Part (a): The function f is analytic in the whole plane with positive imaginary part. What can it be?

Part (b): What if all you know is that the imaginary part of f tends to 0 at infinity?

what we know:

For part(a): Write f(z) = u(x,y) + iv(x,y)

u(x,y) and v(x,y) are both harmonic; in particular, they are harmonic conjugates to each other. So, Uxx+Uyy=0, and similarly for v.

The Cauchy-Riemann equations hold: $$Ux = Vy$$ $$Uy=-Vx$$

Since f has positive imaginary part, then v(x,y) $>/=$ 0, and its partial derivatives, Vx and Vy are both non-negative.

What more can I say about this function, for part(a)?

Some natural (or maybe not so natural?) guesses would be that f is a constant, or a polynomial, or an exponential function, since we know that f is entire. But, at the moment, I can't seem to extract any more information from the question itself.

Thanks in advance,

  • $\begingroup$ Positive imaginary part - that means the image is contained in the upper half-plane. But the upper half-plane is conformally equivalent to an open disk. So after composing with a suitable mapping, $f$ takes its values in a disk. What does that tell you? $\endgroup$ – Mike Nov 18 '14 at 22:35
  • $\begingroup$ Or, if you seem to know only the very definition of complex analytic function (i.e. real and imaginary part are harmonic conjugates), what do you know about non-negative harmonic functions in $\mathbb{R}^n$? $\endgroup$ – wisefool Nov 18 '14 at 22:38
  • $\begingroup$ @Mike, applying your hint and making my best guess, I'd say that, consider the standard mapping phi: upper half plane -> unit disk, compose it with f, then phi(f(z)) takes values in the open disk, which is bounded. I'd now like to argue that since f is entire and bounded, and by Liouville's theorem, we conclude that f is a constant. But, that may not be correct, since, it's really phi(f(z)) that is bounded - not f(z), which lives in the upper half space. Also, f is entire, but I don't think phi(f(z)) is entire. Am I sort of on the right track? $\endgroup$ – User001 Nov 18 '14 at 23:09
  • $\begingroup$ "Lebron", what is the domain of $\phi(f(z))$? $\endgroup$ – Mike Nov 18 '14 at 23:16
  • $\begingroup$ The upper half plane... $\endgroup$ – User001 Nov 18 '14 at 23:24

Part (a) can be done as Mike suggested in the comments (or using the Liouville theorem for harmonic functions), or considering the function $e^{-if}$ which is bounded by $1$ on the whole plane, hence constant.

For part (b), if $\mathsf{Im}(f(z))$ tends to $0$ as $|z|\to\infty$, then there is $R>0$ such that, for $|z|\geq R$, $|\mathsf{Im}(f(z))|\leq 1$.

Take $K=\{z\in\mathbb{C}\ :\ |z|\leq 2R\}$. This is a compact set, hence $|\mathsf{Im}(f(z))|$ achieves a maximum on $K$, say $M>0$, but outside $K$, $|\mathsf{Im}(f(z))|\leq 1$, so $|\mathsf{Im}(f(z))|\leq\max\{M,1\}$ on the whole plane, hence it is constant.

Equivalently, set $g=e^{-if}$, then $|g|\to1$ when $|z|\to\infty$, meaning that $g$ is bounded around $\infty$. By Riemann's extension theorem, the function $h(z)=g(1/z)$, bounded around $0$, can be extended holomorphically to $z=0$; therefore the function $g$ extends holomorphically (or at least continuously if you have problems dealing with Riemann surfaces) to a function $\tilde{g}:\mathbb{C}\cup\{\infty\}\to\mathbb{C}$, i.e. $\tilde{g}:\mathbb{S}^2\to\mathbb{C}$. But as $\mathbb{S}^2$ is compact, $\tilde{g}$ is bounded and so is $g$, which is then constant and so is $f$.

(or if you know enough about holomorphic functions, on compact Riemann surfaces like $\mathbb{CP}^1=\mathbb{C}\cup\{\infty\}$ there are no non-constant holomorphic functions…)

  • $\begingroup$ Ok, got it - thanks so much, @wisefool. Looking at the function e^if is indeed very nice and direct, but I'm glad I got a bit more practice with using conformal mappings, before finding out about your suggestion. Have a great night. $\endgroup$ – User001 Nov 21 '14 at 0:06

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.