Uniqueness results for special Lagrangians and Lagrangian mean curvature flow expanders in \(\mathbb{C}^m\) (Q283529): Difference between revisions

A closed nonsingular Lagrangian \(m\)-fold \(L\) in \(\mathbb C^m\) is called asymptotically conical (AC) with the rate \(\rho<2\) and cone \(C\) if there exist a compact subset \(K\subset L\) and a diffeomorphism \(\phi:\Sigma\times(T,\infty)\to L\setminus K\) for some \(T>0\) such that \[ | \nabla^k(\phi-\iota)| =O(r^{\rho-1-k}), \;r\to \infty, \;k=0,1,2,\ldots. \] (Definition 2.9). Here \(\Sigma=C\cap\mathcal{S}^{2m-1}\), and \(\iota:\Sigma\times(0,\infty))\to \mathbb C^m\) is defined by \(\iota(\sigma,r)=r\sigma\). Alternatively, \(L\) is called a Lagrangian mean curvature flow (LMCF) expander if the mean curvature \(H\) of \(L\) satisfies \(H=\alpha F^\perp\), \(F:L\to \mathbb C^m\) is the inclusion, and \(\alpha>0\) (Definition 3.10). Let \(\Pi_1\), \(\Pi_2\) be transverse special Lagrangian planes in \(\mathbb C^m\). Then \(\Pi_1\cup \Pi_2\) is a Lagrangian cone in \(\mathbb C^m\). The main theorems of this paper are: Theorem 1.1. Let \(L\) be a closed embedded exact AC special Lagrangian in \(\mathbb C^m\), \(m\geq 3\), asymptotic at rate \(\rho<0\) to \(\Pi_1\cap\Pi_2\). Then \(L\) is equivalent under an \(\mathrm{SU}(m)\) rotation, to one of the Lawlor necks \(L_{\phi,A}\). If \(L\) is asymptotic at rate \(\rho<2\), then \(L\) is equivalent to some \(L_{\phi,A}\) under an \(\mathrm{SU}(m)\) rotation and a translation in \(\mathbb C^m\). If instead \(L\) is immersed rather than embedded, the only extra possibilities are that \(L=\Pi_1\cap\Pi_2\) if \(\rho<0\) and \(L=\Pi_1\cap\Pi_2+c\) for \(\rho<2\). Theorem 1.2. If \(L\) is a compact closed embedded exact ACLMCF expander in \(\mathbb C^m\), \(m\geq 3\), satisfying the expander equation \(H=\alpha F^\perp\), and asymptotic rate \(\rho<2\) to \(\Pi_1\cap\Pi_2\). Then \(L\) is equivalent under a \(\mathrm{U}(m)\) rotation to one of the LMCF expanders \(L^\alpha_\phi\). If \(L\) is immersed, then the only extra possibility is \(L=\Pi_1\cap\Pi_2\). The definition of ``Lawlor neck'' [\textit{G. Lawlor}, Invent. Math. 95, No. 2, 437--446 (1989; Zbl 0662.49018)] is given in Example 3.9. A Lawlor neck \(L_{\phi,A}\) is defined as follows: Let \(p(x)=(1+a_1x^2)\cdots(1+a_mx^2)-1\), \(P(x)=\dfrac{p(x)}{x^2}\) and let \[ \phi_k(y)=\int_{-\infty}^y\frac{dx}{(1+a_kx^2)\sqrt{P(x)}}, \;\phi_k=\psi_k(\infty), \;A=\int_{-\infty}^\infty\frac{dx}{2\sqrt{P(x)}}. \] Set \(z_k(y)=e^{t\psi_k(y)}\sqrt{a_k^{-1}+y^2}\), and write \(\phi=(\phi_1,\ldots,\phi_k)\). Then \(L_{\phi,A}\) is defined by \[ L_{\phi,A}=\{(z_1(y)x_1,\ldots,z_m(y)x_m):y\in\mathbb R, x_k\in\mathbb R,x_1^2+\cdots+x_m^2=1\}. \] The \(L^\alpha_\phi\) found in [\textit{D. Joyce} et al., J. Differ. Geom. 84, No. 1, 127--161 (2010; Zbl 1206.53071)] and given as example 3.15. is defined as