数学物理学报, 2024, 44(1): 26-36

关于常曲率空间中子流形$p$-调和$\ell$-形式的一个消灭定理

张友花,

福建师范大学数学与统计学院 福州 350000

A Vanishing Theorem for$p$-harmonic$\ell$-forms in Space with Constant Curvature

Zhang Youhua,

Fujian Normal University, Fuzhou 350000

收稿日期: 2022-10-8   修回日期: 2023-09-28  

基金资助: 国家自然科学基金(2021J01165)

Received: 2022-10-8   Revised: 2023-09-28  

Fund supported: NSFC(2021J01165)

作者简介 About authors

张友花,E-mail:505289305@qq.com

摘要

令$M^{n}(n \geq 3)$是常曲率空间$N^{n+m}(c)$中的、具有平坦法丛的完备非紧致的浸入子流形. 假设$M^{n}(n \geq 3)$满足四个不同的具体几何条件之一时, 该文利用 Bochner-Weitzenböck 公式和 Sobolev 不等式, 通过 Duzaar-Fuchs 截断函数方法, 证明了$M^{n}$上不存在非平凡的$L^{\beta}~p$-调和$\ell$-形式, 其中$\beta\geq p\geq 2$.

关键词: 子流形; $p$-调和形式; 消灭性定理

Abstract

Let$M^{n}(n \geq 3)$be a complete non-compact submanifold immersed in a space with constant curvature$N^{n+m}(c)$with flat normal bundle. By using Bochner-Weitzenböck formula, Sobolev inequality, Moser iteration and Fatou lemma, we prove that every$L^{\beta}~p$-harmonic forms on$M$is trivial if$M^{n}$satisfies some geometic conditions, where$\beta\geq p\geq 2$.

Keywords: Submanifold; $p$-harmonic form; Vanishing theorem

PDF (551KB) 元数据 多维度评价 相关文章 导出 EndNote| Ris| Bibtex  收藏本文

本文引用格式

张友花. 关于常曲率空间中子流形$p$-调和$\ell$-形式的一个消灭定理[J]. 数学物理学报, 2024, 44(1): 26-36

Zhang Youhua. A Vanishing Theorem for$p$-harmonic$\ell$-forms in Space with Constant Curvature[J]. Acta Mathematica Scientia, 2024, 44(1): 26-36

1 背景介绍

黎曼流形上的调和形式的消灭定理的相关研究一直以来都是几何学的一个重要研究方向. 设$(M^n,g)$是$n$维完备定向的黎曼流形, 其黎曼度量为$g$. 定义$\mathrm{d}$为外微分算子, 其对偶算子为

$\begin{aligned}\delta=\ast \mathrm{d}\ast,\end{aligned}$

其中$\ast$是 Hodge 星算子, 则 Hodge-Laplace 算子$\Delta$为

$\begin{aligned}\Delta=-(\mathrm{d} \delta+\delta \mathrm{d}).\end{aligned}$

对$M$上$\ell$-形式$\omega$, 若$\omega$满足$\mathrm{d}\omega=0 \mbox{和} \delta(|\omega|^{p-2}\omega)=0$, 则称$\omega$为$p$-调和$\ell$-形式. 当$p=2$时,$\omega$为调和形式. 当$\ell=0$, 即$\omega$是$p$-调和函数时,$\omega$即为能量泛函的一个临界点. Hodge 理论说明紧致流形上的调和$\ell$-形式同构于它的第$\ell$个 de Rham 上同调群, 也即第$\ell$个 Betti 数等于$p$-调和$\ell$-形式空间的维数. 但是 Hodge 理论不能用于完备非紧致流形, 所以大家都很自然地去考虑$L^2$调和形式. 早在1980年, Li[1]得到了重要的紧致黎曼流形上的 Sobolev 不等式. Tanno[2]证明了欧氏空间上完备可定向稳定的极小超曲面$M$上$L^2$调和$\ell$-形式空间上并不存在非平凡的$L^2$调和$\ell$-形式. 最近, Lin[3]给出了在法丛平坦的前提条件下的高阶微分形式的 Weitzenböck 公式的具体表达式, 同时证明了若第二基本形式的模长平方与平均曲率的平方满足一定的关系式, 则该子流形上不存在非平凡的$L^2$调和$\ell$-形式.

现考虑一般的$p$-调和形式, Zhang[4] 证明了具有非负 Ricci 曲率的完备流形上并不存在非平凡$L^q$($q>0)$可积的$p$-调和 1-形式. 近几年, Lin[5]证明当子流形是极小的或$\frac{n}{n+1}$-稳定时, 单位球空间$\mathbb{S}^{n+1}$中的$n$维完备非紧致超曲面上不存在非平凡$L^q$($q>0)$$p$-调和 1-形式, 亦可称为单位球$\mathbb{S}^{n+1}$空间中的$n$维完备非紧致超曲面上的消灭定理. 随后, Han[6]得出结论:当子流形的全曲率有限时, 球空间中完备非紧致子流形, 当其法丛平坦的时的$L^p$$p$-调和$\ell$-形式的有限性定理和消灭定理. 最近, Lin 和 Yang[7]通过假设某个与无迹第二基本形式相关的 Schrödinger 算子的指标为 0, 得出了 Hardamard 流形中完备非紧致子流形的$L^Q$$p$-调和形式的消灭定理.

受到上述结果的启发,本文将球空间形式推广至一般的常曲率空间, 同时去讨论其子流形上$L^Q$$p$-调和形式的消灭性定理, 得到如下结果.

定理 1.1 设完备非紧致流形$M^{n}(n \geq 3)$等距浸入到常曲率流形$N^{n+m}(c)$中,设$M^{n}$$(n \geq 3)$的法丛平坦, 当且仅当下列条件之一成立时

(i)

$|A|^{2} \leq \frac{n^{2}|H|^{2}-2\ell(\ell-n)c}{C_{n, \ell}}, \quad \operatorname{Vol}(M)=\infty.$

(ii) 当$c\leqslant 0$时

$\frac{n^{2}|H|^{2}+2 \ell(n-\ell)c}{C_{n, \ell}}<|A|^{2} \leqslant \frac{n^{2}|H|^{2}}{C_{n, \ell}} \text {和} \lambda_{1}(M)>-\frac{\ell(n-\ell)c \beta^{2}}{4\left[\beta-1+(p-1)^{2}K_{p, n, \ell}\right]}.$

(iii) 若全曲率$\|A\|_{n}$有上界, 且$M$的第一特征值$\lambda_{1}(M)>0$.

