betaFeedback

WorksMathematics

Large-time behavior of solutions to the 1D nematic liquid crystal with γ>1γ>1

Abstract

We study the large-time behavior of solutions to the system modeling the compressible flow of nematic liquid crystals with a polytropic equation of state P(ρ)=ργP(ρ)=ρ^γ (γ>1)(γ>1) in Lagrangian coordinates. In contrast to the isothermal case γ=1γ=1 treated in [4,5], we establish a new representation formula for the specific volume (Lemma 2.2) and, using Jensen's inequality, derive time-uniform estimates that yield global upper and lower bounds for the specific volume. A key ingredient is the reconstruction of the DD–ZZ structure for γ>1γ>1, which gives an exponential decay of Z(t)Z(t) and thus overcomes the difficulty caused by the power-type pressure.

Introduction

The nematic liquid crystal flow [1, 2, 6] is described by the following coupled system:

ρt+(ρu)x=0,(1)\rho_t+(\rho u)_x=0, \tag*{(1)}
(ρu)t+(ρu2)x+(P(ρ))x=uxx−(∣nx∣2)x,(2)(\rho u)_t+(\rho u^2)_x+(P(\rho))_x=u_{xx}-(|n_x|^2)_x, \tag*{(2)}
nt+unx=nxx+∣nx∣2n.(3)n_t+u n_x=n_{xx}+|n_x|^2n. \tag*{(3)}

Mathematically, the one-dimensional compressible nematic liquid crystal system exhibits a hybrid structure: the mass equation is hyperbolic, the momentum equation is parabolic-hyperbolic, and the orientation equation is a linear parabolic equation with an algebraic constraint ∣n∣=1|n|=1. Here ρ(x,t)\rho(x,t) denotes the density, uu and nn are unknown functions standing for the velocity field, macroscopic average of the orientation field, respectively.

We consider the initial-boundary data for system (1.1) in the domain {(x,t)∣0<x<1, t∈[0,+∞)}\{(x,t)\mid0<x<1,\ t\in[0,+\infty)\}

{u(0,t)=u(1,t)=0,nx(0,t)=nx(1,t)=0,(ρ,u,n)(⋅,0)=(ρ0,u0,n0),(4)\begin{cases} u(0,t)=u(1,t)=0,\quad n_x(0,t)=n_x(1,t)=0,\\ (\rho,u,n)(\cdot,0)=(\rho_0,u_0,n_0), \tag*{(4)} \end{cases}

and the polytropic gas law P(ρ)=ργP(\rho)=\rho^\gamma (γ>1\gamma>1). Under the Lagrangian coordinates, i.e., y=∫0xρ(ξ,t) dξy=\int_0^x\rho(\xi,t)\,\mathrm{d}\xi, system (1.1)–(4) transformed into the following:

vt=uy,(5)v_t=u_y, \tag*{(5)}
ut=(−P+uyv−∣ny∣2v2)y,(6)u_t=\left(-P+\frac{u_y}{v}-\frac{|n_y|^2}{v^2}\right)_y, \tag*{(6)}
nt=1v(nyv)y+∣ny∣2v2n,(7)n_t=\frac{1}{v}\left(\frac{n_y}{v}\right)_y+\frac{|n_y|^2}{v^2}n, \tag*{(7)}
{u(0,t)=u(1,t)=0,ny(0,t)=ny(1,t)=0,(v,u,n)(⋅,0)=(v0,u0,n0),(8)\begin{cases} u(0,t)=u(1,t)=0,\quad n_y(0,t)=n_y(1,t)=0,\\ (v,u,n)(\cdot,0)=(v_0,u_0,n_0), \tag*{(8)} \end{cases}

where v=1/ρv=1/\rho and v0=1/ρ0v_0=1/\rho_0.

In recent years, there have been numerous papers concerning the existence theory of solutions to the liquid crystal system, see References [7, 8, 9, 10]. For the one-dimensional case, Ding-Wang-Wen [3] established in 2011 the global existence of weak solutions in H1×H01×H2H^1\times H_0^1\times H^2 under the isothermal pressure law P(ρ)=ρP(\rho)=\rho. Qin and Huang [4] established the global high-regularity for system (1.1)–(4) and obtained an exact representation formula for the specific volume vv, a key tool for analyzing large-time behavior. Subsequently, Qin-Feng [5] derived time-uniform estimates and the large-time asymptotics in the isothermal case γ=1\gamma=1. However, the isothermal pressure law is physically restrictive, corresponding only to the isothermal process of an ideal gas in a heat bath. The more physically relevant pressure law is the polytropic gas law which describes the adiabatic process of real gases. The extension of Qin-Feng [5] to the polytropic regime (P(ρ)=ργP(\rho)=\rho^\gamma with γ>1\gamma>1) remains an open and non-trivial problem.

