Sobolev embeddings with weights in complete Riemannian manifolds

https://doi.org/10.1016/j.indag.2020.02.007Get rights and content

Abstract

We prove Sobolev embedding Theorems with weights for vector bundles in a complete Riemannian manifold. We also get general Gaffney’s inequality with weights. As a consequence, under a “weak bounded geometry” hypothesis, we improve classical Sobolev embedding Theorems for vector bundles in a complete Riemannian manifold. We also improve known results on Gaffney’s inequality in a complete Riemannian manifold.

Introduction

In the following M(M,g) will be a C smooth connected complete Riemannian manifold without boundary unless otherwise stated. We shall just say “Riemannian manifold” to mean it.

The Sobolev inequalities in M for functions play a major role in the study of differential operators and nonlinear functional analysis. They are valid in Rn or if M is compact. More generally we have:

Theorem 1.1

Let M be a Riemannian manifold of dimension n with Ricci curvature bounded from below. The Sobolev embeddings for functions are valid for M if and only if there is a uniform lower bound for the volume of balls which is independent of their center, namely if and only if infxMVol(B(x,1))>0.

The necessity is a well known fact, see p.18 in [13] and was generalized by Carron [5].

The sufficiency of Theorem 1.1 was done by Varopoulos [22], see also Theorem 3.18, p. 37 in [13], based on the work of Coulhon and Saloff-Coste [7].

Let M be a Riemannian manifold and let GM be a smooth complex vector bundle over M with a metric connection G on G, i.e. for any vector field X on M we have X(u,v)=(Xu,v)+(u,Xv) for any smooth sections u,v of G. We shall call a vector bundle with these properties an adapted vector bundle.

We shall introduce weights, given by the geometry, on the Riemannian manifold M, in order to have Sobolev embeddings on vector bundles, always valid with these weights, without any curvature conditions nor volume control.

In order to state the results, we need some definitions.

Definition 1.2

Let M be a Riemannian manifold and xM. We shall say that the geodesic ball B(x,R) is (0,ϵ)-admissible if there is a chart (B(x,R),φ) such that:

  • ()

    (1ϵ)δijgij(1+ϵ)δij in B(x,R) as bilinear form.

    We shall say that the geodesic ball B(x,R) is (1,ϵ)-admissible if moreover

  • ()

    |β|=1Rsupi,j=1,,n,yBx(R)|βgij(y)|ϵ.

We shall denote with A(0,ϵ) the set of (0,ϵ)-admissible balls on M and A(1,ϵ) the set of (1,ϵ)-admissible balls on M.

The (0,ϵ)-admissible balls will be adapted to functions and the (1,ϵ)-admissible balls will be adapted to sections of the vector bundle G. We shall use the shortcut s.t. to mean such that in formula.

Definition 1.3

Let xM, we set R(x)sup{R>0s.t.B(x,R)A(ϵ)}. We shall say that Rϵ(x)min(1,R(x)2) is the ϵ-admissible radius at x.

Here A(ϵ) will be either A(0,ϵ) or A(1,ϵ) depending on the context, functions or sections of G. The notation Rϵ(x) will mean either R0,ϵ(x) or R1,ϵ(x) depending on the choice of A(ϵ).

Let w be a weight on M, i.e. a positive measurable function on M. If kN and r1 are given, we denote by CGk,r(M,w) the space of smooth sections ω of G in C(M) such that |jω|Lr(M,w) for j=0,,k with the pointwise modulus associated to the pointwise scalar product and where is the connection associated to G. Let WGk,r(M,w) be the completion of CGk,r(M,w). See the precise Definition 4.1.

Our general result with weights is:

Theorem 1.4

Let M be a Riemannian manifold of dimension n. Let G be an adapted vector bundle over M. Let r1,mk0 and 1s=1r(mk)n>0. Let w(x)Rϵ(x)γ and wRϵ(x)ν with νs(2+γr). Then WGm,r(M,w) is embedded in WGk,s(M,w) and: uWGm,r(M,w),uWGk,s(M,w)CuWGm,r(M,w).

Because the admissible radius Rϵ(x) is always smaller than 1, to get rid of the weights we just need that there exists a δ>0 such that for any xM, we have R(x)δ.

This is precisely what is given by a Theorem of Hebey–Herzlich  [14, Corollary, p. 7], which has the easy Corollary:

Theorem 1.5

Let M be a Riemannian manifold. If the injectivity radius verifies rinj(x)i>0 and the Ricci curvature verifies Rc(M,g)(x)λgx for some λR and all xM, then there exists a positive constant δ>0, depending only on n,ϵ,λ,i such that for any xM,R0,ϵ(x)δ.

If moreover we have |Rc(M,g)(x)|C for all xM, then there exists a positive constant δ>0, depending only on n,ϵ,i and C, such that for any xM,R1,ϵ(x)δ.

So the following results will be consequences of Theorem 1.4 and of the Theorem of Hebey–Herzlich.

To state them precisely we shall need to weaken the definition of bounded geometry.