(a) 当截面曲率$c\leqslant0$时

$\|A\|_{n}^{2}<\min \left\{\frac{n^{2}}{C_{n, \ell} C_{1}(n)}, \frac{2}{C_{n, \ell} C_{1}(n)}\left[\frac{4\left[\beta-1+(p-1)^{2}K_{p, n, \ell}\right]}{\beta^{2}}+\frac{\ell(n-\ell)c}{\lambda_{1}(M)}\right]\right\};$

(b) 当截面曲率$c=1$时

$\|A\|_{n}^{2}<\min \left\{\frac{n^{2}}{C_{n, \ell} C_{2}(n)}, \frac{8\left[\beta-1+(p-1)^{2}K_{p, n, \ell}\right]}{C_{n,\ell} C_{2}(n)\beta^{2}(1+\frac{n^2}{\lambda_{1}(M)})}\right\},$

其中${C_{n, \ell}} \text{=} \max\{\ell,n-\ell\}$,$C_{1}(n)$,$C_{2}(n)$为仅与$n$有关的 Sobolev 常数.

(iv) 若第二基本形式模长的平方$\sup _{M}|A|^{2}<\infty$, 第一特征值$\lambda_{1}(M)$满足下列不等式关系

$\lambda_1(M)>\max \left\{\frac{\beta^2 C_{n, \ell} \sup _M|A|^2}{8\left[\beta-1+(p-1)^2 K_{p, n, \ell}\right]}, \frac{\beta^2\left[2 \ell(\ell-n) c+C_{n, \ell} \sup _M|A|^2\right]}{8\left[\beta-1+(p-1)^2 K_{p, n, \ell}\right]}\right\},$

则$M^{n}$上不存在非平凡的$L^{\beta} p$-调和$\ell$-形式, 即$H^{\ell}(L^{\beta}(M))=\{0\}$, 其中$\beta\geqslant p\geqslant2$.

2 准备知识

记$A^\ell(M)$表示$M^n$上$\ell$-次外微分形式空间, 则$M^n$上的$L^\beta$$p$-调和$\ell$-形式空间为

$\begin{aligned} H^{\ell}(L^{\beta}(M))=\Big\{\omega\in A^{\ell}(M)\mid \int_{M}|\omega|^{\beta} \mathrm{d}v<\infty, \mathrm{d}\omega=0 \mbox{和} \delta(|\omega|^{p-2}\omega)=0 \Big\}, \end{aligned} $

其中$\mathrm{d}v$为$M^n$的体积元 (以下为方便起见, 积分省去体积元符号).

设$M^{n}$是等距浸入到$N^{n+m}$的完备子流形. 对$M^n$上任意的点$x$, 设$\left\{e_{1},\cdots,e_{n+m}\right\}$是$N^{n+m}$的一组局部正交基, 其中$\left\{e_{1}, \cdots, e_{n}\right\}$是$M$的正交标架场, 记$\left\{\theta^{1}, \cdots, \theta^{n}\right\}$为其对偶标架场. 对任意$\alpha, n+1 \leqslant \alpha \leqslant n+m$, 第二基本形式$A$定义为

$A(X, Y)=\sum_{\alpha}\left\langle\bar{\nabla}_{X} Y, e_{\alpha}\right\rangle e_{\alpha}.$

这里$X, Y$是$M$上的切向量,$\overline{\nabla}$是$N^{n+m}$上的黎曼联络. 设$h_{i j}^{\alpha}=\left\langle A_{\alpha} e_{i}, e_{j}\right\rangle$为第二基本形式的系数. 第二基本形式模长的平方$|A|^{2}$和平均曲率向量$H$定义为

$\begin{aligned} &|A|^{2}=\sum_{i j \alpha}\left(h_{i j}^{\alpha}\right)^{2}, \\ &H=\sum_{\alpha} H^{\alpha} e_{\alpha}=\frac{1}{n} \sum_{i \alpha} h_{i i}^{\alpha} e_{\alpha}. \end{aligned} $

为了证明本文的主要结论, 需要以下引理.

引理 2.1[8] (Bochner-Weitzenböck 公式) 对于$M^n$上的$\ell$-形式$\omega$, 有下列的公式成立

$\begin{equation} \frac{1}{2} \Delta|\omega|^{2}=\langle\Delta \omega, \omega\rangle+|\nabla \omega|^{2}+\langle K_{\ell}(\omega), \omega\rangle, \end{equation}$

其中$K_{\ell}(\omega)=\sum\limits_{j, k=1}^{n} \theta^{k} \wedge i_{e_{j}} R\left(e_{k}, e_{j}\right) \omega$,$R$为曲率算子.

引理 2.2[9-11](Kato 不等式) 对于$M^n$上的$p$-调和$\ell$-形式$\omega$, 有

$\begin{equation} |\nabla(|\omega|^{p-2} \omega)|^{2} \geqslant(1+K_{p, n, \ell})|\nabla| \omega|^{p-1}|^{2}, \end{equation}$

其中

$\begin{equation*} K_{p, n, \ell}= \begin{cases} \frac{1}{\max \{\ell, n-\ell\}} & \text { 若 } p=2 \text { 且 } 1<\ell \leqslant n-1, \\[3mm] \frac{1}{(p-1)^{2}} \min \{1, \frac{(p-1)^{2}}{n-1}\} & \text { 若 } p>2 \text { 且 } \ell=1,\\ 0 & \text { 若 } p>2 \text { 且 } 1<\ell \leqslant n-1.\end{cases} \end{equation*}$

引理 2.3[12] (Hoffman-Spruck 不等式) 设$M^{n}$是非正截面曲率流形$N^{n+m}$的完备浸入子流形, 则对于任意的$\varphi \in C_{0}^{\infty}(M)$都有

$\begin{aligned} & \left(\int_{M}|\varphi|^{\frac{n}{n-1}} \right)^{\frac{n-1}{n}} \leqslant & C^{\prime}(n) \int_{M}(|\nabla \varphi|+n|H||\varphi|), \end{aligned}$

其中$C^{\prime}(n)$仅依赖于$n$.

由引理 2.3, 可得下列$L^2$-Sobolev 不等式如下

$\begin{aligned} & \bigg(\int_{M}|\varphi|^{\frac{2n}{n-2}} \bigg)^{\frac{n-2}{n}} \leqslant & C_{1}(n) \int_{M}(|\nabla \varphi|^{2}+|H|^{2} \varphi^{2}), \end{aligned}$

其中$C_{1}(n)$为仅与$n$有关的 Sobolev 正常数.

引理 2.4[12,13] 设$M^{n}(n \geqslant 3)$是球空间$\mathbb{S}^{n+m}$中的完备非紧浸入子流形, 则对任意$\varphi \in C_{0}^{\infty}(M)$, 有

$\begin{equation} \left(\int_{M}|\varphi|^{\frac{2 n}{n-2}} \right)^{\frac{n-2}{n}} \leqslant C_{2}(n)\left(\int_{M}|\nabla \varphi|^{2} +n^{2} \int_{M}\left(|H|^{2}+1\right)\varphi^{2} \right), \end{equation}$

其中$C_{2}(n)$为仅与$n$有关的 Sobolev 正常数.

