\documentclass[reqno]{amsart} \usepackage{hyperref} \AtBeginDocument{{\noindent\small \emph{Electronic Journal of Differential Equations}, Vol. 2016 (2016), No. 85, pp. 1--9.\newline ISSN: 1072-6691. URL: http://ejde.math.txstate.edu or http://ejde.math.unt.edu \newline ftp ejde.math.txstate.edu} \thanks{\copyright 2016 Texas State University.} \vspace{9mm}} \begin{document} \title[\hfilneg EJDE-2016/85\hfil Wave map and related systems] {Regularity criteria for the wave map and related systems} \author[J. Fan, Y. Zhou \hfil EJDE-2016/85\hfilneg] {Jishan Fan, Yong Zhou} \address{Jishan Fan \newline Department of Applied Mathematics, Nanjing Forestry University, Nanjing 210037, China} \email{fanjishan@njfu.com.cn} \address{Yong Zhou (corresponding author) \newline School of Mathematics, Shanghai University of Finance and Economics, \newline Shanghai 200433, China. \newline Department of Mathematics, King Abdulaziz University, Jeddah 21589, Saudi Arabia} \email{yzhou@sufe.edu.cn} \thanks{Submitted February 18, 2016. Published March 29, 2016.} \subjclass[2010]{35K55, 35Q35, 70S15} \keywords{Regularity criterion; wave map; liquid crystals; Hall-MHD} \begin{abstract} We obtain some regularity criteria for the wave map, a liquid crystals model, and the Hall-MHD with ion-slip effect. \end{abstract} \maketitle \numberwithin{equation}{section} \newtheorem{theorem}{Theorem}[section] \allowdisplaybreaks \section{Introduction} First, we consider the $n$D wave maps $d:\mathbb{R}^{1+n}\to\mathbb{S}^m\subset\mathbb{R}^{1+m}$ which obey the nonlinear wave equation \begin{equation} \partial_t^2d-\Delta d=d(|\nabla d|^2-|\partial_td|^2)\label{1.1} \end{equation} with the initial conditions \begin{equation} (d,\partial_td)(\cdot,0)=(d_0,d_1),\; d_0\in\mathbb{S}^m,\quad d_0\cdot d_1=0.\label{1.2} \end{equation} Wave maps have wide applications in physics from the harmonic gauge in general relativity to the nonlinear $\sigma$-models in particle physics. Local well-posedness of \eqref{1.1} \eqref{1.2} has been proved by Tao \cite{1}. Shatah \cite{2} showed that solutions to the Cauchy problem for wave maps may blow up in finite time. However, some smallness assumption on the initial data or integrability condition on the solution itself are sufficient to guarantee the regularity. Fan and Ozawa \cite{3} obtained the regularity criterion \begin{equation} \nabla d,\partial_t d\in L^1(0,T;{\dot{B}^0_{\infty,\infty}}(\mathbb{R}^n))\label{1.3} \end{equation} when $n=2$. The first aim of this article is to prove a following regularity criterion when $n\geq3$. \begin{theorem} \label{thm1.1} Let $n\geq3$ and $(\nabla d_0,d_1)\in H^{1+s}(\mathbb{R}^n)$ with $s>\frac{n}{2}$, $|d_0|=1$, $d_0\cdot d_1=0$ and $d$ be a smooth solution of \eqref{1.1}, \eqref{1.2}. If \eqref{1.3} and $\partial_td\in L^\infty(0,T;L^n(\mathbb{R}^n))$ hold true with $00$. \end{theorem} Next, we consider the liquid crystals model \cite{4,5,6,7}: \begin{gather} \partial_tu+u\cdot\nabla u+\nabla\pi-\Delta u =-\nabla\cdot(\nabla d\odot\nabla d),\label{1.4}\\ \partial_td+u\cdot\nabla d-\Delta d=d|\nabla d|^2,\quad |d|=1,\label{1.5}\\ \operatorname{div} u=0,\label{1.6}\\ (u,d)(\cdot,0)=(u_0,d_0)\quad \text{in }\mathbb{R}^n,\; |d_0|=1.