I am trying to prove the following conjecture.
Let $(R, +,\times)$ be a finite ring with an identity. Let $G$ be a subgroup of $(R,\times)$ with order $d$. Then $-1\in G$, if and only if $2\mid d$ or $\text{Char}(R)=2$.
My attempt:
$\Leftarrow$: Case 1:If $\text{Char}(R)=2$, then $-1=1$. Since $G$ is a multiplicative group, the indentity $1\in G$. Hence $-1\in G$.
Case 2: If $2\mid d$, then ...(I do not know how to prove in this case.)
$\Rightarrow$: Suppose that $-1 \in G$. Then $(-1)^2=1\in G$. The multiplicative order of $-1$ is either $2$ or $1$. If $\text{Ord}(-1)=1$, then $-1=1$. Consequently, $\text{Char}(R)=2$. If $\text{Ord}(-1)=2$, then it must have $\text{Ord}(-1) \mid \text{Ord}(G)$, i.e., $2\mid d$.
===============
So does $2\mid d$ implies that $-1\in G$ ? If the conjecture does not hold, then can we have a sufficient and necessary condition of $-1 \in G$?
Thanks to Arthur, we have the following result using the fact that $-1$ is the only element of order $2$.
Let $(R, +,\times)$ be an integral domain. Let $G$ be a subgroup of $(R,\times)$ with order $d$. Then $-1\in G$, if and only if $2\mid d$ or $\text{Char}(R)=2$.
Counterexample: let $R=\Bbb Z_{12}$ with standard addition and multiplication, and $G=\{1,5\}$.
As for sufficient conditions, if $R$ is an integral domain with characteristic different from $2$, then the equation $x^2=1$ has exactly two solutions (rewrite to $(x-1)(x+1)=0$ and use the definition of integral domains). Thus $-1$ is the only element with multiplicative order $2$, and by Cauchy's theorem must be in $G$ if $2\mid ord(G)$.