Abstract
In this paper, we obtain weighted Sobolev type inequalities with explicit constants that extend the inequalities obtained by Guo et al. (Math Res Lett 28(5):1419–1439, 2021) in the Riemannian setting. As an application, we prove some new logarithmic Sobolev type inequalities in some smooth metric measure spaces.
Similar content being viewed by others
1 Introduction
The study of Sobolev inequalities with sharp constants has a long tradition in analysis and geometry. For example, on the unit sphere \(S^{n-1}\) endowed with its standard metric, Escobar [7] classified all positive solutions of
by an integral method and hence [8] proved that, for all \(u\in C^{\infty }({\overline{B}}^{n})\),
where \(d\nu \) and \(\textrm{vol}(S^{n-1})\) are respectively the Riemannian measure and the Riemannian volume of \(S^{n-1}\). This inequality plays an important role in the study of the Yamabe problem on Riemannian manifolds. Note that, using harmonic analysis, Beckner [1] derived a family of inequalities
provided \(1<q< \infty \), if \(n=2\), and \(1<q\le \frac{n}{n-2}\), if \(n\ge 3\). The corresponding Euler–Lagrange equation to (1.3) is
It is apparent that the case \(n\ge 2\) and \(q=\frac{n}{n-2}\) of (1.3) and (1.4) are just (1.2) and (1.1) respectively. Also, in the same paper, Beckner [1] confirmed
provided \(1<q< \infty \), if \(n=2\) or 3, and \(1<q\le \frac{n+1}{n-3}\), if \(n\ge 3\). By considering the Euler–Lagrange equation (1.4) and using integral methods, Bidaut-Véron and Véron [2] were able to give another proof of the inequality (1.5). More results about the Sobolev inequalities on the unit sphere can be found in [5, 10, 11].
Guo et al. [9] recently generalized the spherical inequality (1.3) to any smooth compact Riemannian manifolds with nonnegative sectional curvature and strictly convex boundary. They proved the following Sobolev inequality.
Theorem 1.1
[9] Let \((M^{n},g)\) be a smooth compact Riemannian manifold with nonnegative sectional curvature and \(\uppercase {ii}\ge 1\) on the boundary \(\partial M\). Assume \(2\le n\le 8\) and \(1<q\le \frac{4n}{5n-9}\). Then for any \(u\in C^{\infty }(M)\), we have
In the limiting case Theorem 1.1 implies the following logarithmic Sobolev inequality.
Corollary 1.2
[9] Let \((M^{n},g)\) be a smooth compact Riemannian manifold with nonnegative sectional curvature and \(\uppercase {ii}\ge 1\) on the boundary \(\partial M\). Assume \(2\le n\le 8\). Then for any \(u\in C^{\infty }(M)\) with \(\frac{1}{\textrm{vol}(\partial M)}\int _{\partial M}u^{2} \,d\Sigma =1\), we have
The purpose of the present paper is to adapt the technique that has been used in [9] to the setting of smooth metric measure spaces with nonnegative sectional curvature and strictly convex boundary. We generalize Theorem 1.1 and Corollary 1.2 to the smooth metric measure space. Let us fix some required notations before stating our results.
Let (M, g) be a smooth compact n-dimensional Riemannian manifold and \(\phi \) be a \(C^2(M)\) function. We denote \(\nabla \), \(\Delta \) and \(\nabla ^{2}\) the gradient, Laplacian and Hessian operator on M with respect to g, respectively. \(\textrm{Ric}\) and R denote Ricci curvature and scalar curvature, respectively. An n-dimensional smooth metric measure space \((M, g, d\sigma =e^{-\phi } \,d\Omega )\) is a smooth compact n-dimensional Riemannian manifold (M, g) endowed with a weighted measure \(e^{-\phi } \,d\Omega \) and \(\,d\Omega \) is the Riemannian volume element of the metric g. On a smooth metric measure space \((M, g, d\sigma =e^{-\phi } \,d\Omega )\), we let
stand for the Bakry–Émery Ricci curvature which is also called \(\infty \)-Bakry–Émery Ricci curvature, i.e., the \(m =\infty \) case of the following m-Bakry–Émery Ricci curvature defined by
with some constant \(m\ge n\), and \(m=n\) if and only if \(\phi \) is a constant. The equation \(\textrm{Ric}_{\phi }=\lambda g\) for some constant \(\lambda \) is just the gradient Ricci soliton equation, which plays an important role in the study of Ricci flow. The equation \(\textrm{Ric}_{\phi }^{m}=\lambda g\) corresponds to the quasi-Einstein equation [4], which has been studied by many mathematicians. In recent years, the smooth metric measure space received much attention from many mathematicians, see [6, 12,13,14,15,16, 19,20,21,22, 24,25,26] and the references therein.
