Strichartz and Spectral Projection Estimates on Asymptotically Conic Manifolds

Zhexing Zhang · 2026

Read Strichartz and Spectral Projection Estimates on Asymptotically Conic Manifolds by Zhexing Zhang (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-07629.txt.

Introduction

Zhexing Zhang proves dispersive estimates for wave-like equations on curved surfaces with cone-like or flat far ends. Strichartz estimates bound how solutions spread out over time, and spectral projection bounds do related work band by band. Trapping, where rays linger in a region, is the usual obstacle. Curvature is what controls it.

Under the assumption that a large enough region around the trapped set has negative curvature, Zhang proves a lossless Strichartz theorem on unit time spans for asymptotically conic surfaces. For flatter surfaces with nonpositive curvature, the spectral projection theorem is proved for a range of exponents. If the negatively curved region reaches far enough out, the critical and subcritical bounds follow without a sign assumption near infinity.

Catalog record

CREATORZhexing Zhang
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE93 KB
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.07629

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Strichartz and Spectral Projection Estimates on Asymptotically Conic Manifolds Zhexing Zhang Abstract . We prove the lossless unit interval Strichartz theorem on asymptotically conic surfaces, assuming that a large enough neighborhood of its trapped set has negative curvature. We also prove the spectral projection theorem on surfaces with Euclidean ends and nonpositive curvature, assuming a large enough neighborhood of its trapped set has negative curvature. We also discuss the spectral projection theorem on asymptotically Euclidean manifolds with dimension greater than or equal to 3, assuming some local smoothing estimates. 1. Introduction Assume (M,g) is an asymptotically conic manifold of the form M=M_{0}\cup M_{\infty} , where M_{0} is compact, and on M_{\infty}=[A,\infty)\times\partial M , the metric is dr^{2}+\sum_{i,j}g_{ij}(r,\theta)d\theta_{i}d\theta_{j}, for some A\geq 0 and some g_{ij}\in C^{\infty}([A,\infty)\times\partial M) . Let \Delta_{g} denote the Laplacian operator on M. Define p:=\det(g_{ij})^{1/2} , and we define an effective potential \displaystyle q(r,\theta):=(2p)^{-2}(\partial_{r}p)^{2}+(2p)^{-2}\sum_{i,j}\frac{\partial p}{\partial\theta_{i}}\frac{\partial p}{\partial\theta_{j}}g^{ij}+2^{-1}p\Delta_{M_{\infty}}(p^{-1}). Here, we let (1) \displaystyle\Delta_{M_{\infty}}:=\Delta_{g}|_{M_{\infty}}=-\partial_{r}^{2}-\frac{\partial_{r}p}{p}\partial_{r}+\Delta_{\partial M_{\infty}}, where \displaystyle\Delta_{\partial M_{\infty}}=-p^{-1}\sum_{i,j}\partial_{\theta_{i}}(pg^{ij}\partial_{\theta_{j}}). We now make the same assumptions as (1.2) and (1.3) of [ 4 ] . Assume there exists a constant C>0, such that (2) \displaystyle|q(r,\theta)|<C\,\,\,,\frac{\partial q}{\partial r}<Cr^{-1-\delta_{0}} for some \delta_{0}>0 and (3) \displaystyle-\frac{\partial}{\partial r}\sum_{i,j}g^{ij}(r,\theta)\xi_{i}\xi_{j}\geq\frac{C}{r}\sum_{i,j}g^{ij}(r,\theta)\xi_{i}\xi_{j} for (\theta,\xi)\in T^{*}\partial M. Notice that ( 3 ) implies that on M_{\infty}, the geodesic flow is convex in terms of r . Meanwhile, notice that in 2-dimensional surfaces, assume the metric is dr^{2}+f(r,\theta)^{2}d\theta^{2} on M_{\infty} for some f\in C^{\infty}([A,\infty)\times\partial M) , then ( 3 ) is equivalent to (4) \displaystyle-\frac{\partial f(r,\theta)^{-2}}{\partial r}\geq\frac{C}{r}f(r,\theta)^{-2}. Burq, Guillarmou