5
$\begingroup$

How does one get from this $$\zeta(s)=\sum_{k=1}^{\infty}\frac1{k^s}$$ to the integral representation $$\zeta(s)=\frac1{\Gamma(s)}\int_{0}^\infty \frac{x^{s-1}}{e^x-1}dx$$ of the Riemann Zeta function?

I can see that it can be rewritten as

$$\Gamma(s)\zeta(s)=\int_{0}^\infty \frac{x^{s-1}}{e^x-1}dx$$ and the Gamma function as an integral yields

$$\zeta(s)\int_{0}^\infty \frac{x^{s-1}}{e^x}dx=\int_{0}^\infty \frac{x^{s-1}}{e^x-1}dx$$

But this approach does not work as the right integral does not converge. So how does one go from the summation to the integral representation?

$\endgroup$
  • 2
    $\begingroup$ A reference: This is derived on page nine of Riemann's Zeta Function by Edwards. $\endgroup$ – apnorton Jan 13 '15 at 20:23
10
$\begingroup$

Recall that for $t>1$, $$\frac{1}{t-1}=\sum_{n=1}^\infty t^{-n}$$

Then substitute $t=e^{x}$ for $x>0$.

Then substituting $x=\frac{v}{n}$ in the $n$th term of the integral, you get:

$$\int_0^{\infty} x^{s-1}e^{-nx}\,dx=\frac{1}{n^s}\int_0^\infty v^{s-1}e^{-v}dv = \frac{1}{n^s}\Gamma(s)$$

$\endgroup$
  • $\begingroup$ Okay. So. Now you sum over $n$ on both sides and you have a geometric series in the integrand? $\endgroup$ – Eleven-Eleven Jan 14 '15 at 0:17
  • $\begingroup$ Yes. My real analysis memory is somewhat eroded, but you'll need an argument to show why you can switch the sum and integral. $\endgroup$ – Thomas Andrews Jan 14 '15 at 2:46
0
$\begingroup$

$$\mathscr{L}\{t^k\}=\int_0^\infty t^ke^{-st} = \dfrac{k!}{s^{k+1}}$$ is the well-known laplace transform of $t^{k}$

then like above in the first answer, sum over $s$ and swap integration and summation

$\endgroup$

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.