I want to prove $\left(i\sqrt {\frac{21 \ + \ 5\sqrt{5}}{2}}\right)^n$ does not lie in $\Bbb{Q}(ζ_{11})$ for all positive integer $n$.
This problem arises from arithmetic geometry, but this problem itself is purely a field theoretic one.
For $n$ is even, this is equivalent to $\sqrt{5}\notin \Bbb{Q}(ζ_{11})$. But if $\sqrt{5}\in \Bbb{Q}(ζ_{11})$, prime ideal above $5$ will not be unratified, but from class field theory, prime ideal above $5$ does not ramify, this is a contradiction.
But I'm sticking with the case that $n$ is odd.
If $\left(i\sqrt {\frac{21 \ + \ 5\sqrt{5}}{2}}\right)^n\in \Bbb{Q}(\zeta_{11})$ then $\left(i\sqrt {\frac{21 \ + \ 5\sqrt{5}}{2}}\right)^{2n}=(-\frac{21 \ + \ 5\sqrt{5}}{2})^n \in \Bbb{Q}(\zeta_{11})$.
$Gal(\Bbb{Q}(\zeta_{11})/\Bbb{Q})$ is cyclic so $\Bbb{Q}(\zeta_{11})$ has a unique quadratic subfield, which is $\Bbb{Q}(\sqrt{-11})$.
So we get that $(21 \ + \ 5\sqrt{5})^n\in \Bbb{Q}(\zeta_{11})\cap \Bbb{Q}(\sqrt5)=\Bbb{Q}$.
Take the non-trivial automorphism $\sigma\in Gal(\Bbb{Q}(\sqrt5)/\Bbb{Q})$ such that $ \sigma(21 \ + \ 5\sqrt{5})=21 \ - \ 5\sqrt{5}$.
If $(21 + 5\sqrt{5})^n\in \Bbb{Q}$ then $(21 + 5\sqrt{5})^n=\sigma((21 + 5\sqrt{5})^n)=(21 - 5\sqrt{5})^n$ which is absurd as $|21 - 5\sqrt{5}|\ne |21 + 5\sqrt{5}|$.