(K),0<αβ,

then

λ1(D)α24α24csch2αρ+B,

where the quantity B is a positive number depending on α, β, ρ, R.

We also obtain the upper bound of the first eigenvalue. By Cheng’s eigenvalue comparison theorem ([17], Theorem 1.1), we obtain a stronger upper bound of the first eigenvalue λ1 (Theorem 4.1).

The Bonnesen-style isoperimetric inequalities

Let DrM denote the geodesic disc of radius r on the complete simply connected surface of constant Gauss curvature KM. Let ArM,LrM be, respectively, the area and the length of boundary of DrM. Then the explicit expressions for these quantities are

M=α2<0:LrM=2πsinhαrα,ArM=4πsinh2αr2α2=2πcoshαr1α2;M=0:LrM=2πr,ArM=πr2;M=α2>0:LrM=2πsinαrα,ArM=4πsin2αr2α2=2π1cosαrα2.

For the geodesic disc, the following equation can be easily verified in all three cases:

(LrM)24πArM+M(ArM)2=0. 2.1

The isoperimetric inequality on a surface of constant curvature KM is

L24πA+MA20. 2.2

Namely, given a domain D of area A, if r is chosen so that ArM equals A, then (2.1) and (2.2) imply LLrM, so that the disc DrM has minimum boundary length among all domains of the same area.

Osserman considered the isoperimetric inequality of two-dimensional complete surface with bounded Gauss curvature [5].

Let D be a simply connected domain whose Gauss curvature K satisfies KM. Let L and A be the boundary length and the area of D, respectively. Then

L24πA+MA20, 2.3

where the equality holds if and only if KM and D is a geodesic disc.

Osserman also obtained the following Bonnesen-style isoperimetric inequalities.

Theorem A

[5]

Let D be a simply connected domain whose Gauss curvature K satisfies KM. Let ρ be the inradius of D, A be the area of D and L be the length of its boundary. Then the following inequalities are equivalent:

LLρM+MAAρM2π(A+AρM), 2.4
L24πA+MA2(LLρMAρMA)2, 2.5
L24πA+MA2(LLρM)2+M(AAρM)2. 2.6

Moreover, if MA<4π, then these inequalities are equivalent to

L24πA+MA2(2πLρM(AAρM))2. 2.7

Osserman estimated lower bounds of the first eigenvalue by Cheeger’s isoperimetric constants as follows.

Theorem B

[5]

Let S be a simply connected complete surface with Gauss curvature K, Kα2, α>0. Then, for any domain DS of circumradius R,

λ1(D)α24α24(cschαR)2. 2.8

If D is simply connected and ρ is its inradius, then

λ1(D)α24α24(cschαρ)2. 2.9

The lower bound of λ1

In this section, we give some lower bounds of the first eigenvalue λ1 by Cheeger’s isoperimetric constants and Bonnesen-style isoperimetric inequalities. We need the following lemmas.

Lemma 3.1

Let f(r) be continuously differentiable on the interval 0rr0. Suppose that, except at a finite number of the points in the interval, f(r) exists and satisfies

f(r)+cf(r)0,f(0)=0,f(0)=a 3.1

for some constants a, c. Let h(r) be the unique solution of

h(r)+ch(r)=0,h(0)=0,h(0)=1. 3.2

Let s be any number such that h(r)>0 for 0<r<s, and let r1=min{r0,s}. Then

f(r)ah(r) 3.3

for 0rr1.

Proof

Let ϕ(r)=f(r)h(r). Then by (3.1) and (3.2)

(h2ϕ)=(fhfh)=fhfh0,

except at the singular points. By the mean value theorem, h2ϕ is a weakly monotone decreasing function, and hence

(h2ϕ)(r)(h2ϕ)(0)=0,0rr1.

That is, ϕ(r)0, and hence

f(r)h(r)=ϕ(r)ϕ(0)=limr0f(r)h(r)=a

for 0rr1. □

Lemma 3.2

Let Dρ be a geodesic disc of radius ρ, and let Aρ be the area of Dρ. If MK0 on Dρ, then

AρAρM, 3.4

where equality holds if and only if KM on Dρ.

