Well here's another proof that \(\displaystyle \sum _{ n=1 }^{ \infty }{ \frac { 1 }{ { n }^{ 2 } } }= \frac{{\pi}^{2}}{6}\)

We will start from the integral \(\displaystyle I=\int _{ 0 }^{ \pi /2 }{ ln(cos(x))dx } \)

now \(cos(x)=\dfrac{e^{ix}+e^{-ix}}{2}\)

\(\Rightarrow I=\displaystyle \int _{ 0 }^{ \pi /2 }{ ln(\frac { { e }^{ ix }+{ e }^{ -ix } }{ 2 } )dx } \)

\( I = \displaystyle \int _{ 0 }^{ \pi /2 }{ ln(1+{ e }^{ 2ix })dx } - \int _{ 0 }^{ \pi /2 }{ (ln(2)+ix)dx } \)

Considering taylor expansion of \(ln(1+x)\) we have :

\(\displaystyle I = \int _{ 0 }^{ \pi /2 }{ \sum _{ r=1 }^{ \infty }{ \frac { { (-1) }^{ r-1 } }{ r } { e }^{ 2irx } } dx } -(\dfrac{\pi ln(2)}{2}+\frac{{\pi}^{2}i}{8})\)

Changing the order of integration and summation we have :

\(\displaystyle I = \sum _{ r=1 }^{ \infty }{ \frac { { (-1) }^{ r-1 } }{ r } \int _{ 0 }^{ \pi /2 }{ { e }^{ 2irx }dx } } -(\dfrac{\pi ln(2)}{2}+\frac{{\pi}^{2}i}{8})\)

Integrating and putting limits we get :

\(\displaystyle I= \sum _{ r=1 }^{ \infty }{ \frac { { (-1) }^{ r-1 } }{ 2i{ r }^{ 2 } } ( } { e }^{ \pi ir }-1) - (\dfrac{\pi ln(2)}{2}+\frac{{\pi}^{2}i}{8}) \)

Now \({e}^{\pi ir} = {(-1)}^{r} \)

Hence \(\displaystyle I = (\sum _{ r=1 }^{ \infty }{ \frac { 1-{ (-1) }^{ r } }{ 2{ r }^{ 2 } } } )i - (\dfrac{\pi ln(2)}{2}+\frac{{\pi}^{2}i}{8})\)

We will be calculating the value of our summation.

\(S=(1+\dfrac { 1 }{ { 3 }^{ 2 } } +\dfrac { 1 }{ { 5 }^{ 2 } } +......)\)

\(S=(1+\dfrac { 1 }{ { 2 }^{ 2 } } +\dfrac { 1 }{ { 3 }^{ 2 } } +\dfrac { 1 }{ { 4 }^{ 2 } } +......)-(\dfrac { 1 }{ { 2 }^{ 2 } } +\dfrac { 1 }{ { 4 }^{ 2 } } +....)\)

\(S=(1+\dfrac { 1 }{ { 2 }^{ 2 } } +\dfrac { 1 }{ { 3 }^{ 2 } } +\dfrac { 1 }{ { 4 }^{ 2 } } +......)-\dfrac { 1 }{ 4 } (\dfrac { 1 }{ { 1 }^{ 2 } } +\dfrac { 1 }{ { 2 }^{ 2 } } +....)\)

\(S=(\dfrac { 3 }{ 4 } )(1+\dfrac { 1 }{ { 2 }^{ 2 } } +\dfrac { 1 }{ { 3 }^{ 2 } } +\dfrac { 1 }{ { 4 }^{ 2 } } +......)\)

\(S=(\dfrac { 3 }{ 4 } )\zeta (2)\)

Putting the value of \(S\) in our integral we have :

\(I=\dfrac { -\pi ln(2) }{ 2 } +(\dfrac { 3\zeta (2) }{ 4 } -\dfrac { { \pi }^{ 2 } }{ 8 } )i\)

Now since the integral is real hence imaginary part of our integral is \(0\) hence :

\(\dfrac { 3\zeta (2) }{ 4 } =\dfrac { { \pi }^{ 2 } }{ 8 } \)

Finally :

\(\zeta(2) = \dfrac{{\pi}^{2}}{6}\)

## Comments

Sort by:

TopNewestVery slick! :D

Just curious though: what motivated you to start with that particular integral?

Log in to reply

This is amazing!!!

Log in to reply

wow Ingenious .. and much impressive ! You have really genius brain . I'am much impress with your skills ! Hat's off

Log in to reply

Nicely Done \(\ddot\smile\)

Log in to reply

Is this an original proof? If not, where did you get the proof from? It is a remarkable and elegant one.

Log in to reply

It is an original proof.

Log in to reply

That's wonderful! If you find analogues for \(2n\) please do share with us.

Log in to reply

@Jake Lai

Hi dude , can't help but smile when I see your Profile pic xd .

Just curious to know why you changed your profile pic ?

Log in to reply

Wow wonderful !

Log in to reply

Its a superb proof.

Please tell the motivation :D

Log in to reply

wow ! .....

imaginary part must be zero...... i would have been foolish enough to leave it seeing that we got a unreal value .... its great tht u carried it over ........ i learned from this :) ...... thanks for a beautiful derivationLog in to reply

Hi Ronak , is your B'day on the 16th of March ?

Log in to reply

Nope on 16th March Result of INCHO will be announced. I am very much excited for that.

Log in to reply

Ok , so when is your B'day then ?

Log in to reply

Log in to reply

@Ronak Agarwal

My birthday is on 14th July!!!Log in to reply

Log in to reply

Can this be done? -

\(\displaystyle I=\int _{ 0 }^{ \pi /2 }{ ln^2(cos(x))dx } \) if yes, can this also - \(\displaystyle I=\int _{ 0 }^{ \pi /2 }{ ln^n(cos(x))dx } \) \( n \to N\)

Log in to reply

Yes this can be done I will be posting this as a problem soon.

Log in to reply

Comment deleted Feb 23, 2015

Log in to reply

As @Ronak Agarwal said, there is no closed form for \(\zeta(3),\) and there isn't a closed form for any \(\zeta(2k+1)\) given positive integer \(k\).

An interesting result I found a while back is that all of the \(\zeta(2k)\) values with all of their \(\pi^{2k}\) numerators and strange denominators can be manipulated to sum to a nice result: \[\sum_{k=1}^\infty(\zeta(2k)-1)=\dfrac{3}{4}\] It just goes to show how looking at a problem from a new perspective can lead you to a surprisingly simple result.

Log in to reply

I believe it must be \(\dfrac{3}{4}\) , I think you have quoted the wrong value as you can see \(\zeta{(2)}-1=\dfrac{{\pi}^{2}}{6} -1 > 0.5 \)

Log in to reply

Log in to reply

Ha Ha Ha!!! you know very well that till date no such closed form of \(\zeta{(3)}\) has been found

Log in to reply

@Ronak Agarwal Could you tell me when can we change the order of summation and integration.

Log in to reply