Let $n \ge 2$. Set $\dot{H}^1(\mathbb{R}^n)$ to be the homogeneous Sobolev space, defined as the Hilbert completion of $C_0^\infty(\mathbb{R}^n)$ with respect to the norm $\| \varphi \|^2_{\dot{H}^1} \equiv \int_{\mathbb{R}^n} |\nabla \varphi(x) |^2dx = \int_{\mathbb{R}^n} |\xi|^2|\hat{\varphi}(x) |^2dx$, where the "hat" denotes Fourier transform.

Let $R_0(\lambda) = (-\Delta - \lambda^2)^{-1}$, $\text{Im }\lambda >0$ denote the resolvent of the free Laplacian. $R_0(\lambda)$ can be defined as a bounded operator $\dot{H}^1 \to \dot{H}^1$ by first giving the usual definition of $R(\lambda) \varphi$ via the Fourier transform of $\varphi \in \mathcal{S}(\mathbb{R}^n)$, and then noticing that \begin{equation} \label{define free resolv dot H one} \| R_0(\lambda) \varphi \|^2_{\dot{H}^1} = \int_{\mathbb{R}^n} \frac{|\xi|^2 |\hat{\varphi}(\xi)|^2}{|\xi|^2- \lambda^2}d\xi \le \sup_{r \in \mathbb{R}}{\frac{1}{r - \lambda^2}} \int_{\mathbb{R}^n} |\xi|^2 |\hat{\varphi}(\xi)|^2 = C(\lambda) \| \varphi \|^2_{\dot{H}^1}. \end{equation}
So then $R_0(\lambda)$ may be extended to all of $\dot{H}^1$ by density.

I would like to know whether the cutoff resolvent $\chi R_0(\lambda) \chi,$ ($\chi \in C_0^\infty$) can be meromorphically continued from $\text{Im } \lambda > 0$ through the real spectrum and into all of $\mathbb{C}$ (or to the Logarithmic cover of $\mathbb{C}$ in the case of even dimensions), and if $\dot{H}^1 \to \dot{H}^1$ estimates can be obtained for the continuation.

For dimension $n \ge 3$, I believe the answer is yes because of the following fact, which is a consequence of combining the Gagliardo-Nirenberg-Sobolev inequalty along with Holder's inequality:

If $n \ge 3$ and $\chi \in C_0^\infty(\mathbb{R}^n)$, then there exists a constant $C_\chi> 0$ such that for all $u \in C_0^\infty(\mathbb{R}^n)$, we have $$\| \chi u \|_{L^2(\mathbb{R}^n)} \le C_\chi \| \nabla u \|_{L^2(\mathbb{R}^n)}.$$

This estimate allows one to take $u \in C_0^\infty(\mathbb{R}^n)$ and compute $$ \nabla \chi R_0(\lambda) \chi u= (\nabla\chi) \tilde{\chi} R_0(\lambda) \tilde{\chi} \chi u + \chi \tilde{\chi} R_0(\lambda) \tilde{\chi} \nabla \chi u + \chi \tilde{\chi} R_0(\lambda) \tilde{\chi} \chi \nabla u, $$

where $\tilde{\chi} \in C_0^\infty(\mathbb{R}^n)$ and $\tilde{\chi} \equiv 1$ on the support of $\chi$. By applying the above estimate when $n \ge 3$, we may lean on the $L^2 \to L^2$ bounds that are known for the continuation of $\tilde{\chi} R_0(\lambda) \tilde{\chi}$ (as shown in Chapter 3 of the book in progress by Dyatlov and Zworski) to achieve the desired bounds $\dot{H}^1 \to \dot{H}^1.$

In dimension two, I think there are counterexamples to the Sobolev-type inequality given above, and so I feel we cannot so easily obtain the continuation from $\text{Im } \lambda > 0$. I am currently stuck, and would like to know if anyone knows of reference where this or something similar is worked out.

Thank you very much for any suggestions.


Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Browse other questions tagged or ask your own question.