数学物理学报  2016, Vol. 36 Issue (6): 1027-1039   PDF    
扩展功能
加入收藏夹
复制引文信息
加入引用管理器
Email Alert
RSS
本文作者相关文章
邱望华
侯中华
伪黎曼乘积空间中具有平行平均曲率向量的曲面
邱望华1, 侯中华2     
1. 九江学院理学院 江西九江 332005 ;
2. 大连理工大学数学科学学院 辽宁大连 116024
摘要:Batista在M2c)×$\mathbb{R}$中的具有常平均曲率的曲面上引入了一个特殊的(1,1)型张量S.之后,Fetcu和Rosebberg将张量S推广到Mnc)×$\mathbb{R}$中的具有平行平均曲率向量的曲面上.该文将张量S推广到了伪黎曼乘积空间中的曲面上,并研究了S的Pinching问题,得到了若干Pinching常数.特别地,对外围空间是黎曼乘积空间的情况,得到的Pinching常数优于Batista得到的相应的Pinching常数.
关键词平行平均曲率向量     乘积空间     Simons型方程     Pinching常数    
Surfaces in Products of Semi-Riemannian Space Forms with Parallel Mean Curvature Vector
Qiu Wanghua1, Hou Zhonghua2     
1. College of Sciences, Jiujiang University, Jiangxi Jiujiang 332005 ;
2. Institute of Mathematics, Dalian University of Technology, Liaoning Dalian 116024
Abstract: Batista introduced a special (1, 1) tensor S on a CMC immersed surface Σ2 in M2(c$\mathbb{R}$. Later on, Fetcu and Rosebberg extended (1, 1) tensor S to PMC surface Σ2Mn-1(c$\mathbb{R}$. In the present paper, the authors consider a more general tensor S on PMC immersed surface Σ2 in Lorentzian product spaces (Mn-1(c$\mathbb{R}$, g-1) and Riemmannian product spaces (Mn-1(c$\mathbb{R}$, g+1). The authors compute the Simons type equations of |S|2, and characterize CMC surfaces in (M2(c$\mathbb{R}$, gε). For case ε=+1, we obtain several pinching constants greater than that given by Batista.
Key words: Parallel mean curvature vector     Simons type equation     Product spaces     Pinching constants    
1 引言

近些年来, 越来越多的研究者们致力于研究乘积空间中的几何问题, 特别是对黎曼乘积空间$M^{n-1}(c)\times {\Bbb R}$和洛伦兹乘积空间$M^{n-1}(c)\times {\Bbb R}_1$的研究甚是广泛.例如:陈航, 陈刚毅与李海中[11]研究了${\Bbb S}^m(1)\times {\Bbb R}$中的紧致极小子流形, 并分别得到了关于Ricci曲率和截面曲率的Pinching定理; Dillen等人[15-16]$M^2(c)\times {\Bbb R}$中的常角曲面进行了分类, 在文献[14, 17]中, 他们刻画了具有典型主方向的曲面.富宇等人[23]完全刻画了洛伦兹乘积空间$M^2(c)\times {\Bbb R}_1$中的具有常角性质或典型主方向的类空和类时曲面.de Lima等人[26]利用极值原理研究了$M^n(c)\times {\Bbb R}_1$中的完备类空超曲面.更多相关的结论请参考文献[4-7, 9-10, 13, 18-21, 24-25, 27].

Abresch和Rosenberg[1]$M^2(c)\times {\Bbb R}$中具有常平均曲率(简记为CMC)的曲面上引进了一个Hopf型全纯微分Q(2, 0), 其中Q(2, 0)是二次型Q的(2, 0)部分.Q的定义如下

$ Q(X, Y)=2H h(X, Y)-c\left\langle X, {\partial}_t\right\rangle \left\langle Y, {\partial}_t\right\rangle, $

其中, ${\partial}_t$${\Bbb R}$方向上的单位切向量.进而, 作者将$M^2(c)\times {\Bbb R}$中满足Q(2, 0)=0的完备CMC曲面分成了四类.后来学者们将满足Q(2, 0)为零的曲面称为Abresch-Rosenberg曲面.

Alencar, do Carmo和Tribuzy[2]将上述的Hopf型微分推广到$ M^n(c)\times {\Bbb R}$中的具有平行平均曲率向量(简记为PMC)曲面上, 如下

$ \begin{equation}\label{eq-1-0} Q(X, Y)=2\langle h(X, Y), \overrightarrow{H}\rangle -c\left\langle X, {\partial}_t\right\rangle\left\langle Y, {\partial}_t\right\rangle. \end{equation} $ (1.1)

他们证明了

定理1.1[2-3]   设$\Sigma^2$$M^2(c)\times {\Bbb R}$中的亏格为零的紧致浸入曲面.如果

$ |dH|\le g|Q^{(2, 0)}|, $

其中|dH|表示平均曲率微分dH的范数, g是一个连续非负的实函数.则Q(2, 0)恒为0.

Batista[8]$M^2(c)\times {\Bbb R}$中的CMC曲面上引进了一个张量S, 定义如下

$ SX=2HAX-c\left\langle X, T\right\rangle T+\frac{c}{2}|T|^2 X -2H^2 X, $

其中, T${\partial}_t$在切空间上的投影.作者证明了S=0与Q(2, 0)=0是等价的.因此, S=0的曲面就是Abresh-Rosenberg曲面.作者还分别计算了张量S以及第二基本形式无迹部分$\varphi$的Simons型方程, 并研究了$|S|^2$$|\varphi|^2$的Pinching问题.

随后, Fetcu和Rosebberg[22]$M^n(c)\times {\Bbb R}$中的PMC曲面上引入了如下的(1, 1)型张量S

$ \begin{equation}\label{eq-1-1} S=2\sum\limits_\alpha {H^\alpha} A_\alpha- c\langle T, \cdot\rangle T+\frac{c}{2} |T|^2 I - 2H^2 I, \end{equation} $ (1.2)

其中, I是单位变换.它是文献[8]中张量S的自然推广.作者利用文献[12]的结论给出了张量S的Simons型方程, 这个Simons型方程是内蕴的.

本文中, 我们将上述张量S进一步推广.为了同时处理黎曼乘积空间和洛伦兹乘积空间的情况, 我们把外围空间记为

$ \overline{M}^n_\varepsilon:=(M^{n-1}(c)\times {\Bbb R}, \overline{g}_\varepsilon), \varepsilon=\pm 1, $

其中, $\overline{g}_\varepsilon=g|_{M^{n-1}(c)}+\varepsilon dt^2 :=g+\varepsilon dt^2$.度量$\overline{g}_{-1}$($ \overline{g}_{+1}$)是洛伦兹度量(黎曼度量), 对应的流形称为洛伦兹乘积流形(黎曼乘积流形).对洛伦兹乘积流形, 本文只考虑其中的类空曲面.类似地, 对PMC浸入曲面$\Sigma^2 \looparrowright \overline{M}^{n}_\varepsilon$, 我们定义了更一般的Hopf型微分Q(参考3.1式)和张量S(参考3.2式).我们得到了关于ST的Simons型方程, 并刻画了($M^{2}(c)\times {\Bbb R}, \overline{g}_\varepsilon$)中的CMC曲面.对$\varepsilon=+1$的情形, 我们得到的若干个Pinching常数比文献[8]中相应的Pinching常数更佳.具体来说, 我们证明了

定理1.2   设$\Sigma^2$$({\Bbb S}^{2}(1)\times {\Bbb R}, \overline{g}_{-1})$中的完备类空CMC曲面且$0<H^2<1$.则|S|=0, 即曲面$\Sigma^2$是Abresch-Rosenberg型曲面.