This paper extends these results to polytropic gases, (P(ρ)=ργP(\rho)=\rho^\gamma with γ>1\gamma>1). Our main contribution is to carry the large-time analysis over to γ>1\gamma>1: by combining the representation formula of Lemma 2.2 with energy estimates and Jensen’s inequality, we obtain time-uniform estimates that yield global upper and lower bounds for vv and the associated large-time control. The key ingredient is a γ\gamma-dependent reconstruction of the v,D,Zv,D,Z representation which yields exponential decay of Z(t)Z(t), thereby resolving the nonlinear difficulties caused by the power-type pressure.

Theorem 1.1. Suppose the initial data (v0,u0,n0)∈Hi(0,1)×H0i(0,1)×Hi+1(0,1)(v_0,u_0,n_0) \in H^i(0,1) \times H^i_0(0,1) \times H^{i+1}(0,1) (i=1,2)(i=1,2) satisfying the compatibility conditions. Then the system (1.3)–(8) possesses a unique global smooth solution (v(t),u(t),n(t))∈Hi(0,1)×H0i(0,1)×Hi+1(0,1)(v(t),u(t),n(t)) \in H^i(0,1) \times H^i_0(0,1) \times H^{i+1}(0,1), and there exists a constant C>0C>0, such that

0<C1−1≤v(y,t)≤C,∀(y,t)∈[0,1]×[0,+∞),(9)0<C_1^{-1}\leq v(y,t)\leq C,\quad\forall(y,t)\in[0,1]\times[0,+\infty), \tag*{(9)}

and

∥v(t)∥Hi2+∥u(t)∥Hi2+∥n(t)∥Hi+12+∫0∞(∥uy(t)∥Hi2+∥ny(t)∥Hi+12)(τ) dτ≤C.(10)\|v(t)\|_{H^i}^2+\|u(t)\|_{H^i}^2+\|n(t)\|_{H^{i+1}}^2+\int_0^\infty\left(\|u_y(t)\|_{H^i}^2+\|n_y(t)\|_{H^{i+1}}^2\right)(\tau)\,d\tau\leq C. \tag*{(10)}

Moreover, as t→+∞t\to+\infty, we have

∥v(t)−vˉ∥Hi→0,∥u(t)∥Hi→0,∥n(t)−nˉ∥Hi+1→0,(11)\|v(t)-\bar{v}\|_{H^i}\to0,\quad\|u(t)\|_{H^i}\to0,\quad\|n(t)-\bar{n}\|_{H^{i+1}}\to0, \tag*{(11)}

where vˉ=∫01v(y,t) dy=vˉ0,nˉ=lim⁡t→∞∫01n(y,t) dy\bar{v}=\int_0^1v(y,t)\,dy=\bar{v}_0,\bar{n}=\lim_{t\to\infty}\int_0^1n(y,t)\,dy.

Uniform estimates

Lemma 2.1. Under the assumptions in Theorem 1.1, the following estimates hold in Lagrangian coordinates, for any t>0t>0,

∫01[u2+v1−γγ−1+∣ny∣2v](y,t) dy+∫0t∫01(uy2v+1v∣n×(nyv)y∣2) dy ds≤C1,(12)\int_0^1\left[u^2+\frac{v^{1-\gamma}}{\gamma-1}+\frac{|n_y|^2}{v}\right](y,t)\,dy+\int_0^t\int_0^1\left(\frac{u_y^2}{v}+\frac{1}{v}\left|n\times\left(\frac{n_y}{v}\right)_y\right|^2\right)\,dy\,ds\leq C_1, \tag*{(12)}
∫01∣ny∣2v dy+∫0t∫011v[(nyv)y]2 dy ds≤C1.(13)\int_0^1\frac{|n_y|^2}{v}\,dy+\int_0^t\int_0^1\frac{1}{v}\left[\left(\frac{n_y}{v}\right)_y\right]^2\,dy\,ds\leq C_1. \tag*{(13)}

