Cylindrical contact homology of 3-dimensional Brieskorn manifolds (Q2133952)

The authors of the paper under review present a complete computation of cylindrical contact homology over the rationals of Brieskorn 3-manifolds equipped with a certain hypertight contact structure. A hypertight contact form is one which admits no contractible Reeb orbits, and cylindrical contact homology of such contact forms is defined due to work of \textit{E. Bao} and \textit{K. Honda} [J. Topol. 11, No. 4, 1002--1053 (2018; Zbl 1409.53064)] and \textit{M. Hutchings} and \textit{J. Nelson} [J. Symplectic Geom. 14, No. 4, 983--1012 (2016; Zbl 1369.53064)]. Let \(p,q,r \geq 2\) be integers. The Brieskorn 3-manifold associated to the triple \((p,q,r)\) is defined as \[ \Sigma(p,q,r) = \left\{(x,y,z) \in \mathbb C^3 \mid x^p + y^q + z^r = 0 \right\} \cap \left\{|x|^2 + |y|^2 + |z|^2 = 1\right\}. \] The \(S^1\)-action on \(\mathbb C^3\) given by \(\zeta \cdot (x,y,z) = (\zeta^{qr}x, \zeta^{pr}y, \zeta^{pq}z)\) gives that \(\Sigma(p,q,r)\) is a Seifert fibered manifold over \(\left\{(x,y,z) \in \mathbb C^3 \mid x^p + y^q + z^r = 0 \right\}\) with three Seifert fibers \(\gamma_1\), \(\gamma_2\) and \(\gamma_3\) corresponding to stabilizers of order \(p\), \(q\) and \(r\), respectively. The authors prove the following. Let \(1/p + 1/q + 1/r < 1\) and assume that \(p\), \(q\) and \(r\) are pairwise relatively prime. There is a hypertight contact form \(\xi_{p,q,r}\) on \(\Sigma(p,q,r)\) such that the cylindrical contact homology of \((\Sigma(p,q,r), \xi_{p,q,r})\) is isomorphic as a relatively graded \(\mathbb Q\)-vector space to \[ \mathrm{HC}_\ast(\varSigma(p,q,r),\xi_{p,q,r}) = \bigoplus_{n \geq 1} G(p,q,r)[-2dn] \oplus \bigoplus_{n\geq 1} H_\ast(S^2)[-2dn-2], \] where \(d = pqr-qr-pr-pq\). Here \(G(p,q,r)\) is the graded \(\mathbb Q\)-vector space \[ G(p,q,r) = \bigoplus_{k=1}^{p-1} \mathbb Q \left\langle \gamma_1^k\right\rangle \oplus \bigoplus_{k=1}^{q-1} \mathbb Q \left\langle \gamma_2^k\right\rangle \oplus \bigoplus_{k=1}^{r-1} \mathbb Q \left\langle \gamma_3^k\right\rangle, \] where the \(\gamma_j^k\) are generators whose grading is given by \[ \mathrm{gr}(\gamma_1^k) = -2 \left\lfloor \frac{kqr}{s}\right\rfloor-2, \quad \mathrm{gr}(\gamma_2^k) = -2 \left\lfloor \frac{kpr}{s}\right\rfloor-2, \quad \mathrm{gr}(\gamma_1^k) = -2 \left\lfloor \frac{kpq}{s}\right\rfloor-2. \] Geometrically the \(\gamma_j^k\) are iterates of the simple Reeb orbits corresponding to the exceptional Seifert fibers of \(\Sigma(p,q,r)\). To prove the main theorem the authors consider the one-form on \(\mathbb C^3\) given by \[ \alpha_{p,q,r} := \frac 14 i (p (z_1 d \overline z_1 - \overline z_1 dz_1) + q (z_2 d \overline z_2 - \overline z_2 dz_2)+ r (z_3 d \overline z_3 - \overline z_3 dz_3)). \] Then \(\lambda_{p,q,r} := \left.\alpha_{p,q,r}\right|_{\Sigma(p,q,r)}\) is a contact form on \(\Sigma(p,q,r)\) whose Reeb vector field is the tangent field of the \(S^1\)-action on \(\Sigma(p,q,r)\). It is immediate that \(\lambda_{p,q,r}\) is hypertight but degenerate. The main technical work in the paper under review is about finding a perturbation of \(\lambda_{p,q,r}\) which is hypertight and nondegenerate. To achieve this, assuming \(1/p + 1/q + 1/r < 1\), the authors use that \(\Sigma(p,q,r)\) are certain quotients of the universal cover of \(\mathrm{PSL}(2,\mathbb R)\).