定理1.3   设$\Sigma^2 $$({\Bbb H}^{2}(-1)\times {\Bbb R}, \overline{g}_{-1})$中的完备类空CMC曲面且$H\ne 0$, $g(ST, T)\geq 0$.如果

$ \mathop{\inf}\limits_{\Sigma^2}|T|^2> 8H^2+4\sqrt{H^2(5H^2+1)}, $

则|S|=0, 即曲面∑2是Abresch-Rosenberg型曲面.

定理1.4   设∑2$(M^{2}(c)\times {\Bbb R}, \overline{g}_{-1})$, c>0中的紧致类空CMC曲面.则∑2是全测地的且$|T|^2=0$.此时, 曲面∑2自然就是Abresch-Rosenberg型曲面.

定理1.5   设∑2$({\Bbb S}^{2}(1)\times {\Bbb R}, \overline{g}_{+1})$中的完备CMC曲面, 满足$H>{1}/{2}$.如果

$ \sup\limits_{\Sigma^2}|S|<\frac{\sqrt{2}}{2}(4H^2-1), $

则|S|=0, 即曲面∑2是Abresch-Rosenberg曲面.

定理1.6   设∑2$({\Bbb S}^{2}(1)\times {\Bbb R}, \overline{g}_{+1})$中的完备CMC曲面, 且满足H>1/2, 则不存在这样的曲面使得$|S|=\frac{\sqrt{2}}{2}(4H^2-1)$.

通过增加一个条件, 我们能得到比定理1.5中更佳的Pinching常数.

定理1.7   设∑2$({\Bbb S}^{2}(1)\times {\Bbb R}, \overline{g}_{+1})$中的完备CMC曲面且H>1/2, g(ST, T)≤0.若

$ \sup\limits_{\Sigma^2}|S|<\sqrt{8H^4-\frac{1}{2}}, $

则|S|=0, 即曲面∑2是Abresch-Rosenberg曲面.

定理1.8  设∑2$({\Bbb H}^{2}(-1)\times {\Bbb R}, \overline{g}_{+1})$中的完备CMC曲面, 满足H>1.如果

$ \sup\limits_{\Sigma^2}|S|<\frac{\sqrt{2}}{2}[\sqrt{16H^2(H^2-1)+1}-1], $

则|S|=0, 即曲面∑2是Abresch-Rosenberg曲面.

定理1.9  设∑2$({\Bbb H}^{2}(-1)\times {\Bbb R}, \overline{g}_{+1})$中的完备CMC曲面, 满足H>1, 则不存在这样的曲面使得$|S|=\frac{\sqrt{2}}{2}[\sqrt{16H^2(H^2-1)+1}-1]$.

注1.1  定理1.5中的Pinching常数$\frac{\sqrt{2}}{2}(4H^2-1)$比文献[8, 定理1.3]的$\sqrt{8H^4-2H^2+1}$ -1更大.定理1.8中的Pinching常数$\frac{\sqrt{2}}{2}[\sqrt{16H^2(H^2-1)+1}-1]$大于文献[8, 定理1.6]的Pinching常数$\sqrt{8H^4-12H^2}-1$, 且条件H>1比$H>\sqrt{\frac{{3+\sqrt{11}}}{4}}$更弱.

2 预备知识

$\overline M^n=(M^{n - 1}(c)\times {\Bbb R}, \overline{g}_{\varepsilon})$, 其中$M^{n-1}(c)$是(n-1)维截面曲率为$c\ne 0$的空间形式, $\varepsilon=\pm 1$.设$\Sigma^2$$\overline M^n$中的浸入曲面.记${\partial}_t$是切于${\Bbb R} $$\varepsilon$单位向量场, 即$ \overline{g}_{\varepsilon}({\partial}_t, {\partial}_t)=\varepsilon$.对$ \varepsilon=-1$的情形, 浸入曲面∑2是类空的曲面, 即∑2上的诱导度量是黎曼度量.

本文中, 我们约定指标范围如下

$ 1 \le i, j, k, \cdots \le 2;\quad 3 \le \alpha, \beta, \gamma, \cdots \le n;\quad 1 \le A, B, C, \cdots \le n. $

沿着曲面∑2, 我们在$ \overline M^n $上选取标准正交标架场$ \{e_1, e_2, \cdots, e_n \}$, 使得限制在∑2上时, $ \{ e_1, e_2 \} $切于∑2, 其它的法于 $\Sigma^2 $, 并且满足$ \overline{g}_{\varepsilon}(e_A, e_B)=\varepsilon_A \delta_{AB}$, 其中$\varepsilon_n=\varepsilon$, 其它的$\varepsilon_A=1$.设$\{\omega^1, \omega^2, \cdots, \omega^n\}$为相应的对偶标架场.记$\nabla$($\overline \nabla $)是∑2($\overline M^n$)的Levi-Civita联络.

$\overline M^n$的结构方程为

$ \begin{array}{*{20}{l}} {d{\omega ^A}=- \sum\limits_B {{\varepsilon _B}} \omega _B^A \wedge {\omega ^B}, }&{\omega _B^A+\omega _A^B=0, }\\ {d\omega _B^A=- \sum\limits_C {{\varepsilon _C}} \omega _C^A \wedge \omega _B^C+\bar \Omega _B^A, }&{\bar \Omega _B^A+\bar \Omega _A^B=0, }\\ {\bar \Omega _B^A=\frac{1}{2}\sum\limits_{C, D} {{\varepsilon _C}} {\varepsilon _D}\bar R_{BCD}^A{\omega ^C} \wedge {\omega ^D}, }&{\bar R_{BCD}^A+\bar R_{BDC}^A=0.} \end{array} $ (2.1)

限制到曲面∑2上, 我们有:对任意的α, $\omega^\alpha=0$.从而可以得到∑2的结构方程

$ \begin{array}{l} d{\omega ^i}=- \sum\limits_j {\omega _j^i} \wedge {\omega ^j}, \omega _j^i+\omega _i^j=0, \\ d\omega _j^i=- \sum\limits_k {\omega _k^i} \wedge \omega _j^k+\Omega _j^i, \Omega _j^i+\Omega _i^j=0, \\ \omega _\beta ^\alpha=- \sum\limits_\gamma {{\varepsilon _\gamma }} \omega _\gamma ^\alpha \wedge \omega _\beta ^\gamma+\Omega _\beta ^\alpha, \Omega _\beta ^\alpha+\Omega _\alpha ^\beta=0, \end{array} $ (2.2)
$ \begin{equation} \Omega_j^i={\overline\Omega}_j^i +\sum\limits_\alpha \varepsilon_\alpha {\omega_i^\alpha}\wedge\omega_j^\alpha, \quad \Omega_\beta^\alpha={\overline\Omega}_\beta^\alpha +\sum\limits_{i}{\omega_i^\alpha}\wedge\omega_i^\beta, \label{eq-2-3-4} \end{equation} $ (2.3)
$ \begin{equation} d\omega^\alpha=-\sum\limits_j{\omega_j^\alpha}\wedge\omega^j=0, \quad \omega_j^\alpha+\omega_\alpha^j=0, \label{eq-2-3-1} \end{equation} $ (2.4)
$ \begin{equation} d\omega_i^\alpha=-\sum\limits_j{\omega_j^\alpha}\wedge\omega_i^j-\sum\limits_{\gamma}\varepsilon_\gamma {\omega_\gamma^\alpha}\wedge\omega_i^\gamma+{\overline\Omega}_i^\alpha.\label{eq-2-3-2} \end{equation} $ (2.5)

