2
$\begingroup$

Let $\mathbb D : = \left \{z \in \mathbb C\ :\ \left \lvert z \right \rvert < 1 \right \}.$ Prove that there is no continuous function $f : \overline {\mathbb D} \longrightarrow \mathbb C$ such that $f$ is holomorphic on $\mathbb D$ and $f(z) = \overline z$ for all $z \in \partial \mathbb D.$

My Attempt $:$ By Cauchy Integral Formula we get $f^{(k)} (0) = 0$ for any $k \geq 0.$ But since $f(z) = \sum\limits_{k = 0}^{\infty} \frac {f^{(k)} (0)} {k!} z^k$ we have $f(z) \equiv 0$ on $\mathbb D$ and this cannot happen since $f$ is continuous on $\overline {\mathbb D}$ and $f(z) = \overline z$ for all $z \in \partial \mathbb D.$

Could anyone please have a look at my solution? Thanks in advance.

$\endgroup$
4
  • 1
    $\begingroup$ How did you get $f^{(k)}(0) = 0$? $\endgroup$
    – copper.hat
    Commented May 12 at 19:30
  • $\begingroup$ (The solution is fine, but the usual Cauchy integral formula requires the path to lie in the region of analyticity, so some elaboration is needed.) $\endgroup$
    – copper.hat
    Commented May 12 at 19:38
  • $\begingroup$ Could also use Cauchy's integral theorem to avoid derivatives, though in both cases one should reason why one is able to use them (continuity, compact, blabla). $\endgroup$
    – user1318062
    Commented May 12 at 19:41
  • $\begingroup$ @copper.hat$:$ Is it due to homotopy version of Cauchy's theorem. $\endgroup$
    – Anacardium
    Commented May 12 at 21:00

2 Answers 2

2
$\begingroup$

I don't know whether you can just apply Cauchy integral formula on the boundary. The version that I know does not allow it.

However, the function $g(z) := zf(z)$ for all $z \in \overline{\mathbb{D}}$ fulfils the following for all $z \in \partial\mathbb{D}$: $$ g(z) = z f(z) = z\bar z = \lvert z \rvert^2 = 1 $$ As a holomorphic function that is real on the boundary, it has to be constant. Therefore $zf(z) = 0%$ for all $z \in \mathbb{D}$ as $g(0)=0$. From this we conclude that $f=0$. This contradicts the boundary condition.

$\endgroup$
3
  • $\begingroup$ The linked statement seems to suppose that $f$ is analytic on the boundary as well. OP does not suppose that. $\endgroup$ Commented May 15 at 5:25
  • $\begingroup$ In my opinion, the answer does not use the assumption that you mentioned. So everything still works, even with only continuous boundary $\endgroup$ Commented May 15 at 7:22
  • 1
    $\begingroup$ Yes you're right, the answer there also works in the more general setting we have here. $\endgroup$ Commented May 15 at 17:57
1
$\begingroup$

By Cauchy-Goursat, $\int_{B_r(0)} f(z) dz =0$ for every radius $r<1$. Convince yourself that by continuity, this entails $\int_{B_1(0)} f(z) dz =0$ as well. (This is the nontrivial part, you need that the smaller circles uniformly approach the big one, and actual epsilon-estimates. Cf. Cauchy's integral theorem and domain boundaries).

But it's very well known that $\int_{B_1(0)} \bar z dz = \int_{B_1(0)} z^{-1} dz =2\pi i\neq 0$.

$\endgroup$
1

Not the answer you're looking for? Browse other questions tagged .