Let $L:=\mathbb{Q}\left(\zeta_{n}\right)$, where $\zeta_{n}=e^{2\pi i/n}$ for some prime $n\geq 3$. Let $K:=L \cap \mathbb{R}$.
Find $a\in L$ such that $K=\mathbb{Q}(a)$.
What is known:
The degree $[L:K]$ equals $2$.
I can take $a=\cos(2\pi/n)$, as for $b=i\sin(2\pi/n)$, one has the equality $L=\mathbb{Q}(a,b)$:
- "$\subseteq$" is clear, as $\zeta_{n}=a+b$;
- "$\supseteq$" follows from the equalities $\frac{\zeta_{n}+\zeta_{n}^{-1}}{2}=a$ and $\frac{\zeta_{n}-\zeta_{n}^{-1}}{2}=b$.
Now $p-1=[L:\mathbb{Q}]=[L:\mathbb{Q}(a)]\cdot [\mathbb{Q}(a):\mathbb{Q}]$, where the second factor equals $\frac{p-1}{2}$. Hence $[L:\mathbb{Q}(a)]=2$.
As $\mathbb{Q}(a)\subseteq K$ (again by the argument that $\frac{\zeta_{n}+\zeta_{n}^{-1}}{2}=a$) and $[L:\mathbb{Q}(a)]=2=[L:K]$, we must have $\mathbb{Q}(a)=K$.
The problem is that I have no proof for the fact that $[\mathbb{Q}(a):\mathbb{Q}]=\frac{p-1}{2}$. I have found online that the degree of the minimal polynomial of $a=\cos(2\pi/n)$ over $\mathbb{Q}$ equals $\frac{p-1}{2}$.
There is no need to prove it using this. I just wanted to argue why I know that this $a$ works.