Lemma 2.2. Let

U(y,t)=∫0yu(ξ,t) dξ,vˉ0=∫01v0(y) dy,(14)U(y,t)=\int_0^y u(\xi,t)\,d\xi,\qquad\bar{v}_0=\int_0^1v_0(y)\,dy, \tag*{(14)}

and define

D(y,t)=v0(y)exp⁡{U(y,t)−U0(y)−1vˉ0[∫01vU dy−∫01v0U0 dy]},(15)D(y,t)=v_0(y)\exp\left\{U(y,t)-U_0(y)-\frac{1}{\bar{v}_0}\left[\int_0^1vU\,dy-\int_0^1v_0U_0\,dy\right]\right\}, \tag*{(15)}
Z(t)=exp⁡{−1vˉ0∫0t∫01(u2+v1−γ+∣ny∣2v) dy ds}.(16)Z(t)=\exp\left\{-\frac{1}{\bar{v}_0}\int_0^t\int_0^1\left(u^2+v^{1-\gamma}+\frac{|n_y|^2}{v}\right)\,dy\,ds\right\}. \tag*{(16)}

Then for any t≥0t\geq0, the solution v(y,t)v(y,t) to problems (1.3)–(8) possesses the following expression:

vγ(y,t)=Dγ(y,t)Zγ(t)exp⁡{γ∫0t∣ny(y,s)∣2v2(y,s) ds}×[1+γ∫0tD−γ(y,s)Z−γ(s)exp⁡{−γ∫0s∣ny(y,τ)∣2v2(y,τ) dτ} ds],(17)v^\gamma(y,t)=D^\gamma(y,t)Z^\gamma(t)\exp\left\{\gamma\int_0^t\frac{|n_y(y,s)|^2}{v^2(y,s)}\,ds\right\}\times\left[1+\gamma\int_0^tD^{-\gamma}(y,s)Z^{-\gamma}(s)\exp\left\{-\gamma\int_0^s\frac{|n_y(y,\tau)|^2}{v^2(y,\tau)}\,d\tau\right\}\,ds\right], \tag*{(17)}

and

C−1≤D(y,t)≤C.(18)C^{-1}\leq D(y,t)\leq C. \tag*{(18)}

Proof. Since u(0,t)=u(1,t)=0u(0,t)=u(1,t)=0, integrate equation (5) over yy, there holds

∫01v(y,t) dy=vˉ0.(19)\int_0^1v(y,t)\,dy=\bar{v}_0. \tag*{(19)}

(1) We first prove that DD is uniformly bounded. By the Cauchy-Schwarz inequality and Lemma 2.1, we get

∣∫01vU dy∣≤∥U∥L∞∫01v dy≤∥u(t)∥L2∫01v dy≤Cvˉ0,\left|\int_0^1vU\,dy\right|\leq\|U\|_{L^\infty}\int_0^1v\,dy\leq\|u(t)\|_{L^2}\int_0^1v\,dy\leq C\bar{v}_0,

similarly,

∣∫01v0U0 dy∣≤C.(20)\left|\int_0^1v_0U_0\,dy\right|\leq C. \tag*{(20)}

Thus there exists a constant CD>1C_D>1, such that

CD−1≤D(y,t)≤CD.(21)C_D^{-1}\leq D(y,t)\leq C_D. \tag*{(21)}

(2) We next prove the exponential decay rate of Z(t)Z(t). Define

G(t)=∫01(u2+v1−γ+∣ny∣2v) dy,(22)G(t)=\int_0^1\left(u^2+v^{1-\gamma}+\frac{|n_y|^2}{v}\right)\,dy, \tag*{(22)}

by Lemma 2.1, we have

G(t)≤C,andZ(t)≥e−CZt.(23)G(t)\leq C,\quad\text{and}\quad Z(t)\geq e^{-C_Zt}. \tag*{(23)}

Note that G(t)≥∫01v1−γ dyG(t) \ge\int_0^1 v^{1-\gamma}\,dy. Since γ>1\gamma> 1, the function f(v)=v1−γf(v)=v^{1-\gamma} satisfies f′′(v)=γ(γ−1)v−γ−1>0f^{\prime\prime}(v)=\gamma(\gamma-1)v^{-\gamma-1}>0, ff is convex. Using Jensen’s inequality and mass conservation equation (19), we obtain

G(t)≥∫01v1−γ dy≥(∫01v dy)1−γ≥vˉ01−γ=:c0>0.(24)G(t) \ge\int_0^1 v^{1-\gamma}\,dy \ge\left(\int_0^1 v\,dy\right)^{1-\gamma} \ge\bar{v}_0^{1-\gamma}=:c_0>0. \tag*{(24)}

