If the sum $$S=\frac14+\frac1{10}+\frac1{82}+\frac1{6562}+\cdots+\frac1{3^{2^{100}}+1}$$ is expressed in the form $\frac pq,$ where $p,q\in\mathbb N$ and $\gcd(p, q) =1.$ Then what is smallest prime factor of $p$ ?
We have: $$S=\sum_{k=0}^{100} \frac1{3^{2^k}+1}.$$ Please give me some hints to evaluate such kind of sums, in general.
Remark.
Since $4$ divides the denominator of the zeroth term ($\frac14$) of $S$, but does not divide the denominator of any other term, we can see that $2\nmid p$.
Note that each term of $S$ is $\equiv 1\pmod{3}$. Therefore, $S\equiv 101\cdot 1\equiv 2\pmod{3}$, so $3\nmid p$.
Since $5$ divides the denominator of the first term ($\frac1{10}$) of $S$, but not the denominator of any other term, we conclude that $5\nmid p$.
I'm not sure what kind of problem this is.
It seems that the problem is not intended to be done by hand.
If we denote by $S(n)$ the number $\displaystyle\sum_{k = 0}^n \frac 1{3^{2^k} + 1}$, then the first several values of the numerator of $S(n)$ look like this:
Of course, it is easy to show that $3$ divides the numerator if $n \equiv 2 \pmod 3$. This would be a much more reasonable exercise in elementary number theory.
However we have here $n = 100$. This leads to something without a pattern.
With the help of some computer algebra system, I am able to find that the smallest prime factor of the numerator is $37$. This is done by checking all primes up to $37$ one by one. So it is not really doable with paper and pencil.