应用Cartan引理和(2.4)式, 我们得到

$ \begin{equation}\label{eq-2-1-1} \omega_i^\alpha= h_{ij}^\alpha\omega^j, \quad h_{ij}^\alpha=h_{ji}^\alpha. \end{equation} $ (2.6)

第二基本形式h和平均曲率向量${\overrightarrow H}$分别定义如下

$ \begin{equation}\label{eq-2-2} h=\sum\limits_{\alpha, i, j}\varepsilon_\alpha\omega^i\omega_i^\alpha e_\alpha =\sum\limits_{\alpha, i, j}\varepsilon_\alpha h_{ij}^\alpha\omega^i\omega^j e_\alpha, \quad {\overrightarrow H}=\sum\limits_\alpha \varepsilon_\alpha {H^\alpha} e_\alpha, \end{equation} $ (2.7)

其中$H^\alpha=(\sum\limits_ih_{ii}^\alpha)/2$.平均曲率H定义为

$ \begin{equation}\label{eq-2-2-3} H=|{\overrightarrow H}|=\bigg[\sum\limits_{\alpha} {\left({H^\alpha}\right)^2}\bigg]^{1/2}.\end{equation} $ (2.8)

$ \begin{array}{*{20}{l}} {\Omega _j^i=\frac{1}{2}\sum\limits_{k, l} {R_{jkl}^i} {\omega ^k} \wedge {\omega ^l}, \;R_{jkl}^i+R_{jlk}^i=0;}\\ {\Omega _\beta ^\alpha=\frac{1}{2}\sum\limits_{k, l} {R_{\beta kl}^\alpha } {\omega ^k} \wedge {\omega ^l}, \;R_{\beta kl}^\alpha+R_{\beta lk}^\alpha=0.} \end{array} $ (2.9)

将(2.6)和(2.9)式代入(2.3)式, 我们可以得到Gauss方程和Ricci方程

$ \begin{eqnarray} &R_{jkl}^i={\overline R}_{jkl}^i+\sum\limits_{\alpha}\varepsilon_\alpha(h_{ik}^\alpha h_{jl}^\alpha-h_{il}^\alpha h_{jk}^\alpha), &R_{\beta kl}^\alpha={\overline R}_{\beta kl}^\alpha+\sum\limits_{i} (h_{ik}^\alpha h_{il}^\beta-h_{il}^\alpha h_{ik}^\beta).\label{eq-2-4} \end{eqnarray} $ (2.10)

定义h的协变微分如下

$ \nabla h_{ij}^\alpha:=dh_{ij}^\alpha -\sum\limits_k h_{ik}^\alpha{\omega_j^k}-\sum\limits_k{h_{kj}^\alpha\omega^k_i} +\sum\limits_{\gamma}\varepsilon_\gamma{h_{ij}^\gamma\omega_\gamma^\alpha}:= \sum\limits_{k}h_{ijk}^\alpha\omega^k. $

对(2.6)式外微分, 并利用(2.2)和(2.5)式, 我们可以得到Codazzi方程

$ \begin{eqnarray} h_{ijk}^\alpha-h_{ikj}^\alpha={\overline R}_{ikj}^\alpha.\label{eq-2-4-2} \end{eqnarray} $ (2.11)

${\partial}_t$分解为切部分T和法部分N, 即${\partial}_t=T+N$, 其中

$ \begin{equation}\label{eq-2-5} T=\sum\limits_iT^i e_i, \quad T^i=\overline{g}_\varepsilon(e_i, {\partial}_t);\quad N=\sum\limits_\alpha \varepsilon_\alpha N^\alpha e_\alpha, \quad N^\alpha=\overline{g}_\varepsilon(e_\alpha, {\partial}_t). \end{equation} $ (2.12)

对任意的$X\in \Gamma(T{\overline M}^n)$, 记

$ \widetilde X=X - \varepsilon\overline{g}_\varepsilon( X, {\partial}_t){\partial}_t, $

则有

$ \widetilde{e}_i=e_i- \varepsilon T^i{\partial}_t, \quad \widetilde{e}_\alpha=e_\alpha- \varepsilon N^\alpha{\partial}_t. $

我们有

$ {\overline R}_{BCD}^A =c[\overline{g}_\varepsilon(\widetilde{e}_A, \widetilde{e}_C)\overline{g}_\varepsilon(\widetilde{e}_B, \widetilde{e}_D) -\overline{g}_\varepsilon(\widetilde{e}_A, \widetilde{e}_D)\overline{g}_\varepsilon(\widetilde{e}_B, \widetilde{e}_C)], $

从而

$ \begin{equation} {\overline R}_{jkl}^i=c\{\delta_{ik}\delta_{jl}-\delta_{il}\delta_{jk}+\varepsilon\delta_{il}T^jT^k +\varepsilon\delta_{jk}T^iT^l-\varepsilon\delta_{ik}T^jT^l-\varepsilon\delta_{jl}T^iT^k\}, \label{eq-2-7} \end{equation} $ (2.13)
$ \begin{equation} {\overline R}_{ikj}^\alpha=c \varepsilon N^\alpha(T^j\delta_{ik}-T^k\delta_{ij}), \quad {\overline R}_{\beta kl}^\alpha=0.\label{eq-2-7-1} \end{equation} $ (2.14)

由Gauss方程, 我们可以计算出曲面∑2的Gauss曲率K

$ \begin{eqnarray}\label{eq-2-7-2} 2K=4\sum\limits_\alpha \varepsilon_\alpha(H^\alpha)^2-\sum\limits_{\alpha, i, j}\varepsilon_\alpha(h_{ij}^\alpha)^2+2c(1-\varepsilon |T|^2). \end{eqnarray} $ (2.15)
3 推广的张量S与Simons型方程

设∑2$(M^{n - 1}(c)\times {\Bbb R}, \overline{g}_{\varepsilon})$中的PMC浸入曲面.我们在曲面∑2上定义推广的二次型Q和张量S如下

$ \begin{equation} Q(X, Y)=2\overline{g}_\varepsilon(h(X, Y), \overrightarrow{H}) -c\varepsilon\overline{g}_\varepsilon(X, {\partial}_t)\overline{g}_\varepsilon(Y, {\partial}_t), \label{eq-3-0-1} \end{equation} $ (3.1)
$ \begin{eqnarray} g(SX, Y)&=&2\overline{g}_\varepsilon(h(X, Y), \overrightarrow{H}) -c\varepsilon\overline{g}_\varepsilon(X, {\partial}_t)\overline{g}_\varepsilon(Y, {\partial}_t)\nonumber\\ &&+\frac{c}{2}\varepsilon |T|^2 g(X, Y)-2\overline{g}_\varepsilon(\overrightarrow{H}, \overrightarrow{H})g(X, Y), \label{eq-3-0-2} \end{eqnarray} $ (3.2)

其中X, Y是∑2上的切向量.

$\varepsilon=1$的情形, 我们定义的二次型Q和张量S恰好是文献[2]中定义的Q和文献[22]中定义的张量S.通常称满足Q(2, 0)=0的浸入曲面$\Sigma^2 \looparrowright(M^2(c)\times {\Bbb R}, \overline{g}_{+1})$为Abresch-Rosenberg曲面.Alencar, do Carmo和Tribuzy[2]证明了Q(2, 0)是全纯的.通过这一证明过程[2, 第2节], 容易可以看出:对$\varepsilon=-1$的情形, Q(2, 0)也是全纯的.我们称满足Q(2, 0)=0的浸入曲面$\Sigma^2 \looparrowright(M^2(c)\times {\Bbb R}, \overline{g}_{-1})$为Abresch-Rosenberg型曲面.