Definition 1.6

A Riemannian manifold M has k-order weak bounded geometry if:

the injectivity radius r(x) at xM is bounded below by some constant δ>0 for any xM

for 0jk, the covariant derivatives jRc of the Ricci curvature tensor are bounded in L(M) norm.

Recall the classical Theorem of Cantor [4]:

Theorem 1.7

Let (M,g) be a Riemannian manifold. Let GM be a complex smooth adapted vector bundle over M. Suppose moreover that M verifies:

C1: The injectivity radius of M is bounded away from zero.

C2: There is a δ such that for each xM and V,WTxM, the sectional curvature satisfies |Kx(V,W)|<δ.

Let 0k<m and 1s=1r(mk)n. Let uWGm,r(M). Then we have uWGk,s(M) with the control: uWGk,s(M)sCuWGm,r(M)r.

Here we prove:

Theorem 1.8

Let M be a Riemannian manifold. Let GM be a complex smooth adapted vector bundle over M. Suppose M has a 0-order weak bounded geometry. Let 0k<m and 1s=1r(mk)n. Let uWGm,r(M). Then we have uWGk,s(M) with the control: uWGk,s(M)sCuWGm,r(M)r.In the case of functions instead of sections of G, the conditions are weaker: If the injectivity radius of M is bounded away from zero and if the Ricci curvature verifies Rc(M,g)(x)λgx for some λR and all xM, we get uWk,s(M)sCuWm,r(M)r.

This improves the Theorem by Cantor because he used the hypothesis that all the sectional curvatures are bounded and here we need only that the Ricci curvature be bounded.

We also prove a global Gaffney’s type inequality:

Theorem 1.9

Let M be a Riemannian manifold. Let d be the exterior derivation on M and d its formal adjoint. Let ω be a p-differential form in C0(Λp(M)). If M has a 0-order weak bounded geometry, then we have, with r1: ωWp1,r(M)C(d(ω)Lp+1r(M)+d(ω)Lp1r(M)+ωLpr(M)).

And, as an easy corollary:

Corollary 1.10

Let M be a Riemannian manifold with a 0-order weak bounded geometry. Let r1 and let ω be a p-differential form in C0(Λp(M)). We have: ωLps(M)C(d(ω)Lp+1r(M)+d(ω)Lp1r(M)+ωLpr(M))with 1s=1r1n>0.

N. Lohoué [17] proved the same result under the stronger hypothesis that M has a 2-order bounded geometry plus some other hypotheses on the Laplacian and the range of r. Here already 0-order weak bounded geometry is enough.

This work is presented as follows:

In the next Section we study the main properties of the ϵ-admissible balls.

In Section 3 we define the vector bundle G we are interested in and precise the metric connection G on it we shall use.

In Section 4 we define the Sobolev spaces of smooth sections of G, with weights. We prove in this Section a generalization of a nice result of T. Aubin [3] which says that, in order to prove Sobolev embeddings for sections of G with weights, we have just to prove them at the first level. This is crucial for our estimates.

In Section 5 we prove the local estimates, i.e. for sections of G in the ϵ-admissible balls.

In Section 5.3 we also prove a local Gaffney type inequality, using a result by C. Scott [19].

Then in order to get global results we group the ϵ-admissible balls via a Vitali type covering in Section 6.

In Section 7 we prove the global estimates for functions, sections of G and Gaffney type in Lr.

In Section 8 we improve the classical Sobolev embeddings to the case of Riemannian manifolds with weak bounded geometry, by use of a Theorem of Hebey–Herzlich [14]. This implies the validity of Sobolev embeddings for vector bundles in compact Riemannian manifold without boundary.

In Section 8.3 we deduce from the compact case without boundary the validity of Sobolev embeddings for vector bundles in compact Riemannian manifold with smooth boundary. We use here the method of the “double” manifold.

Finally in Section 8.4 we introduce the lifted doubling property and we study the case of hyperbolic manifolds.

Section snippets

Admissible balls

Recall the definition of the admissible radius:

Definition 2.1

Let M be a Riemannian manifold. Let xM, we set R(x)sup{R>0s.t.B(x,R)A(ϵ)}. We shall say that Rϵ(x)min(1,R(x)2) is the ϵ-admissible radius at x.

Then clearly if B(x,R)A(ϵ), i.e. is ϵ-admissible, then so is B(x,S) if SR.

Remark 2.2

Let x,yM. Suppose that R(x)>dg(x,y), where dg(x,y) is the Riemannian distance between x and y. Consider the ball B(y,ρ) of center y and radius ρR(x)dg(x,y). This ball is contained in B(x,R(x)) hence, by definition of R(x

Vector bundle

Let M be a Riemannian manifold and let GM be an adapted complex vector bundle over M. Recall that this means that G has a smooth scalar product (,) and a metric connection G:C(M,G)C(M,GTM), i.e. verifying d(u,v)=(Gu,v)+(u,Gv), where d is the exterior derivative on M. See [21, Section 13].