According to (16), for any 0≤s≤t0\le s\le t, it holds

e−CZ(t−s)≤Z(t)Z(s)=exp⁡[−1vˉ0∫stG(τ) dτ]≤e−cZ(t−s),CZ=Cvˉ0, cZ=c0vˉ0.(25)e^{-C_Z(t-s)}\le\frac{Z(t)}{Z(s)}=\exp\left[-\frac{1}{\bar{v}_0}\int_s^t G(\tau)\,d\tau\right]\le e^{-c_Z(t-s)},\qquad C_Z=\frac{C}{\bar{v}_0},\ c_Z=\frac{c_0}{\bar{v}_0}. \tag*{(25)}

(3) Derive the representation formula for vv. Let

A=uyv−v−γ−∣ny∣2v2,(26)A=\frac{u_y}{v}-v^{-\gamma}-\frac{|n_y|^2}{v^2}, \tag*{(26)}

then the momentum equation (6) reads

ut=Ay.(27)u_t=A_y. \tag*{(27)}

According to (14) and (27), we have

Ut=∫0yut(ξ,t) dξ=A(y,t)−A(0,t).(28)U_t=\int_0^y u_t(\xi,t)\,d\xi=A(y,t)-A(0,t). \tag*{(28)}

On the one hand, using equation (5) and (26), we get

(ln⁡v)t=vtv=A+v−γ+∣ny∣2v2.(29)(\ln v)_t=\frac{v_t}{v}=A+v^{-\gamma}+\frac{|n_y|^2}{v^2}. \tag*{(29)}

Multiplying (28) by vv, and integrate over [0,1][0,1],

∫01vA dy=A(0,t)vˉ0+∫01vUt dy.(30)\int_0^1 vA\,dy=A(0,t)\bar{v}_0+\int_0^1 vU_t\,dy. \tag*{(30)}

For the left-hand side of (30), by equation (26), we have

∫01vA dy=∫01(uy−v1−γ−∣ny∣2v) dy=−∫01v1−γ dy−∫01∣ny∣2v dy.(31)\int_0^1 vA\,dy=\int_0^1\left(u_y-v^{1-\gamma}-\frac{|n_y|^2}{v}\right)\,dy=-\int_0^1 v^{1-\gamma}\,dy-\int_0^1\frac{|n_y|^2}{v}\,dy. \tag*{(31)}

Next we consider the right-hand side of (30), since

ddt∫01vU dy=∫01vtU dy+∫01vUt dy=∫01uyU dy+∫01vUt dy,\frac{d}{dt}\int_0^1 vU\,dy=\int_0^1 v_tU\,dy+\int_0^1 vU_t\,dy=\int_0^1 u_yU\,dy+\int_0^1 vU_t\,dy,

using integrate by parts and Uy=uU_y=u, we obtain ∫01uyU dy=−∫01u2 dy\int_0^1 u_yU\,dy=-\int_0^1 u^2\,dy, then

∫01vUt dy=ddt∫01vU dy+∫01u2 dy.(32)\int_0^1 vU_t\,dy=\frac{d}{dt}\int_0^1 vU\,dy+\int_0^1 u^2\,dy. \tag*{(32)}

Substitute (30)–(32):

A(0,t)vˉ0+ddt∫01vU dy+∫01u2 dy=−∫01v1−γ dy−∫01∣ny∣2v dy,A(0,t)\bar{v}_0+\frac{d}{dt}\int_0^1 vU\,dy+\int_0^1 u^2\,dy=-\int_0^1 v^{1-\gamma}\,dy-\int_0^1\frac{|n_y|^2}{v}\,dy,

therefore

A(0,t)=−1vˉ0ddt∫01vU dy−1vˉ0G(t).(33)A(0,t)=-\frac{1}{\bar{v}_0}\frac{d}{dt}\int_0^1 vU\,dy-\frac{1}{\bar{v}_0}G(t). \tag*{(33)}

