Edit: on a second thought, I don't think it's possible since
$$ f(z) = \sqrt {z\sin z} = e^{\large \frac{1}{2} \log z}e^{\large \frac{1}{2} \log\sin z}$$
$$e^{\large \frac{1}{2} (\ln|z| + iArg(z))}e^{\large \frac{1}{2} (\ln|\sin(z)| +iArg(sin(z))}$$
is undefined at the origin -- we'd get a $ln|sin(0)| = ln|0|$ factor. What do you think? Thanks,
The problem statement is:
Consider the function $ f(z) = \sqrt {z\sin z}$ . Can $f(z)$ be defined near the origin as a single valued analytic function?
What will be the radius of convergence of the power series expansion of $f$ around $z=0$?
My thoughts:
It is tempting to say "no", since $z=0$ seems like a branch point of $f(z)$.
But a closer look at the function
$$ f(z) = \sqrt {z\sin z} = e^{\large \frac{1}{2} \log z}e^{\large \frac{1}{2} \log\sin z}$$
shows that we need to pick two branch cuts. And if we choose both cuts to be $\mathbb R^- \cup \{0\}$, i.e., choose the principal branch of the logarithm for both factors, then each factor is discontinuous and not defined on the negative real axis (plus the origin), and each jumps by a factor of $e^{\frac{1}{2} 2 \pi i} = e^{i\pi} = -1$.
However, considering both factors together, we get a total "jump" of $e^{2\pi i} =1$, and I think that now $f(z)$ has been made single-valued and analytic on the negative real axis, including the origin, and thus we get an analytic continuation onto the negative axis and the origin, so that $f(z)$ can in fact be defined near the origin as a single-valued analytic function.
What do you think?
This seems a bit too generous, though. I feel like I have claimed that this function is entire, which I am almost certain cannot be true...because of the complex logarithm used to define $f(z)$.
Any ideas are welcome.
Thanks,
HINT:
$$z \sin z = z^2 \left( 1 + (\frac{\sin z}{z} -1) \right) $$
and $\phi(z) = \frac{\sin z}{z} -1$ is entire, taking value $0$ at $0$, so one can define $\sqrt{1 + \phi(z)}$ for $z$ close enough to $0$.