While solving another problem I have come across this integral which I am unable to evaluate. Can someone please evaluate the following integral? Thank you.


  • $\begingroup$ Are you familiar with methods of complex contour integration? (residue theorem). It should be fairly straight-forward. $\endgroup$ – orion May 26 '14 at 15:46
  • $\begingroup$ no i am not familiar with complex analytical methods $\endgroup$ – rockstar123 May 26 '14 at 15:59

Consider the following integral: $$I(a)=\int_0^{\pi} \frac{d\theta}{a+\cos\theta}$$ Rewrite $\cos\theta=\frac{1-\tan^2(\theta/2)}{1+\tan^2(\theta/2)}$ to get: $$I(a)=\int_0^{\pi} \frac{\sec^2(\theta/2)}{a+1+\tan^2(\theta/2)(a-1)}d\theta$$ Use the substitution $\tan(\theta/2)=t \Rightarrow \sec^2(\theta/2)d\theta=2\,dt$ to get: $$I(a)=\int_0^{\infty} \frac{2}{a+1+(a-1)t^2}\,dt$$ The above is easy to evaluate and gives: $$I(a)=\frac{\pi}{\sqrt{a^2-1}}$$ Hence, $$\int_0^{\pi} \frac{d\theta}{a+\cos\theta}=\frac{\pi}{\sqrt{a^2-1}}$$ Differentiate both the sides wrt $a$ to get: $$\int_0^{\pi} \frac{d\theta}{(a+\cos\theta)^2}=\frac{a\pi}{(a^2-1)^{3/2}}\,\,\,\,\,(*)$$ Getting back to the problem, you have, $$\int_0^{2\pi}\frac{d\theta}{(2+\cos\theta)^2}=2\int_0^{\pi} \frac{d\theta}{(2+\cos\theta)^2}$$ From $(*)$ and with $a=2$, the final answer is: $$2\frac{2\pi}{(3\sqrt{3})}=\boxed{\dfrac{4\pi}{3\sqrt{3}}}$$


Here's the contour integration approach.

$$\int_{0}^{2 \pi} \frac{1}{(2+ \cos \theta)^{2}} \ d \theta = \int_{0}^{2 \pi} \frac{1}{(2+\frac{e^{i \theta} + e^{- i \theta}}{2})^{2}} \ d \theta$$

Let $z=e^{i \theta}$.


$$ \begin{align}\int_{0}^{2 \pi} \frac{1}{(2+ \cos \theta)^{2}} \ d \theta &= \int_{|z|=1} \frac{1}{(2+\frac{z+z^{-1}}{2})^{2}} \frac{dz}{iz} \\ &= \frac{4}{i}\int_{|z|=1} \frac{z}{(z^{2}+4z+1)^{2}} \ d z \\ &=\frac{4}{i} \int_{|z|=1}\frac{z}{[(z-\sqrt{3}+2)(z+\sqrt{3}+2)]^{2}} \ dz \end{align}$$

The only pole inside the unit circle is at $z= \sqrt{3}-2$.


$$ \begin{align} \int_{0}^{2 \pi} \frac{1}{(2+ \cos \theta)^{2}} \ d \theta &= 2 \pi i \frac{4}{i} \text{Res} \left[ \frac{z}{[(z-\sqrt{3}+2)(z+\sqrt{3}+2)]^{2}}, \sqrt{3}-2 \right] \\ &= 8 \pi \lim_{z \to \sqrt{3}-2} \frac{d}{dz} \frac{z}{(z+\sqrt{3}+2)^{2}} \\ &= 8 \pi \lim_{z \to \sqrt{3}-2} \frac{(z+\sqrt{3}+2)^{2}-2z(z+\sqrt{3}+2)}{(z+\sqrt{3}+2)^{4}} \\ &= 8 \pi \lim_{z \to \sqrt{3}-2} \frac{-z+ \sqrt{3} + 2}{(z+\sqrt{3}+2)^{3}} \\ &= 8 \pi \frac{4}{24 \sqrt{3}} = \frac{4 \pi}{3 \sqrt{3}} \end{align}$$

  • $\begingroup$ @juantheron Thanks. $\endgroup$ – Random Variable May 31 '14 at 3:34

Here is an ugly way of doing it. Pranav's solution is much more elegant.

First substitute $u = \tan(\tfrac{\theta}{2})$ and $\mathrm{d}u = \tfrac{1}{2} \sec^2(\tfrac{\theta}{2})\mathrm{d}\theta$. Then, transform the integrand using the substitution $\sin(\theta) = \frac{2u}{u^2+1}$, $\cos(\theta) = \frac{1-u^2}{u^2+1}$ and $\mathrm{d}\theta = \frac{2\mathrm{d}u}{u^2+1}$ to get $$I = \int \frac{1}{(2+\cos(\theta))^2}\mathrm{d}\theta = \int \frac{2}{(u^2 + 1)\left(\tfrac{1-u^2}{u^2+1} +2\right)^2} \mathrm{d}u = 2\int \frac{u^2+1}{u^4 + 6u^2 +9} \mathrm{d}u.$$