Proof

We introduce geodesic polar coordinates in Dρ. The metric can be written as ds2=dr2+g(r,θ)dθ2, where for each θ, the function f(r)=g(r,θ) satisfies f(0)=0, f(0)=1. Since K0, the geodesic disc of radius ρ always exists. Then with the fact K=1g2r2g and the condition MK0, f(r) satisfies (3.1), with a=1, c=M, r0=ρ. By (3.3), we have

Aρ=02π0ρg(r,θ)drdθ2π0ρh(r)dr=AρM. 3.5

Since h(r) can be written explicitly as h(r)=12πLρM, it satisfies (3.2) and Aρ=0ρL(r)dr. The equality holds if and only if g(r,θ)h(r), hence KM. □

Theorem 3.1

Let S be a simply connected complete surface with Gauss curvature K0 everywhere. For any simply connected domain DS, let A, L, R be the area, the boundary length and the circumradius of D, respectively. Let

α2=infD(K),β2=supD(K),0<αβ, 3.6

then

λ1(D)α24β24(cschβR2)2. 3.7

Proof

By Lemma 3.2 and (3.6), we have

ARβ2ARA. 3.8

Via (3.6), the isoperimetric inequality (2.3) can be rewritten as

(LA)24π1A+α2.

Then, by (3.8), we have

(LA)24π1A+α24π1ARβ2+α2=β2(cschβR2)2+α2. 3.9

By (1.6) and (3.9), then

λ1(D)α24β24(cschβR2)2.

We complete the proof. □

Since xsinhx is monotonically decreasing for x0, hence (3.7) is stronger than (2.8) if 12β<α. By (2.4) we obtain a lower bound of λ1 that is stronger than the one in (2.9).

Theorem 3.2

Let S be a simply connected complete surface with Gauss curvature K0 everywhere. For any simply connected domain DS, let A, L, ρ, R be the area, the boundary length, the inradius and the circumradius of D, respectively. Let

α2=infD(K),β2=supD(K),0<αβ, 3.10

then

λ1(D)α24α24(cschαρ)2+B, 3.11

where

B=14(βsinhαρ2αsinhβR2)4(αsinhαρ)2+12(βsinhαρ2αsinhβR2)2α2cothαρsinhαρ.

Proof

By (2.4) and (3.10), we have

LALρα22π(1+Aρα2A)+α2Aρα2,

and hence,

LAAρα2Aαsinhαρ+αcothαρ. 3.12

By (3.4) and (3.10), (3.12) can be rewritten as

LAAρα2Aαsinhαρ+αcothαρAρα2ARβ2αsinhαρ+αcothαρ=(βsinhαρ2αsinhβR2)2αsinhαρ+αcothαρ.

By (1.5) and (1.6), we have

λ1(D)14{(βsinhαρ2αsinhβR2)2αsinhαρ+αcothαρ}2=α24(cothαρ)2+B=α24+α24(cschαρ)2+B,

and

B=14(βsinhαρ2αsinhβR2)4(αsinhαρ)2+12(βsinhαρ2αsinhβR2)2α2cothαρsinhαρ.

 □

Since B0, hence inequality (3.11) is stronger than inequality (2.9). Let R=ρ in Theorem 3.2, that is, let D be a geodesic disc with radius ρ on S. Then let ρ in (3.11), we immediately obtain the following.

Corollary 3.1

Let S be a simply connected complete surface with Kα2, α>0 everywhere, then

λ1(S)α24.

Next, we give a lower bound of the first eigenvalue λ1.

Theorem 3.3

Let D be a complete simply connected surface and A denote the area of D. Then

λ1(D)4πA.

Proof

By Sobolev inequality (1.2) and Hölder’s inequality, we have

4πDf2dx(D|f|dx)2D12dxD|f|2dx=AD|f|2dx,

that is,

D|f|2dxDf2dx4πA.

By (1.4), we have

λ1=inff0D|f|2dxDf2dx4πA.

 □

By Lemma 3.2, we obtain a lower bound of λ1.

Corollary 3.2

Let S be a simply connected complete surface with Gauss curvature K, β2K0 everywhere. Suppose DS, then

λ1(D)β2(cschβR2)2.

Here R is the circumradius of D.

The upper bound of λ1

In this section, we consider the upper bound of the first eigenvalue λ1. We start with the following eigenvalue comparison theorem proved by Cheng in [17]. Denote the open geodesic ball of radius r with center x by D(x;r). Denote by Vn(M;r) the geodesic ball of radius r in the n-dimensional simply connected space form with constant sectional curvature M. We write λ1(D(x;r)) as λ1(D(x;r)).

