5
$\begingroup$

Let $h$ be a real valued harmonic function on the twice punctured plane $Ω = \text{C \ {0, 1}}$. Show that there exist unique real numbers $a_0$, $a_1$ such that $u(z) = h(z) − a_0 \log |z| − a_1 \log |z − 1|$ is the real part of a holomorphic function on $Ω$ .

I tried to show that $u$ has a harmonic conjugate. I assumed there exists such a conjugate $v$ exist, then used Cauchy Riemann equations on $u$ and $v$. After integrating two equations, I tried to use the result to find $a_0$ and $a_1$. However, I failed to cancel out the extra terms. Is there any other direction I should go?

$\endgroup$
2
$\begingroup$

Define a function $g = h'_x - ih'_y$. Since $h$ is harmonic (and hence $C^2$), $g$ satisfies Cauchy-Riemann's equations on $\Omega$. Indeed: $$ (h'_x)'_x = h''_{xx} = -h''_{yy} = (-h'_y)'_y $$ and $$ (h'_x)'_y = h''_{xy} = h''_{yx} = -(-h'_y)'_x. $$ In other words, $g$ is holomorphic on $\Omega$. Assume for a moment that $g$ admits an anti-derivative $G$ on $\Omega$. If $G = U+iV$, Cauchy-Riemann shows that $G' = U'_x-iU'_y$, but $G' = g = h'_x-ih'_y$, so $\nabla U = \nabla h$, i.e. $U=h+C$, and we may as well take $C=0$ by adjusting our choice of $G$. Hence $h = U = \operatorname{Re} G$.

Now, the problem is that $\Omega$ is not simply connected, so we can't guarantee that $g$ admits an anti-derivative. On the other hand, we know that $g$ has an anti-derivative on $\Omega$ if and only if $\int_\gamma g(z)\,dz = 0$ for every simple closed curve in $\Omega$, and it's not difficult to see that this condition is equivalent to the fact that the residues of $g$ at $0$ and $1$ should both be $0$.

Let $a_0 = \operatorname{Res}\limits_{z=0} g$ and $a_1 = \operatorname{Res}\limits_{z=1} g$, and put $u(z) = h(z)-a_0\log|z|-a_1\log|z-1|$.

Repeat the above argument to get a holomorphic function $f=u'_x-iu'_y$ on $\Omega$. It's straight-forward to verify that $$ f(z) = g(z) - \frac{a_0}{z} - \frac{a_1}{z-1} $$ so by construction, the residues of $f$ at $0$ and $1$ vanish, and from the discussion above, we are done.

This takes care of existence. For uniqueness, assume that $u(z) = h(z)-a_0\log|z|-a_1\log|z-1|$ and $\tilde u(z) = h(z)-\tilde a_0\log|z|-\tilde a_1\log|z-1|$ are two such functions. Then $u(z) = \tilde u(z) = (a_0-\tilde a_0) \log|z|-(a_1-\tilde a_1)\log|z-1|$ is the real part of a holomorphic function on $\Omega$. It's a standard exercise to check that this can only happen if $a_0 = \tilde a_0$ and $a_1 = \tilde a_1$.

$\endgroup$

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.