$\iint_{\Sigma} \frac{d\sigma}{\sqrt{x^2+y^2+(z+R)^2}}$ with $\Sigma$ the upper half of the sphere $x^2+y^2+z^2=R^2$

My attempt:

$$K = \{(r,\theta): 0 \le r \le R, 0 \le \theta \le 2\pi \}$$. I chose following parameterization: $$\vec{\varphi}(r,\theta)=(r\cos\theta, r\sin\theta,\sqrt{R^2-r^2}).$$ And after further calculations, I got $$\left\|\frac{\partial{\vec{\varphi}}}{\partial r} \times\frac{\partial{\vec{\varphi}}}{\partial \theta}\right\| = \frac{rR}{\sqrt{R^2-r^2}}.$$

And now, $$\iint_{\Sigma} \frac{d\sigma}{\sqrt{x^2+y^2+(z+R)^2}} = \int_0^{2\pi}d\theta\int_0^R \frac{r^2R}{\sqrt{(2R^2+2R\sqrt{R^2-r^2})(R^2-r^2)}}dr.$$ Applying the substitution $$t = \sqrt{R^2-r^2}$$, we get $$2\pi R\int_R^0\frac{R^2-t^2}{\sqrt{(2R^2+2Rt)t^2}}\frac{-t}{\sqrt{R^2-t^2}}dt=\sqrt{2R}\pi\int_0^R\sqrt{R-t}dt=\frac23\sqrt2\pi R^2.$$

The correct answer, however, should be $$2\pi R(2-\sqrt2)$$. Can somebody spot my mistake?

• When you wrote $\int_0^R \frac{r^2}{\sqrt{(2 R^2+2 R\sqrt{R^2-r^2} ) (R^2-r^2)}}$ the numerator should have been $rR$ rather than $r^2$. I think you made another error when you did the $t$ substitution, but I'm not sure yet. – irchans Mar 20 at 15:23
• @zachary If you choose to use polar coordinates here, note that the projection of an elemental surface onto the $x-y$ plane results in the transformation $$d\sigma=\frac{dx\,dy}{\hat z\cdot\left( \frac{\hat xx+\hat yy+\hat zz}{\sqrt{x^2+y^2+z^2}}\right)}=\frac{R\,dx\,dy}{\sqrt{R^2-r^2}}=\frac{Rr\,dr\,d\theta}{\sqrt{R^2-r^2}}$$ – Mark Viola Mar 20 at 16:35


• (+1) This is the easier approach. – Mark Viola Mar 20 at 16:19
• @MarkViola That's true. Thanks, Mark. – Felix Marin Mar 20 at 16:28

When you wrote $$\int_0^R \frac{r^2}{\sqrt{(2 R^2+2 R\sqrt{R^2-r^2} ) (R^2-r^2)}}$$ the numerator should have been $$rR$$ rather than $$r^2$$. You can evaluate the integral using your substitution $$t=\sqrt{R^2-r^2}$$ as follows: $$\int_0^{2 \pi}d\theta\int_0^R \frac{rR}{\sqrt{(2 R^2+2 R\sqrt{R^2-r^2} ) (R^2-r^2)}}dr$$ $$= 2 \pi \int_R^0 \frac{R \sqrt{R^2-t^2}}{\sqrt{(2 R^2+2 Rt ) t^2}} \frac{-t}{\sqrt{R^2-t^2}} dt$$ $$= 2 \pi \int_R^0 \frac{R }{\sqrt{(2 R^2+2 Rt ) t^2}} (-t) dt$$ $$= 2 \pi \int_0^R \frac{R }{\sqrt{2 R^2+2 Rt }} dt$$ $$= 2\sqrt2 \pi \int_0^R \frac{R }{2\sqrt{R^2+ Rt }} dt$$ $$= 2\sqrt2 \pi (\sqrt{R^2+ R^2 } - R)$$ $$= 2\sqrt2 \pi R (\sqrt{2} - 1)$$ $$= 2 \pi R (2 - \sqrt2).$$

• You might consider adding the reason that your first statement is correct. If you choose to use polar coordinates here, note that the projection of an elemental surface onto the $x-y$ plane results in the transformation $$d\sigma=\frac{dx\,dy}{\hat z\cdot\left( \frac{\hat xx+\hat yy+\hat zz}{\sqrt{x^2+y^2+z^2}}\right)}=\frac{R\,dx\,dy}{\sqrt{R^2-r^2}}=\frac{Rr\,dr\,d\theta}{\sqrt{R^2-r^2}}$$ – Mark Viola Mar 20 at 16:39
• @MarkViola I assumed that Felix just had a typo. I liked how he computed the determinant of the Jacobian $\left\|\frac{\partial{\vec{\varphi}}}{\partial r} \times\frac{\partial{\vec{\varphi}}}{\partial \theta}\right\| = \frac{rR}{\sqrt{R^2-r^2}}.$ Glad that you posted the spherical coordinate solution which is cleaner. – irchans Mar 20 at 16:51
• Actually, the OP is "Zachary" and "Felix Marin" posted the solution that used spherical coordinates. – Mark Viola Mar 20 at 17:02
• Oops! I apologize. – irchans Mar 20 at 17:07