3 定理的证明

设完备子流形$M^{n}$浸入到具有常截面曲率$c$的黎曼流形$N^{n+m}(c)$中. 若子流形$M^n$有平坦的法丛, 则对于$M^{n}$上的$\ell$-形式$\omega$, 当$n\geqslant4,2\leqslant \ell\leqslant n-2,\mbox{或} n=3,\ell=2$时, 由文献[3,5,6]有

$\begin{matrix} \langle K_{\ell}(\omega), \omega\rangle&=\ell(n-\ell)c|\omega|^{2}+\frac{1}{2}[n^{2}|H|^{2}-\max\{\ell,n-\ell\}|A|^{2}]|\omega|^{2} \\&=\ell(n-\ell)c|\omega|^{2} +\frac{1}{2}(n^{2}|H|^{2}-{C_{n, \ell}}|A|^{2})|\omega|^{2}, \end{matrix}$

其中${C_{n, \ell}} \text{=}\max\{\ell,n-\ell\}$. 对于$M^{n}$上的$L^{\beta}$$p$-调和$\ell$-形式$\omega$, 由引理 2.1 的 Bochner-Weitzenböck 公式有

$\begin{aligned} \frac{1}{2} \Delta|\omega|^{2(p-1)} &=\left|\nabla\left(|\omega|^{p-2} \omega\right)\right|^{2}-\left\langle(\delta \mathrm{d}+\mathrm{d} \delta)\left(|\omega|^{p-2} \omega\right),|\omega|^{p-2} \omega\right\rangle+ \langle K_{\ell}(|\omega|^{p-2} \omega),|\omega|^{p-2} \omega\rangle \\ &=\left|\nabla\left(|\omega|^{p-2} \omega\right)\right|^{2}-\left\langle\delta \mathrm{d}\left(|\omega|^{p-2} \omega\right),|\omega|^{p-2} \omega\right\rangle+|\omega|^{2(p-2)} \langle K_{\ell}(\omega), \omega\rangle. \end{aligned}$

进一步化简为

$\begin{equation*} |\omega|^{p-1} \Delta|\omega|^{p-1}=(|\nabla(|\omega|^{p-2} \omega)|^{2}-|\nabla| \omega|^{p-1}|^{2})-|\omega|^{p-2}\langle\delta \mathrm{d}(|\omega|^{p-2} \omega), \omega\rangle+|\omega|^{2p-2} \langle K_{\ell}(\omega), \omega\rangle. \end{equation*}$

利用引理 2.2 中的 Kato 不等式进一步化简得

$\begin{equation} |\omega| \Delta|\omega|^{p-1} \geqslant(p-1)^{2}K_{p, n, \ell}|\omega|^{p-2}|\nabla| \omega||^{2}-\left\langle\delta \mathrm{d}\left(|\omega|^{p-2} \omega\right), \omega\right\rangle+|\omega|^{p}\langle K_{\ell}(\omega), \omega\rangle. \end{equation}$

因此将 (3.2) 式左右同乘$|\omega|^{q}$, 其中$q\geqslant0$待定, 可得

$\begin{equation} |\omega|^{q+1} \Delta|\omega|^{p-1} \geqslant(p-1)^{2}K_{p, n, \ell}|\omega|^{p+q-2}|\nabla| \omega||^{2}-\left\langle\delta \mathrm{d}\left(|\omega|^{p-2} \omega\right),|\omega|^{q} \omega\right\rangle+|\omega|^{p+q}\langle K_{\ell}(\omega), \omega\rangle. \end{equation}$

将 (3.3) 式两边同乘$\varphi^2$, 其中$\varphi \in C_0^\infty(M)$, 并在$M$上积分可以得到

$\begin{matrix} \int_{M} \varphi^{2}|\omega|^{q+1} \Delta|\omega|^{p-1} \geqslant\ & (p-1)^{2}K_{p, n, \ell} \int_{M} \varphi^{2}|\omega|^{p+q-2}|\nabla| \omega||^{2} \\& -\int_{M}\langle\delta \mathrm{d}(|\omega|^{p-2} \omega),|\omega|^{q} \omega\rangle\varphi^{2}+\int_{M} \varphi^{2}|\omega|^{p+q}\langle K_{\ell}(\omega), \omega\rangle. \end{matrix}$

对(3.4)式左边进行分部积分后可得

$\begin{matrix} \int_{M}\left\langle\nabla\left(\varphi^{2}|\omega|^{q+1}\right), \nabla|\omega|^{p-1}\right\rangle \leqslant& -(p-1)^{2}K_{p, n, \ell} \int_{M} \varphi^{2}|\omega|^{p+q-2}|\nabla| \omega||^{2} \\ &+\int_{M}\langle\delta \mathrm{d}(|\omega|^{p-2} \omega),|\omega|^{q} \omega\rangle\varphi^{2}-\int_{M} \varphi^{2}|\omega|^{p+q}\langle K_{\ell}(\omega), \omega\rangle. \end{matrix}$

对(3.5)式左边利用 Cauchy-Schwarz 不等式化简得

$\begin{matrix} \int_{M}\left\langle\nabla\left(\varphi^{2}|\omega|^{q+1}\right), \nabla|\omega|^{p-1}\right\rangle\geqslant& -2(p-1) \int_{M} \varphi|\omega|^{p+q-1}|\nabla| \omega|||\nabla \varphi| \\ &+(q+1)(p-1) \int_{M} \varphi^{2}|\omega|^{p+q-2}|\nabla| \omega|| ^{2}. \end{matrix}$

对(3.5)式右边第二项化简有

$\begin{matrix} \int_{M}\langle\delta \mathrm{d}(|\omega|^{p-2} \omega),|\omega|^{q} \omega\rangle\varphi^{2}&= \int_{M}\left\langle \mathrm{\mathrm{d}}\left(|\omega|^{p-2} \omega\right), \mathrm{d}\left(\varphi^{2}|\omega|^{q} \omega\right)\right\rangle \\ & =\int_{M}\left\langle \mathrm{d}\left(|\omega|^{p-2}\right) \wedge \omega, \mathrm{d}\left(\varphi^{2}|\omega|^{q}\right) \wedge \omega\right\rangle \\ &\leqslant \int_{M}\left|\mathrm{d}\left(|\omega|^{p-2}\right) \wedge \omega\right| \cdot\left|\mathrm{d}\left(\varphi^{2}|\omega|^{q}\right) \wedge \omega\right| \\ &\leqslant \int_{M}\left|\nabla\left(|\omega|^{p-2}\right)\right||\omega| \cdot\left|\nabla\left(\varphi^{2}|\omega|^{q}\right)\right||\omega| \\ &\leqslant (p-2) q \int_{M} \varphi^{2}|\omega|^{p+q-2}|\nabla| \omega||^{2} \\ & \ +2(p-2) \int_{M} \varphi|\omega|^{p+q-1}|\nabla| \omega|||\nabla \varphi|. \end{matrix}$