Let \(\mathbf {\nu }\) be the unit outward normal of \(\partial M\). Define the second fundamental form of \(\partial M\) by \(\uppercase {ii}(X,Y)=\langle \nabla _{X}\mathbf {\nu }, Y\rangle \) for any two tangent vector fields X and Y on M, and the mean curvature by \(H = tr(\uppercase {ii})\). The f-mean curvature (see [22, p. 398]) at a point \(x\in M\) with respect to \(\mathbf {\nu }\) is given by \(H_{\phi }(x)=H(x)-\langle \nabla \phi (x), \mathbf {\nu }(x)\rangle \), where \(\langle \cdot , \cdot \rangle \) denotes the Riemannian metric g.
On \((M, g, d\sigma =e^{-\phi } \,d\Omega )\), we consider the weighted Laplacian as follows:
where \(\nabla \) denotes the Levi-Civita connection, \(div=tr(\nabla \cdot )\) denotes the Riemannian divergence operator, and \(\Delta =div \nabla \) is the Laplace–Beltrami operator. Notice that the Green formula (the integration by parts formula)
holds provided u or h belongs to \(C^{2}(M)\), where \(u_{\mathbf {\nu }}=\langle \mathbf {\nu }, \nabla u\rangle \), and \(\,dv= e^{-\phi } \,d\Sigma \) and \(\,d\Sigma \) is the volume form on \(\partial M\).
The following is one of our main results.
Theorem 1.3
Let \((M, g, d\sigma =e^{-\phi } \,d\Omega )\) be a smooth compact metric measure space with nonnegative sectional curvature and \(II\ge c\) for a positive constant c on the boundary \(\partial M\). Let \(\phi \) be a potential function such that \(\nabla ^{2}\phi -\frac{1}{m-n}d\phi \otimes d\phi \ge 0\) on M. Assume \(2\le m\le 8\) and \(1<q\le \frac{4m}{5m-9}\). Then for any \(u\in C^{\infty }(M)\), we have
where \(\textrm{vol}_{\phi }(\partial M)\) is the weighted area of \(\partial M\).
Note that Theorem 1.3 recovers Theorem 1.1 obtained by Guo–Hang–Wang. Moreover, Theorem 1.3 implies the logarithmic type Sobolev inequality.
Corollary 1.4
Let \((M, g, d\sigma =e^{-\phi } \,d\Omega )\) be a smooth compact metric measure space with nonnegative sectional curvature and \(\uppercase {ii}\ge c\) for a positive constant c on the boundary \(\partial M\). Let \(\phi \) be a potential function such that \(\nabla ^{2}\phi -\frac{1}{m-n}d\phi \otimes d\phi \ge 0\) on M. Assume \(2\le m\le 8\). Then for any \(u\in C^{\infty }(M)\) with \(\frac{1}{\textrm{vol}_{\phi }(\partial M)}\int _{\partial M}u^{2} \,dv=1\), we have
The proof of Theorem 1.3 is based on the uniqueness results, which we state in this setting as follows.
Theorem 1.5
Let \((M, g, d\sigma =e^{-\phi } \,d\Omega )\) be a smooth compact metric measure space with boundary \(\partial M\). Assume that the sectional curvature is nonnegative on M, and the second fundamental form \(\uppercase {ii}\ge c\) for a positive constant c on \(\partial M\). Let \(\phi \) be a potential function such that \(\nabla ^{2}\phi -\frac{1}{m-n}d\phi \otimes d\phi \ge 0\) on M. Let u be a positive solution of the following system:
Then the only positive solution to the Eq. (1.11) is constant if \(\lambda \le \frac{c}{q-1}\), provided \(2 \le m \le 8\) and \(1 <q \le \frac{4m}{5m-9}\).
