# Rationality of sum of reciprocals of irreducible polynomial

(Sorry for my poor english..)

I have a question in number theory. (Just my curiosity)

Let $$f(z)\in \mathbb{Z}[x]$$ be an irreducible polynomial with degree is larger than or equal to 2. Then is the number $$$$\sum_{n=1}^{\infty} \frac{1}{f(n)}$$$$ irrational? Or there is an irreducible polynomial $$f(x)\in \mathbb{Z}[x]$$ such that this number is rational?

1. More generally, if for all $$n\in \mathbb{N}$$, $$f(x)$$ and $$f(x+n)$$ are relativlely prime in $$\mathbb{Z}[x]$$, then is it true..?

I also know that proving the irrationality is very difficult, for example $$\zeta(2k+1)$$...

I could not find references about this problem.. Has this problem been proved? Or, is there a name for this problem?

• To give an example showing that this is probably a very deep question in general: when $f(n)=n^2+1$, the series evaluates to $\big( \pi(e^{2\pi}+1)/(e^{2\pi}-1) - 1\big)/2$ (says Mathematica); and I don't believe that number is known to be irrational...! – Greg Martin Sep 26 '18 at 17:53
• @GregMartin You would think so, but you would actually be wrong! Algebraic relations between $\pi$ and $e^\pi$, or rather lack thereof, are much better understood that those between $\pi$ and $e$ - indeed, they are algebraically independent, rendering that number (provably!) transcendental. See Wikipedia – Wojowu Sep 26 '18 at 18:01
• Is there any expectation that every absolutely convergent series of this form is an (exponential) period? Or a trick to show this? – Marty Sep 26 '18 at 21:07
• @Marty: For $f$ even and free of multiple roots, I think all is explained by the partial fraction expansion of the cotangent function. On all of the complex plane $\mathbb{C} \ni z$, we have $\pi \cot{\pi z = \lim_{N \to \infty}} \sum_{n = -N}^N \frac{1}{n+z}$. – Vesselin Dimitrov Sep 26 '18 at 21:51
• So $\sum_{n=-\infty}^\infty \frac{1}{f(n)}$ instead of $\sum_{n=1}^\infty \frac{1}{f(n)}$ @VesselinDimitrov – reuns Sep 26 '18 at 23:22