Theorem C

Suppose that S is a complete Riemannian manifold and Ricci curvature of S(n1)M, n=dimS. Then, for xS, we have

λ1(D(x;r))λ1(Vn(M;r))

and equality holds if and only if D(x;r) is isometric to Vn(M;r).

In particular, the eigenvalue comparison theorem is also valid when S is a two-dimensional complete simply connected surface.

Corollary 4.1

Suppose that S is a complete simply connected surface with Gauss curvature KM. Let Dr be a geodesic disc with radius r on S, then

λ1(Dr)λ1(DrM),

where equality holds if and only if KM on S.

The next lemma will be needed in proving our theorem.

Lemma 4.1

[5]

Suppose that S is a simply connected complete surface with Gauss curvature K0 everywhere. Let Dρ be a geodesic disc of radius ρ. If

α2=infDρ(K),β2=supDρ(K),0<αβ, 4.1

then

λ1(Dρ)(β22αcothαρ)2+(πρ)2. 4.2

Combining Corollary 4.1 and Lemma 4.1 immediately yields the following.

Theorem 4.1

Suppose that S is a simply connected complete surface with Gauss curvature K, β2K0 everywhere. Let Dρ be a geodesic disc of radius ρ, then

λ1(Dρ)(β2cothβρ)2+(πρ)2. 4.3

Proof

Since Dρβ2 satisfies the hypotheses of Lemma 4.1 when α=β, hence

λ1(Dρβ2)(β2cothβρ)2+(πρ)2.

By Corollary 4.1, we immediately obtain (4.3). We complete the proof of Theorem 4.1. □

Since the function xcothx is monotonically increasing for x0, hence inequality (4.3) is stronger than (4.2). Let ρ in (4.3), we can easily have the following corollary.

Corollary 4.2

Let S be a simply connected complete surface with Gauss curvature K, β2K0 everywhere. Then

λ1(S)β24.

Acknowledgements

The authors would like to thank anonymous referees for helpful comments and suggestions that directly led to the improvement of the original manuscript. This paper is supported in part by Natural Science Foundation Project (grant number: #11671325).

Footnotes

Competing interests

The authors declare that they have no competing interests.

Authors’ contributions

All authors read and approved the final manuscript.

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Contributor Information

Niufa Fang, Email: nfafang@163.com.

Jiazu Zhou, Email: zhoujz@swu.edu.cn.