If we take \(\phi =constant\), then Theorem 1.5 becomes Theorem 2 proved by Guo et al. [9].
The rest of this paper is organized as follows. In Sect. 2, we establish some elementary lemmas (Lemmas 2.1, 2.2, 2.3). The uniqueness results (Theorem 1.5) are discussed in Sect. 3. Finally, Sect. 4 is dedicated to the proof of Theorem 1.3 and Corollary 1.4, respectively.
2 Preliminaries
In this section, we drive some useful lemmas that will be used later.
Lemma 2.1
(Weighted Reilly formula) Let \((M, g, d\sigma =e^{-\phi } \,d\Omega )\) be a smooth compact metric measure space with boundary \(\partial M\) and \(V:M\rightarrow R\) be a twice differential function. Given a smooth function f on M, we have
where J:= \(|\nabla ^{2} f|^{2}-\frac{1}{m}\left( {\mathbb {L}}_{\phi }f\right) ^{2}+\textrm{Ric}_{\phi }(\nabla f,\nabla f)\), \(\overline{{\mathbb {L}}_{\phi }}\cdot ={\overline{\Delta }}-g({\overline{\nabla }}\phi \),\({\overline{\nabla }}\cdot )\), \({\overline{\nabla }}\) and \({\overline{\Delta }}\) are respectively the gradient operator and the Laplace operator on \(\partial M\).
Proof
Now in the calculations that follow (at a point \(x\in \partial M\)), we will use an orthonormallocal frame \(\{e_{1},\ldots , e_{n}\}\) such that \(e_{1},\ldots , e_{n-1}\) are tangent to the boundary \(\partial M\) and \(e_{n}=\mathbf {\nu }\) is the outward unit normal to \(\partial M\).
We use the integration by parts and the Ricci identity to derive that
Using the integration by parts again, we obtain
and
Inserting (2.3) and (2.4) into (2.2), we attain
By the integration by parts, we have
Substituting (2.6) into (2.5), we can get
We easily infer from (2.7) the following:
Applying the integration by parts, we get
Taking (2.9) into (2.8), we have
Note that \(H_{\phi } = H -\langle \nabla \phi , \mathbf {\nu }\rangle \). From the Gauss–Weingarten formula
and
we have
Putting (2.11) and (2.12) into (2.13), we get
Again applying the integration by parts shows that
Then combining (2.14) and (2.15), we obtain
Substituting (2.16) into (2.10), we can get
which is equivalent to
Therefore, by the definition of J, we conclude the proof of Lemma 2.1. \(\square \)
Let \((M, g, d\sigma =e^{-\phi } \,d\Omega )\) be a smooth compact metric measure space with boundary \(\partial M\) and \(f \in C^{\infty }(M)\) be a positive function. Let \(\omega \) be another smooth function on M satisfying the following boundary conditions
We take \(V=f^{b}\omega \), for any \(b\in {\mathbb {R}}\).
Lemma 2.2
Proof
A direct calculation gives
Plugging these equations into (2.1), we have
Using the boundary conditions for \(\omega \) in (2.19) yields
Using the integration by parts and \(\omega |_{\partial M}=0\), we get
Substituting (2.23) into (2.22), we attain
Reorganizing yields the desired equality (2.20). Therefore the proof of Lemma 2.2 is completed. \(\square \)
Lemma 2.3
(Weighted Pohozave identity)
Proof
For any smooth vector field \(\nabla \omega \), we can get
Note that \(\nabla \omega =-\mathbf {\nu }\) on \(\partial M\). Multiplying both sides of the above identity by \(f^{b}\) and integrating yields
By a straightforward computation, we obtain
Substituting (2.26) into (2.25), we get
Thus, we have
which completes the proof of Lemma 2.3. \(\square \)
Lemma 2.4
[17] Let u be a smooth function on M. For every \(m\ge n\), we have
Moreover, the equality in (2.27) holds if and only if
3 Proof of Theorem 1.5
In this section, we shall prove Theorem 1.5.
