Quantum cohomology as a deformation of symplectic cohomology (Q2155484)

In this paper, the authors study the relationship between the quantum cohomology of a monotone closed symplectic manifold \(M\) and the symplectic cohomology of the complement of a simple normal crossings (or SC) divisor \(D\), mainly under the assumption that the pair \((M,D)\) is logarithmic effective. They prove that the former is a deformation of the latter in a suitable sense. They also prove some rigidity results about the isotropic skeleton of \(X=M-D\). A closed symplectic manifold \((M,\omega)\) is called monotone if the cohomology class \([\omega]\) is a multiple of \(c_1(TM)\). In particular, \([\omega]\) is a rational cohomology class, and Donaldson's construction of a (smooth) symplectic divisor Poincare dual to a large multiple of \([\omega]\) applies. In this paper, the authors consider an SC divisor \(D=\bigcup_{i=1}^N D_i\) such that \(c_1(TM)\) is a weighted sum \[ c_1(TM)=\textbf{D}:=\sum_{i=1}^N a_i D_i \] with positive rational weights (Note that \(\lambda_i=2a_i\) in the paper's notation). The last identity and the monotonicity imply that \(X\) is a convex exact symplectic manifold with a vanishing first Chern class. Furthermore, if \([\omega]\) is a positive multiple of \(c_1(TM)\), the positivity of \(D\) implies that \(X\) is a Weinstein domain and retracts to an isotropic skeleton. In Hypothesis A, the authors assume that \(a_i \leq 1\), which means \[ K_M+D=D-\textbf{D}=\sum_{i=1}^n (1-a_i)D_i \] is effective (in the algebraic geometric language). Here, \(K_M\) is the canonical line bundle of \((M,\omega)\) and \(K_M+D\) is the logarithmic canonical line bundle of the pair \((M,D)\). The borderline case is \(a_i\equiv 1\), where \((X,D)\) is a symplectic log Calabi-Yau pair. The vague conjecture in the literature has been that under certain assumptions on \((M,D)\), the Novikov field-valued quantum cohomology \(QH^*(M,\Lambda)\) of \(M\) coincides the cohomology of a natural deformation \(\big(SC^*(X;\Lambda),\partial\big)\) of the symplectic cochain complex \(\big(SC^*(X;\Lambda),d\big)\). The former is defined using closed pseudoholomorphic curves, and the latter involves periodic orbits of a Hamiltonian function on \(X\) and ceratin pseudoholomorphic-type curves connecting them. In Theorem C, the authors prove this conjecture when \(M\) is monotone and \((M,D)\) is log-effective. Furthermore, they prove that the spectral sequence of the deformed filtered complex converges on the first page to the quantum cohomology. As a corollary (Corollary 1.7), they conclude that \(X\) has non-vanishing symplectic cohomology. Over the course of the proof, they study compact subsets of \(X\) and their symplectic homology relative to \(M\). For instance, if \((M,D)\) is log-effective, they prove that the isotropic skeleton \(L\) of \(X\) is SH-full, meaning that the symplectic homology of any compact set \(K\) in \(X-L\) is zero. They conclude that \(L\) can not be displaced from any ``Floer-theoretically essential monotone'' Lagrangian in \(X\).