By the definition of DD, ZZ, and equation (28), it follows that

DtD=Ut−1vˉ0ddt∫01vU dy=A−A(0,t)−1vˉ0ddt∫01vU dy,ZtZ=−G(t)vˉ0,(34)\frac{D_t}{D}=U_t-\frac{1}{\bar{v}_0}\frac{d}{dt}\int_0^1 vU\,dy=A-A(0,t)-\frac{1}{\bar{v}_0}\frac{d}{dt}\int_0^1 vU\,dy,\qquad\frac{Z_t}{Z}=-\frac{G(t)}{\bar{v}_0}, \tag*{(34)}

and added equation (29), using (33), so we get

ddtln⁡vDZ=A+v−γ+∣ny∣2v2−[A−A(0,t)−1vˉ0ddt∫vU]+Gvˉ0=v−γ+∣ny∣2v2.(35)\frac{d}{dt}\ln\frac{v}{DZ}=A+v^{-\gamma}+\frac{|n_y|^2}{v^2}-\left[A-A(0,t)-\frac{1}{\bar{v}_0}\frac{d}{dt}\int vU\right]+\frac{G}{\bar{v}_0}=v^{-\gamma}+\frac{|n_y|^2}{v^2}. \tag*{(35)}

On the other hand, set

W(y,t)=v(y,t)D(y,t)Z(t),W(y,t)=\frac{v(y,t)}{D(y,t)Z(t)},

then

Wt=(v−γ+∣ny∣2v2)W.(36)W_t=\left(v^{-\gamma}+\frac{|n_y|^2}{v^2}\right)W. \tag*{(36)}

and we have W(y,0)=1W(y,0)=1. Let X=WγX=W^\gamma. By using (36), since v−γX=(DZ)−γv^{-\gamma}X=(DZ)^{-\gamma}, therefore

Xt=γ(DZ)−γ+γ∣ny∣2v2X,(37)X_t=\gamma(DZ)^{-\gamma}+\gamma\frac{|n_y|^2}{v^2}X, \tag*{(37)}

this is a linear ODE of XX! Therefore

Wγ(y,t)=exp⁡{γ∫0t∣ny(y,s)∣2v(y,s)2 ds}×[1+γ∫0t(D(y,s)Z(s))−γexp⁡{−γ∫0s∣ny(y,τ)∣2v(y,τ)2 dτ} ds].(38)W^\gamma(y,t)=\exp\left\{\gamma\int_0^t\frac{|n_y(y,s)|^2}{v(y,s)^2}\,ds\right\}\times\left[1+\gamma\int_0^t(D(y,s)Z(s))^{-\gamma}\exp\left\{-\gamma\int_0^s\frac{|n_y(y,\tau)|^2}{v(y,\tau)^2}\,d\tau\right\}\,ds\right]. \tag*{(38)}

This yields the Lemma 2.2:

vγ(y,t)=Dγ(y,t)Zγ(t)eγ∫0t∣ny∣2v2 ds×[1+γ∫0tD−γ(y,s)Z−γ(s)e−γ∫0s∣ny∣2v2 dτ ds].v^\gamma(y,t)=D^\gamma(y,t)Z^\gamma(t)e^{\gamma\int_0^t\frac{|n_y|^2}{v^2}\,ds}\times\left[1+\gamma\int_0^tD^{-\gamma}(y,s)Z^{-\gamma}(s)e^{-\gamma\int_0^s\frac{|n_y|^2}{v^2}\,d\tau}\,ds\right].

Lemma 2.3. There holds

0<C1−1≤v(y,t)≤C1,∀ (y,t)∈[0,1]×[0,+∞),(39)0<C_1^{-1}\le v(y,t)\le C_1,\quad\forall\,(y,t)\in[0,1]\times[0,+\infty), \tag*{(39)}
∫0t∥ny(s)∥L∞2 ds≤C1,∀ t>0.(40)\int_0^t\|n_y(s)\|_{L^\infty}^2\,ds\le C_1,\quad\forall\,t>0. \tag*{(40)}

Proof. (1) First note that eγ∫0t∣ny∣2v2 ds≥1e^{\gamma\int_0^t\frac{|n_y|^2}{v^2}\,ds}\ge1. Thus by (38), we have Wγ(y,t)≥γ∫0t(DZ)−γ(y,s) dsW^\gamma(y,t)\ge\gamma\int_0^t(DZ)^{-\gamma}(y,s)\,ds, hence

vγ(y,t)≥γDγ(y,t)Zγ(t)∫0tD−γ(y,s)Z−γ(s) ds.v^\gamma(y,t)\ge\gamma D^\gamma(y,t)Z^\gamma(t)\int_0^tD^{-\gamma}(y,s)Z^{-\gamma}(s)\,ds.

