I have a problem with my solution. $$\int_{0}^{\infty}\frac{2x}{e^{x}-e^{-x}}dx=\int_{0}^{a}\frac{2x}{e^{x}-e^{-x}}dx+\int_{a}^{\infty}\frac{2x}{e^{x}-e^{-x}}dx$$

So in first integral, if I compare it with : $ \frac{1}{\sqrt x}$ I have: $$\lim_{x \to 0^+ }\frac{\frac{2x}{e^{x}-e^{-x}}}{\frac{1}{\sqrt x}} =0$$ and I also know that $\int_{0}^{a}\frac{1}{\sqrt x}$ converges. And I can conclude that my integral also converges. But if I choose to compare it with $\frac{1}{x^2}$ , again I get $$\lim_{x \to 0^+ }\frac{\frac{2x}{e^{x}-e^{-x}}}{\frac{1}{x^2}} =0$$ but now I conclude that my integral diverges together with $\int_{0}^{a}\frac{1}{x^2}$. What the right solution and what is wrong with my solution??! Thanks a lot!


Expand the denominator as follows:

$$\begin{align}\int_0^{\infty} dx \frac{2 x}{e^x-e^{-x}} &= 2 \int_0^{\infty} dx \frac{x \, e^{-x}}{1-e^{-2 x}} \\ &= 2 \sum_{k=0}^{\infty} \int_0^{\infty} dx\, x \, e^{-(2 k+1) x}\\ &= 2 \sum_{k=0}^{\infty} \frac1{(2 k+1)^2}\\ &= 2 \left [\sum_{k=1}^{\infty} \frac1{k^2} - \sum_{k=1}^{\infty}\frac1{(2 k)^2} \right ] \\ &= \frac{3}{2} \sum_{k=1}^{\infty} \frac1{k^2} \\ &= \frac{\pi^2}{4}\end{align}$$

  • $\begingroup$ Thank you, but what is wrong with my solution? $\endgroup$ – Laura Feb 10 '14 at 7:08
  • $\begingroup$ $$e^x - e^{-x} \sim 2 x \quad (x \to 0)$$ $\endgroup$ – Ron Gordon Feb 10 '14 at 7:10
  • $\begingroup$ Thanks again, can you please explain why? $\endgroup$ – Laura Feb 10 '14 at 7:20
  • $\begingroup$ @Laura: from the Taylor expansion of an exponential: $e^x = 1+x+x^2/2+\cdots$ $\endgroup$ – Ron Gordon Feb 10 '14 at 7:42

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.