\label{1.7} \end{gather} Here $u$ is the velocity, $\pi$ is the pressure, $d$ is the direction vector, and $(\nabla d\odot\nabla d)_{i,j}:=\sum_k\partial_id_k\partial_jd_k$, and hence $$ \nabla\cdot(\nabla d\odot\nabla d) =\sum_k\Delta d_k\nabla d_k+\frac12\nabla|\nabla d|^2. $$ If $u=0$, then \eqref{1.5} is the harmonic heat flow. Fan-Gao-Guo \cite{8} proved the blow-up criterion \begin{equation} u,\nabla d\in L^2(0,T;{\dot{B}^0_{\infty,\infty}})\label{1.8} \end{equation} when $n=3$. One can find other related results in \cite{FAHNZ-AA, ZF} and references therein. We will prove the following theorem. \begin{theorem} \label{thm1.2} Let $n\geq3$ and $s>\frac{n}{2}$ be an integer. Let $u_0$ and $d_0$ satisfy $u_0,\nabla d_0\in H^s, \operatorname{div} u_0=0$, and $|d_0|=1$ in $\mathbb{R}^n$. Let $(u,d)$ be a local strong solution to the problem \eqref{1.4}-\eqref{1.7}. If $\nabla u$ and $\nabla^2d$ satisfy \begin{equation} \nabla u,\nabla^2d\in L^\frac{2}{2-\alpha}(0,T;\dot B_{\infty,\infty}^{-\alpha} (\mathbb{R}^n))\label{1.9} \end{equation} with $0<\alpha<1$ and $00$. \end{theorem} Also we consider the incompressible MHD with the Hall or ion-slip system \begin{gather} \partial_tu+u\cdot\nabla u+\nabla\Big(\pi+\frac12|b|^2\Big) -\Delta u=b\cdot\nabla b,\label{1.10}\\ \partial_tb+u\cdot\nabla b-b\cdot\nabla u+h\operatorname{rot}(\operatorname{rot} b\times b) -\gamma\operatorname{rot}[(\operatorname{rot} b\times b)\times b]=\Delta b,\label{1.11}\\ \operatorname{div} u=\operatorname{div} b=0,\label{1.12}\\ (u,b)(\cdot,0)=(u_0,b_0)\quad \text{in }\mathbb{R}^3.\label{1.13} \end{gather} Here $b$ is the magnetic field. $h$ is the Hall effect coefficient, and $\gamma\geq0$ the ion-slip effect coefficient, respectively. Applications of the Hall-MHD system cover a very wide range of physical subjects, such as, magnetic reconnection in space plasmas, star formation, neutron stars, and geo-dynamos. Very recently, Zhang \cite{9} obtained the regularity criterion \begin{equation} u\in L^\frac{2}{1-\alpha}(0,T;\dot B_{\infty,\infty}^{-\alpha}),\quad \nabla b\in L^\frac{2}{1-\beta}(0,T;\dot B_{\infty,\infty}^{-\beta})\label{1.14} \end{equation} with $-1<\alpha<1$ and $0<\beta<1$ when $h=1$ and $\gamma=0$. Local well-posedness of strong solutions to \eqref{1.10}-\eqref{1.13} has been proved by Fan, Jia, Nakamura and Zhou \cite{10}, they also obtained the regularity criterion \begin{equation} u\in L^\frac{2q}{q-3}(0,T;L^q),\quad b\in L^\infty(0,T;L^\infty),\quad \nabla b\in L^\frac{2p}{p-3}(0,T;L^p)\label{1.15} \end{equation} with $30$. \end{theorem} In the following proofs, we use the logarithmic Sobolev inequality \cite{11}: \begin{gather} \|\nabla d\|_{L^\infty}\leq C(1+\|\nabla d\|_{\dot{B}^0_{\infty,\infty}} \log(e+\|\nabla d\|_{H^{1+s}})),\label{1.16}\\ \|\partial_td\|_{L^\infty}\leq C(1+\|\partial_td\|_{\dot{B}^0_{\infty,\infty}} \log(e+\|\partial_td\|_{H^{1+s}}))\label{1.17} \end{gather} for $s>\frac{n}{2}-1$, and the bilinear product and commutator estimates due to Kato-Ponce \cite{12}: \begin{gather} \|\Lambda^s(fg)\|_{L^p}\leq C(\|\Lambda^sf\|_{L^{p_1}} \|g\|_{L^{q_1}}+\|f\|_{L^{p_2}}\|\Lambda^sg\|_{L^{q_2}}),\label{1.18}\\ \|\Lambda^s(fg)-f\Lambda^sg\|_{L^p}\leq C(\|\nabla f\|_{L^{p_1}} \|\Lambda^{s-1}g\|_{L^{q_1}}+\|\Lambda^sf\|_{L^{p_2}}\|g\|_{L^{q_2}}),\label{1.19} \end{gather} with $s>0$, $\Lambda:=(-\Delta)^\frac12$ and $\frac1p=\frac{1}{p_1}+\frac{1}{q_1}=\frac{1}{p_2}+\frac{1}{q_2}$. We also use the Gagliardo-Nirenberg inequalities \begin{gather} \|\nabla d\|_{L^{2p}}^2\leq C\|d\|_{L^\infty}\|\nabla^2d\|_{L^p},\label{1.20}\\ \|\nabla^2d\|_{L^p}\leq C\|\nabla d\|_{L^\infty}^{1-\theta} \|\Lambda^{2+s}d\|_{L^2}^\theta,\label{1.21}\\ \|\Lambda^{1+s}d\|_{L^\frac{2p}{p-2}}\leq C\|\nabla d\|_{L^\infty}^\theta \|\Lambda^{2+s}d\|_{L^2}^{1-\theta}\label{1.22} \end{gather} with $p:=2s+2$ and $\theta:=1/(1+s)$. We also use the improved Gagliardo-Nirenberg inequalities \cite{13,14,15}: \begin{gather} \|\nabla u\|_{L^{q_1}}\leq C\|\nabla u\|_{\dot B_{\infty,\infty}^{-\alpha} }^{1-\theta_1}\|u\|_{\dot H^{s+\alpha}}^{\theta_1},\label{1.23}\\ \|\Lambda^su\|_{L^\frac{2q_1}{q_1-2}} \leq C\|\nabla u\|_{\dot B_{\infty,\infty}^{-\alpha}}^{\theta_1} \|u\|_{\dot H^{s+\alpha}}^{1-\theta_1},\label{1.24} \end{gather} with $q_1:=\frac{2(s-1+2\alpha)}{\alpha}$ and $\theta_1:=2/q_1$, and \begin{gather} \|\nabla d\|_{L^{q_2}} \leq C\|\nabla d\|_{\dot B_{\infty,\infty}^{-\alpha}}^{1-\theta_2} \|\nabla d\|_{\dot H^{s+\alpha}}^{\theta_2},\label{1.25}\\ \|\Lambda^s\nabla d\|_{L^\frac{2q_2}{q_2-2}} \leq C\|\nabla d\|_{\dot B_{\infty,\infty}^{-\alpha}}^{\theta_2} \|\nabla d\|_{\dot H^{s+\alpha}}^{1-\theta_2},\label{1.26} \end{gather} with $q_2:=\frac{2(s+2\alpha)}{\alpha}$ and $\theta_2:=\frac{2}{q_2}$, \begin{equation} \|\nabla d\|_{\dot B_{\infty,\infty}^{-\alpha}} \leq C\|d\|_{L^\infty}^\frac{1}{2-\alpha} \|\nabla^2d\|_{\dot B_{\infty,\infty}^{-\alpha}}^\frac{1-\alpha}{2-\alpha},\label{1.27} \end{equation} and \begin{gather} \|D^ku\|_{L^{p_k}} \leq C\|\nabla u\|_{\dot B_{\infty,\infty}^{-\alpha}}^{1-\tilde\theta_k} \|u\|_{\dot H^{s+\alpha}}^{1-\tilde\theta_k}, \label{1.28}\\ \|D^{s+2-k}d\|_{L^\frac{2p_k}{p_k-2}} \leq C\|\nabla^2d\|_{\dot B_{\infty,\infty}^{-\alpha}}^{\tilde\theta_k} \|\nabla d\|_{\dot H^{s+\alpha}}^{1-\tilde\theta_k}, \label{1.29} \end{gather} with $p_k:=\frac{2}{\tilde\theta_k}$ and $\tilde\theta_k:=\frac{k+\alpha-1}{s+2\alpha-1}$, and \begin{gather} \|\nabla u\|_{L^3}^3\leq C\|u\|_{\dot B_{\infty,\infty}^{-\alpha}}\|u\|_{\dot H^\frac{3+\alpha}{2}}^2\ \ \mathrm{with}\ \ -1<\alpha<1,\label{1.30}\\ \|u\|_{\dot H^\frac{3+\alpha}{2}}\leq C\|\nabla u\|_{L^2}^\frac{1-\alpha}{2}\|\Delta u\|_{L^2}^\frac{1+\alpha}{2}\ \ \mathrm{with}\ \ -1<\alpha<1,\label{1.31} \end{gather} and \begin{gather} \|\nabla b\|_{L^4}^2\leq C\|\nabla b\|_{\dot B_{\infty,\infty}^{-\beta}} \|b\|_{\dot H^{1+\beta}}\quad \text{with } 0<\beta<1,\label{1.32}\\ \|b\|_{\dot H^{1+\beta}} \leq C\|\nabla b\|_{L^2}^{1-\beta}\|\Delta b\|_{L^2}^\beta\quad \text{with } 0<\beta<1.\label{1.33} \end{gather} \section{Proof of Theorem \ref{thm1.1}} Testing \eqref{1.1} by $\partial_td$ and using $|d|=1$ and $d\cdot\partial_td=0$, we easily get the conservation of the energy: \begin{equation} \frac{d}{dt}\int(|\partial_td|^2+|\nabla d|^2) dx=0.