(i) 若$|A|^{2} \leqslant \frac{n^{2}|H|^{2}+2 \ell(n-\ell)c}{C_{n, \ell}}$, 即$\langle K_{\ell}(\omega), \omega\rangle \geqslant 0$. 则由(3.5)式可得

$\begin{aligned} \int_{M}\langle\nabla(\varphi^{2}|\omega|^{q+1}), \nabla|\omega|^{p-1}\rangle \leqslant &-(p-1)^{2}K_{p, n, \ell} \int_{M} \varphi^{2}|\omega|^{p+q-2}|\nabla| \omega||^{2} \\ &+\int_{M}\langle\delta \mathrm{d}(|\omega|^{p-2} \omega), \varphi^{2}|\omega|^{q} \omega\rangle. \end{aligned} $

因此, 由(3.6)和(3.7)式可得

$\begin{aligned} & -2(p-1) \int_{M} \varphi|\omega|^{p+q-1}|\nabla| \omega|||\nabla \varphi|+(q+1)(p-1) \int_{M} \varphi^{2}|\omega|^{p+q-2}|\nabla| \omega||^{2} \\ \leqslant &-(p-1)^{2}K_{p, n, \ell} \int_{M} \varphi^{2}|\omega|^{p+q-2}|\nabla| \omega \|^{2} \\ &+(p-2) q \int_{M} \varphi^{2}|\omega|^{p+q-2}|\nabla| \omega \|^{2}+2(p-2) \int_{M} \varphi|\omega|^{p+q-1}|\nabla| \omega|||\nabla \varphi|. \end{aligned} $

整理后有

$[p+q-1+(p-1)^{2}K_{p, n, \ell}]\int_{M} \varphi^{2}|\omega|^{p+q-2}|\nabla| \omega||^{2} \leqslant 2(2p-3) \int_{M} \varphi|\omega|^{p+q-1}|\nabla| \omega|||\nabla \varphi|.$

利用基本不等式$2 A B \leqslant \varepsilon A^{2}+\frac{1}{\varepsilon}B^{2}$, 对任意的$\varepsilon>0$有

$\begin{aligned} 2\int_{M} \varphi|\omega|^{p+q-1}|\nabla| \omega|||\nabla \varphi| \leqslant\varepsilon \int_{M} \varphi^{2}|\omega|^{p+q-2}|\nabla| \omega||^{2}+\frac{1}{\varepsilon} \int_{M}|\omega|^{p+q}|\nabla \varphi|^{2}. \end{aligned}$

结合上述两个不等式可以得到

$\begin{equation} [p+q-1+(p-1)^{2}K_{p,n,\ell}-{(2p-3)}{\varepsilon}]\int_{M}\varphi^{2}|\omega|^{p+q-2}|\nabla|\omega||^{2} \leqslant\frac{2p-3}{\varepsilon}\int_{M}|\omega|^{p+q}|\nabla \varphi|^{2}. \end{equation}$

取$q$的值使其满足$\beta:=p+q\geq2$, 当取足够小$\varepsilon>0$时, 由$\beta-1+(p-1)^{2}K_{p, n, \ell}>0$, 可得存在常数$K_1, K_2>0$使

$\begin{aligned} K_1\int_{M}\varphi^{2}|\omega|^{\beta-2}|\nabla|\omega||^{2} \leqslant K_2 \int_{M}|\omega|^{\beta}|\nabla \varphi|^{2}, \end{aligned}$

其中

$K_1=p+q-1+(p-1)^{2}K_{p, n,\ell}-{(2p-3)}{\varepsilon}, K_2=\frac{2p-3}{\varepsilon}.$

使用文献[14]中的 Duzaar-Fuchs 截断函数方法. 设函数$\zeta \in C_0^\infty(M)$,定义

$\varphi_{\varepsilon}=\zeta^2\psi_{\varepsilon},$

其中对于$\varepsilon>0$,$\psi_{\epsilon}=\mathrm{min}\ \{\frac{|\omega|}{\epsilon}, 1\}$, 则$\varphi_{\varepsilon}$是$M$上的紧支撑函数. 将$\varphi_{\varepsilon}$替换式(3.10) 中$\varphi$可得

$\begin{aligned} K_1\int_{M}|\omega|^{\beta-2}|\nabla|\omega||^2\zeta^4\psi_{\varepsilon}\leqslant 8K_2\int_{M}|\omega|^\beta\zeta^2|\nabla\zeta|^2\psi_{\varepsilon}^2+2K_2\int_{M}|\omega|^Q|\nabla\psi_\varepsilon|^2\zeta^4. \end{aligned}$

由于有

$\begin{aligned} \int_{M}|\omega|^\beta|\nabla\psi_\varepsilon|^2\zeta^4\leqslant \varepsilon^{\beta-2}\int_{M}|\nabla|\omega||^2\zeta^4\chi\{|\omega|\leqslant\varepsilon\}. \end{aligned}$

因为$|\nabla|\omega||\in L^2_{loc}(M)$且$\beta\geqslant2$, 所以(3.11)式右边第二项会在$\varepsilon\to 0$时消失. 对式(3.11)令$\varepsilon\to 0$同时运用法图引理可得

$\begin{aligned}K_1\int_{M}|\omega|^{\beta-2}|\nabla|\omega||^2\zeta^4\leqslant 8K_2\int_{M}|\omega|^\beta\zeta^2|\nabla\zeta|^2, \end{aligned}$

其中$\zeta\in C_0^\infty(M)$. 取$M^n$上光滑截断函数$\zeta$, 满足$0\leqslant\zeta(x)\leqslant1$,$|\nabla\zeta|\leqslant\frac{2}{r}$且

$\begin{equation} \zeta(x)=\left\{ \begin{aligned} 1, \quad &x\in B_{x_{0}}(r),\\ 0, \quad & x\in M\backslash B_{x_{0}}(2r), \end{aligned} \right. \end{equation}$

其中$x_{0}$是$M$中任意一点,$B_{x_{0}}(r)$是以$x_{0}$为中心且半径为$r$的测地球. 则由上(3.14)式中$\zeta$的取值和(3.13)式可得