Proof of Theorem 1.5
Let u be a positive solution of (1.11). Let a be a nonzero real number to be determined later and take \(u=f^{-a}\). Then \(f> 0\) satisfies the following equation
For any \(s\in {\mathbb {R}}\), multiplying the first equation in (3.1) by \(f^{s}\) and integrating over M yields
Inserting (3.1) into (2.20), we obtain
Taking (3.1) into (2.24), we have
In order to eliminate the term
we multiply (3.4) by \(\frac{\frac{b}{2}}{a+1+\frac{b}{2}}\) to get
Substituting (3.5) into (3.3), we can get
We choose \(b=-\frac{4}{3}(a+1)\). Then
\(\square \)
By arguing as in [23], we consider a weight function \(\psi :=\rho -c\frac{\rho ^{2}}{2}\), where \(\rho =d(\cdot ,\partial M)\) denotes the distance function to the boundary \(\partial M\). Notice that \(\psi \) is smooth near \(\partial M\) and satisfies
From now on we assume that M has nonnegative sectional curvature and \(\uppercase {ii}\ge c\) for a positive constant c on \(\partial M\). By the Hessian comparison theorem [18] \(\rho \le \frac{1}{c}\) hence \(\psi \ge 0\) and \(-\nabla ^{2}\psi \ge cg\) in the support sense. To overcome the difficulty that \(\psi \) is not smooth, we also need
Proposition 3.1
[23] Fix a neighborhood C of \(Cut(\partial M)\) in the interior of M, with \(Cut(\partial M)\) the cut-locus of points at the boundary \(\partial M\). Then for any \(\varepsilon >0\), there exists a smooth nonnegative function \(\psi _{\varepsilon }\) on M such that \(\psi _{\varepsilon }= \psi \) on \(M\backslash C\) and \(-\nabla ^{2}\psi _{\varepsilon }\ge (c-\varepsilon )g\).
Taking the weight \(\omega =\psi _{\varepsilon }\) in (3.6) yields
By letting \(\varepsilon \rightarrow 0\) and shrinking the neighborhood, we get the following
Since the function \(\psi \) is smooth and \(-\nabla ^{2}\psi \ge cg\) on \(M\backslash C\), we obtain
Applying (3.2) and the boundary condition for f in (3.1), we have
which is equivalent to
where, with \(x=a^{-1}\)
By choosing a such that \(A,B,C\le 0\), we can get
Direct computation gives
The selection is possible when \(\frac{3}{2}q-1\le x\le \frac{3}{2}\frac{c}{\lambda }+\frac{1}{2}\) and \(\frac{3}{2}q-1\le \frac{m+9}{5m-9}\le 0\), that is, \((q-1)\lambda \le c\) and \(q\le \frac{4\,m}{5\,m-9}\). Since \(q>1\) we must have \(2\le m\le 8\). Then when \(q\le \frac{4m}{5m-9}\) and \((q-1)\lambda \le c\), take \(\frac{1}{a}=\frac{3}{2}q-1\), we get
Therefore, the left hand side of (3.8) is nonpositive while the right hand side is nonnegative. Thus, both sides of (3.8) are zero and we must have
From (3.9), we obtain
If \(q<\frac{4m}{5m-9}\) or \((q-1)\lambda < c\) we have \(A<0\) or \(B<0\), respectively and thus f must be constant. It needs to verify that f must also be constant when
With the assumption (3.11), we can get
Since \(\textrm{Ric}^{m}_{\phi }=0\) and \(\nabla ^{2}\phi -\frac{1}{m-n}d\phi \otimes d\phi \ge 0\) on M, we have \(\textrm{Ric}(\nabla f,\cdot )=0\). We denote
By (3.10), we obtain \(\nabla ^{2}f=\xi g\). Working with a local orthonormal frame we differentiate
Thus \(\xi _{j}=0\) and \(\xi \) is constant. To continue, recall that we have
Differentiating both sides we get
Thus, we have
By taking inner product on both sides with \(\nabla f\) and applying \(f>0\), we obtain \(\frac{(25m-12n+9)(m+9)n}{(12m-6n)^{2}}\xi ^{2}=0\). Since \(m\ge n\), we have \(\xi =0\) and thus \(\nabla f=0\) and f must be a constant function. This finishes the proof of Theorem 1.5.
