So I was scrolling through Youtube again when I found this video by Blackpenredpen which defined all real and imaginary solutions for$$x^x=y^y$$which got me interested, because it got me asking$$\text{Is there another solution for }i^i=x^x\text{?}$$
$$\text{Here is my attempt at finding another solution to }i^i=x^x$$
$$i^i=x^x$$$$i\log_xi=x$$$$\log_xi=-ix\text{ because }\frac{x}{i}=\frac{x}{\sqrt{-1}}=\frac{x\sqrt{-1}}{-1}=-ix$$$$x^{log_xi}=x^{-ix}$$$$i=\frac{1}{x^{ix}}$$$$ix^{ix}=1$$$$ix^{ix}-2=e^{i\pi}\text{ or }ix^{ix}=e^{i\pi}+2$$$$x^{ix}=-ie^{i\pi}-2i$$$$ix\ln(x)=\ln(-ie^{i\pi}-2i)$$$$x\ln(x)=-i\ln(-ie^{i\pi}-2i)$$$$x\ln(x)=-i\ln(-i(-1)-2i)$$$$x\ln(x)=-i\ln(i-2i)$$$$x\ln(x)=-i\ln(-i)$$$$x^x=(-i)^{-i}$$$$\therefore i^i=(-i)^{-i}$$My question
Is it correct that $i^i$ does in fact equal $(-i)^{-i}$, or is there a flaw in my proof?
This is equivalent to solving $z^z=e^{-\pi/2}$
$z\ln z=-\pi/2$.
$r(\cos \theta +i\sin\theta)(\ln r+i\theta)=-\pi/2$
$r[(\cos \theta \ln r -\theta \sin \theta)+i(\sin\theta \ln r+\theta\cos \theta)]=-\pi/2$
$\sin\theta \ln r+\theta\cos\theta=0\implies r=e^{(-\theta \cot\theta)}$
$e^{-\theta \cot \theta}(-\theta \cos^2 \theta \csc \theta-\theta\sin \theta)=-\pi/2$
$e^{-\theta \cot \theta}=(\pi /2) \sin \theta / \theta$
$-\theta \cos \theta/\sin \theta = \ln(\pi /2)+\ln (\sin \theta /\theta) $
I don't think that last line can be solved analytically.