$\begin{equation} K_1\int_{B_{x_{0}}(r)\cap M}|\omega|^{\beta-2}|\nabla|\omega||^2\leqslant \frac{32K_2}{r^2}\int_{B_{x_0}(2r)\cap M}|\omega|^\beta. \end{equation}$

由$\int_{M}|\omega|^\beta<\infty$得

$\begin{aligned}\lim\limits_{r\to\infty}\frac{32K_2}{r^2}\int_{M}|\omega|^\beta=0. \end{aligned}$

结合(3.15)式和(3.16)式有

$\begin{aligned}K_1\int_{ B_{x_0}(r)\cap M}|\omega|^{\beta-2}|\nabla|\omega||^2\leqslant 0.\end{aligned}$

从而$|\omega|^{\beta-2}=0$或者$|\nabla|\omega||^2=0$. 若$\nabla|\omega|=0$, 则在$M$上$|\omega|=$const (常数),即$|\omega|$在整个$M$上为常数. 又由假设$M$的体积无穷大, 可得在$M$上$|\omega|=0$. 即$M$上不存在非平凡的$L^{\beta}$$p$-调和$\ell$-形式.

(ii) 当$c\leqslant 0$时, 设$\frac{n^{2}|H|^{2}+2\ell(n-\ell)c}{C_{n, \ell}}<|A|^{2} \leqslant \frac{n^{2}|H|^{2}}{C_{n, \ell}}$. 结合式(3.5)可得

$\begin{matrix} \int_{M}\langle\nabla(\varphi^{2}|\omega|^{q+1}), \nabla|\omega|^{p-1}\rangle \leqslant &-(p-1)^{2}K_{p, n, \ell} \int_{M}|\omega|^{p+q-2}|\nabla| \omega||^{2} \\ & +\int_{M}\langle\delta \mathrm{d}(|\omega|^{p-2} \omega), \varphi^{2}|\omega|^{q} \omega\rangle-\ell(n-\ell)c\int_{M} \varphi^{2}|\omega|^{p+q}. \end{matrix}$

根据第一特征值$\lambda_{1}(M)$的定义和(3.8)式, 对任意$\varepsilon>0$, 可得

$\begin{matrix} \lambda_{1}(M) \int_{M} \varphi^{2}|\omega|^{p+q}& \leqslant \int_{M}|\nabla(\varphi|\omega|^{\frac{p+q}{2}})|^{2} =\int_{M }\left(\frac{p+q}{2}|\omega|^{\frac{p+q-2}{2}} \varphi \nabla|\omega|+|\omega|^{\frac{p+q}{2}} \nabla \varphi\right)^{2} \\ &\leqslant(1+\varepsilon) \frac{(p+q)^{2}}{4} \int_{M} \varphi^{2}|\omega|^{p+q-2}|\nabla| \omega||^{2}+\left(1+\frac{1}{\varepsilon}\right) \int_{M}|\omega|^{p+q}|\nabla \varphi|^{2}. \end{matrix}$

由(3.6), (3.7), (3.8), (3.17) 式和(3.18)式可得

$C_{\varepsilon} \int_{M} \varphi^{2}|\omega|^{p+q-2}|\nabla| \omega||^{2} \leqslant D_{\varepsilon} \int_{M}|\omega|^{p+q}|\nabla \varphi|^{2},$

其中

$\begin{aligned} C_{\varepsilon}: &=p+q-1+(p-1)^{2}K_{p, n, \ell} -(2 p-3)\varepsilon+(1+\varepsilon) \frac{\ell(n-\ell)c(p+q)^{2}}{4 \lambda_{1}(M)} \\ &=\beta-1+(p-1)^{2}K_{p, n, \ell} -(2p-3)\varepsilon+(1+\varepsilon) \frac{\ell(n-\ell)c \beta^{2}}{4 \lambda_{1}(M)}, \end{aligned}$

$D_{\varepsilon}:=\frac{2 p-3}{\varepsilon}-\frac{\ell(n-\ell)c}{\lambda_{1}(M)}\left(1+\frac{1}{\varepsilon}\right);$

其中$\beta:=p+q \geqslant 2$. 由假设

$\lambda_{1}(M)>-\frac{\ell(n-\ell)c \beta^{2}}{4[\beta-1+(p-1)^{2}K_{p, n, \ell}]},$

可得$M$体积无穷大. 取$\varepsilon>0$足够小, 使得常数$C_{\varepsilon}, D_{\varepsilon}>0$. 即存在常数$K_3, K_4>0$使

$\begin{aligned} K_3\int_{M}\varphi^{2}|\omega|^{\beta-2}|\nabla|\omega||^{2} \leqslant K_4 \int_{M}|\omega|^{\beta}|\nabla \varphi|^{2}. \end{aligned}$

利用 (i) 中相同的讨论方法可得$\omega=0$.

(iii) 由(3.5)式可得

$\begin{matrix} & \int_{M}\langle\nabla(\varphi^{2}|\omega|^{q+1}), \nabla|\omega|^{p-1}\rangle+\frac{n^{2}}{2} \int_{M} \varphi^{2}|H|^{2}|\omega|^{p+q} \\ \leqslant &-(p-1)^{2}K_{p, n, \ell} \int_{M} \varphi^{2}|\omega|^{p+q-2}|\nabla| \omega||^{2} \\ &+\int_{M}\langle\delta \mathrm{d}(|\omega|^{p-2} \omega), \varphi^{2}|\omega|^{q} \omega\rangle+\frac{C_{n, \ell}}{2} \int_{M} \varphi^{2}|A|^{2}|\omega|^{p+q}-\ell(n-\ell)c\int_{M} \varphi^{2}|\omega|^{p+q}. \end{matrix}$

(a) 当截面曲率$c$非正时, 利用 Hölder 不等式和引理 2.3 得

$\begin{matrix} \int_{M} \varphi^{2}|A|^{2}|\omega|^{p+q}& \leqslant\|A\|_{n}^{2}\left(\int_{M}\left(\varphi|\omega|^{\frac{p+q}{2}}\right)^{\frac{2 n}{n-2}}\right)^{\frac{n-2}{n}} \\ & \leqslant C_{1}(n)\|A\|_{n}^{2}\left(\int_{M}\left|\nabla\left(\varphi|\omega|^{\frac{p+q}{2}}\right)\right|^{2}+\int_{M} \varphi^{2}|H|^{2}|\omega|^{p+q}\right), \end{matrix}$

