164F129D254D
(Lawlor-neck desingularizations of grim cylinders) Fix $m\geq 3$. Let
Choose $\phi _ j\in (0,\pi )$ with $\sum _{j=1}^ m\phi _ j=\pi $, and $\lambda \in \mathbb R$. Define
The two grim cylinders meet at one transverse point and satisfy the tangent-plane angle criterion. Su proves that for all sufficiently small $\varepsilon >0$ their intersection can be replaced by a Lawlor neck and then perturbed to a smooth embedded Lagrangian translator $L_\varepsilon $, diffeomorphic to $S^{m-1}\times \mathbb R$. It is complete and proper, with two ends approaching the original cylinder ends in the weighted construction. Its dimension and codimension are $m$. With $e_ m$ the real unit vector in the last complex coordinate,
The parameter $\varepsilon $ is a gluing scale, not flow time. As $\varepsilon \downarrow 0$ the translators converge as currents to $L_0\cup P_{\phi ,\lambda }(L_0)$. The displayed formula is for the building blocks; the perturbed translator has no closed formula. Properness and completeness follow from the small decaying graph on the complete proper ends and the compact neck region.
Source: [Su2022], Main Theorem (p. 3), Sections 3.2–3.4 (pp. 8–10), Section 4.1 (pp. 11–12), and Theorem 6.10 (p. 28). Its Corollary 1.1 also gives examples with arbitrarily many ends and noncontractible loops by finite admissible arrangements of grim cylinders. Those more general topologies are not assigned to this two-ended case. The construction requires the source's $T$-finite condition; it does not establish gluing arbitrary planes to grim cylinders. No uniqueness, entropy value, or dynamical stability is claimed. The pictures specialize to $m=3$, $\phi _1=\phi _2=\phi _3=\pi /3$, and $\lambda =0$. The common point is
An equal-angle Lawlor neck centered at $p$ can be written, up to its scale, as
where $0<\arg w_\varepsilon <\pi /3$. The identity $\operatorname {Im}(w_\varepsilon ^3)=\varepsilon ^3$ gives the symmetric special Lagrangian neck; it is a local minimal ingredient, not itself the translator. The drawing uses the slice $\omega _2=0$ and an illustrative interpolation to the exact grim cylinders. This coordinate interpolation does not implement Su's potential gluing or final correction. The projected slice retains $(\operatorname {Re}(e^{-i\beta }z_1),\operatorname {Im}z_3,\operatorname {Re}z_3)$; omitted coordinates can create apparent crossings. The profile view selects $\omega _1=\omega _2=0$, and does not determine the topology of the full translator.
At every time the picture translates by $t$ in $\operatorname {Re}z_3$ with $\varepsilon $ fixed. The frames $t=-1,0,1$ therefore have identical shapes and neck sizes. The existence theorem only supplies a sufficiently small gluing range; the displayed $\varepsilon =0.22$ is illustrative, not a certified member of that range.
Comments
Comments are not enabled yet.