I'm trying to prove following identity

$$\int_0^\infty \frac{\Xi(t)}{t^2 + \frac{1}{4}} \cos(xt) dt = \frac{1}{2} \pi (e^{\frac{1}{2}x} - 2e^{-\frac{1}{2}x} \psi(e^{-2x}))$$


$$\psi(x)=\sum_{n = 1}^{\infty} e^{-n^2 \pi x}$$


$$\Xi(t) = \xi(\frac{1}{2} + it)$$

is the xi function on the critical line.


I only don't understand the following equality

$$-\frac{1}{4 i \sqrt{y}} \int_{\frac{1}{2} - i\infty}^{\frac{1}{2} + i\infty} \Gamma(\frac{1}{2}s) \pi^{-\frac{1}{2}s} \zeta(s) y^s ds = -\frac{\pi}{\sqrt{y}} \psi(\frac{1}{y^2}) + \frac{1}{2} \pi \sqrt{y}.$$

Where does latter summand $\frac{1}{2} \pi \sqrt{y}$ come from? When I write $(\frac{1}{y})^{-s}$ and substitute with $s = 2w$ I only get the first summand. We know that by the Mellin transform theorem $\psi(x)$ can be recovered by the Mellin transform integral over $\Gamma(s) \pi^{-s} \zeta(2s)$.


This is from Titchmarsh's book "The Theory of the Riemann Zeta-function" p. 35−36

enter image description here

  • $\begingroup$ you are allowed to upvote and accept my answer :) (the second part makes everything rigorous) and you easily get $\int_0^\infty \frac{\Xi(t)}{t^2 + \frac{1}{4}} \cos(xt) dt = \frac{1}{2} \pi (e^{\frac{1}{2}x} - 2e^{-\frac{1}{2}x} \psi(e^{-2x}))$ from it, since $\pi\int_0^\infty \frac{\Xi(t)}{t^2 + \frac{1}{4}} \cos(xt) dt =\frac{1}{2 i \pi} \int_{1/2-i\infty}^{1/2+i \infty} E(s) e^{x (s-1/2)} ds$ $\endgroup$
    – reuns
    Jun 29, 2016 at 3:28

1 Answer 1


Let $f(x) = 2\sum_{n = 1}^\infty e^{- \pi n^2 x^2}$ and $E(s) = \pi^{-s/2} \Gamma(s/2)\zeta(s)$.

For $Re(s) >0$ : $2\int_0^\infty x^{s-1} e^{-\pi n^2 x^2} dx = n^{-s} \pi^{-s/2}\Gamma(s/2) $ so we have for $Re(s) > 1$ :

$$E(s) = \Gamma(s/2) \pi^{-s/2} \zeta(s) = 2\sum_{n=1}^\infty \int_0^\infty x^{s-1} e^{-\pi n^2 x^2} dx = \int_0^\infty f(x) x^{s-1} dx \quad (1)$$

By inverse Mellin transform : $$f(x) = \frac{1}{2 i \pi}\int_{\sigma-i\infty}^{\sigma+i\infty} E(s) x^{-s} ds$$

but this is only true for $\sigma > 1$, since $E(s)$ has a pole at $s=1$.

Note that for $Re(s) > 1$ : $\displaystyle\quad\frac{1}{s-1} = \int_0^1 x^{s-2} dx = \int_0^\infty \frac{1_{x < 1}}{x} x^{s-1} dx $,

and for $Re(s) < 1$ : $\displaystyle\quad-\frac{1}{s-1} = \int_1^\infty x^{s-2} dx = \int_0^\infty \frac{1_{x > 1}}{x} x^{s-1} dx $

Hence, at least for $Re(s) > 1$ :

$$E(s) - \frac{1}{s-1} = \int_0^\infty \left( f(x)- \frac{1_{x < 1}}{x}\right) x^{s-1} dx \qquad (2) $$

indeed, this is true also for $Re(s) > 0$ (see below) and we get, for $Re(s) \in (0,1)$ :

$$E(s) = \int_0^\infty \left( f(x) - \frac{1}{x}\right) x^{s-1} dx$$

Finally, by inverse Mellin transform, for $\sigma \in (0,1)$ : $$f(x) - \frac{1}{x} = \frac{1}{2 i\pi}\int_{\sigma -i \infty}^{\sigma +i \infty} E(s) x^{-s}dx$$ and $\displaystyle f(1/y) - y = \frac{1}{2 i\pi} \int_{\sigma -i \infty}^{\sigma +i \infty} E(s) y^{s}dy$ as expected.

You can show $(2)$ converges for $Re(s) > 0$ by proving $\displaystyle\theta(x) = 1+ f(x) = \sum_{n =- \infty}^\infty e^{-\pi n^2 x^2}$ fulfills the functional equation $$\theta(1/x) = x\,\theta(x)$$ using the Poisson summation formula (see this proof).

So that $\displaystyle f(1/x) = \theta(1/x)-1= x \,\theta(x)-1 = x(f(x)+1)-1$ and : $$\int_0^1 \left(f(x)- \frac{1}{x}\right) x^{s-1} dx = \int_1^\infty (f(1/y) - y) y^{-s+1} \frac{dy}{y^2} $$ $$= \int_1^\infty (y (f(y)+1)-1-y) y^{-s-1} dy = \int_1^\infty (y f(y)-1) y^{-s-1} dy$$ whence $$E(s) - \frac{1}{s-1} = \int_0^1+\int_1^\infty \left(f(x)-\frac{1_{x<1}}{x}\right) x^{s-1}dx $$ $$= \int_1^\infty \left( f(x) (x^{s-1} + x^{-s}) - x^{-s-1}\right) dx$$ converges for $Re(s) > 0$.

Riemann obtained the functional equation $E(s) = E(1-s)$ from $$E(s) - \frac{1}{s-1} + \frac{1}{s} = \int_1^\infty f(x) (x^{s-1} + x^{-s}) dx$$ which is entire.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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