在这一节, 我们首先证明了S满足的一些性质, 然后得到了关于ST的Simons型方程.首先, 我们知道子流形的基本公式为

$ \begin{equation} \overline\nabla_{e_i } e_j=\nabla_{e_i} e_j+h(e_i, e_j), \label{eq-3-1} \end{equation} $ (3.3)
$ \begin{equation} \overline\nabla_{e_i } e_\alpha =-A_\alpha e_i+\nabla_{e_i}^\bot e_\alpha.\label{eq-3-2} \end{equation} $ (3.4)

引理3.1[11, 24-25]  设∑2$\overline M^n$中的曲面.则

$ \begin{equation}\nabla T^i =\sum\limits_{\alpha, j}\varepsilon_\alpha{N^\alpha}h_{ij}^\alpha\omega^j, \label{eq-3-3} \end{equation} $ (3.5)
$ \begin{equation} \nabla^\bot N^\alpha=-\sum\limits_{i, j} {T^i}h_{ij}^\alpha\omega^j.\label{eq-3-4} \end{equation} $ (3.6)

  由${\partial}_t$$\overline M^n$上是平行的, 故

$ \begin{eqnarray} 0&=&\overline\nabla{\partial}_t=\overline\nabla (T+N)=\overline\nabla(T^i e_i) +\overline\nabla(\varepsilon_\alpha N^\alpha e_\alpha)\nonumber\\ &=&dT^i\otimes e_i+T^i\omega_i^j\otimes e_j+T^i\varepsilon_\alpha \omega_i^\alpha \otimes e_\alpha+\varepsilon_\alpha dN^\alpha\otimes e_\alpha+\varepsilon_\alpha N^\alpha(\omega_\alpha^i\otimes e_i+\varepsilon_\beta\omega_\alpha^\beta\otimes e_\beta)\nonumber\\ &=&(dT^i+T^j\omega_j^i+\varepsilon_\alpha N^\alpha\omega_\alpha^i)\otimes e_i +\varepsilon_\alpha(dN^\alpha+T^i\omega_i^\alpha+\varepsilon_\beta N^\beta\omega_\beta^\alpha)\otimes e_\alpha. \label{eq-3-5} \end{eqnarray} $ (3.7)

因此, 我们可以得到

$ \begin{equation} 0=dT^i+T^j\omega_j^i+\varepsilon_\alpha N^\alpha\omega_\alpha^i =\nabla T^i+\varepsilon_\alpha N^\alpha\omega_\alpha^i, \label{eq-3-6} \end{equation} $ (3.8)
$ \begin{equation} 0=dN^\alpha+T^i\omega_i^\alpha+\varepsilon_\beta N^\beta\omega_\beta^\alpha =\nabla^\bot N^\alpha+T^i\omega_i^\alpha, \label{eq-3-7} \end{equation} $ (3.9)

从而

$ \begin{eqnarray*} \nabla T^i=\sum\limits_\alpha\varepsilon_\alpha{N^\alpha\omega_i^\alpha} =\sum\limits_{\alpha, j} \varepsilon_\alpha {N^\alpha} h_{ij}^\alpha\omega^j, \quad \nabla^\bot N^\alpha=-\sum\limits_i {T^i}\omega_i^\alpha =-\sum\limits_{i, j}{T^i}h_{ij}^\alpha\omega^j. \end{eqnarray*} $

证毕.

利用引理3.1, 我们可以得到

定理3.1  设∑2$\overline{M}^{n}_\varepsilon$中具有PMC的浸入曲面, 则

$ \begin{eqnarray}\label{eq-3-7-1} \frac{1}{2}\Delta {|T|}^2=c\varepsilon {|T|}^2 \sum\limits_{\alpha}\varepsilon_\alpha(N^\alpha)^2-\sum\limits_{\alpha, k}\varepsilon_\alpha\Big(\sum\limits_{i}T^i h^\alpha_{ik} \Big)^2+|\nabla T|^2, \end{eqnarray} $ (3.10)

其中,

$ \begin{eqnarray}\label{eq-3-7-2} |\nabla T|^2=\sum\limits_{i}|\nabla T^i|^2=\sum\limits_{i, j}\Big(\sum\limits_{\alpha} \varepsilon_\alpha N^\alpha h^\alpha_{ij} \Big)^2. \end{eqnarray} $ (3.11)

  由$\sum\limits_{k}\overline{R}^\alpha_{kik}=-c\varepsilon T^i N^\alpha $$ \sum\limits_{k} h^\alpha_{kki}=0 $, 我们有

$ \begin{eqnarray*} \Delta T^i &=&\sum\limits_{k}\nabla_{e_k}(\nabla_{e_k}T^i)\\ &=&\sum\limits_{\alpha, k}\nabla_{e_k}( \varepsilon_\alpha N^\alpha h^\alpha_{ik}) =\sum\limits_{\alpha, k}\varepsilon_\alpha\big[(\nabla^\bot_{e_k} N^\alpha)h^\alpha_{ik}+N^\alpha h^\alpha_{ikk}\big] \\ &=&-\sum\limits_{\alpha, k, l}\varepsilon_\alpha T^l h^\alpha_{lk}h^\alpha_{ik}+\sum\limits_{\alpha, k}\varepsilon_\alpha N^\alpha(h^\alpha_{kki}-\overline{R}^\alpha_{kik}) \\ &=&-\sum\limits_{\alpha, k, l}\varepsilon_\alpha T^l h^\alpha_{lk}h^\alpha_{ik}+c\varepsilon T^i \sum\limits_{\alpha}\varepsilon_\alpha(N^\alpha)^2, \end{eqnarray*} $

从而

$ \begin{eqnarray*} \frac{1}{2}\Delta |T|^2 &=&|\nabla T|^2+\sum\limits_{i}T^i\Delta T^i\\ &=&|\nabla T|^2-\sum\limits_{\alpha, i, k, l}\varepsilon_\alpha T^i T^l h^\alpha_{lk}h^\alpha_{ik}+c\varepsilon |T|^2 \sum\limits_{\alpha}\varepsilon_\alpha(N^\alpha)^2 \\ &=&|\nabla T|^2-\sum\limits_{\alpha, k}\varepsilon_\alpha\Big(\sum\limits_{i}T^i h^\alpha_{ik} \Big)^2+c\varepsilon |T|^2 \sum\limits_{\alpha}\varepsilon_\alpha(N^\alpha)^2. \end{eqnarray*} $

根据(3.2)式, 张量S的分量$S_{ij}$

$ \begin{equation} S_{ij}=2\sum\limits_\alpha \varepsilon_\alpha{H^\alpha} h_{ij}^\alpha-c\varepsilon T^iT^j +\frac{c}{2}\varepsilon |T|^2\delta_{ij}-2\sum\limits_\alpha \varepsilon_\alpha(H^\alpha)^2\delta_{ij}, \label{eq-3-9} \end{equation} $ (3.12)

$S_{ij}$的一阶协变微分$S_{ijk}$和二阶协变微分$S_{ijkl}$分别定义如下

$ \begin{equation} \sum\limits_k {S_{ijk}}\omega^k =dS_{ij}-\sum\limits_k{S_{ik}}\omega_j^k-\sum\limits_k{S_{kj}}\omega_i^k, \label{eq-3-a} \end{equation} $ (3.13)
$ \begin{equation} \sum\limits_l{S_{ijkl}}\omega^l =dS_{ijk}-\sum\limits_l{S_{ljk}}\omega_i^l-\sum\limits_l{S_{ilk}}\omega_j^l -\sum\limits_l{S_{ijl}}\omega_k^l.\label{eq-3-b} \end{equation} $ (3.14)

利用(3.12), (3.13)和(2.2)式, 我们可以得到Ricci恒等式

$ \begin{equation}\label{eq-3-c} S_{ijkl}-S_{ijlk}=\sum\limits_m {S_{im}} R^m_{jkl}+\sum\limits_m {S_{mj}} R^m_{ikl}.\end{equation} $ (3.15)

证毕.