其中$C_{1}(n)$为仅与$n$有关的 Sobolev 常数. 上式结合 (3.18) 式, 对任意$\varepsilon>0$有

$\begin{matrix} \int_{M} \varphi^{2}|A|^{2}|\omega|^{p+q} \leqslant\ & C_{1}(n)\|A\|_{n}^{2}(1+\varepsilon) \frac{(p+q)^{2}}{4} \int_{M} \varphi^{2}|\omega|^{p+q-2}|\nabla| \omega||^{2} \\ & +C_{1}(n)\|A\|_{n}^{2}(1+\frac{1}{\varepsilon}) \int_{M}|\omega|^{p+q}|\nabla \varphi|^{2}+C_{1}(n)\|A\|_{n}^{2} \int_{M} \varphi^{2}|H|^{2}|\omega|^{p+q}. \end{matrix}$

联立 (3.6), (3.7) 式和(3.22)式得到

$E_{\varepsilon}\int_{M} \varphi^{2}|\omega|^{p+q-2}|\nabla| \omega \|^{2}+\frac{1}{2}\left(n^{2}-C_{n, \ell} C_{1}(n)\|A\|_{n}^{2}\right) \int_{M} \varphi^{2}|H|^{2}|\omega|^{p+q} \leqslant F_{\varepsilon} \int_{M}|\omega|^{p+q}|\nabla \varphi|^{2}$
$\begin{eqnarray*} E_{\varepsilon}:& =& p+q-1+(p-1)^{2}K_{p, n, \ell}-(2 p-3)\varepsilon\\ &&-(1+\varepsilon) \frac{(p+q)^{2}}{4}\left(\frac{-\ell(n-\ell)c}{\lambda_{1}(M)}+\frac{C_{n, \ell} C_{1}(n)\|A\|_{n}^{2}}{2}\right) \\ & = & \beta-1+(p-1)^{2}K_{p, n, \ell}-(2 p-3)\varepsilon-(1+\varepsilon) \frac{\beta^{2}}{4}\left(\frac{-\ell(n-\ell)c}{\lambda_{1}(M)}+\frac{C_{n, \ell} C_{1}(n)\|A\|_{n}^{2}}{2}\right); \\ F_{\varepsilon}:&=&\frac{2 p-3}{\varepsilon}+\left(1+\frac{1}{\varepsilon}\right)\left(\frac{-\ell(n-\ell)c}{\lambda_{1}(M)}+\frac{C_{n, \ell} C_{1}(n)\|A\|_{n}^{2}}{2}\right); \end{eqnarray*}$

其中$\beta:=p+q \geqslant 2$.由于

$\|A\|_{n}^{2}<\min \left\{\frac{n^{2}}{C_{n, \ell} C_{1}(n)}, \frac{2}{C_{n, \ell} C_{1}(n)}\left[\frac{4(\beta-1+(p-1)^{2}K_{p, n, \ell})}{\beta^{2}}+\frac{\ell(n-\ell)c}{\lambda_{1}(M)}\right]\right\},$

所以存在足够小的$\varepsilon>0$, 使得常数$E_{\varepsilon}, ~F_{\varepsilon}>0$. 从而存在常数$K_5, K_6>0$使

$\begin{aligned} K_5\int_{M}\varphi^{2}|\omega|^{\beta-2}|\nabla|\omega||^{2} \leqslant K_6 \int_{M}|\omega|^{\beta}|\nabla \varphi|^{2}. \end{aligned}$

利用 (i) 相同的证明方法得证在整个$M$上$|\omega|$都为常数, 根据文献[15,引理 2.1]{Lin2} 可得$M$的体积无穷大, 这与已知条件$\int_{M}|\omega|^p {\rm d}v<\infty$矛盾, 所以在$M$上有$|\omega|=0$, 即$M$上不存在非平凡的$L^{\beta}$$p$-调和$\ell$-形式.

(b) 当截面曲率$c=1$时,(3.20)式可写为

$\begin{matrix} & \int_{M}\langle\nabla(\varphi^{2}|\omega|^{q+1}), \nabla|\omega|^{p-1}\rangle+\frac{n^{2}}{2} \int_{M} \varphi^{2}|H|^{2}|\omega|^{p+q} \\ \leqslant &-(p-1)^{2}K_{p, n, \ell} \int_{M} \varphi^{2}|\omega|^{p+q-2}|\nabla| \omega||^{2} \\ &+\int_{M}\langle\delta \mathrm{d}(|\omega|^{p-2} \omega), \varphi^{2}|\omega|^{q} \omega\rangle+\frac{C_{n, \ell}}{2} \int_{M} \varphi^{2}|A|^{2}|\omega|^{p+q}. \end{matrix}$

利用 Hölder 不等式和引理 2.4 可得

$\begin{matrix} \int_{M} \varphi^{2}|A|^{2}|\omega|^{p+q} & \leqslant\|A\|_{n}^{2}\left(\int_{M}\left(\varphi|\omega|^{\frac{p+q}{2}}\right)^{\frac{2 n}{n-2}}\right)^{\frac{n-2}{n}} \\ & \leqslant C_{2}(n)\|A\|_{n}^{2}\left(\int_{M}\left|\nabla\left(\varphi|\omega|^{\frac{p+q}{2}}\right)\right|^{2}+\int_{M} \varphi^{2}n^2(1+|H|^{2})|\omega|^{p+q}\right), \end{matrix}$

其中$C_{2}(n)$为仅与$n$有关的 Sobolev 常数. 上式结合 (3.7) 式, 对任意$\varepsilon>0$有

$\begin{matrix} \int_{M} \varphi^{2}|A|^{2}|\omega|^{p+q} \leqslant\ & C_{2}(n)\|A\|_{n}^{2}(1+\varepsilon) \frac{(p+q)^{2}}{4} \int_{M} \varphi^{2}|\omega|^{p+q-2}|\nabla| \omega||^{2} \\&+C_{2}(n)\|A\|_{n}^{2}(1+\frac{1}{\varepsilon}) \int_{M}|\omega|^{p+q}|\nabla \varphi|^{2}+C_{2}(n)\|A\|_{n}^{2} n^2\int_{M} \varphi^{2}|H|^{2}|\omega|^{p+q} \\& +C_{2}(n)\|A\|_{n}^{2}n^2\int_{M} \varphi^{2}|\omega|^{p+q}. \end{matrix}$

联立 (3.6), (3.7), (3.18) 式和(3.24)式得到

$\begin{aligned} \widetilde{E_{\varepsilon}}\int_{M} \varphi^{2}|\omega|^{p+q-2}|\nabla| \omega \|^{2}+\frac{1}{2}\left(n^{2}-C_{n, \ell} C_{2}(n)\|A\|_{n}^{2}n^2\right) \int_{M} \varphi^{2}|H|^{2}|\omega|^{p+q} \leqslant \widetilde{F_{\varepsilon} } \int_{M}|\omega|^{p+q}|\nabla \varphi|^{2}, \end{aligned}$