4 Proof of Theorem 1.3 and Corollary 1.4
In this section, we shall prove Theorem 1.3 and Corollary 1.4.
Proof of Theorem 1.3
We suppose \(m>2\), and consider the following family of functionals \({{\mathcal {J}}}_{q}\), which is defined by
and consider \(\mu _{q}:= \inf \{{\mathcal {J}}_{q}(\eta ), \eta \in {\mathcal {H}}_{q}\}\), where
Another important key here is that the real-valued function
is decreasing. So
as \(n < m\). Using the compactness of the inclusions
for any \(q+1<\frac{2m}{m-2}\), we can confirm that \(\mu _{q}\) is realized by a positive function \(\psi _{q}\in {\mathcal {H}}_{q}\) and hence we can easily check that \(\psi _{q}\) verifies weakly the following system:
The regularity result of [3, Theorem 1] indicates that \(\psi _{q}\) is smooth, so by applying Theorem 1.5, we can infer that \(\psi _{q}\) is constant. Since \(\psi _{q}\in {\mathcal {H}}_{q}\), we have
Hence, recalling the definition of \(\mu _{q}\), we can get
Simple calculation can be obtained,
Considering any \(\eta \in H^{2}_{1}(\,d\sigma )\) satisfies
Hence,
Thus we complete the proof of Theorem 1.3 for \(m>2\). The case where \(m = 2\) (i.e., \(n = 2\) and \(\phi \) is constant), (1.10) can be obtained from [9, Corollary 1]. Therefore the proof of Theorem 1.3 is completed. \(\square \)
Using Theorem 1.3 we can prove Corollary 1.4.
Proof of Corollary 1.4
Under the assumption on u (4.3) can be written as
Let \(F(q):=\left( \frac{1}{\textrm{vol}_{\phi }(\partial M)}\int _{\partial M}|u|^{q+1} \,dv\right) ^{\frac{2}{q+1}}-1\). We can get
Taking limit \(q\downarrow 1\) and applying L’Hospital’s rule yields
Substituting (4.5) into (4.4), we get the desired inequality. we complete the proof. \(\square \)
Data Availability
Data sharing is not applicable to this article as no new data were created or analyzed in this study.
References
Beckner, W.: Sharp Sobolev inequalities on the sphere and the Moser–Trudinger inequality. Ann. Math. (2) 138(1), 213–242 (1993)
Bidaut-Véron, M., Véron, L.: Nonlinear elliptic equations on compact Riemannian manifolds and asymptotics of Emden equations. Invent. Math. 106(3), 489–539 (1991)
Cherrier, P.: Problèmes de Neumann non linéaires sur les variétés riemanniennes. J. Funct. Anal. 57(2), 154–206 (1984)
Case, J., Shu, Y., Wei, G.: Rigidity of quasi-Einstein metrics. Differ. Geom. Appl. 29(1), 93–100 (2011)
Chen, L., Lu, G., Tang, H.: Sharp stability of log-Sobolev and Moser–Onofri inequalities on the sphere. J. Funct. Anal. 285, 110022 (2023)
Du, F., Mao, J., Wang, Q., Xia, C.: Estimates for eigenvalues of weighted Laplacian and weighted \(p\)-Laplacian. Hiroshima Math. J. 51(3), 335–353 (2021)
Escobar, J.: Sharp constant in a Sobolev trace inequality. Indiana Univ. Math. J. 37(3), 687–698 (1988)
Escobar, J.: Uniqueness theorems on conformal deformation of metrics, Sobolev inequalities, and an eigenvalue estimate. Commun. Pure Appl. Math. 43(7), 857–883 (1990)
Guo, Q., Hang, F., Wang, X.: Liouville type theorems on manifolds with nonnegative curvature and strictly convex boundary. Math. Res. Lett. 28(5), 1419–1439 (2021)
Guo, Q., Wang, X.: Uniqueness results for positive harmonic functions on \(\overline{{\mathbb{B} }_{n}}\) satisfying a nonlinear boundary condition. Calc. Var. Partial Differ. Equ. 59, 146 (2020)