Batista在文献[8]中证明了$\Sigma^2\looparrowright(M^2(c)\times {\Bbb R}, \overline{g}_{+1})$上的张量S满足"Codazzi"方程.下面我们将利用不一样的方法证明广义张量S满足下面的性质.

引理3.2  设∑2$(M^n(c)\times {\Bbb R}, \overline{g}_{\varepsilon})$中具有PMC的浸入曲面, 则张量S是对称无迹的, 且满足"Codazzi"方程, 即

$ \begin{equation}\label{eq-3-d} S_{ij}=S_{ji}, \qquad \sum\limits_i {S_{ii}}=0, \qquad S_{ijk}=S_{ikj}.\end{equation} $ (3.16)

  (3.16)式中前两个式子是显而易见的.下面证明第三个方程.由于$\nabla^\bot\overrightarrow H=0$, 对任意的k$\alpha$$H^\alpha_{, k}=0$.利用相容性方程(3.5), 有

$ \begin{array}{l} {S_{ijk}}=2\sum\limits_\alpha {{\varepsilon _\alpha }} {H^\alpha }h_{ijk}^\alpha - c\varepsilon {T^i}{\nabla _k}{T^j} - c\varepsilon {T^j}{\nabla _k}{T^i}+\frac{c}{2}\varepsilon {\delta _{ij}}{\nabla _k}(|T{|^2})\\ \;\;\;\;\;\;=2\sum\limits_\alpha {{\varepsilon _\alpha }} {H^\alpha }h_{ijk}^\alpha - c\varepsilon {T^i}\sum\limits_\alpha {{\varepsilon _\alpha }} {N^\alpha }h_{jk}^\alpha - c\varepsilon {T^j}\sum\limits_\alpha {{\varepsilon _\alpha }} {N^\alpha }h_{ik}^\alpha \\ \;\;\;\;\;\;+c\varepsilon {\delta _{ij}}\sum\limits_{\alpha, l} {{\varepsilon _\alpha }} {N^\alpha }{T^l}h_{lk}^\alpha. \end{array} $ (3.17)

$ \begin{eqnarray} S_{ijk}-S_{ikj}&=&2\sum\limits_\alpha \varepsilon_\alpha{H^\alpha}(h_{ijk}^\alpha-h_{ikj}^\alpha) +c\varepsilon\Big\{T^k\sum\limits_\alpha\varepsilon_\alpha{N^\alpha}h_{ij}^\alpha - T^j\sum\limits_\alpha\varepsilon_\alpha{N^\alpha}h_{ik}^\alpha \nonumber\\ &&+\delta_{ij}\sum\limits_{\alpha, l}\varepsilon_\alpha{N^\alpha T^l}h_{lk}^\alpha -\delta_{ik}\sum\limits_{\alpha, l}\varepsilon_\alpha{N^\alpha T^l}h_{lj}^\alpha\Big\}.\label{eq-3-e} \end{eqnarray} $ (3.18)

根据(2.11)和(2.14)式, 有

$ \begin{equation}\label{eq-3-f} h_{i12}^\alpha-h_{i21}^\alpha=-\bar R_{i12}^\alpha=c \varepsilon N^\alpha(T^1\delta_{i2}-T^2\delta_{i1}).\end{equation} $ (3.19)

对任意的i, 有恒等式

$ \begin{equation}\label{eq-3-g} h_{i1}^\alpha=\delta_{i1} h_{11}^\alpha+\delta_{i2} h_{12}^\alpha, \qquad h_{i2}^\alpha=\delta_{i1} h_{12}^\alpha+\delta_{i2} h_{22}^\alpha, \end{equation} $ (3.20)

将(3.19)和(3.20)式代入(3.18)式, 并令j=1, k=2, 可得

$ \begin{eqnarray*} S_{i12}-S_{i21}&=&c\varepsilon\Big\{ (\delta_{i2}T^1-\delta_{i1}T^2)\sum\limits_\alpha\varepsilon_\alpha{N^\alpha}(h_{11}^\alpha+h_{22}^\alpha) \nonumber\\ &&+T^2\sum\limits_\alpha\varepsilon_\alpha{N^\alpha}(\delta_{i1} h_{11}^\alpha+\delta_{i2} h_{12}^\alpha) -T^1\sum\limits_\alpha\varepsilon_\alpha{N^\alpha}(\delta_{i1} h_{12}^\alpha+\delta_{i2} h_{22}^\alpha)\nonumber\\ &&+\delta_{i1}\sum\limits_{\alpha, l}\varepsilon_\alpha{N^\alpha T^l} h_{l2}^\alpha-\delta_{i2} \sum\limits_{\alpha, l}\varepsilon_\alpha{N^\alpha T^l} h_{l1}^\alpha\Big\}\nonumber\\ &=&c\varepsilon\delta_{i1}\sum\limits_\alpha \varepsilon_\alpha{N^\alpha}\Big\{T^2 h_{11}^\alpha-T^1 h_{12}^\alpha -T^2(h_{11}^\alpha+h_{22}^\alpha)+\sum\limits_l{T^l}h_{l2}^\alpha\Big\} \nonumber\\ &&+c\varepsilon\delta_{i2}\sum\limits_\alpha\varepsilon_\alpha{N^\alpha}\Big\{T^2 h_{12}^\alpha-T^1 h_{22}^\alpha +T^1(h_{11}^\alpha+h_{22}^\alpha)-\sum\limits_l{T^l}h_{l1}^\alpha\Big\} \nonumber\\ &=&0.\nonumber \end{eqnarray*} $

证毕.

Cheng和Yau[12]得到了任意满足"Codazzi"方程的张量的Simons型方程, 它是一个内蕴的结论.Fetcu和Rosebberg在文献[22]利用这一结论得到了关于S的Simons型方程.事实上, 由于(3.2)式所定义的算子S也满足"Codazzi"方程, 从而我们有以下的Simons型方程(这与文献[22]中的结论在形式上式一致的).

定理3.2[22]  设∑2$(M^n(c)\times {\Bbb R}, \overline{g}_{\varepsilon})$中具有PMC的浸入曲面, 则

$ \begin{equation}\label{eq-3-h} \frac{1}{2}\Delta(|S|^2)=|\nabla S|^2+2K |S|^2, \end{equation} $ (3.21)

其中$|\nabla S|^2=\sum\limits_{i, j, k} {S_{ijk}^2 }$ K是曲面∑2的高斯曲率.

4 $(M^{n-1}(c)\times {\Bbb R}, \overline{g}_{\varepsilon})$中的曲面的几何特征

在这一节, 我们将利用Simons型方程研究$|S|^2$$|T|^2$的Pinching现象, 并得到曲面$\Sigma^2 \looparrowright \overline{M}^{n}_\varepsilon $的一些几何特征.

由(3.1)和(3.2)式, 可知

$ \begin{equation}\label{eq-4-0-1} g(SX, Y)=Q(X, Y)-\frac{{\rm tr} Q}{2} g(X, Y), \end{equation} $ (4.1)

由此, 我们可以得到下面的引理.

