a) Let x be an invertible element and let n be a positive integer such that an=0. Since a+x=x(x−1a+1), it is enough to show that x−1a+1 is invertible. Set b=x−1a. Then bn=x−nan=0, (because A is commutative), hence b2n+1=0. Consequently,
1=b2n+1+1=(b+1)(b2n−b2n−1+⋯−b+1),
i.e., b+1 is invertible.
b) Induct on n, n≥1, to prove that an−1 is invertible. For x=−1 we get that a−1 is invertible. Assume that b=an−1 is invertible. From the hypothesis it follows that a−b−1 is also invertible, and so is ab−1=(a−b−1)b. Moreover,
an+1−1=a+(a(an−1)−1)=a+(ab−1)
is invertible. Since A is finite, there exist two integers q>p≥1 such that ap=aq, i.e. ap(aq−p−1)=0. Since aq−p−1 is invertible, it follows that ap=0.