\label{2.1} \end{equation} Applying the operator $\Lambda^{1+s}$ to equation \eqref{1.1}, testing by $\Lambda^{1+s}\partial_td$, using \eqref{1.18}, \eqref{1.16}, \eqref{1.17}, \eqref{1.20}, \eqref{1.21} and \eqref{1.22}, we reach \begin{align*} &{\frac12\frac{d}{dt}\int}(|\Lambda^{1+s}\partial_td|^2+|\Lambda^{2+s}d|^2)dx\\ &= \int\Lambda^{1+s}(d|\nabla d|^2-d|\partial_td|^2)\Lambda^{1+s}\partial_t d dx\\ &\leq (\|\Lambda^{1+s}(d|\nabla d|^2)\|_{L^2} +\|\Lambda^{1+s}(d|\partial_td|^2)\|_{L^2})\|\Lambda^{1+s}\partial_td\|_{L^2}\\ &\leq C(\|d\|_{L^\infty}\|\Lambda^{1+s}(|\nabla d|^2)\|_{L^2} +\|\nabla d\|_{L^{2p}}^2\|\Lambda^{1+s}d\|_{L^\frac{2p}{p-2}})\|\Lambda^{1+s} \partial_td\|_{L^2}\\ &\quad +C(\|d\|_{L^\infty}\|\Lambda^{1+s}(|\partial_td|^2)\|_{L^2} +\|\partial_td\|_{L^{2n}}^2\|\Lambda^{1+s}d\|_{L^\frac{2n}{n-2}}) \|\Lambda^{1+s}\partial_td\|_{L^2}\\ &\leq C(\|\nabla d\|_{L^\infty}\|\Lambda^{2+s}d\|_{L^2} +\|\nabla^2d\|_{L^p}\|\Lambda^{1+s}d\|_{L^\frac{2p}{p-2}})\|\Lambda^{1+s} \partial_td\|_{L^2}\\ &\quad +C(\|\partial_td\|_{L^\infty}\|\Lambda^{1+s}\partial_td\|_{L^2} +\|\partial_td\|_{L^n} \|\partial_td\|_{L^\infty}\|\Lambda^{2+s}d\|_{L^2}) \|\Lambda^{1+s}\partial_td\|_{L^2}\\ &\leq C\|\nabla d\|_{L^\infty}\|\Lambda^{2+s}d\|_{L^2} \|\Lambda^{1+s}\partial_td\|_{L^2}\\ &\quad +C\|\partial_td\|_{L^\infty}\|\Lambda^{1+s}\partial_td\|_{L^2}^2 +C\|\partial_td\|_{L^\infty} \|\Lambda^{2+s}d\|_{L^2}\|\Lambda^{1+s} \partial_td\|_{L^2}\\ &\leq C(\|\nabla d\|_{L^\infty}+\|\partial_td\|_{L^\infty}) (\|\Lambda^{2+s}d\|_{L^2}^2 +\|\Lambda^{1+s}\partial_td\|_{L^2}^2)\\ &\leq C(1+\|\nabla d\|_{\dot{B}^0_{\infty,\infty}}+\|\partial_td\|_{\dot{B}^0_{\infty,\infty}})\log(e+y^2)y^2, \end{align*} with $y^2:=\|\Lambda^{1+s}\partial_td\|_{L^2}^2+\|\Lambda^{2+s}d\|_{L^2}^2$, which gives $$ \sup_{0\leq t\leq T}(\|\Lambda^{1+s}\partial_td\|_{L^2}^2 +\|\Lambda^{2+s}d\|_{L^2}^2)\leq C. $$ This completes the proof. \section{Proof of Theorem \ref{thm1.2}} Since it is easy to prove that there are $T_0>0$ and a unique strong solution $(u,\pi,d)$ to the problem \eqref{1.4}-\eqref{1.7} in $[0,T_0]$, we only need to prove a priori estimates. Testing \eqref{1.4} by $u$ and using \eqref{1.6}, we see that \begin{equation} {\frac12\frac{d}{dt}\int}|u|^2 dx+\int|\nabla u|^2 dx=-\int(u\cdot\nabla)d\cdot\Delta d dx.\label{3.1} \end{equation} Testing \eqref{1.5} by $-\Delta d$, using $d\Delta d=-|\nabla d|^2$ and $|d|=1$, we find that \begin{equation} \begin{aligned} {\frac12\frac{d}{dt}\int}|\nabla d|^2 dx+\int|\Delta d|^2 dx &= \int(u\cdot\nabla)d\cdot\Delta d dx+\int(d\Delta d)^2 dx\\ &\leq \int(u\cdot\nabla)d\cdot\Delta d dx+\int|\Delta d|^2 dx. \end{aligned} \label{3.2} \end{equation} Summing up \eqref{3.1} and \eqref{3.2}, we have \begin{equation} \int(|u|^2+|\nabla d|^2)dx\leq\int(|u_0|^2+|\nabla d_0|^2)dx.\label{3.3} \end{equation} Applying $D^s$ to \eqref{1.4}, testing by $D^su$, using \eqref{1.6}, \eqref{1.18}, \eqref{1.19}, \eqref{1.23}, \eqref{1.24}, \eqref{1.25}, \eqref{1.26} and \eqref{1.27}, we obtain \begin{equation} \begin{aligned} &{\frac12\frac{d}{dt}\int}|D^su|^2 dx+\int|D^{1+s}u|^2 dx\\ &= -\int(D^s(u\cdot\nabla u)-u\nabla D^su)D^s u dx +\int D^s(\nabla d\odot\nabla d):\nabla D^su dx\\ &\leq C\|\nabla u\|_{L^{q_1}}\|D^su\|_{L^\frac{2q_1}{q_1-2}}\|D^su\|_{L^2} +C\|\nabla d\|_{L^{q_2}}\|D^s\nabla d\|_{L^\frac{2q_2}{q_2-2}}\|\nabla D^su\|_{L^2}\\ &\leq C\|\nabla u\|_{\dot B_{\infty,\infty}^{-\alpha}}\|D^{s+\alpha}u\|_{L^2}\|D^su\|_{L^2} +C\|\nabla d\|_{\dot B_{\infty,\infty}^{-\alpha}}\|D^{s+\alpha}\nabla