Proving Zeta Relations Without Direct Evaluation

81 Views Asked by At

Is it possible to derive the following $\zeta$ relations without actually finding the values themselves?

\begin{align*} 2 \zeta(2)^2 &= 5\zeta(4)\\ 4\zeta(2)\zeta(4) &= 7\zeta(6)\\ 3\zeta(2)\zeta(6) &= 5\zeta(8) \end{align*}

These small integer relations make it look like there is a nice relation between these values.

However, the pattern breaks at the next one and the simplest relation I can find is $4+6=10$

$$10\zeta(4)\zeta(6) = 11\zeta(10)$$

I thought of using the integral representation of $\zeta(s)$ but I do not see an obvious way forward.

1

There are 1 best solutions below

0
On

$\mathbf{\text{Hint:}}$(after seeing that you are much interested for even zeta values)

The pattern you look towards is seen from the fact that $$\zeta(2n)=\frac{(-1)^{n+1}(2\pi)^{2n}B_{2n}}{2(2n)!}$$ Where $B_n$ are the Bernoulli numbers and $n\in\Bbb{N}$