For p=2 we have b2a2=611, and therefore gcd(2,b2)=2.
For p an odd prime we propose two solutions to this problem:
First solution. We have
bpap=(p2−1)!∑k=1p2−1k(p2−1)!.
Because the only multiples of p appearing in 1×2×⋯×(p2−1) are p,2p,…,(p−1)p, and p is a prime number, pp−1 is the largest power of p dividing (p2−1)!.
For 1≤k≤p2−1, if k is not a multiple of p then pp−1 divides k(p2−1)!. If k is a multiple of p then pp−2 is the largest power of p dividing k(p2−1)!.
Summing up the terms where k is a multiple of p we obtain
j=1∑p−1jp(p2−1)!=p!(p2−1)!j=1∑p−1j(p−1)!
Notice that all these fractions are integers. Using the fact that for each integer 1≤j≤p−1 there is a unique inverse integer 1≤ψ(j)≤p−1 modulo p (i.e. jψ(j)≡1modp) and that this inverse map j→ψ(j) is a permutation of {1,2,…,p−1}, we get
j=1∑p−1j(p−1)!≡j=1∑p−1j(p−1)!jψ(j)≡(p−1)!j=1∑p−1ψ(j)≡(p−1)!j=1∑p−1j≡(p−1)!2p(p−1)≡0,(modp)
This proves that pp−1 divides ∑j=1p−1jp(p2−1)! and so it divides ∑k=1p2−1k(p2−1)!.
We deduce that pp−1 cancels from the denominator and therefore
gcd(p,bp)=1.
Second solution. Because p is odd, we have
bpap=k=1∑2p2−1(k1+p2−k1)=k=1∑2p2−1k(p2−k)p2
For 1≤k≤2p2−1, if k is not a multiple of p then the denominator k(p2−k) is not divisible by p. If k is a multiple of p then
k(p2−k)p2=j(p−j)1,
where k=jp for some integer 1≤j≤2p−1. But the denominator j(p−j) is not divisible by p. Therefore bpap is a sum of fractions where none of the denominators is divisible by p and, since bpap is irreducible, then gcd(p,bp)=1.