Calculate the following definite integral

$$\int_{- \infty}^{\infty}\frac{\sin 3x}{x(x^2+1)}dx$$

using the residue theorem

This was my attempted solution:

Solution: $$\int_{- \infty}^{\infty}\frac{\sin 3x}{x(x^2+1)}dx = \int_{-\infty}^{\infty} \operatorname{Im}\left[\frac{e^{i3z}}{z(z^2+1)}\right]dz = \operatorname{Im}\left[\int_{- \infty}^{\infty} \frac{e^{i3z}}{z(z^2+1)} dz\right]$$

Now note that $f(z) = \frac{e^{i3z}}{z(z^2+1)}$ has singularities at $z_0 = 0$ and $z_1 = i$ and $z_2 = -i$ then using the residue theorem we have $$\int_{- \infty}^{\infty} \frac{e^{i3z}}{z(z^2+1)} dz = 2\pi i \sum_{k=0}^2 \operatorname{Res}_{z = z_k} f(z)$$

Then calculating the residues at $z_0$, $z_1$ (we don't need to calcualte $z_2$ since it's not in the upper half-plane) we get $\operatorname{Res}_{z = z_0} f(z) = 1$ and $\operatorname{Res}_{z = z_1} f(z) = \frac{1}{-2e^3}$

Hence $$ \sum_{k=0}^2 \operatorname{Res}_{z = z_k} f(z) = 1- \frac{1}{2e^3}$$ and we have that $$\operatorname{Im}\left[\int_{- \infty}^{\infty} \frac{e^{i3z}}{z(z^2+1)} dz\right] = \pi\left(2 -\frac{1}{e^3} \right)$$ and thus $$\int_{- \infty}^{\infty}\frac{\sin 3x}{x(x^2+1)}dx = \pi\left(2 -\frac{1}{e^3} \right)$$

However the solution that my lecturer gave was $\pi\left(1 -\frac{1}{e^3} \right)$, is his solution incorrect or is mine incorrect and if so why?


The singularity at $z=0$ is a simple pole on the contour, so it contributes half its residue.

To avoid the singularity, use an contour indented with a small semicircle of radius $\varepsilon$, expand the integrand about the pole using the Laurent expansion, and evaluate the integral around the small semicircle using this, and take $\varepsilon \downarrow 0$. (And this gives the half in the simple pole case.)

It won't always just be half the residue for higher-order poles, but you can still do the calculation with an indented contour to get the correct answer.

  • $\begingroup$ It seems like the first singularity always contributes half, is this always the case? $\endgroup$ – Perturbative Jun 2 '18 at 15:23
  • $\begingroup$ What do you mean by "first"? $\endgroup$ – Chappers Jun 2 '18 at 15:29
  • $\begingroup$ Sorry I explained that poorly, disregard that last comment of mine. What I meant was this, take this integral $$\int_{-\infty}^{\infty}\frac{\sin x -x}{x^3} dx$$. To calculate that I'll take the imaginary part of the integral $$\int_{-\infty}^{\infty} \frac{e^{iz}-z}{z^3}dz$$ and apply the residue theorem. Then note that $z_0 = 0$ is a pole (the only pole) of order $n =3$, however it also contributes half it's residue. (This is judging by the answer I get which is 2x the actual answer). I don't see how your line of reasoning extends to this case. $\endgroup$ – Perturbative Jun 2 '18 at 15:39
  • $\begingroup$ Ah, right. The point is that writing down half the residue won't always work. The indentation and take the limit technique will work. $\endgroup$ – Chappers Jun 2 '18 at 15:45
  • $\begingroup$ Could you explain that technique please in your answer if possible? $\endgroup$ – Perturbative Jun 2 '18 at 15:57

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.