Lemma 3.1

The ϵ-admissible balls B(x,Rϵ(x)) trivialize the bundle G.

Proof

Because if B(x,R) is an ϵ-admissible ball, we have by Remark 2.3 that Rrinj(x). Then, one can choose a local frame field for G on B(x,R) by

Sobolev spaces for sections of G with weight

We have seen that G:C(M,G)C(M,GTM). On the tensor product of two Hilbert spaces we put the canonical scalar product (uω,vμ)(u,v)(ω,μ), with uωGTM, and completed by linearity to all elements of the tensor product. On TM we have the Levi-Civita connection M, which is of course a metric one, and on G we have the metric connection G so we define a connection on the tensor product GTM: GTM(uω)=(Gu)ω+u(TMω)by asking that this connection be a derivation. We get easily that G

Sobolev comparison estimates for functions

Lemma 5.1

We have the Sobolev comparison estimates, where B(x,R) is a (0,ϵ)-admissible ball in M and φ:B(x,R)Rn is the admissible chart relative to B(x,R), uW1,r(B(x,R)),uW1,r(B(x,R))Cuφ1W1,r(φ(B(x,R))),and, with Be(0,t) the Euclidean ball in Rn centered at 0 and of radius t, vW1,r(Be(0,(1ϵ)R))CuW1,r(B(x,R)).The constants C,C depending only on ϵ,n and not on BA(ϵ).

Proof

We have to compare the norms of u,u, with the corresponding ones for vuφ1 in Rn. First we have because (1ϵ)δijgij(1+ϵ

Vitali covering

Lemma 6.1

Let F be a collection of balls {B(x,r(x))} in a metric space, with B(x,r(x))F,0<r(x)R. There exists a disjoint subcollection G of F with the following properties: every ball B in F intersects a ball C in G and B5C.

This is a well known lemma, see for instance [11], section 1.5.1.

Fix ϵ>0 and let xM,r(x)Rϵ(x)10, where Rϵ(x) is the admissible radius at x. We built a Vitali covering with the collection F{B(x,r(x))}xM. The previous lemma gives a disjoint subcollection G such that every ball B

Global estimates for sections of G

Lemma 7.1

We have, for any section f of G, with w(x)R(x)μ and B(x)B(x,R(x)), with xD(ϵ) and R(x)Rϵ(x), that: τ1,fLGτ(M,w)τxD(ϵ)R(x)μfLGτ(B(x))τ.

Proof

We consider the function |f|. We set w(x)R(x)μ to get an adapted weight. Let xD(ϵ), then B(x)B(x,R(x))Cϵ.

We have, because Cϵ is a covering of M, fLτ(M,w)τM|f(x)|τw(x)dv(x)xD(ϵ)B(x)|f(y)|τR(y)μdv(y). Because, by Lemma 2.4, we have that for any yB,R(y)2R(x), then we get xD(ϵ)B(x)|f(y)|τR(y)μdv(y)xD(ϵ)2μR(x)μB(x)|f(y)|τdv(y)2μxD(

Applications

We shall give some examples where we have classical estimates using that for any xM,Rϵ(x)δ>0, via [14, Corollary, p. 7] (see also Theorem 1.3 in the book by Hebey [13]).

We have:

Corollary 8.1

Let M be a Riemannian manifold. If the injectivity radius verifies rinj(x)i>0 and the Ricci curvature verifies Rc(M,g)(x)λgx for some λR and all xM, then there exists a positive constant δ>0, depending only on n,ϵ,λ,α,i such that for any xM,rH(1+ϵ,0,α)(x)δ.

Hence R0,ϵ(x)δ for all xM.

If we have |Rc(M,g)(x)|C for

References (23)

  • GuneysuB. et al.

    Calderon-Zygmund inequality and Sobolev spaces on noncompact riemannian manifolds

    Adv. Math.

    (2015)
  • AmarE.

    Lr Solutions of elliptic equation in a complete riemannian manifold

    J. Geom. Anal.

    (2018)
  • AmarE.

    Sobolev solutions of parabolic equation in a complete riemannian manifold

    (2019)
  • AubinT.
  • CantorM.

    Sobolev inequalities for riemannian bundles

    Bull. Amer. Math. Soc.

    (1974)
  • CarronG.

    Inégalités Isopérimétriques sur les Variétés Riemanniennes

    (1994)
  • CheegerJ. et al.

    Finite propagation speed, kernel estimates for functions of the Laplace operator, and the geometry of complete riemannian manifolds

    J. Differential Geom.

    (1982)
  • CoulhonT. et al.

    Isopérimétrie pour les groupes et les variétés

    Rev. Math. Iberoam.

    (1993)
  • do CarmoM.P.
  • DuffG.

    Differential forms in manifolds with boundary

    Ann. Math.

    (1952)
  • ElderingJ.

    Persistence of Noncompact Normally Hyperbolic Invariant Manifolds in Bounded Geometry

    (2012)
  • Cited by (0)

    View full text