[Math] How does a simple elliptic integral solve this monster

contour-integrationelliptic integralsintegration

During some electromagnetics calculation regarding a loop antenna I stumbled across the following integral

$$\int_0^{\pi/2} \frac{d\phi}{\big(1+\frac{k}{k-2}\cos(2\phi)\big)^{3/2}}$$

and Mathematica gave me the solution

$$\frac{(1-\frac{k}{2})^{3/2}}{k-1} E\left( k \right)$$

where $E(k)$ is the complete elliptic integral of the second kind. Cool so far. Now I am not interested in the full solution process, but the following bothers me: When I write the above slightly differently and use the definition of $E(k)$, I get

$$\int_0^{\pi/2} \sqrt{1-k\sin^2(\phi)}d\phi = \frac{k-1}{(1-\frac{k}{2})^{3/2}} \int_0^{\pi/2} \sqrt{\frac{1}{\big(1+\frac{k}{k-2}\cos(2\phi)\big)^3}}d\phi$$

which totally overstrain or even contradicts my calculus intuition. I am wondering how on earth the simple elliptic integral on the LHS is able to account for all the intricacies of the involved, tedious monster with a third-order polynomial of cosines in the denominator of a square root on the RHS. Is there any concise argument that explains how $E(k)$ pulls off such a feat? If someone with a lot of experience in advanced integration could shed some light on this matter and give me some intuition, that'd be great.

Best Answer

Following the Hanning Makholm's suggestion, we have $$ I = \left(1-\frac{k}{2}\right)^{3/2}\int_{0}^{\pi/2}\frac{d\theta}{(1-k\sin^2\theta)^{3/2}},\tag{1}$$ on the other hand: $$ \begin{eqnarray*}\int_{0}^{\pi/2}\frac{d\theta}{(1-k\sin^2\theta)^{3/2}}&=&\int_{0}^{\pi/2}\frac{d\theta}{(1-k\cos^2\theta)^{3/2}}=\int_{0}^{+\infty}\frac{(1+t^2)^{1/2}}{((1-k)+t^2)^{3/2}}\,dt\\&=&\frac{1}{1-k}\int_{0}^{+\infty}\frac{(1+(1-k)u^2)^{1/2}}{(1+u^2)^{3/2}}\,du\\&=&\frac{1}{1-k}\int_{0}^{\pi/2}\left(\cos^2\theta+(1-k)\sin^2\theta\right)^{1/2}\,d\theta\\&=&\frac{1}{1-k}\int_{0}^{\pi/2}\sqrt{1-k\sin^2\theta}\,d\theta = \color{red}{\frac{E(k)}{1-k}}. \end{eqnarray*}$$

Related Question