其中

$\begin{align*} \widetilde{E_{\varepsilon}}& =p+q-1+(p-1)^{2}K_{p, n, \ell}-(2 p-3)\varepsilon-(1+\varepsilon)\frac{(p+q)^{2}}{4}\frac{C_{n, \ell} C_{2}(n)\|A\|_{n}^{2}}{2}(1+\frac{n^2}{\lambda_{1}(M)}) \\ &=\beta-1+(p-1)^{2}K_{p, n, \ell}-(2 p-3)\varepsilon-(1+\varepsilon) \frac{\beta^{2}}{4}\frac{C_{n, \ell} C_{2}(n)\|A\|_{n}^{2}}{2}(1+\frac{n^2}{\lambda_{1}(M)}); \\ \widetilde{F_{\varepsilon}}&:=\frac{2 p-3}{\varepsilon}+(1+\frac{1}{\varepsilon})\frac{C_{n, \ell} C_{2}(n)\|A\|_{n}^{2}}{2}(1+\frac{n^2}{\lambda_{1}(M)}); \end{align*}$

其中$\beta:=p+q \geqslant 2$.由

$\|A\|_{n}^{2}<\min \left\{\frac{n^{2}}{C_{n, \ell} C_{2}(n)}, \frac{8\left[\beta-1+(p-1)^{2}K_{p, n, \ell}\right]}{C_{n,\ell} C_{2}(n)\beta^{2}(1+\frac{n^2}{\lambda_{1}(M)})}\right\},$

所以存在足够小的$\varepsilon>0$, 使得常数$\widetilde{E_{\varepsilon}}, \widetilde{F_{\varepsilon}}>0$. 从而存在常数$K_7, K_8>0$使

$\begin{aligned} K_7\int_{M}\varphi^{2}|\omega|^{\beta-2}|\nabla|\omega||^{2} \leqslant K_8 \int_{M}|\omega|^{\beta}|\nabla \varphi|^{2}. \end{aligned}$

利用 (i) 相同的证明方法证得在整个$M$上$|\omega|$都为常数. 由假设$M$的第一特征值$\lambda_{1}(M)>0$,可得$M$的体积无穷大, 这与已知条件$\int_{M}|\omega|^p {\rm d}v<\infty$矛盾, 所以在$M$上有$|\omega|=0$,即$M$上不存在非平凡的$L^{\beta}$$p$-调和$\ell$-形式.

(iv) 设$\sup _{M}|A|^{2}<\infty$. 则利用式 (3.18) 结合题设$\lambda_{1}(M)>0$, 可得对任意的$\varepsilon>0$有

$\begin{align*} \int_{M} \varphi^{2}|A|^{2}|\omega|^{p+q} & \leqslant \sup _{M}|A|^{2} \int_{M} \varphi^{2}|\omega|^{p+q} \leqslant \frac{\sup _{M}|A|^{2}}{\lambda_{1}(M)} \int_{M}|\nabla(\varphi|\omega|^{\frac{p+q}{2}})|^{2} \\ & \leq\frac{\sup _{M}|A|^{2}}{\lambda_{1}(M)}(1+\varepsilon) \frac{(p+q)^{2}}{4} \int_{M} \varphi^{2}|\omega|^{p+q-2}|\nabla| \omega||^{2} \\\ & +\frac{\sup _{M}|A|^{2}}{\lambda_{1}(M)}(1+\frac{1}{\varepsilon}) \int_{M}|\omega|^{p+q}|\nabla \varphi|^{2}. \end{align*}$

当截面曲率$c\geqslant 0$时, (3.20)式可写为

$\begin{matrix} & \int_{M}\langle\nabla(\varphi^{2}|\omega|^{q+1}), \nabla|\omega|^{p-1}\rangle+\frac{n^{2}}{2} \int_{M} \varphi^{2}|H|^{2}|\omega|^{p+q} \\ \leqslant &-(p-1)^{2}K_{p, n, \ell} \int_{M} \varphi^{2}|\omega|^{p+q-2}|\nabla| \omega||^{2} \\ & +\int_{M}\langle\delta \mathrm{d}(|\omega|^{p-2} \omega), \varphi^{2}|\omega|^{q} \omega\rangle+\frac{C_{n, \ell}}{2} \int_{M} \varphi^{2}|A|^{2}|\omega|^{p+q}. \end{matrix}$

联立 (3.6), (3.7), (3.8) 和(3.20)式, 对任意的$\varepsilon>0$有

$G_{\varepsilon} \int_{M} \varphi^{2}|\omega|^{p+q-2}|\nabla| \omega||^{2}+\frac{n^{2}}{2} \int_{M} \varphi^{2}|H|^{2}|\omega|^{p+q} \leqslant J_{\varepsilon} \int_{M}|\omega|^{p+q}|\nabla \varphi|^{2},$

其中

$\begin{align*} G_{\varepsilon}:&=p+q-1+(p-1)^{2}K_{p, n, \ell}-(2 p-3)\varepsilon-(1+\varepsilon) \frac{(p+q)^{2}}{4 \lambda_{1}(M)}\frac{C_{n, \ell} \sup _{M}|A|^{2}}{2} \\ &=\beta-1+(p-1)^{2}K_{p, n, \ell}-(2 p-3)\varepsilon-(1+\varepsilon) \frac{\beta^{2}}{4 \lambda_{1}(M)}\frac{C_{n, \ell} \sup _{M}|A|^{2}}{2}; \end{align*}$
$J_{\varepsilon}:=\frac{2 p-3}{\varepsilon}+\left(1+\frac{1}{\varepsilon}\right) \frac{1}{\lambda_{1}(M)}\frac{C_{n, \ell} \sup _{M}|A|^{2}}{2}.$

当$\lambda_{1}(M)>\frac{\beta^{2}C_{n, \ell} \sup _{M}|A|^{2}}{8[\beta-1+(p-1)^{2}K_{p, n, \ell}]}$时, 有$G_{\varepsilon}>0$.当截面曲率$c\leqslant 0$时,联立 (3.6), (3.7), (3.8) 式和(3.20)式, 对任意的$\varepsilon>0$有

$\widetilde{G_{\varepsilon}} \int_{M} \varphi^{2}|\omega|^{p+q-2}|\nabla| \omega||^{2}+\frac{n^{2}}{2} \int_{M} \varphi^{2}|H|^{2}|\omega|^{p+q} \leqslant \widetilde{J_{\varepsilon} }\int_{M}|\omega|^{p+q}|\nabla \varphi|^{2},$