Gu, P., Li, H.: A proof of Guo-Wang’s conjecture on the uniqueness of positive harmonic functions in the unit ball. arXiv:2306.15565v1
Huang, G., Zhu, M.: Some geometric inequalities on Riemannian manifolds associated with the generalized modified Ricci curvature. J. Math. Phys. 63, 111508 (2022)
Huang, G., Ma, B.: Sharp bounds for the first nonzero Steklov eigenvalues for \(f\)-Laplacians. Turk. J. Math. 40(4), 770–783 (2016)
Huang, G., Ma, B.: Eigenvalue estimates for submanifolds with bounded \(f\)-mean curvature. Proc. Indian Acad. Sci. Math. Sci. 127, 375–381 (2017)
Huang, G., Li, Z.: Liouville type theorems of a nonlinear elliptic equation for the \(V\)-Laplacian. Anal. Math. Phys. 8(1), 123–134 (2018)
Huang, Q., Ruan, Q.: Applications of some elliptic equations in Riemannian manifolds. J. Math. Anal. Appl. 409(1), 189–196 (2014)
Ilias, S., Shouman, A.: Sobolev inequalities on a weighted Riemannian manifold of positive Bakry–Émery curvature and convex boundary. Pac. J. Math. 294(2), 423–451 (2018)
Kasue, A.: A Laplacian comparison theorem and function theoretic properties of a complete Riemannian manifold. Jpn. J. Math. (N.S.) 8(2), 309–341 (1982)
Li, X.: Liouville theorems for symmetric diffusion operators on complete Riemannian manifolds. J. Math. Pures Appl. 84(10), 1361–1995 (2005)
Li, H., Wei, Y.: \(f\)-minimal surface and manifold with positive \(m\)-Bakry–Émery Ricci curvature. J. Geom. Anal. 25, 421–435 (2015)
Ma, L., Du, S.: Extension of Reilly formula with applications to eigenvalue estimates for drifting Laplacians. C. R. Math. Acad. Sci. Paris 348(21–22), 1203–1206 (2010)
Wei, G., Wylie, W.: Comparison geometry for the Bakry–Émery Ricci tensor. J. Differ. Geom. 83(2), 377–405 (2009)
Xia, C., Xiong, C.: Escobar’s Conjecture on a sharp lower bound for the first nonzero Steklov eigenvalue. Peking Math. J. (2023). https://doi.org/10.1007/s42543-023-00068-2
Zeng, F.: Gradient estimates for a nonlinear parabolic equation on complete smooth metric measure spaces. Mediterr. J. Math. (N.S.) 18(4), 21 (2021)
Zeng, L., Sun, H.: Eigenvalues of the drifting Laplacian on smooth metric measure spaces. Pac. J. Math. 319(2), 439–470 (2022)
Zhang, L.: Global lower bounds on the first eigenvalue for a diffusion operator. Bull. Malays. Math. Sci. Soc. 43(5), 3847–3862 (2020)
Funding
This work was supported by NSFC, National Natural Science Foundation of China (no. 12101530), Scientific and Technological Key Projects of Henan Province (no. 232102310321), the Key Scientific Research Program in Universities of Henan Province (nos. 21A110021, 22A110021) and Nanhu Scholars Program for Young Scholars of XYNU (no. 2023).
Author information
Authors and Affiliations
Contributions
PYW conceived the idea of the study and wrote the paper. HTC discussed the results and revised the manuscript.
Corresponding author
Ethics declarations
Conflict of interest
The authors declare that they have no conflict of interest.
Ethics approval and consent to participate
Not applicable.
Consent for publication
Not applicable.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Wang, P., Chang, H. Weighted Sobolev Type Inequalities in a Smooth Metric Measure Space. J Nonlinear Math Phys 31, 5 (2024). https://doi.org/10.1007/s44198-024-00168-2
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s44198-024-00168-2