Take the last unit time interval. If t≥1t\ge1, using (2.10) and (2.14), we obtain

(D(t)D(s))γ≥CD−2γ,vγ(y,t)≥γ∫t−1t(D(y,t)D(y,s))γ(Z(t)Z(s))γ ds≥γCD−2γe−γCZ,\left(\frac{D(t)}{D(s)}\right)^\gamma\ge C_D^{-2\gamma},\qquad v^\gamma(y,t)\ge\gamma\int_{t-1}^t\left(\frac{D(y,t)}{D(y,s)}\right)^\gamma\left(\frac{Z(t)}{Z(s)}\right)^\gamma\,ds\ge\gamma C_D^{-2\gamma}e^{-\gamma C_Z},

thus

v(y,t)≥c∗>0,t≥1.(41)v(y,t)\ge c_*>0,\qquad t\ge1. \tag*{(41)}

For 0≤t≤10\le t\le1, using (36), we have W(y,t)≥1W(y,t)\ge1. Thus

v(y,t)=D(y,t)Z(t)W(y,t)≥CD−1e−CZt≥CD−1e−CZ.v(y,t)=D(y,t)Z(t)W(y,t)\ge C_D^{-1}e^{-C_Zt}\ge C_D^{-1}e^{-C_Z}.

Combining the above estimates, we obtain

v(y,t)≥C1−1,(y,t)∈[0,1]×[0,∞).(42)v(y,t)\ge C_1^{-1},\qquad(y,t)\in[0,1]\times[0,\infty). \tag*{(42)}

(2) Let f=nyvf=\frac{n_y}{v}. By the boundary conditions (8), we have f(0,t)=f(1,t)=0f(0,t)=f(1,t)=0. According to the Sobolev inequality yields

∥f(t)∥L∞2≤C(∥f(t)∥L22+∥fy(t)∥L22)=C(∫01∣ny∣2v2 dy+∫01∣(nyv)y∣2 dy).(43)\|f(t)\|_{L^\infty}^2\le C\left(\|f(t)\|_{L^2}^2+\|f_y(t)\|_{L^2}^2\right)=C\left(\int_0^1\frac{|n_y|^2}{v^2}\,dy+\int_0^1\left|\left(\frac{n_y}{v}\right)_y\right|^2\,dy\right). \tag*{(43)}

Since v≥C1−1v\ge C_1^{-1}, and by Lemma 2.1, we have

∫0t∥nyv∥L∞2 ds≤C1∀ t>0.(44)\int_0^t\left\|\frac{n_y}{v}\right\|_{L^\infty}^2\,ds\le C_1\qquad\forall\,t>0. \tag*{(44)}

(3) Using (44) and (38), we obtain

vγ(y,t)≤Dγ(y,t)Zγ(t)eγC1+γDγ(y,t)Zγ(t)∫0tD−γ(y,s)Z−γ(s)eγC1 ds.v^\gamma(y,t)\le D^\gamma(y,t)Z^\gamma(t)e^{\gamma C_1}+\gamma D^\gamma(y,t)Z^\gamma(t)\int_0^tD^{-\gamma}(y,s)Z^{-\gamma}(s)e^{\gamma C_1}\,ds.

Using the upper and lower bounds of DD, and the upper bound in (25), then we lead to

vγ(y,t)≤C+C∫0te−γcZ(t−s) ds.v^\gamma(y,t)\le C+C\int_0^t e^{-\gamma c_Z(t-s)}\,ds.

Note that ∫0te−γcZ(t−s) ds≤1γcZ\int_0^t e^{-\gamma c_Z(t-s)}\,ds\le\frac{1}{\gamma c_Z}, therefore vγ(y,t)≤Cv^\gamma(y,t)\le C, namely

v(y,t)≤C.(45)v(y,t)\le C. \tag*{(45)}

Combined with (2.31), (2.33), the proof is complete.

Higher-order estimates and long-time behaviors

In this section, based on the previous basic energy estimates and uniform bounds of vv we further establish the uniform higher-order estimates of the solution and prove its long-time asymptotic behavior. Lemma 2.1-2.3 serve as the foundation for the subsequent higher-order estimates.