其中

$\begin{align*} \widetilde{G_{\varepsilon}}:& =p+q-1+(p-1)^{2}K_{p, n, \ell}-(2 p-3)\varepsilon-(1+\varepsilon) \frac{(p+q)^{2}}{4 \lambda_{1}(M)}\left[\ell(\ell-n)c+\frac{C_{n, \ell} \sup _{M}|A|^{2}}{2}\right] \\ &=\beta-1+(p-1)^{2}K_{p, n, \ell}-(2 p-3)\varepsilon-(1+\varepsilon) \frac{\beta^{2}}{4 \lambda_{1}(M)}\left[\ell(\ell-n)c+\frac{C_{n, \ell} \sup _{M}|A|^{2}}{2}\right]; \\ \widetilde{J_{\varepsilon}}:& =\frac{2 p-3}{\varepsilon}+\left(1+\frac{1}{\varepsilon}\right) \frac{1}{\lambda_{1}(M)}\left[\ell(\ell-n)c+\frac{C_{n, \ell} \sup _{M}|A|^{2}}{2}\right]. \end{align*}$

结合截面曲率$c\leqslant 0$, 当$\lambda_1(M)>\frac{\beta^2\left[2 \ell(\ell-n) c+C_{n, \ell} \sup _M|A|^2\right]}{8\left[\beta-1+(p-1)^2 K_{p, n, \ell}\right]}$时, 有$\widetilde{G_{\varepsilon}}>0$.

综上可得当

$\lambda_1(M)>\max \left\{\frac{\beta^2 C_{n, \ell} \sup _M|A|^2}{8\left[\beta-1+(p-1)^2 K_{p, n, \ell}\right]}, \frac{\beta^2\left[2 \ell(\ell-n) c+C_{n, \ell} \sup _M|A|^2\right]}{8\left[\beta-1+(p-1)^2 K_{p, n, \ell}\right]}\right\},$

可得$M$体积无穷大. 另外, 由假设可知存在足够小的$\varepsilon>0$, 使得常数$G_{\varepsilon}, J_{\varepsilon}, \widetilde{G_{\varepsilon}}, \widetilde{J_{\varepsilon}}>0$. 从而存在常数$K_9, K_{10}>0$使

$\begin{aligned} K_9\int_{M}\varphi^{2}|\omega|^{\beta-2}|\nabla|\omega||^{2} \leqslant K_{10} \int_{M}|\omega|^{\beta}|\nabla \varphi|^{2}. \end{aligned}$

利用 (i) 和 (ii) 相同的证明方法可得$\omega=0$. 定理得证.

参考文献

Li P.

On the Sobolev constant and the$p$-spectrum of a compact riemannian manifold

Annales scientifiques de l'école Normale Supérieure, 1980, 13(4): 451-468

DOI:10.24033/asens.1392      URL     [本文引用: 1]

Tanno S.

$L^{2}$harmonic forms and stability of minimal hypersurfaces

Tokyo Sugaku Kaisya Zasshi, 1996, 48(4): 761-768

[本文引用: 1]

Lin H Z.

On the structure of submanifolds in Euclidean space with flat normal bundle

Results in Mathematics, 2015, 68(3): 313-329

DOI:10.1007/s00025-015-0435-5      URL     [本文引用: 2]

Zhang X.

A note on$p$-harmonic l-forms on complete manifolds

Canadian Mathematical Bulletin, 2001, 44: 376-384

DOI:10.4153/CMB-2001-038-2      URL     [本文引用: 1]

In this paper we prove that there is no nontrivial Lq-integrably p-harmonic 1-form on a complete manifold with nonnegatively Ricci curvature (0 &lt; q &lt; ∞).

Lin H Z.

Vanishing theorems for hypersurfaces in the unit sphere

Glasgow Mathematical Journal, 2018, 60(3): 661-671

DOI:10.1017/S0017089517000350      URL     [本文引用: 2]

Let Mn, n ≥ 3, be a complete hypersurface in $\\mathbb{S}$n+1. When Mn is compact, we show that Mn is a homology sphere if the squared norm of its traceless second fundamental form is less than $\\frac{2(n-1)}{n}$. When Mn is non-compact, we show that there are no non-trivial L2 harmonic p-forms, 1 ≤ p ≤ n − 1, on Mn under pointwise condition. We also show the non-existence of L2 harmonic 1-forms on Mn provided that Mn is minimal and $\\frac{n-1}{n}$-stable. This implies that Mn has only one end. Finally, we prove that there exists an explicit positive constant C such that if the total curvature of Mn is less than C, then there are no non-trivial L2 harmonic p-forms on Mn for all 1 ≤ p ≤ n − 1.

Han Y B.

$p$-Harmonic$\ell$-forms on complete noncompact submanifolds in sphere with flat normal bundle

Bulletin of the Brazilian Mathematical Society, New Series, 2018, 49(1): 107-122

DOI:10.1007/s00574-017-0051-y      URL     [本文引用: 2]

Lin H Z, Yang B G.

The$p$-eigenvalue estimates and$L^q$$p$-harmonic forms on submanifolds of Hadamard manifolds

Journal of Mathematical Analysis and Applications, 2020, 488(1): 124018

DOI:10.1016/j.jmaa.2020.124018      URL     [本文引用: 1]

Li P. Lecture Notes on Geometric Analysis. Seoul: Global Analysis Research Cewter, 1993

[本文引用: 1]

Calderbank D M J, Gauduchon P, Herzlich M.

Refined Kato inequalities and conformal weights in Riemannian geometry

Journal of Functional Analysis, 2000, 173(1): 214-255

DOI:10.1006/jfan.2000.3563      URL     [本文引用: 1]

Dung N T, Tien P T.

Vanishing properties of$p$-harmonic$\ell$-forms on Riemannian manifolds

Journal of the Korean Mathematical Society, 2018, 55(5): 1103-1129

[本文引用: 1]

Dung N T, Seo K.

$p$-harmonic functions and connectedness at infinity of complete Riemannian manifolds

Annali Di Matematica Pura Ed Applicata, 2017, 196(4): 1489-1511

DOI:10.1007/s10231-016-0625-0      [本文引用: 1]

Hoffman D, Spruck J.

Sobolev and isoperimetric inequalities for Riemannian submanifolds

Communications on Pure and Applied Mathematics, 1974, 27(6): 715-727

DOI:10.1002/cpa.v27:6      URL     [本文引用: 2]

Fu H P, Xu H W.

Total curvature and$L^2$harmonic 1-forms on complete submanifolds in space forms

Geometriae Dedicata, 2010, 144(1): 129-140

DOI:10.1007/s10711-009-9392-z      URL     [本文引用: 1]

Duzaar F, Fuchs M.

On removable singularities of$p$-harmonic maps

Ann Inst H Poincar'e Anal Non Lin'eaire, 1990, 7: 385-405

[本文引用: 1]

Lin H Z.

On the srtucture of conformally flat Riemannian manifolds

Nonlinear Analysis, 2015, (123/124): 115-125

[本文引用: 1]

/