引理4.1[8, 22]  设∑2$(M^{n-1}(c)\times {\Bbb R}, \overline{g}_{\varepsilon})$中完备的PMC浸入曲面, 则|S|=0当且仅当$|Q^{(2, 0)}|=0$.因此, 满足|S|=0的曲面是Abresh-Rosenberg型曲面.

由定理3.2和引理4.1, 我们可以得到

定理4.1  设∑2$(M^{n-1}(c)\times {\Bbb R}, \overline{g}_{\varepsilon})$中完备的PMC浸入曲面, 且高斯曲率满足$\mathop{\inf}\limits_{\Sigma^2} K >0 $, 则|S|=0, 即∑2是Abresch-Rosenberg型曲面.

  条件假设和Bonnet-Myers定理表明了曲面∑2是紧致的.又由定理3.2, 容易知道$|S|^2$是下调和的, 从而$|S|^2$是调和的.再根据定理3.2, 可知$|S|=0$.

注4.1  对$\varepsilon=+1$的情形, 定理4.1包含于文献[22]中的定理1.2.

接下来, 我们分$\varepsilon=-1$$\varepsilon=+1$两种情况来研究$(M^2(c)\times {\Bbb R}, \overline{g}_{\varepsilon})$中的CMC曲面.首先, 我们在曲面∑2上选取适当的标准正交切标架$\{e_1, e_2\}$, 满足$S_{ij}=\mu_i\delta_{ij}$.由于S是无迹, 故可设$\mu_1=-\mu_2=\mu$.根据(2.13)和(2.15)式, 有

$ \begin{equation}\label{eq-4} 2K=2\varepsilon H^2-2\varepsilon(h_{12})^2-\frac{1}{2}\varepsilon(h_{11}-h_{22})^2+2c(1-\varepsilon |T|^2). \end{equation} $ (4.2)
4.1 $(M^2(c)\times {\Bbb R}, \overline{g}_{-1})$中的类空曲面

定理4.2  设∑2$({\Bbb S}^{2}(1)\times {\Bbb R}, \overline{g}_{-1})$中的完备类空CMC曲面且$0<H^2<1$.则|S|=0, 即曲面∑2是Abresch-Rosenberg型曲面.

  由(3.12)式知

$ \begin{equation}\label{eq-4-2-1} S_{ij}= -\Big(\sum\limits_{k}h_{kk}\Big)h_{ij}+cT^iT^j-\frac{c}{2}|T|^2\delta_{ij}+\frac{1}{2} \Big(\sum\limits_{k}h_{kk}\Big)^2\delta_{ij}, \end{equation} $ (4.3)

从而

$ \mu=\mu_1=-\Big(\sum\limits_{k}h_{kk}\Big)h_{11}+c(T^1)^2-\frac{c}{2}|T|^2+2H^2, $
$ -\mu=\mu_2=-\Big(\sum\limits_{k}h_{kk}\Big)h_{22}+c(T^2)^2-\frac{c}{2}|T|^2+2H^2, \quad 0=-\Big(\sum\limits_{k}h_{kk}\Big)h_{12}+cT^1T^2, $

由此可得

$ \begin{array}{*{20}{l}} {{{({h_{11}}- {h_{22}})}^2}=\frac{1}{{4{H^2}}}\{ 4{\mu ^2}- 4c\mu [{{({T^1})}^2}-{{({T^2})}^2}]+{c^2}{{[{{({T^1})}^2}-{{({T^2})}^2}]}^2}\}, }\\ {{{({h_{12}})}^2}=\frac{{{c^2}}}{{4{H^2}}}{{({T^1}{T^2})}^2}.} \end{array} $ (4.4)

将(4.4)式代入(4.2)式, 并取$\varepsilon=-1$, 可得

$ \begin{equation}\label{eq-4-2-4} 2K=\frac{1}{4H^2} \Big\{-8H^4+8cH^2(1+|T|^2)+\frac{c^2}{2}|T|^4+|S|^2-2c\mu\Big[(T^1)^2-(T^2)^2\Big]\Big\}. \end{equation} $ (4.5)

在(4.15)式中令c=1, 有

$ 2K \ge \frac{1}{{4{H^2}}}\{-8{H^4}+8{H^2}(1+|T{|^2})+\frac{1}{2}|T{|^4}+|S{|^2}-2|\mu | \cdot |T{|^2}\} $ (4.6)
$ =\frac{1}{{4{H^2}}}\{-8{H^4}+8{H^2}(1+|T{|^2})+{(|S|-\frac{{|T{|^2}}}{{\sqrt 2 }})^2}\} $ (4.7)
$ \ge 2(1-{H^2})> 0. $ (4.8)

利用定理4.1立即有|S|=0, 曲面∑2是Abresch-Rosenberg型曲面.

定理4.3  设∑2$({\Bbb H}^{2}(-1)\times {\Bbb R}, \overline{g}_{-1})$中的完备类空CMC曲面且$H\ne 0$, $g(ST, T)\geq 0$.若

$ \begin{equation}\label{eq-4-2-8} \mathop{\inf}\limits_{\Sigma^2}|T|^2> 8H^2+4\sqrt{H^2(5H^2+1)}, \end{equation} $ (4.9)

则|S|=0, 即曲面∑2是Abresch-Rosenberg型曲面.

  根据前提条件, 有

$ g(ST, T)=\mu\big[(T^1)^2-(T^2)^2\big]\geq 0. $

将上式代入到(4.5)式, 并令c=-1, 我们得到

$ \begin{eqnarray} 2K&=&\frac{1}{4H^2} \Big\{-8H^4-8H^2(1+|T|^2)+\frac{1}{2}|T|^4+|S|^2+2\mu\big[(T^1)^2-(T^2)^2\big]\Big\}\nonumber\\ &\ge&\frac{1}{4H^2}\Big\{|S|^2+\frac{1}{2}|T|^4-8H^2|T|^2-8H^2-8H^4\Big\}.\label{eq-4-2-9} \end{eqnarray} $ (4.10)

$f(x)=\frac{1}{2}x^2-8H^2 x-8H^2-8H^4$, 则由条件(4.9)式, 我们有

$ f(|T|^2)=\frac{1}{2}|T|^4-8H^2|T|^2-8H^2-8H^4\geq f(\mathop{\inf}\limits_{\Sigma^2}|T|^2)>0, $

从而$\mathop{\inf}\limits_{\Sigma^2} K>0$.因此|S|=0, 曲面∑2是Abresch-Rosenberg型曲面.

现在, 我们利用关于$|T|^2$的Simons型方程研究$(M^{2}(c)\times {\Bbb R}, \overline{g}_{-1})$, c>0中的紧致类空的CMC曲面.

定理4.4  设∑2$(M^{2}(c)\times {\Bbb R}, \overline{g}_{-1})$, c>0中的紧致类空CMC曲面.则∑2是全测地的且$|T|^2=0$.此时, 曲面∑2自然就是Abresch-Rosenberg型曲面.

  在(3.10)式中令$\varepsilon=-1 $可得

$ \begin{array}{l} \frac{1}{2}\Delta |T{|^2}=c|T{|^2} \cdot |N{|^2}+\sum\limits_k(\sum\limits_i {{T^i}} {h_{ik}}{)^2}+|N{|^2}\sum\limits_{i, j} {{{({h_{ij}})}^2}} \\ \;\;\;\;\;\;\;\;\;\;\;\;\;\ge |N{|^2}(c|T{|^2}+\sum\limits_{i, j} {{{({h_{ij}})}^2}})\ge 0, \end{array} $ (4.11)

其中$|N|^2=\sum\limits_{\alpha}(N^\alpha)^2$.由于$|N|^2=|T|^2+1>0$及紧致性, 可知$\sum\limits_{i, j}(h_{ij})^2=|T|^2=0$.因此, ∑2是全测地的.再由(3.12)式, 可知|S|=0, 因此∑2是Abresch-Rosenberg型曲面.

