How can we proof the following argument:

Given an analytic function $f(z)$ ($z=x+iy, x,y\in \mathbb R$) in the closed upper half plane such that $|f(z)|<1$ for all $z$ in the open upper half plane, and $|f(x)|=1$ for all $x\in \mathbb R$, then:

there is an open simply connected set $S$ with $\mathbb R \subseteq S \subseteq \mathbb C$ such that $f(z)$ does not vanish on $S$, and there exists an analytic function $\psi:S \rightarrow \mathbb C$ such that $f(z)=\exp(i\psi(z))$

  • $\begingroup$ Is that meant to say $\mathbb R \subseteq S \subseteq \mathbb C$? $\endgroup$ – joriki Dec 23 '11 at 3:35
  • $\begingroup$ Yes, I fixed it. $\endgroup$ – Eva Md Dec 23 '11 at 20:52

As far as I can tell this involves little to no complex analysis. Also, I'm assuming you want $S$ to be open in the upper half plane, not necessarily in all of $\mathbb{C}$.

To construct $S$, consider any $[a,b] \times [0, R] \subseteq \mathbb{H}$ (that is, $\{z: a \leqslant Re z \leqslant b, 0 \leqslant Im z \leqslant R \}$, for $R >> 0$, by compactness and hence uniform continuity there exists a subset $S_{a,b}$ of the form $[a,b] \times [0, \epsilon)$ on which $f$ is nonvanishing.

Now taking $[a,b]$ to be say $\ldots [-1,0], [0,1], [1,2], \ldots$, and patching your $S_{a,b}$ together nicely at the endpoints gives $S$.

$S$ is clearly simply connected (it deformation retracts onto the real line via "squishing").

Lastly, to construct $\psi$, we have $\Omega := S - \mathbb{R}$ is open and simply connected, $f$ is nonvanishing on $\Omega$, and hence $f = e^{g(z)}$ for $g$ holomorphic. Take $\psi := g/i$.

If this is unclear (or wrong!) lemme know.

  • $\begingroup$ To extend $f$ holomorphically into a neighborhood of the real line you can take $f(\bar{z}) = 1/\overline{f(z)}$ for $z$ in the lower half plane by the Schwarz reflection principle; then your argument gives the desired neighborhood of the real line. $\endgroup$ – t.b. Dec 24 '11 at 1:15

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.