Lemma 3.1. Under the assumptions of Theorem 1.1. Let (v,u,n)(v,u,n) be a smooth solution to the system (1.3)-(8), there exists a constant C>0C>0, such that

sup⁡t≥0(∥v(t)∥H12+∥u(t)∥H12+∥n(t)∥H22)+∫0∞(∥uy(t)∥H12+∥ny(t)∥H22) dt≤C.(46)\sup_{t\geq0}\left(\|v(t)\|_{H^1}^2+\|u(t)\|_{H^1}^2+\|n(t)\|_{H^2}^2\right)+\int_0^\infty\left(\|u_y(t)\|_{H^1}^2+\|n_y(t)\|_{H^2}^2\right)\,\mathrm{d}t\leq C. \tag*{(46)}

Proof. Using (1.3)a, we rewrite (1.3)b as

(u−vyv)t=γvyvγ+1+2∣ny∣2vyv3−2ny⋅nyyv2,(47)\left(u-\frac{v_y}{v}\right)_t=\frac{\gamma v_y}{v^{\gamma+1}}+\frac{2|n_y|^2v_y}{v^3}-\frac{2n_y\cdot n_{yy}}{v^2}, \tag*{(47)}

multiplying the equation (47) by (u−vyv)\left(u-\frac{v_y}{v}\right), integrate over (0,1)×[0,t](0,1)\times[0,t], and using Lemma 2.1-2.3, we deduce

∫01∣u−vyv∣2(t) dy+∫0t∫01(vy2+∣ny∣2vy2v4+∣nyy∣2v3) dyds≤C1.(48)\int_0^1\left|u-\frac{v_y}{v}\right|^2(t)\,\mathrm{d}y+\int_0^t\int_0^1\left(v_y^2+\frac{|n_y|^2v_y^2}{v^4}+\frac{|n_{yy}|^2}{v^3}\right)\,\mathrm{d}y\mathrm{d}s\leq C_1. \tag*{(48)}

Multiplying the equation (1.3)b by uyyu_{yy}, integrate over (0,1)(0,1) about yy, we deduce

12ddt∫01∣uy∣2 dy+∫01∣uyy∣2v dy=−γ∫01vyuyyvγ+1 dy+∫01uyvyuyyv2 dy+2∫01(ny⋅nyy)uyyv2 dy−2∫01∣ny∣2vyuyyv3 dy,(49)\frac{1}{2}\frac{\mathrm{d}}{\mathrm{d}t}\int_0^1|u_y|^2\,\mathrm{d}y+\int_0^1\frac{|u_{yy}|^2}{v}\,\mathrm{d}y=-\gamma\int_0^1\frac{v_yu_{yy}}{v^{\gamma+1}}\,\mathrm{d}y+\int_0^1\frac{u_yv_yu_{yy}}{v^2}\,\mathrm{d}y+2\int_0^1\frac{(n_y\cdot n_{yy})u_{yy}}{v^2}\,\mathrm{d}y-2\int_0^1\frac{|n_y|^2v_yu_{yy}}{v^3}\,\mathrm{d}y, \tag*{(49)}

according to ∥u∥L∞≤C(∥u∥L2+∥uy∥L2)\|u\|_{L^\infty}\leq C(\|u\|_{L^2}+\|u_y\|_{L^2}) and a32b≤εa2+Cεb4a^{\frac{3}{2}}b\leq\varepsilon a^2+C_\varepsilon b^4, integrate (49) over (0,t)(0,t) about ss, we obtain

∫01∣uy∣2 dy+∫0t∫01∣uyy∣2v dyds≤C.(50)\int_0^1|u_y|^2\,\mathrm{d}y+\int_0^t\int_0^1\frac{|u_{yy}|^2}{v}\,\mathrm{d}y\mathrm{d}s\leq C. \tag*{(50)}

Multiplying the equation (1.3)c by v2v^2, taking ∥⋅∥L2\|\cdot\|_{L^2}, we get

∥nyy∥L22≤C1(∥nt∥L22+1).(51)\|n_{yy}\|_{L^2}^2\leq C_1\left(\|n_t\|_{L^2}^2+1\right). \tag*{(51)}

Take the tt derivative of the equation (1.3)c,

ntt=1v(ntyv)y−uyv2(nyv)y+2(ny⋅nty)v2n−2∣ny∣2uyv3n+∣ny∣2v2nt,(52)n_{tt}=\frac{1}{v}\left(\frac{n_{ty}}{v}\right)_y-\frac{u_y}{v^2}\left(\frac{n_y}{v}\right)_y+\frac{2(n_y\cdot n_{ty})}{v^2}n-\frac{2|n_y|^2u_y}{v^3}n+\frac{|n_y|^2}{v^2}n_t, \tag*{(52)}

multiply the equations by ntn_t, we have

∫01∣nt(t)∣2 dy+∫0t∫01∣nty(t)∣2 dyds≤C1.(53)\int_0^1|n_t(t)|^2\,\mathrm{d}y+\int_0^t\int_0^1|n_{ty}(t)|^2\,\mathrm{d}y\mathrm{d}s\leq C_1. \tag*{(53)}