follows: Let \(P(0)=\alpha+a_1+\cdots+a_m\) and \[ P(x)=\frac{1}{x^2}\bigl(e^{\alpha x^2}\prod_{k=1}^m(1+a_kx^2)-1\bigr), \;x\not=0. \] Set \(\psi_k(y)=a_k\int_{-\infty}^y \dfrac{dx}{(1+a_kx^2)\sqrt{P(x)}}\), \(\phi_k=\psi_k(\infty)\), \(\phi=(\phi_1,\ldots,\phi_m)\) and \(z_k(y)=e^{i\psi_k(y)}\sqrt{a_k^{-1}+y^2}\), then \[ L^\alpha_\phi=\{(z_1(y)x_1,\ldots,z_m)y)x_m):y\in\mathbb R, x_k\in\mathbb R, x_1^2+\cdots+x_m^2=1 \}. \] The authors say that the restriction \(m\geq 3\) comes from results on the derived Fukaya category \(D^b\mathcal{F}(T^\ast\mathcal{S}^m\sharp T^\ast\mathcal{S}^m)\) by \textit{M. Abouzaid} and \textit{I. Smith} [Geom. Funct. Anal. 22, No. 4, 785--831 (2012; Zbl 1266.53073)], used in the proofs. But analogues of Theorems 1.1 and 1.2 for \(m=2\) can be proved. They are sketched in Remark 1.3. The proofs use deep results from symplectic geometry concerning \(J\)-homomorphic curves, Lagrangian Floer cohomology and Fukaya category. These are explained in \S2 together with other important notions such as symplectic Calabi-Yau manifolds, and proved assuming that \((M,\omega)\) is a symplectic Calabi-Yau Liouville manifold of real dimension \(2m\) and that \(L,L'\) are transversally intersecting, exact, graded Lagrangian submanifolds in \(M\). Then each intersection point \(p\in L\cap L'\) has a degree \(\mu_{L,L'}(p)\in\mathbb Z\). If \(L,L'\) are compact, then they are objects in the derived Fukaya category \(D^b\mathcal{F}(M)\). If they are isomorphic in \(D^b\mathcal{F}(M)\), then for any generic almost-complex structure \(J\) on \(M\) compatible with \(\omega\), there exist points \(p,q\in L\cap L'\) with \(\mu_{L,L'}(p)=0\) and \(\mu_{L,L'}(q)=m\), and a \(J\)-holomorphic disk \(\Sigma\) in \(M\) with boundary \(L\cup L'\) and corners at \(p,q\) (Theorem 2.15, cf. Figure 1). Section 3 discusses Calabi-Yau manifolds, special Lagrangian submanifolds and LMCF expanders, and gives detailed studies on Lawlor necks and \(A^\alpha_\phi\). In \S4, if \(M\) is a Calabi-Yau \(m\)-fold, and let \(L,L'\) be transversally intersecting special Lagrangian \(m\)-folds in \(M\), it is shown that \(\mu_{L,L'}(q)-\mu_{L,L'}(p)<m\) (Theorem 4.3), and if \(L,L'\) are transversally intersecting, graded LMCF expanders in \(\mathbb C^m\), and \(J\) is any almost complex structure on \(\mathbb C^m\) compatible with \(\omega\) and \(\Sigma\) is a \(J\)-holomorphic disc in \(\mathbb C^m\) with boundary \(L\cap L'\) and corners at \(p,q\in L\cup L'\) as in Theorem 2.15, then \(\mu_{L,L'}(q)-\mu_{L,L7}(p)<m\) (Theorem 4.4). Theorems 1.1 and 1.2 are proved by using these facts in \S4. The authors say that the methods which they use to prove Theorems 1.1 and 1.2 are new and at least as interesting as the theorems themselves, and can very probably also used to prove uniqueness of other classes of special Lagrangians and solitons for Lagrangian MCF.