d\|_{L^2}\|D\Lambda^su\|_{L^2}\\ &\leq C\|\nabla u\|_{\dot B_{\infty,\infty}^{-\alpha}}\|D^su\|_{L^2}^{2-\alpha}\|D^{1+s}u\|_{L^2}^\alpha\\ &\quad +C\|\nabla^2d\|_{\dot B_{\infty,\infty}^{-\alpha}}^\frac{1-\alpha}{2-\alpha} \|D^{s+1}d\|_{L^2}^{1-\alpha}\|D^{s+2}d\|_{L^2}^\alpha\|D^{1+s}u\|_{L^2}\\ &\leq \frac18\|D^{1+s}u\|_{L^2}^2+\frac18\|D^{s+2}d\|_{L^2}^2 +C\|\nabla u\|_{\dot B_{\infty,\infty}^{-\alpha}}^\frac{2}{2-\alpha}\|D^su\|_{L^2}^2\\ &\quad+C\|\nabla^2d\|_{\dot B_{\infty,\infty}^{-\alpha}}^\frac{2}{2-\alpha}\|D^{s+1}d\|_{L^2}^2. \end{aligned} \label{3.4} \end{equation} Applying $D^{s+1}$ to \eqref{1.5}, testing by $D^{s+1}d$ and using \eqref{1.6}, we obtain \begin{equation} \begin{aligned} &{\frac12\frac{d}{dt}\int}|D^{s+1}d|^2 dx+\int|D^{s+2}d|^2 dx\\ &=\int D^{s+1}(d|\nabla d|^2)D^{s+1}d dx\\ &\quad -\int(D^{s+1}(u\cdot\nabla d)-u\nabla D^{s+1}d)D^{s+1}d \,dx =:I_1+I_2. \end{aligned}\label{3.5} \end{equation} Using \eqref{1.18}, $|d|=1$, \eqref{1.25}, \eqref{1.26}, and \eqref{1.27}, we bound $I_1$ as follows. \begin{align*} I_1&\leq \|D^{s+1}(d|\nabla d|^2)\|_{L^\frac{2q_2}{q_2+2}} \|D^{s+1}d\|_{L^\frac{2q_2}{q_2-2}}\\ &\leq C(\|d\|_{L^\infty}\|D^{s+1}(|\nabla d|^2)\|_{L^\frac{2q_2}{q_2+2}} +\|\nabla d\|_{L^{q_2}}^2\|D^{s+1}d\|_{L^\frac{2q_2}{q_2-2}}) \|D^{s+1}d\|_{L^\frac{2q_2}{q_2-2}}\\ &\leq C(\|\nabla d\|_{L^{q_2}}\|D^{s+2}d\|_{L^2} +\|\nabla d\|_{L^{q_2}}^2\|D^{s+1}d\|_{L^\frac{2q_2}{q_2-2}}) \|D^{s+1}d\|_{L^\frac{2q_2}{q_2-2}}\\ &\leq C\|\nabla d\|_{L^{q_2}}^2\|D^{s+1}d\|_{L^\frac{2q_2}{q_2-2}}^2 +\frac{1}{16}\|D^{s+2}d\|_{L^2}^2\\ &\leq C\|\nabla d\|_{\dot B_{\infty,\infty}^{-\alpha}}^2\|\nabla d\|_{\dot H^{s+\alpha}}^2 +\frac{1}{16}\|D^{s+2}d\|_{L^2}^2\\ &\leq \frac18\|D^{s+2}d\|_{L^2}^2+C\|\nabla^2d\|_{\dot B_{\infty,\infty}^{-\alpha}}^\frac{2}{2-\alpha} \|D^{s+1}d\|_{L^2}^2. \end{align*} %3.6 Using the Leibniz rule, we write $I_2$ as follows. \begin{equation} \begin{aligned} I_2 &= -\int(C_1DuD^{s+1}d+\sum_{k=2}^sC_kD^kuD^{s+2-k}d +C_{s+1}D^{s+1}u\cdot\nabla d)D^{s+1}d dx\\ &=:I_2^1+\sum_{k=2}^sI_2^k+I_2^{s+1}. \end{aligned} \label{3.7} \end{equation} By the same calculations as that of $I_1$, we have \begin{equation} \begin{aligned} I_2^{s+1} &\leq C\|\nabla d\|_{L^{q_2}}\|D^{s+1}d\|_{L^\frac{2q_2}{q_2-2}}\|D^{s+1}u\|_{L^2}\\ &\leq \frac{1}{16}\|D^{s+1}u\|_{L^2}^2 +C\|\nabla^2d\|_{\dot B_{\infty,\infty}^{-\alpha}}^\frac{2}{2-\alpha}\|D^{s+1}d\|_{L^2}^2. \end{aligned} \label{3.8} \end{equation} Using \eqref{1.23} and \eqref{1.24}, we bound $I_2^1$ as follows. \begin{equation} \begin{aligned} I_2^1 &\leq C\|\nabla u\|_{L^{q_1}}\|D^{s+1}d\|_{L^\frac{2q_1}{q_1-2}}\|D^{s+1}d\|_{L^2}\\ &\leq C\|\nabla u\|_{\dot B_{\infty,\infty}^{-\alpha}}^{1-\theta_1}\|u\|_{\dot H^{s+\alpha}}^{\theta_1} \cdot\|\nabla^2d\|_{\dot B_{\infty,\infty}^{-\alpha}}^{\theta_1}\|\nabla d\|_{\dot H^{s+\alpha}}^{1-\theta_1} \|D^{s+1}d\|_{L^2}\\ &\leq C(\|\nabla u\|_{\dot B_{\infty,\infty}^{-\alpha}}+\|\nabla^2d\|_{\dot B_{\infty,\infty}^{-\alpha}})(\|u\|_{\dot H^{s+\alpha}} +\|\nabla d\|_{\dot H^{s+\alpha}})\|D^{s+1}d\|_{L^2}\\ &\leq \frac{1}{16}\|D^{s+1}u\|_{L^2}^2+\frac{1}{16}\|D^{s+2}d\|_{L^2}^2\\ &\quad +C(\|\nabla u\|_{\dot B_{\infty,\infty}^{-\alpha}}^\frac{2}{2-\alpha} +\|\nabla^2d\|_{\dot