and Hassell [ 3 ] proved the Strichartz estimates on a large class of asymptotically conic manifolds with hyperbolic trapped set and satisfying the pressure condition. Recently, Huang, Sogge, Tao and the author [ 14 ] proved the Strichartz estimates on the unit time interval on asymptotically hyperbolic surfaces with negative curvature and bounded geometry without assuming the pressure condition. By constructing an asymptotically hyperbolic background manifold of M , which agrees with M on M_{0} , we can use the result in [ 14 ] to prove the following Strichartz estimate without assuming the pressure condition. We denote the Laplace operator on M associated with the metric g by \Delta_{g} . Theorem 1 . Assume (M,g) is a 2-dimensional asymptotically conic surface in the form M=M_{0}\cup M_{\infty} as above, satisfying ( 2 ) and ( 4 ). Meanwhile, assume M_{0} has negative sectional curvature. Then, we have the following Strichartz estimate. (5) \displaystyle\|e^{-it\Delta_{g}}u_{0}\|_{L^{p}_{t}L^{q}_{x}(M\times[0,1])}\leq C\|u_{0}\|_{L^{2}(M)}, for (p,q) satisfying the Keel-Tao condition, (6) \displaystyle 2(1/2-1/q)=2/p,\,\,\,\,\,\,\,\text{ for }p\in(2,\infty). Stein [ 17 ] and Tomas [ 19 ] proved the spectral measure estimates for Euclidean spaces. Sogge [ 15 ] proved the sharp spectral projection with spectral window of unit size on compact manifolds. Huang and Sogge [ 13 ] proved the spectral projection in logarithmic scale for compact manifolds with nonpositive curvature. Then, Huang, Sogge, Tao and the author [ 14 ] proved the spectral projection in logarithmic scale for manifolds with nonpositive curvature and bounded geometry, as well as the spectral projection estimates on even asymptotically hyperbolic surfaces with strictly negative curvature. Let \mathbf{1}_{S} be the indicator function on the set S. Let I\subset\mathbb{R} be an interval. Then, \mathbf{1}_{I}(\sqrt{-\Delta_{g}}) is the spectral projection operator on the spectral window I . We may follow [ 14 ] to obtain the following spectral projection estimate on surfaces with Euclidean ends. Theorem 2 . Assume (M,g) is a 2-dimensional surface with nonpositive curvature and Euclidean ends. Meanwhile, M=M_{0}\cup M_{\infty} as above, satisfying ( 2 ) and ( 4 ), and M_{0} has negative sectional curvature. For q>2 , denote (7) \mu(q)=\begin{cases}2(\tfrac{1}{2}-\tfrac{1}{q})-\frac{1}{2},\quad q\geq 6,\\ \tfrac{1}{2}(\tfrac{1}{2}-\tfrac{1}{q}),\quad q\in(2,6].\end{cases} Then, we have the following spectral projection estimate. (8) \displaystyle\|\mathbf{1}_{[\lambda,\lambda+\delta]}(\sqrt{-\Delta_{g}})f\|_{L^{q}(M)}\lesssim\begin{cases}\lambda^{\mu(q)}\delta^{1/2}\|f\|_{L^{2}(M)}&\text{for }q\in[6,\infty],\\ (\lambda\delta)^{\mu(q)}\|f\|_{L^{2}(M)}&\text{for }q\in(2,6],\end{cases} if \lambda\gg 1 and \delta\in(0,1] . Notice that Theorem 2 agrees with the Stein-Thomas restriction theorem, [ 17 ] and [ 19 ] , and is equivalent to it for \mathbb{R}^{2}. The reason we only consider surfaces with Euclidean ends is that we lack some desired estimates on the spectral measure for other nontrapped asymptotically conic surfaces. Hassell, Guillarmou and Sikora [ 10 ] stated that the difficulty is to obtain a suitable inverse for the model operator at the zf face in the construction of Section 3 of [ 11 ] , which is not invertible as an operator on the L^{2} space of a two dimensional asymptotically conic surface. One might be able to use the method in this paper to generalize Theorem 2 to asymptotically conic surfaces if the 2-dimensional nontrapped case is explored. Burq, Gérard and Tzvetkov [ 2 ] proved the sharp spectral projection estimate with spectral window unit size restricted to submanifolds of compact manifolds. Then, Chen [ 5 ] refined the unit band estimate in [ 2 ] to a sharp \log -scale estimate on compact manifolds with nonpositive curvature for supercritical exponents. The author proved the logarithmic scale projection estimate restricted to uniformly embedded submanifolds of manifolds of bounded geometry with nonpositive sectional curvature for supercritical exponents, as well as the lossless spectral projection estimate restricted to nontrapped geodesics of asymptotically hyperbolic surface with strictly negative curvature [ 21 ] . Let \gamma be a nontrapped geodesic of M , and let R_{\gamma} be the restriction operator from M to \gamma. Define (9) \displaystyle\kappa(q)=\frac{1}{2}-\frac{1}{q},\,\,\,\,\, \displaystyle\text{ if }q>4. We can prove the supercritical spectral projection estimate restricted to nontrapped geodesics of manifold with Euclidean ends. Theorem 3 . Let (M,g) be as in Theorem 2 . Let q>4, \lambda\gg 1 and \delta\in(0,1] . Let \gamma be a nontrapped geodesic in M , then (10) \displaystyle||R_{\gamma}\mathbf{1}_{[\lambda,\lambda+\delta]}(\sqrt{-\Delta_{g}})f||_{L^{q}(\gamma)}\lesssim{\lambda^{\kappa(q)}}{\delta^{1/2}}||f||_{L^{2}(M)}. Let S^{*}M be the cosphere bundle of M and denote the principal symbol of P by p(x,\xi) . Let (x(t),\xi(t))=e^{tH_{p}}(x,\xi) , where e^{tH_{p}} denotes the geodesic flow on the cotangent bundle. Define (11) \displaystyle\Gamma_{\pm}:=\{(x,\xi)\in S^{*}M:x(t)\not\to\infty\text{ as }t\to\pm\infty\}. We denote K=\Gamma_{+}\cap\Gamma_{-}. Define \pi:S^{*}M\to M with \pi(x,\xi)=x. The trapped set of M is defined to be \pi(K) . Hassell, Guillarmou and Sikora [ 10 ] proved the restriction theorem on nontrapping asymptotically conic manifolds with dimension n\geq 3 and no conjugate points. Then, Chen [ 6 ] removed the condition on conjugate points in their results. Hassell, Guillarmou and Sikora [ 10 ] also proved the restriction theorem on the nontrapped region of possibly trapping asymptotically conic manifolds with dimension n\geq 3. This inspired us to consider spectral projection on such manifolds. However, our proof requires the local smoothing estimate on the trapped set with at most logarithmic loss in dimension n\geq 3. So, we have to make some further assumptions. Theorem 4 . For n\geq 3, assume (M,g) is an n -dimensional asymptotically Euclidean manifold with nonpositive sectional curvature. For q>2 , denote (12) \mu(q)=\begin{cases}n(\tfrac{1}{2}-\tfrac{1}{q})-\frac{1}{2},\quad q\geq q_{c}=\tfrac{2(n+1)}{n-1},\\ \tfrac{n-1}{2}(\tfrac{1}{2}-\tfrac{1}{q}),\quad q\in(2,q_{c}].\end{cases} Further assume that if \chi is a compact cutoff function on M , then (13) \displaystyle\|\chi e^{-it\Delta_{g}}\beta(\sqrt{-\Delta_{g}}/\lambda)u_{0}\|_{L^{2}_{t,x}(M\times\mathbb{R})}\lesssim\lambda^{-1/2}(\log\lambda)^{1/2}\|u_{0}\|_{L^{2}(M)}. Then, we have the following spectral projection estimate. (14) \displaystyle\|\mathbf{1}_{[\lambda,\lambda+\delta]}(\sqrt{-\Delta_{g}})f\|_{L^{q}(M)}\lesssim\begin{cases}\lambda^{\mu(q)}\delta^{1/2}\|f\|_{L^{2}(M)}&\text{for }q\in[q_{c},\infty),\\ (\lambda\delta)^{\mu(q)}\|f\|_{L^{2}(M)}&\text{for }q\in(2,q_{c}],\end{cases} if \lambda\gg 1 and \delta\in(0,1] . The local smoothing estimate with logarithmic loss on the trapped set, ( 13 ), is satisfies for a large class of asymptotically Euclidean manifolds. For example, from [ 3 ] and [ 8 ] , we know ( 13 ) is satisfied if a neighborhood of the trapped set of M has negative sectional

Read the complete text