MHB Evaluation of an Integral (using contours or otherwise)

  • Thread starter Thread starter roninpro
  • Start date Start date
  • Tags Tags
    Integral
roninpro
Messages
2
Reaction score
0
Hello everyone - this is my first post here since the move. I had to deal with a few engineering students yesterday; they asked me how to exactly evaluate the integral

[math]\int_{0}^\infty \frac{x^3}{e^x-1} \text{d}x[/math]

It looked like this integral would best be handled by treating it as complex-valued and setting up a contour so that the complex parts dropped out. However, the students did not know any complex analysis and were looking for a more elementary approach. It wasn't obvious how to do it using real integrals (and I tried to use tricks as in the evaluation of [math]\int_{-\infty}^\infty e^x \text{d}x[/math]). I also discussed this with another graduate student, and we were actually even unable to select a decent contour to take the integral. (Though we did put it into Mathematica which computed [math]\frac{\pi^4}{15}[/math].) We're basically baffled. I was hoping to find some advice about this integral here. Any help would be appreciated!
 
Physics news on Phys.org
roninpro said:
Hello everyone - this is my first post here since the move. I had to deal with a few engineering students yesterday; they asked me how to exactly evaluate the integral

[math]\int_{0}^\infty \frac{x^3}{e^x-1} \text{d}x[/math]

It looked like this integral would best be handled by treating it as complex-valued and setting up a contour so that the complex parts dropped out. However, the students did not know any complex analysis and were looking for a more elementary approach. It wasn't obvious how to do it using real integrals (and I tried to use tricks as in the evaluation of [math]\int_{-\infty}^\infty e^x \text{d}x[/math]). I also discussed this with another graduate student, and we were actually even unable to select a decent contour to take the integral. (Though we did put it into Mathematica which computed [math]\frac{\pi^4}{15}[/math].) We're basically baffled. I was hoping to find some advice about this integral here. Any help would be appreciated!
Write it as $\displaystyle\int_{0}^\infty \frac{x^3}{e^x-1} \,dx = \int_{0}^\infty \frac{x^3}{e^x(1-e^{-x})}\,dx = \int_{0}^\infty x^3e^{-x}\sum_{n=0}^\infty e^{-nx}\,dx$ (binomial series). Then (if you can justify integrating term by term) this is equal to $\displaystyle\sum_{n=1}^\infty \int_{0}^\infty x^3e^{-nx}\,dx.$

Integrate by parts several times to find that $\displaystyle \int_0^\infty x^3e^{-nx}\,dx = \frac 6{n^4}.$ The result then follows from the fact that $\displaystyle\sum_{n=1}^\infty\frac1{n^4} = \frac{\pi^4}{90}.$
 
We actually originally tried to interpret the term [math]\frac{1}{e^x-1}[/math] as a geometric series, but convergence was an issue. Your rearrangement fixes that issue. Thanks very much, this works for me!
 
With systematic application of integration by part You arrive to the general formula...

$\displaystyle \int t^{m}\ \ln ^{n} t\ dt = n!\ t^{m+1}\ \sum_{k=0}^{n}\frac{(-1)^{k}\ \ln^{n-k} t}{(m+1)^{k}\ (n-k)!} +c$ (1)

... and from (1) with simple steps...

$\displaystyle \zeta(n)= \sum_{k=1}^{\infty} \frac{1}{k^{n}}= \frac{(-1)^{n-1}}{(n-1)!} \int_{0}^{1} \frac{\ln^{n-1} t}{1-t}\ dt$ (2)

... where $\zeta(*)$ is the so called 'Riemann Zeta Function' and $n>1$. Setting in (2) $\ln t=x$ You obtain the equivalent relation...

$\displaystyle \zeta(n)= \frac{1}{(n-1)!} \int_{0}^{\infty} \frac{x^{n-1}}{e^{x}-1}\ dx$ (3)

Kind regards

$\chi$ $\sigma$
 
Back
Top