B_{\infty,\infty}^{-\alpha}}^\frac{2}{2-\alpha})(\|D^su\|_{L^2}^2+\|D^{s+1}d\|_{L^2}^2). \end{aligned}\label{3.9} \end{equation} Using \eqref{1.28} and \eqref{1.29}, we bound $\sum_{k=2}^sI_2^k$ as follows. \begin{equation} \begin{aligned} \sum_{k=2}^sI_2^k &\leq C\|\nabla u\|_{\dot B_{\infty,\infty}^{-\alpha}}^{1-\tilde\theta_k} \|u\|_{\dot H^{s+\alpha}}^{\tilde\theta_k}\|\nabla^2d\|_{\dot B_{\infty,\infty}^{-\alpha}}^{\tilde\theta_k} \|\nabla d\|_{\dot H^{s+\alpha}}^{1-\tilde\theta_k}\|D^{s+1}d\|_{L^2}\\ &\leq C(\|\nabla u\|_{\dot B_{\infty,\infty}^{-\alpha}}+\|\nabla^2d\|_{\dot B_{\infty,\infty}^{-\alpha}})(\|u\|_{\dot H^{s+\alpha}} +\|\nabla d\|_{\dot H^{s+\alpha}})\|D^{s+1}d\|_{L^2}\\ &\leq C(\|\nabla u\|_{\dot B_{\infty,\infty}^{-\alpha}}+\|\nabla^2d\|_{\dot B_{\infty,\infty}^{-\alpha}})(\|u\|_{\dot H^{s+\alpha}} +\|\nabla d\|_{\dot H^{s+\alpha}})\\ &\quad\times (\|D^su\|_{L^2}+\|D^{s+1}d\|_{L^2})\\ &\leq \frac{1}{16}\|D^{s+1}u\|_{L^2}^2+\frac{1}{16}\|D^{s+2}d\|_{L^2}^2\\ &\quad +C(\|\nabla u\|_{\dot B_{\infty,\infty}^{-\alpha}}^\frac{2}{2-\alpha} +\|\nabla^2d\|_{\dot B_{\infty,\infty}^{-\alpha}}^\frac{2}{2-\alpha})(\|D^su\|_{L^2}^2 +\|D^{s+1}d\|_{L^2}^2). \end{aligned}\label{3.10} \end{equation} Inserting the above estimates into \eqref{3.5} and combining with \eqref{3.4} and using the Gronwall inequality, we arrive at $$ \|D^su\|_{L^\infty(0,T;L^2)}+\|D^{s+1}d\|_{L^\infty(0,T;L^2)}\leq C. $$ This completes the proof. \section{Proof of Theorem \ref{thm1.3}} We only need to show a priori estimates. For simplicity, we will take $h=\gamma=1$. First, testing \eqref{1.10} by $u$ and using \eqref{1.12}, we see that \begin{equation} {\frac12\frac{d}{dt}\int}|u|^2 dx+\int|\nabla u|^2 dx=\int(b\cdot\nabla)b\cdot u dx.\label{4.1} \end{equation} Testing \eqref{1.11} by $b$ and using \eqref{1.12}, we find that \begin{equation} {\frac12\frac{d}{dt}\int}|b|^2 dx+\int|\nabla b|^2 dx+\int|b\times\operatorname{rot} b|^2 dx =\int(b\cdot\nabla)u\cdot b dx.\label{4.2} \end{equation} Summing up \eqref{4.1} and \eqref{4.2}, we obtain \begin{equation} {\frac12\frac{d}{dt}\int}(|u|^2+|b|^2) dx+\int(|\nabla u|^2+|\nabla b|^2+|b\times\operatorname{rot} b|^2) dx=0.\label{4.3} \end{equation} Testing \eqref{1.10} by $-\Delta u$, using \eqref{1.12}, \eqref{1.30} and \eqref{1.31}, we infer that \begin{equation} \begin{aligned} &{\frac12\frac{d}{dt}\int}|\nabla u|^2 dx+\int|\Delta u|^2 dx\\ &= \int(u.\nabla)u\cdot\Delta u dx-\int(b\cdot\nabla)b\cdot\Delta u dx\\ &= -\sum_{i,j}\int\partial_ju_i\partial_iu\partial_ju dx -\int(b\cdot\nabla)b\cdot\Delta u dx\\ &\leq C\|\nabla u\|_{L^3}^3+\|b\|_{L^\infty}\|\nabla b\|_{L^2}\|\Delta u\|_{L^2}\\ &\leq C\|u\|_{\dot B_{\infty,\infty}^{-\alpha}}\|u\|_{\dot H^\frac{3+\alpha}{2}}^2 +C\|\nabla b\|_{L^2}\|\Delta u\|_{L^2}\\ &\leq C\|u\|_{\dot B_{\infty,\infty}^{-\alpha}}\|\nabla u\|_{L^2}^{1-\alpha}\|\Delta u\|_{L^2}^{1+\alpha} +C\|\nabla b\|_{L^2}\|\Delta u\|_{L^2}\\ &\leq \frac18\|\Delta u\|_{L^2}^2 +C\|u\|_{\dot B_{\infty,\infty}^{-\alpha}}^\frac{2}{1-\alpha}\|\nabla u\|_{L^2}^2+C\|\nabla b\|_{L^2}^2. \end{aligned}\label{4.4} \end{equation} Testing \eqref{1.11} by $-\Delta b$ and using \eqref{1.12}, we deduce that \begin{equation} \begin{aligned} &{\frac12\frac{d}{dt}\int}|\nabla b|^2 dx+\int|\Delta b|^2 dx\\ &=\int(u\cdot\nabla)b\cdot\Delta