Let $(X,Y)$ have a bivariate normal density centered at the origin with $\mathbb{E}(X^2)=\mathbb{E}(Y^2)=1$ and $\mathbb{E}(XY)=\rho$ with $|\rho|<1$. In polar coordinates, $(X,Y)$ becomes $(R,\Phi)$ where $X=R\cos\Phi$ and $Y=R\sin\Phi$ . Then it can be shown that $\Phi$ has a density given by $$f_{\Phi}(\phi)=\frac{\sqrt{1-\rho^2}}{2\pi(1-\rho\sin2\phi)},\quad 0<\phi<2\pi$$
But I am having difficulty proving that this is indeed a density, more specifically the fact that $$\int_0^{2\pi}f_{\Phi}(\phi)\,\mathrm{d}\phi=1$$ Applying the usual substitution $t=\tan\phi$ in the integral $\int_0^{2\pi}\frac{1}{1-\rho \sin{2 \phi }}\,\mathrm{d}\phi$ seems to be misleading. Should there be any change in the integration limits for proving that this a density. If so, why?
Another related integral is $$I=\int_0^{\pi/2}\frac{\mathrm{d}\theta}{1-\rho\sin{2\theta}}=\frac{\pi-\arccos\rho}{\sqrt{1-\rho^2}}\,,$$ which arises while calculating $\mathbb{P}(X>0,Y>0)$ for the above distribution once I transform to polar coordinates.