Take the yy derivative of the equation (1.3)c,

nyyy=2vvynt+v2nty+nyyvy+nyuyyv−nyvy2v2−2(ny⋅nyy)n−∣ny∣2ny,(54)n_{yyy}=2vv_yn_t+v^2n_{ty}+\frac{n_{yy}v_y+n_yu_{yy}}{v}-\frac{n_yv_y^2}{v^2}-2(n_y\cdot n_{yy})n-|n_y|^2n_y, \tag*{(54)}

taking ∥⋅∥L2\|\cdot\|_{L^2}, and using the above inequalities, we get

∫0t∥nyyy∥L22 ds≤C1.(55)\int_0^t\|n_{yyy}\|_{L^2}^2\,\mathrm{d}s\leq C_1. \tag*{(55)}

This completes the proof of the lemma.

:::

Lemma 3.2. Under the assumptions of Theorem 1.1. Let (v,u,n)(v,u,n) be a smooth solution to the system (1.3)-(8), there exists a constant C>0C>0, such that

sup⁡t≥0(∥v(t)∥H22+∥u(t)∥H22+∥n(t)∥H32)+∫0∞(∥uy(t)∥H22+∥ny(t)∥H32) dt≤C,(56)\sup_{t\geq0}\left(\|v(t)\|_{H^2}^2+\|u(t)\|_{H^2}^2+\|n(t)\|_{H^3}^2\right)+\int_0^\infty\left(\|u_y(t)\|_{H^2}^2+\|n_y(t)\|_{H^3}^2\right)\,\mathrm{d}t\leq C, \tag*{(56)}

and

lim⁡t→∞∥v(t)−vˉ∥Hi=0,lim⁡t→∞∥u(t)∥Hi=0,lim⁡t→∞∥n(t)−nˉ∥Hi+1=0,i=1,2.(57)\lim_{t\to\infty}\|v(t)-\bar v\|_{H^i}=0,\qquad\lim_{t\to\infty}\|u(t)\|_{H^i}=0,\qquad\lim_{t\to\infty}\|n(t)-\bar n\|_{H^{i+1}}=0,\qquad i=1,2. \tag*{(57)}

Proof. By virtue of the boundary conditions, using Lemmas 8-11 in [5] and Lemmas 12-21 in [5], we obtain the assertion. This completes the proof of the lemma.

References

  1. [1]F. M. Leslie, Some constitutive equations for liquid crystals, Arch. Ration. Mech. Anal. 28 (1968) 265-283.
  2. [2]F. M. Leslie, Theory of flow orientation of nematic liquid crystals, Adv. Math. 6 (1971) 339-417.
  3. [3]S. J. Ding, J. Y. Wang, H. Y. Wen, Weak solutions to the compressible Ericksen-Leslie system in one dimension, J. Differential Equations 251 (2011) 3014-3051.
  4. [4]Y. M. Qin, L. Huang, Global existence of smooth solutions to a 1D compressible liquid crystal system, Math. Methods Appl. Sci. 37 (2014) 2543-2565.
  5. [5]Y. M. Qin, B. W. Feng, Large-time behavior of solutions to a 1D liquid crystal system, Math. Methods Appl. Sci. 40 (2017) 7077-7103.DOI
  6. [6]F. H. Lin, C. Liu, Nonparabolic dissipative systems, modelling of nematic liquid crystals, Comm. Pure Appl. Math. 48 (1995) 501-537.DOI
  7. [7]F. H. Lin, C. Liu, Existence of solutions for the Ericksen-Leslie system, Arch. Ration. Mech. Anal. 154 (2000) 135-156.DOI
  8. [8]J. X. Li, Z. P. Xin, Global existence of weak solutions to the compressible Ericksen-Leslie system in 1D, SIAM J. Math. Anal. 47 (2015) 2952-2993.
  9. [9]X. G. Li, D. H. Wang, Global existence of weak solutions to the compressible Ericksen-Leslie system in 2D, J. Differential Equations 263 (2017) 4565-4583.
  10. [10]Y. L. Qin, X. D. Huang, Global well-posedness of 1D compressible nematic liquid crystal system with polytropic pressure, Z. Angew. Math. Mech. 94 (2014) 838-847.

Paper details

vlnc:W_x4y3pmjb
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

Not recorded for this paper

View statement
Looking for related papers…
Contents