b dx-\int(b\cdot\nabla)u\cdot\Delta b dx\\ &\quad +\int(\operatorname{rot} b\times b)\operatorname{rot}\Delta b dx-\int[(\operatorname{rot} b\times b)\times b]\operatorname{rot}\Delta b dx\\ &=:\ell_1+\ell_2+\ell_3+\ell_4. \end{aligned}\label{4.5} \end{equation} We bound $\ell_1$ and $\ell_2$ as follows. \begin{align*} \ell_1&= \sum_{i,j}\int u_i\partial_ib\partial_j^2b dx=-\sum_{i,j}\int\partial_ju_i\partial_ib\partial_jb dx\leq C\|\nabla u\|_{L^2}\|\nabla b\|_{L^4}^2\\ &\leq C\|\nabla u\|_{L^2}\|b\|_{L^\infty}\|\Delta b\|_{L^2}\leq C\|\nabla u\|_{L^2}\|\Delta b\|_{L^2}\leq\frac{1}{16}\|\Delta b\|_{L^2}^2+C\|\nabla u\|_{L^2}^2.\\ \ell_2&\leq \|b\|_{L^\infty}\|\nabla u\|_{L^2}\|\Delta b\|_{L^2}\leq C\|\nabla u\|_{L^2}\|\Delta b\|_{L^2}\leq\frac{1}{16}\|\Delta b\|_{L^2}^2+C\|\nabla u\|_{L^2}^2. \end{align*} Using \eqref{1.32} and \eqref{1.33}, we bound $\ell_3$ and $\ell_4$ as follows. \begin{align*} \ell_3 &= -\sum_i\int(\operatorname{rot} b\times\partial_ib)\partial_i\operatorname{rot} b dx \leq C\|\nabla b\|_{L^4}^2\|\Delta b\|_{L^2}\\ &\leq C\|\nabla b\|_{\dot B_{\infty,\infty}^{-\beta}} \|b\|_{\dot H^{1+\beta}}\|\Delta b\|_{L^2} \leq C\|\nabla b\|_{\dot B_{\infty,\infty}^{-\beta}}\|\nabla b\|_{L^2}^{1-\beta} \|\Delta b\|_{L^2}^{1+\beta}\\ &\leq \frac{1}{16}\|\Delta b\|_{L^2}^2 +C\|\nabla b\|_{\dot B_{\infty,\infty}^{-\beta}}^\frac{2}{1-\beta} \|\nabla b\|_{L^2}^2. \end{align*} \begin{align*} \ell_4 &= \sum_i\int\partial_i[(\operatorname{rot} b\times b)\times b]\partial_i\operatorname{rot} b dx \leq\sum_i\int[(\operatorname{rot} b\times\partial_ib)\times b]\partial_i\operatorname{rot} b dx\\ &\quad +\sum_i\int[(\operatorname{rot} b\times b)\times\partial_ib]\partial_i\operatorname{rot} b dx \leq C\|b\|_{L^\infty}\|\nabla b\|_{L^4}^2\|\Delta b\|_{L^2}\\ &\leq \frac{1}{16}\|\Delta b\|_{L^2}^2 +C\|\nabla b\|_{\dot B_{\infty,\infty}^{-\beta}}^\frac{2}{1-\beta} \|\nabla b\|_{L^2}^2. \end{align*} Inserting the above estimates into \eqref{4.5}, and combining this with \eqref{4.4}, and using the Gronwall inequality, we conclude that \begin{equation} \|\nabla u\|_{L^\infty(0,T;L^2)}+\|\nabla b\|_{L^\infty(0,T;L^2)}\leq C.\label{4.6} \end{equation} This completes the proof by \eqref{1.15}. \subsection*{Acknowledgment} This work is partially supported by NSFC (No. 11171154). The authors would like to thank the referees for their careful reading and helpful suggestions. \begin{thebibliography}{99} \bibitem{CWW} D. Chae, R. Wan, J. Wu; \emph{Local well-posedness for the Hall-MHD equations with fractional magnetic diffusion}, J. Math. Fluid Mech., 17 (2015), 627-638. \bibitem{4} J. L. Ericksen; \emph{Conservation laws for liquid crystals}, Trans. Soc. Rheo., 5 (1961), 23-34. \bibitem{5} J. L. Ericksen; \emph{Continuum thoery of nematic liquid crystals}, Res Mech., 21 (1987), 381-392. \bibitem{6} J. L. Ericksen; \emph{Liquid crystals with varialbe degree of orientation}, Arch. Rational Mech. Anal., 113(2) (1990), 97-120. \bibitem{FAHZ-AMC} J. Fan, B. Ahmad, T. Hayat, Y. Zhou; \emph{On blow-up criteria for a new Hall-MHD system}, Appl. Math. Comput., 274 (2016), 20-24. \bibitem{FAHZ-NONRWA} J. Fan, B. Ahmad, T. Hayat, Y. Zhou; \emph{On well-posedness and blow-up for the full compressible Hall-MHD system}, to appear in