21

I don't understand the final step of an argument I read.

Suppose $f$ is holomorphic in a neighborhood containing the closed unit disk, nonconstant, and $|f(z)|=1$ when $|z|=1$. There is some point $z_0$ in the unit disk such that $f(z_0)=0$.

By the maximum modulus principle, it follows that $|f(z)|<1$ in the open unit disk. Since the closed disk is compact, $f$ obtains a minimum on the closed disk, necessarily on the interior in this situation.

But why does that imply that $f(z_0)=0$ for some $z_0$? I'm aware of the minimum modulus principle, that the modulus of a holomorphic, nonconstant, nonzero function on a domain does not obtain a minimum in the domain. But I'm not sure if that applies here.

Hana Bailey
  • 1,769
  • If $f$ has modulus 1 on the entire unit disk, then the function has to be constant. However, what you write in the body of the question just requires $|f(z)=1|$ on the unit circle. – hmakholm left over Monica Apr 02 '12 at 21:46
  • Dear @HenningMakholm, thanks, I accidentally used the wrong word. I've now changed it to circle. – Hana Bailey Apr 02 '12 at 21:48
  • 3
    "I'm aware of the minimum modulus principle, that the modulus of a holomorphic, nonconstant, nonzero function on a domain does not obtain a minimum in the domain. But I'm not sure if that applies here." Yes, it applies here. Why wouldn't it? If $f$ had no zeros in the unit disk then that theorem would be contradicted. – Jonas Meyer Apr 03 '12 at 00:52

1 Answers1

29

If not, consider $g(z)=\frac 1{f(z)}$ on the closure of the unit disc. We have $|g(z)|=1$ if $|z|=1$ and $|g(z)|>1$ if $|z|<1$. Since $g$ is holomorphic on the unit disk, the maximum modulus principle yields a contradiction.

Davide Giraudo
  • 181,608