f[1, x] = (1 + x)^(1/2);
f[n_ /; n > 1, x] := f[n - 1, x] /. x :> (x + n^n)^(1/n!)
The above function can generate terms of the series. For example
$\quad \quad f(3,\,x)=\sqrt{1+\sqrt[2!]{2^2+\sqrt[3!]{3^3+x}}}$
$\quad \quad f(4,\,x)=\sqrt{1+\sqrt[2!]{2^2+\sqrt[3!]{3^3+\sqrt[4!]{4^4+x}}}}$
f
can be used to study the convergence of the OP's series.
Table[NumberForm[N @ f[i, x] /. x -> 0, 16], {i, 7}]
{1.`, 1.7320508075688772`, 1.8423273788801293`, 1.8430758571923342`,
1.8430759846682`, 1.8430759846685443`, 1.8430759846685443`}
So with MachinePrecision
, the sequence in question converges to 1.8430759846685443 with MachinePreision
.
(#2^#2 + #1)^(1/#2!) &
, notSqrt[#2^#2 + (#1)]^#2! &
. It looks like the limit is1.84307598466...
, not infinity. $\endgroup$Sqrt[1 + ...]
rather than1 + (...)^(1/2!)
? @ChipHurst's suggestion seems "nicer". Maybe you really want the square root of that result, but it could perhaps help to clarify this. $\endgroup$