We show that, if some integer n≥2 satisfies the required condition, then so does n2+a−1. Since n2+a−1>n, it is then sufficient to exhibit a single n≥2 that fits the bill.
Consider the identity
(n2+a−1)2+(n2+a−1)+a=(n2+n+a)(n2−n+a).(∗)
Suppose now that n! is divisible by n2+n+a for some integer n≥2. Since n<n2−n+a<n2+a−1, it follows that (n2+a−1)! is divisible by n!(n2−n+a); and since n! is divisible by n2+n+a, the identity (*) then shows (n2+a−1)! divisible by (n2+a−1)2+(n2+a−1)+a.
Next, set n=a in (*) to get (a2+a−1)2+(a2+a−1)+a=a3(a+2). If a≥3, then a<a+2<a2<a2+a−1, so a2+a−1≥2 and (a2+a−1)! is divisible by a3(a+2)=(a2+a−1)2+(a2+a−1)+a.
If a=2, set n=10, and notice that n!=10! is divisible by 2⋅7⋅8=112=102+10+2=n2+n+2=n2+n+a.
Finally, if a=1, set n=16, and notice that n!=16! is divisible by 3⋅7⋅13=273=162+16+1=n2+n+1.
Alternative solution:
The argument hinges on the fact that the equation u2−4a(a+2)v2=1−4a has infinitely many solutions (uk,vk) in positive integers, where the uk and the vk both form strictly increasing sequences.
Assume this for the moment. Clearly, the uk are all odd, so the nk=21(uk−1) are all integers. Consider first the case a≥2. Since the uk form a strictly increasing sequence of positive integers, so do the nk, and, since a≥2,
nk=21(uk−1)>(a(a+2)uk2+4a−1)21=2vk,
for all but finitely many indices k.
Recalling that the vk also form a strictly increasing sequence of positive integers, it follows that the numbers a(a+2)<vk<2vk all three occur in nk!, so the latter is divisible by a(a+2)⋅vk⋅2vk=2a(a+2)vk2=21(uk2+4a−1)=21(uk−1)2+(uk−1)+2a=2(nk2+nk+a), for all but finitely many indices k, as desired.
Finally, consider the equation u2−4a(a+2)v2=1−4a along with the associated Pell equation x2−4a(a+2)y2=1. The pair (u0,v0)=(2a+1,1) solves the former and the pair (x0,y0)=(2(a+1)2−1,a+1) solves the latter. For each positive integer k, let uk=u0xk+4a(a+2)v0yk and let vk=v0xk+u0yk, where xk=x0xk−1+4a(a+2)yk−1 and yk=y0xk−1+x0yk−1. Since the (xk,yk) solve the Pell equation, the (uk,vk) solve the equation under consideration. Clearly, the xk,yk,uk,vk all form strictly increasing sequences of positive integers. This completes the proof for a≥2.
The case a=1 is dealt with similarly by considering the equation u2−28v2=−3 along with the associated Pell equation x2−28y2=1. The pair (u0,v0)=(5,1) solves the former and the pair (x0,y0)=(127,24) solves the latter. The rest of the argument goes through verbatim.