I am reading The book of involutions. In page 331, (20.13) Examples (3), they said group of $n^{th}$ roots of unity $\mu_n$ is a non-connected group scheme without any explanation, so I check the case when $n=2$, i.e. consider $\mu_2$. I think $\mu_2$ is connected or not related the characterist of base field $F$. Denote the Hopf algebra of $\mu_2$ by A, then A=$F[X]/(X^2-1)$(See 20.4.5 in this book). I want to use Proposition (20.15) to check it ( let nil(A) be the set of all nilpotent elements of $A$, $A/nil(A)$ is denoted by $A_{red}$, Proposition states that $\mu_n$ is connected $\iff$ $A_{red}$ is an integral domain.
CASE1: $char(F)\neq2$. And we have $(X-1)(X+1)=0$, however $(X-1)^n\neq0$ for all $n$ i.e. $(X-1)$ is not a nilpotent element, thus $nil(A)$ is not a prime ideal which conclude $A_{red}$ is not an integral domain, thus $\mu_2$ is not connected.
CASE2: $char(F)=2$. In this case $(X-1)=(X+1)$, and $(X-1)(X+1)=0=(X-1)^2$, thus $(X-1)$ is a nilpotent element, so $(X-1)\subset nil(A)\subset M$ where $M$ is a maximal ideal of $nil(A)$. Now in $F[X]$, $(X^2-1)\subset (X-1)$, by using the third isomorphism theorem we have $(F[X]/(X^2-1))/(X-1/X^2-1)\cong F[X]/(X-1)$. LHS=$A/(X-1)$($X-1$ in A now) and $X-1$ is irreducible in $F[X]$, thus RHS is a field which means LHS is a field, so $(X-1)$ is a maximal ideal, thus $(X-1)=nil(A)=M$, so $A/nil(A)$ is a field also a integral domain. $A_{red}$ is an integral domain $\Rightarrow \mu_2$ is connected.
This is different about what book said. I wonder my proof is true or false. Any helpful answer will be grateful. Thanks for your attention!