Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Is there a known geometric proof for this famous problem? $$\zeta(2)=\sum_{n=1}^\infty n^{-2}=\frac16\pi^2$$

Moreover we can consider possibilities of geometric proofs of the following identity for positive even inputs of the Zeta function: $$ \zeta(2n)=(-1)^{n+1} \frac{B_{2n}(2\pi)^{2n}}{2(2n)!}$$ and negative inputs: $$ \zeta(-n)=-\frac{B_{n+1}}{n+1}$$

share|improve this question
9  
There are many nice proofs at math.stackexchange.com/questions/8337/… . Is one of them geometric enough for you? (David Speyer's answer might fit the bill.) –  Qiaochu Yuan Aug 27 '12 at 1:04
3  
Here's a proof by Prof. Greene at UCLA: math.ucla.edu/~greene/How%20Geometry.pdf –  Isaac Solomon Aug 27 '12 at 2:20
    
thank you Qiaochu one of them is ;) –  progressiveforest Aug 27 '12 at 2:44
    
I wanted to ask this question some time ago. I'm glad to see it posted. (I thought of an elementary geometrical proof) –  Chris's sis Aug 27 '12 at 5:25
    
do you mean that you have an elementary geometric proof? Would you mind sharing it? –  progressiveforest Aug 27 '12 at 5:42

3 Answers 3

I'm not sure what you mean by a geometric proof, but the following should fit the bill, as here the identity is deduced from a comparison of two areas: the first area is

$\displaystyle\int_{[0,1]^2} \frac{1}{1 - xy} \frac{dx dy}{\sqrt{xy}}$

and the second is

$4 \displaystyle \int_{\substack{\xi, \eta>0 \\ \xi \eta \leq 1}} \frac{ d \xi \, d \eta}{(1+\xi^2)(1+\eta^2)}$;

They are equal by a change of variables. For the first quantity, expand $(1-xy)^{-1}$ as a geometric series and integrate term-wise to get $3 \cdot \zeta(2)$. The second can be computed to $\pi^2/2$. This can be found in detail in Kontsevich and Zagier's Periods (bottom of page 8), where they attribute the idea to Calabi. (You should easily be able to hunt down an identity of $\zeta (n)$ being equal to an integral over the unit square in $\mathbb{R}^n$, like the first above. It might be a fun exercise to see if this idea is can be adapted for even $n$.)

share|improve this answer

Squares of side 1/N pack into a rectangle of that area.

Packing Squares with Side 1/n

share|improve this answer
5  
This is not a proof, unless you are able to intrinsically show that the envelope of the finite packings has area $\frac{\pi^2}{6}$ –  pre-kidney Jan 11 '13 at 12:37

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.