Nonlinear Anal. Real World Appl., (2016), DOI: 10.1016/j.nonrwa.2016.03.003. \bibitem{FAHNZ} J. Fan, A. Alsaedi, T. Hayat, G. Nakamura, Y. Zhou; \emph{On strong solutions to the compressible Hall-magnetohydrodynamic system}, Nonlinear Anal. Real World Appl., 22 (2015), 423-434. \bibitem{FAHNZ-AA} J. Fan, F. Alzahrani, T. Hayat, G. Nakamura, Y. Zhou; \emph{Global regularity for the 2D liquid crystal model with mixed partial viscosity}, Anal. Appl. (Singap.), 13 (2015), 185-200. \bibitem{8} J. Fan, H. Gao, B. Guo; \emph{Regularity criteria for the Novier-Stokes-Landus-Lifshitz system}, J. Math. Anal. Appl., 363(1) (2010), 29-37. \bibitem{10} J. Fan, X. Jia, G. Nakamura, Y. Zhou; \emph{On well-posedness and blowup criteria for the magnetohydrodynamics with the Hall and ion-slip effects}, Z. Angew. Math. Phys. 66 (2015), 1695-1706. \bibitem{3} J. Fan, T. Ozawa; \emph{On regularity criterion for the 2D wave maps and the 4D biharmonic wave maps}, GAKUTO International Series, Math. Sci. Appl., 32 (2010), 69-83. \bibitem{13} H. Hajaiej, L. Molinet, T. Ozawa, B. Wang; \emph{Necessary and sufficient conditions for the fractional Gagliardo-Nirenberg inequalities and applications to Navier-Stokes and generalized Boson equations}, RIMS Kokyuroku Bessatsu 26 (2011), 159-175. \bibitem{HAHZ} F. He, B. Ahmad, T. Hayat, Y. Zhou; \emph{On regularity criteria for the 3D Hall-MHD equations in terms of the velocity}, submitted to Nonlinear Anal. Real World Appl. (2015), revised. \bibitem{12} T. Kato, G. Ponce; \emph{Commutator estimates and the Euler and Navier-Stokes equations}. Commun. Pure Appl. Math., 41 (1988), 891-907. \bibitem{11} H. Kozono, T. Ogawa, Y. Taniuchi; \emph{The critical Sobolev inequalities in Besov spaces and regularity criterion to some semilinear evolution equations}. Math. Z. 242 (2002), 251-278. \bibitem{7} F. Leslie; \emph{Theory of Flow Phenomenum in Liquid Crystals}, Springer, New York, NY, USA, 1979. \bibitem{14} S. Machihara, T. Ozawa; \emph{Interpolation inequalities in Besov spaces}, Proc. Am. Math. Soc., 131 (2002), 1553-1556. \bibitem{15} Y. Meyer; \emph{Oscillating patterns in some nonlinear evolution equations}, in: Mathematical Foundation of Turbulent Viscous Flows, Lecture Notes in Mathematics Vol. 1871, edited by M. Cannone and T. Miyakawa (Springer-Verlag, 2006), pp. 101-187. \bibitem{2} J. Shatah; \emph{Weak solutions and development of singularities in the SU(2) $\sigma$-model}, Commun. Pure Appl. Math., 41 (1988), 459-469. \bibitem{1} T. Tao; \emph{Nonlinear Dispersive Equations}, Local and Global Analysis, CBMS Reg. Conf. Ser. Math., vol. 106, AMS Providence RI, 2006. \bibitem{W} R. Wan; \emph{Global regularity for generalized Hall magneto-hydrodynamics systems}, Electron J. Differential Equations 2015 (2015), No. 179, 18 pp. \bibitem{WZ} R. Wan, Y. Zhou; \emph{On global existence, energy decay and blow-up criteria for the Hall-MHD system}, J. Differential Equations., 259 (2015), 5982-6008. \bibitem{9} Z. Zhang; \emph{A remark on the blow-up criterion for the 3D Hall-MHD system in Besov spaces}, preprint (2015). \bibitem{ZF} Y. Zhou, J. Fan; \emph{A regularity criterion for the nematic liquid crystal flows}, J. Inequal. Appl., 2010, Art. ID 589697, 9 pp. \end{thebibliography} \end{document}