Now, using partial fractions, one can rewrite this integral as $$I = \frac{2}{3}\int\frac{1}{\tfrac{u^2}{3}+1}\mathrm{d}u - 4\int\frac{1}{(u^2+3)^2}\mathrm{d}u.$$

For the first integrand, substitute $s = \frac{u}{\sqrt{3}}$ and $\mathrm{d}s = \frac{1}{\sqrt{3}}\mathrm{d}u$ to get $$I = \frac{2}{\sqrt{3}} \int \frac{1}{s^2+1} \mathrm{d}s - 4\int\frac{1}{(u^2+3)^2}\mathrm{d}u.$$ We know how to integrate the left side. This is just $$I = \frac{2\arctan(s)}{\sqrt{3}} - 4\int\frac{1}{(u^2+3)^2}\mathrm{d}u.$$

For the right-hand side, we substitute $u = \sqrt{3}\tan(p)$ and $\mathrm{d}u=\sqrt{3}\sec^2(p)\mathrm{d}p$. Then, we have that $(u^2+3)^2 = (3\tan^2(p) +2)^2 = 9\sec^4(p)$ and $p = \arctan(\tfrac{u}{\sqrt{3}})$. This yields the integral $$I = \frac{2\arctan(s)}{\sqrt{3}} - \frac{4}{3\sqrt{3}} \int \cos^2(p) \mathrm{d}p.$$ Integrating $\cos^2(p)$ is easy by using the identity $\cos^2(p) = \frac{\cos(2p) + 1}{2}$. All that remains now is to integrate this and substitute your way back to express $I$ in terms of $\theta$. Evaluate from $0$ to $2\pi$ to obtain $$I_0^{2\pi} = \frac{4\pi}{3\sqrt{3}}.$$

  • 3
    $\begingroup$ We have same approach bro. :D $\endgroup$ – Tunk-Fey May 26 '14 at 16:06
  • 1
    $\begingroup$ In all honesty, we have Wolfram's approach. ;) $\endgroup$ – Hubble May 26 '14 at 16:07


Using Weierstrass substitution, the integral turns out to be $$ \int_0^{2\pi}\frac{1}{(2+\cos\theta)^2}\,d\theta=2\int_0^{\pi}\frac{1}{(2+\cos\theta)^2}\,d\theta=4\int_0^\infty\frac{1+x^2}{(3+x^2)^2}\ dt. $$

Now, using partial fraction decomposition, the integrand turns out to be $$ \frac{1+x^2}{(3+x^2)^2}=\frac{1}{3+x^2}-\frac{2}{(3+x^2)^2}. $$ The last step, using substitution $x=\sqrt3\tan t$, the integral should be easy to be evaluated.

  • $\begingroup$ The term "Weierstrass substitution" is a misnomer perpetuated by the popularity of Stewart's calculus textbooks (which is probably not the earliest appearance of the term). $\endgroup$ – Michael Hardy Sep 6 '14 at 16:08
  • $\begingroup$ @MichaelHardy Indeed, I know that. You've mentioned it many times here on MSE. Do you prefer tangent-half substitution to Weierstrass substitution? $\ddot\smile$ $\endgroup$ – Tunk-Fey Sep 6 '14 at 16:32
  • $\begingroup$ I do prefer to call it that. History is not mathematics, but still it's not good to confuse people about it. $\endgroup$ – Michael Hardy Sep 6 '14 at 16:47
  • $\begingroup$ @MichaelHardy OK, as you wish Sir, I'll use the term tangent-half sub from now on. $\endgroup$ – Tunk-Fey Sep 6 '14 at 16:58

Of course you can apply the substitution $t=\tan \frac{\theta}{2}$ directly. $$d\theta =\frac{2dt}{1+t^2}$$ and $$\cos \theta =\frac{1-t^2}{1+t^2}$$ So $$2\int_0^{\pi}\frac{d \theta}{(2+\cos\theta)^2}=4\int_0^{\infty} \frac{t^2+1}{(t^2+3)^2}dt$$


Google the term "tangent half-angle substitution".

If $u = \tan\dfrac\theta 2$ then $\cos\theta=\dfrac{1-u^2}{1+u^2}$ and $d\theta = \dfrac{2\,du}{1+u^2}$.

As $\theta$ goes from $0$ to $2\pi$, then $u$ goes from $-\infty$ to $\infty$.

If you remember some trigonometry, you can derive what you see above.

\begin{align} & \int_0^{2\pi}\frac{d\theta}{(2+\cos\theta)^2} = \int_{-\infty}^\infty \frac{\left(\frac{2\,du}{1+u^2}\right)}{\left(2+\frac{1-u^2}{1+u^2}\right)^2} = \int_{-\infty}^\infty \frac{2(1+u^2)\,du}{(2(1+u^2)+(1-u^2))^2} \\[10pt] = {} & \frac 1 2 \int_{-\infty}^\infty \frac{1+u^2}{(3+u^2)^2} \, du \end{align}

There still a lot of work to do after that. Use partial fractions. If your difficulty was in the part that comes after that, this won't help; otherwise maybe it will.

Perhaps this is an argument for using complex variables; residues, etc.: pursuing what you see above is in some ways routine but is onerous.


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.