4.2 $(M^2(c)\times {\Bbb R}, \overline{g}_{+1})$中的曲面

在文献[8]中, 作者得到了若干个关于S的Pinching常数.在这一小节, 我们利用Simons型方程来研究关于S的Pinching问题, 得到了一些更佳的Pinching常数.

定理4.5  设∑2$({\Bbb S}^{2}(1)\times {\Bbb R}, \overline{g}_{+1})$中的完备CMC曲面, 满足H>1/2.如果

$ \begin{eqnarray}\label{eq-4-0} \sup\limits_{\Sigma^2}|S|<\frac{\sqrt{2}}{2}(4H^2-1), \end{eqnarray} $ (4.12)

则|S|=0, 即曲面∑2是Abresch-Rosenberg曲面.

  根据(3.12)式可知

$ \begin{eqnarray}\label{eq-4-1} S_{ij}&= 2Hh_{ij}-cT^iT^j+\frac{c}{2}|T|^2\delta_{ij}-2H^2\delta_{ij}. \end{eqnarray} $ (4.13)

于是

$ \mu=\mu_1=2Hh_{11}-c(T^1)^2+\frac{c}{2}||T||^2-2H^2, $
$ -\mu=\mu_2=2Hh_{22}-c(T^2)^2+\frac{c}{2}||T||^2-2H^2, \quad 0=2Hh_{12}-cT^1T^2, $

从上面两个式子可以得到

$ \begin{array}{*{20}{l}} {{{({h_{11}}- {h_{22}})}^2}=\frac{1}{{4{H^2}}}\{ 4{\mu ^2}+4c\mu [{{({T^1})}^2}-{{({T^2})}^2}]+{c^2}{{[{{({T^1})}^2}-{{({T^2})}^2}]}^2}\}, }\\ {{{({h_{12}})}^2}=\frac{{{c^2}}}{{4{H^2}}}{{({T^1}{T^2})}^2}.} \end{array} $ (4.14)

将(4.14)式代入(4.2)式, 并令$\varepsilon=1$, 可得

$ \begin{equation}\label{eq-4-4} 2K=\frac{1}{4H^2} \Big\{8H^4+8cH^2(1-|T|^2)-\frac{c^2}{2}|T|^4-|S|^2-2c\mu\Big[(T^1)^2-(T^2)^2\Big]\Big\}. \end{equation} $ (4.15)

在(4.15)式中令c=1, 有

$ \begin{array}{l} 2K \ge \frac{1}{{4{H^2}}}\{ 8{H^4}+8{H^2}(1-|T{|^2})-\frac{1}{2}|T{|^4}-|S{|^2}- 2|\mu | \cdot |T{|^2}\} \\ \;\;\;\ge \frac{1}{{4{H^2}}}\{(8{H^4} - \frac{1}{2})- |S{|^2} - \sqrt 2 |S|\}=\frac{1}{{4{H^2}}}(L_H^+- |S|)(|S| - L_H^ -) \end{array} $ (4.16)
$ \;\;\;\ge \frac{{-L_H^-}}{{4{H^2}}}(L_H^+-\mathop {\sup }\limits_{{\Sigma ^2}} |S|)> 0, $ (4.17)

其中$ L^{\pm}_H=\frac{\sqrt{2}}{2}(\pm 4H^2-1)$是下面多项式的两个根

$ P_H(x)=-x^2-\sqrt{2} x+8H^4-\frac{1}{2}. $

在(4.16)式中的前两个不等式分别用到了以下事实

$ \begin{eqnarray} -2\mu[(T^1)^2-(T^2)^2]\ge-2|\mu|\cdot |T|^2=-\sqrt{2}|S|\cdot |T|^2\quad\mbox{和}\quad 0\le|T|^2\le 1. \label{eq-4-6} \end{eqnarray} $ (4.18)

利用式子(4.17)和定理4.1立即可得|S|=0, 曲面∑2是Abresch-Rosenberg曲面.

注4.2  定理1.5中的Pinching常数$\frac{\sqrt{2}}{2}(4H^2-1)$比文献[8, 定理4.3]的Pinching常数$\sqrt{8H^4-2H^2+1}-1$更大.

定理4.6   设∑2$({\Bbb S}^{2}(1)\times {\Bbb R}, \overline{g}_{+1})$中的完备CMC曲面, 满足$H>{1}/{2}$, 则不存在这样的曲面使得$|S|=\frac{\sqrt{2}}{2}(4H^2-1)$.

  当$|S|=\frac{\sqrt{2}}{2}(4H^2-1)$时, 不等式(4.16)成为等式.由(4.18)式知

$ \begin{eqnarray*}(1)\quad T^1=|T|=1, \quad T^2=0, \quad\mu >0;\quad\mbox{或}\quad(2)\quad T^2=|T|=1, \quad T^1=0, \quad\mu <0.\end{eqnarray*} $

对上式中的两种情况, 我们都可以容易看出${\partial}_t=T$.这意味着曲面∑2是一个柱面$\Gamma\times{\Bbb R}$, 其中$\Gamma$${\Bbb S}^{2}(1)$中曲率为常数2H的曲线.根据文献[8]中定理4.4的证明过程知

$ \left({{S_{ij}}} \right)=\left({\begin{array}{*{20}{c}} {2{H^2}+\frac{1}{2}}&0\\ 0&{ - \frac{1}{2} - 2{H^2}} \end{array}} \right). $

因此, $|S|=\frac{\sqrt{2}}{2}(4H^2+1)>\frac{\sqrt{2}}{2}(4H^2-1)$, 这与前提条件矛盾.

通过增加一个额外的条件, 我们还可以进一步改进定理1.5.

定理4.7   设∑2$({\Bbb S}^{2}(1)\times {\Bbb R}, \overline{g}_{+1})$中的完备CMC曲面且$H>{1}/{2}$, $ g(ST, T)\leq 0$.若

$ \begin{eqnarray}\label{eq-4-7} \sup\limits_{\Sigma^2}|S|<\sqrt{8H^4-\frac{1}{2}}, \end{eqnarray} $ (4.19)

则|S|=0, 即曲面∑2是Abresch-Rosenberg曲面.

  由(4.15)式和前提条件$g(ST, T)=\mu\big[(T^1)^2-(T^2)^2\big]\leq 0$, 可得

$ \begin{eqnarray*} 2K &\ge&\frac{1}{4H^2} \Big\{8H^4+8H^2(1-|T|^2)-\frac{1}{2}|T|^4-|S|^2\Big\}\nonumber\\ &\ge&\frac{1}{4H^2}\Big\{\Big(8H^4-\frac{1}{2}\Big)-\sup\limits_{\Sigma}|S|^2\Big\}>0. \end{eqnarray*} $

从而立即知曲面∑2是Abresch-Rosenberg曲面.

定理4.8   设∑2$({\Bbb H}^{2}(-1)\times {\Bbb R}, \overline{g}_{+1})$中的完备CMC曲面, 满足H>1.如果

$ \begin{eqnarray}\label{eq-4-9} \sup\limits_{\Sigma^2}|S|<\frac{\sqrt{2}}{2}[\sqrt{16H^2(H^2-1)+1}-1], \end{eqnarray} $ (4.20)

则|S|=0, 即曲面∑2是Abresch-Rosenberg曲面.