References

  • 1.Osserman R. Isoperimetric inequalities and eigenvalues of the Laplacian; Proceedings of the International Congress of Mathematicians; 1978. [Google Scholar]
  • 2.Blaschke W. Kreis und Kugal. 2. Berlin: de Gruyter; 1956. [Google Scholar]
  • 3.Osserman R. The isoperimetric inequalities. Bull. Am. Math. Soc. 1978;84(6):1182–1238. doi: 10.1090/S0002-9904-1978-14553-4. [DOI] [Google Scholar]
  • 4.Luo M, Xu W, Zhou J. Translative containment measure and symmetric mixed isohomothetic inequalities. Sci. China Math. 2015;58:2593–2610. doi: 10.1007/s11425-015-5074-5. [DOI] [Google Scholar]
  • 5.Osserman R. Bonnesen-style isoperimetric inequalities. Am. Math. Mon. 1978;86:1–29. doi: 10.2307/2320297. [DOI] [Google Scholar]
  • 6.Zhou J. Bonnesen-type inequalities on the plan. Acta Math. Sin. 2007;50:1397–1402. [Google Scholar]
  • 7.Santaló L. Integral Geometry and Geometric Probability. Reading: Addison-Wesley; 1976. [Google Scholar]
  • 8.Santaló L. Integral formulas in Crofton’s style on the sphere and some inequalities referring to spherical curves. Duke Math. J. 1942;9:707–722. doi: 10.1215/S0012-7094-42-00949-9. [DOI] [Google Scholar]
  • 9.Klain D. Bonnesen-type inequalities for surfaces of constant curvature. Adv. Appl. Math. 2007;39:143–154. doi: 10.1016/j.aam.2006.11.004. [DOI] [Google Scholar]
  • 10.Zeng C, Ma L, Zhou J, Chen F. The Bonnesen isoperimetric inequality in a surface of constant curvature. Sci. China Math. 2012;55(9):1913–1919. doi: 10.1007/s11425-012-4405-z. [DOI] [Google Scholar]
  • 11.Zhou J, Chen F. The Bonnesen-type inequalities in a plane of constant curvature. J. Korean Math. Soc. 2007;44:1363–1372. doi: 10.4134/JKMS.2007.44.6.1363. [DOI] [Google Scholar]
  • 12.Wang X, Xu W, Zhou J. Some logarithmic Minkowski inequalities for nonsymmetric convex bodies. Sci. China Math. 2017 [Google Scholar]
  • 13.Zhang Z, Zhou J. Bonnesen-style Wulff isoperimetric inequality. J. Inequal. Appl. 2017;2017(1) doi: 10.1186/s13660-017-1305-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Hersch J. Caractérisation variationnelle d’une somme de valeurs propres consécutives. C. R. Acad. Sci. 1961;252:1714–1716. [Google Scholar]
  • 15.Cheeger J. The relation between the Laplacian and diameter for manifolds of non-negative curvature. Arch. Math. 1968;19:558–560. doi: 10.1007/BF01898781. [DOI] [Google Scholar]
  • 16.Chavel I, Feldman E. The first eigenvalue of Laplacian on manifolds of non-negative curvature. Compos. Math. 1974;29:43–53. [Google Scholar]
  • 17.Cheng S. Eigenvalue comparison theorems and its geometric applications. Math. Z. 1975;143:289–297. doi: 10.1007/BF01214381. [DOI] [Google Scholar]
  • 18.Yau S. Isoperimetric constants and the first eigenvalue of a compact Riemannian manifold. Ann. Sci. Éc. Norm. Supér. 1975;8:487–507. doi: 10.24033/asens.1299. [DOI] [Google Scholar]
  • 19.Cheeger J. Problems in Analysis: A Symposium in Honor of Salomon Bocher. Princeton: Princeton University Press; 1970. A lower bound for the smallest eigenvalue of the Laplacian; pp. 145–199. [Google Scholar]
  • 20.Cheng S. Eigenfunctions and eigenvalues of Laplacian. Proc. Symp. Pure Math. 1975;27:185–193. doi: 10.1090/pspum/027.2/0378003. [DOI] [Google Scholar]
  • 21.Li P, Yau S. Estimate of eigenvalue of compact Riemannian manifold. Proc. Symp. Pure Math. 1980;36:203–239. [Google Scholar]
  • 22.Li P. Eigenvalue estimates on homogenous manifolds. Comment. Math. Helv. 1980;55:347–363. doi: 10.1007/BF02566692. [DOI] [Google Scholar]
  • 23.Chen M, Wang F. General formula for lower bound of the first eigenvalue on Riemannian manifolds. Sci. China Ser. A. 1997;27(1):34–42. [Google Scholar]
  • 24.Hile G, Protter M. Inequalities for eigenvalues of the Laplacian. Indiana Univ. Math. J. 1980;29(4):523–538. doi: 10.1512/iumj.1980.29.29040. [DOI] [Google Scholar]
  • 25.Jia F. Estimate on the first eigenvalue of a compact Riemannian manifold. Chin. Ann. Math., Ser. A. 1991;12(4):496–502. [Google Scholar]
  • 26. Xu, S, Xue, C, Hu, Z, Jin, Y: Modern differential geometry: spectral theory and isospectrum problems, curvature and topological invariants. Pre. of USCT (2009)
  • 27.Xu S, Pang H. Estimate of the first eigenvalue on a compact Riemannian manifold. Math. Appl. 2001;14(1):116–119. [Google Scholar]
  • 28.Yang H. Estimate of the first eigenvalue of a compact Riemannian manifold. Sci. China Ser. A. 1990;33(1):39–51. [Google Scholar]
  • 29.Yau S, Schoen R. Lectures on Differential Geometry. Beijing: Higher Education Press; 2004. [Google Scholar]
  • 30.Zhong J, Yang H. On the estimate of the first eigenvalue of a compact Riemannian manifold. Sci. China Ser. A. 1983;27(12):1265–1273. [Google Scholar]
  • 31.Zhao D. Estimate of the first eigenvalue on a compact Riemannian manifold. Sci. China Ser. A. 1999;29(3):207–214. [Google Scholar]

Articles from Journal of Inequalities and Applications are provided here courtesy of Springer

RESOURCES