  在(4.15)式中令c=-1有

$ \begin{array}{l} 2K=\frac{1}{{4{H^2}}}\{ 8{H^2}({H^2} - 1)+|T{|^2}(8{H^2} - \frac{1}{2}|T{|^2})- |S{|^2}+2\mu [{({T^1})^2}-{({T^2})^2}]\} \\ \;\;\;\;\ge \frac{1}{{4{H^2}}}\{ 8{H^2}({H^2} - 1)- |S{|^2} - 2|\mu | \cdot |T{|^2}\} \\ \;\;\;\;\ge \frac{1}{{4{H^2}}}\{ 8{H^2}({H^2} - 1)- |S{|^2} - \sqrt 2 |S|\}=\frac{1}{{4{H^2}}}(Q_H^+- |S|)(|S| - Q_H^ -) \end{array} $ (4.21)
$ \;\;\;\;\ge \frac{{-Q_H^-}}{{4{H^2}}}(Q_H^+-\mathop {\sup }\limits_{{\Sigma ^2}} |S|)> 0, $ (4.22)

其中$Q^\pm_H=\frac{\sqrt{2}}{2}[-1\pm\sqrt{16H^2(H^2-1)+1}]$是以下多项式的两个根

$ \begin{eqnarray*} q_H(x)=-x^2-\sqrt 2 x+8H^2(H^2-1).\end{eqnarray*} $

容易看出$Q^+_H>0$当且仅当H>1.由(4.22)式和定理4.1即可证明所要的结论.

注4.3  容易看出定理4.8中的Pinching常数$\frac{\sqrt{2}}{2}[\sqrt{16H^2(H^2-1)+1}-1]$大于文献[8]中的定理4.6的Pinching常数$\sqrt{8H^4-12H^2}-1$, 且条件H>1比$H>\sqrt{\frac{{3+\sqrt{11}}}{4}}$更弱.

定理4.9  设∑2$({\Bbb H}^{2}(-1)\times {\Bbb R}, \overline{g}_{+1})$中的完备CMC曲面, 且满足H>1, 则不存在这样的曲面使得$|S|=\frac{\sqrt{2}}{2}[\sqrt{16H^2(H^2-1)+1}-1]$.

  当$|S|=\frac{\sqrt{2}}{2}[\sqrt{16H^2(H^2-1)+1}-1]$时, 不等式(4.21)式成为等式.从而可以看出|T|=1和|T|=0同时成立.矛盾.

参考文献
[1] Abresch U, Rosenberg H. A Hopf differential for conastant mean curvature surfaces in $\mathbb{S}$2×$\mathbb{R}$ and $\mathbb{H}$2×$\mathbb{R}$. Acta Math, 2004, 193: 141–174.
[2] Alencar H, do Carmo M, Tribuzy R. A theorem of Hopf and the Cauchy-Riemann inequality. Comm Anal Geom, 2007, 15: 283–298.
[3] Alencar H, do Carmo M, Isabel F, Tribuzy R. A theorem of Hopf and the Cauchy-Riemann inequality Ⅱ. Bull Braz Math Soc, 2007, 38: 525–532.
[4] Alencar H, do Carmo M, Tribuzy R. A Hopf theorem for ambient spaces of dimensions higher than three. J Diff Geom, 2010, 84: 1–17.
[5] Alencar H, do Carmo M, Tribuzy R. Surfaces of Mκ2×R invariant under a one-parameter group of isometries. Ann Mat Pura Appl, 2004, 193(4): 517–527.
[6] Aquino C P, de Lima H F, Lima Eraldo A. On the angle of complete CMC hypersurfaces in Riemannian product spaces. Diff Geom Appl, 2014, 33: 139–148.
[7] Baek J O, Cheng Q M, Suh Y J. Complete space-like hypersurfaces in locally symmetric Lorentz spaces. J Geom and Phys, 2004, 49: 231–247.
[8] Batista M. Simons type equation in $\mathbb{S}$2×$\mathbb{R}$ and $\mathbb{H}$2×$\mathbb{R}$ and applications. Ann Inst Fourier (Grenoble), 2011, 61: 1299–1322.
[9] do Carmo M. Some recent developments on Hopf's holomorphic form. Results Math, 2011, 60: 175–183.
[10] do Carmo M, Isabel F. A Hopf theorem for open surfaces in product spaces. Forum Math, 2009, 21: 951–963.
[11] Chen H, Chen G Y, Li H Z. Some pinching theorems for minimal submanifolds in $\mathbb{S}$m(1)×$\mathbb{R}$. Sci China Math, 2013, 56: 1679–1688.
[12] Cheng S Y, Yau S T. Hypersurfaces with constant scalar curvature. Math Ann, 1997, 225: 195–204.
[13] Deng Q T. Complete hypersurfaces with constant mean curvature and finite index in hyperbolic spaces. Acta Math Sci, 2011, 31: 353–360.
[14] Dillen F, Fastenakels J, Van der Veken J. Surfaces in $\mathbb{S}$2×$\mathbb{R}$ with a canonical principal direction. Ann Global Anal Geom, 2009, 35: 381–396.
[15] Dillen F, Fastenakels J, Van der Veken J, Vrancken L. Constant angle surfaces in S2×$\mathbb{R}$. Monaths Math Soc, 2007, 152(2): 89–96.
[16] Dillen F, Munteanu M I. Constant angle surfaces in $\mathbb{H}$2×$\mathbb{R}$. Bull Braz Math Soc, 2009, 40(1): 85–97.
[17] Dillen F, Munteanu M I, Nistor A I. Canonical coordinates and principal directions for surfaces in $\mathbb{H}$2×$\mathbb{R}$. Taiwanese J Math, 2011, 15: 2265–2289.
[18] Espinar J M, Rosenberg H. Complete constant mean curvature surfaces in homogeneous spaces. Comment Math Helv, 2011, 86: 659–674.
[19] Ferreira M J, Tribuzy R. Parallel mean curvature surfaces in symmetric spaces. Ark Mat, 2014, 52: 93–98.
[20] Fetcu D, Rosenberg H. Surfaces with parallel mean curvature in $\mathbb{S}$3×$\mathbb{R}$ and $\mathbb{H}$3×$\mathbb{R}$. Mich Math J, 2012, 61: 715–729.
[21] Fetcu D, Rosenberg H. On complete submanifolds with parallel mean curvature in product spaces. Rev Mat Iberoam, 2013, 29: 1283–1306.
[22] Fetcu D, Rosenberg H. A note on Surfaces with parallel mean curvature. C R Acad Sci Paris Ser Ⅰ, 2011, 349: 1195–1197.
[23] Fu Y, Nistor A I. Constant angle property and canonical principal directions for surfaces in $\mathbb{M}$2(c$\mathbb{R}$1. Mediterr J Math, 2013, 10: 1035–1049.
[24] Hou Z H, Qiu W H. Submanifolds with parallel mean curvature vector field in product spaces. Vietnam J Math, 2015, 43: 705–724.
[25] Hou Z H, Qiu W H. A classification theorem for complete PMC surfaces with non-negative Gaussian curvature in Mn×$\mathbb{R}$. Taiwanese J Math, 2016, 20: 205–226.
[26] de Lima H F, Lima Eraldo A. Generalized maximum principles and the unicity of complete spacelike hypersurfaces immersed in a Lorentzian product space. Beitr Algebra Geom, 2014, 55: 59–75.
[27] Su B P, Shu S C, Han Y A. Hypersurfaces with constant mean curvature in a hyperbolic space. Acta Math Sci, 2011, 31: 1091–1102.