Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

The poisson kernel is sometimes written as $$ \frac{1}{2\pi}\int_0^{2\pi} \frac{R^2-r^2}{R^2 - 2Rr\cos(\varphi-\vartheta) + r^2} \mathrm{d}\vartheta = 1 \ , \ \ R>r>0 $$ Where $\varphi$ is some arbitary angle. This much used in complex analysis amongst other fields. Is there some basic, elementary way of showing that the integral is independent on $R$ and $r$?

Splitting the integral at $\int_0^\pi + \int_\pi^{2\pi}$ and using the Weierstrass substitution seems like somewhat ugly and a messy way to approach the problem.

share|cite|improve this question
Maybe a stupid question: it is possible to differentiate with respect to $R$ and $r$? I never tried to. – Siminore Jan 25 '14 at 12:59
Is $d\theta$ at the right place ? Please correct also in the title. – Claude Leibovici Jan 25 '14 at 13:08
If you want solution using real methods, then I don't understand (complex-analysis) tag. – user110822 Jan 25 '14 at 13:42
up vote 6 down vote accepted

Step I. Using integration by parts one can easily show that $$ J_n=\int_0^{2\pi} \cos^{2n}x\,dx=\frac{2\pi}{4^n}\binom{2n}{n}, $$ as for the cosine integral we make use of a recurrence, \begin{align} J_{n+1}&=\int_0^{2\pi} \cos^{2(n+1)}x\,dx=\int_0^{2\pi} (\cos^{2n+1}x)(\sin x)'\,dx =0+(2n+1)\int_0^{2\pi} \cos^{2n}x\,\sin^2 x\,dx \\&= (2n+1)\int_0^{2\pi} \cos^{2n}x\,(1-\cos^2 x)\,dx=(2n+1)J_n-(2n+1)J_{n+1}, \end{align} and thus $J_{n+1}=\frac{2n+1}{2n+2}J_n=2^{-n-1}\prod_{j=0}^{n+1}\frac{2j-1}{j+1}J_0=\frac{2\pi}{4^{n+1}}\frac{(2n+2)!}{(n+1)!(n+1)!}$.

Step II. If $|a|<1$, then $$ \frac{1}{\sqrt{1-a^2}}=(1-a^2)^{-1/2}=\sum_{n=0}^\infty \binom{-\frac{1}{2}}{n}a^{2n}=\cdots=\sum_{n=0}^\infty \binom{2n}{n}\frac{a^{2n}}{2^{2n}}. $$

Step III. Then $$ I=\frac{1}{2\pi}\int_0^{2\pi} \frac{R^2-r^2}{R^2 - 2Rr\cos(\phi-\theta) + r^2} \mathrm{d}\theta =\frac{R^2-r^2}{2\pi(R^2+r^2)}\int_0^{2\pi}\frac{dx}{1-a\cos x}, $$ where $a=\dfrac{2rR}{r^2+R^2}.$ Also $$ \int_0^{2\pi}\frac{dx}{1-a\cos x}=\sum_{n=0}^\infty\int_0^{2\pi} a^{2n}\cos^{2n}x\,dx=\sum_{n=0}^\infty \frac{2\pi\,a^{2n}}{4^n}\binom{2n}{n}=\frac{2\pi}{\sqrt{1-a^2}} $$

Step IV. Finally, \begin{align} I&=\frac{1}{2\pi}\int_0^{2\pi} \frac{R^2-r^2}{R^2 - 2Rr\cos(\phi-\theta) + r^2} \mathrm{d}\theta =\frac{R^2-r^2}{2\pi(R^2+r^2)}\int_0^{2\pi}\frac{dx}{1-a\cos x} \\ &=\frac{R^2-r^2}{2\pi(R^2+r^2)}\cdot \frac{2\pi}{\sqrt{1-\left(\frac{2rR}{r^2+R^2}\right)^2}} =\cdots=1. \end{align}

Note. I prefer the proof of T.A.E, but the question was asking of a proof without the use of complex numbers!

share|cite|improve this answer

Because you're integrating over a full period of $\cos$ and because $R > 0$, your expression becomes $$ \frac{1}{2\pi}\int_{0}^{2\pi}\frac{1-(r/R)^{2}}{1-2(r/R)\cos\theta+(r/R)^{2}}\,d\theta. $$ Let $\rho = r/R$. Then, by assumption, $0 < \rho < 1$, and $$ \begin{align} \frac{1-(r/R)^{2}}{1-2(r/R)\cos\theta+(r/R)^{2}} & =\frac{1-\rho^{2}}{(1-\rho e^{i\theta})(1-\rho e^{-i\theta})} \\ & = \frac{1}{1-\rho e^{i\theta}}+\frac{\rho e^{-i\theta}}{1-\rho e^{-i\theta}} \\ & = \sum_{n=-\infty}^{\infty}\rho^{|n|} e^{in\theta} \\ & = 1 + 2\sum_{n=1}^{\infty}\rho^{n}\cos(n\theta). \end{align} $$ When you integrate the terms in this sum over $0 \le \theta \le 2\pi$, only the integral of the constant term is non-zero. The final integral value is 1, regardless of $0 \le \rho < 1$.

If you don't like the complex exponential in between, just verify directly that $$ F(\rho) = \frac{1}{2\pi}\int_{0}^{2\pi}\frac{1-\rho^{2}}{1-2\rho\cos\theta+\rho^{2}}d\theta $$ satisfies the differential equation $$ F''(\rho)+\frac{1}{\rho}F'(\rho)=0. $$ This a separable equation whose solutions have the form $$ F'(\rho) = C\frac{1}{\rho} \\ F(\rho) = C\ln \rho + D. $$ The constant $C$ must be $0$ in this case for obvious reasons.

share|cite|improve this answer
You know i explicitly asked for solutions using real methods? =) Still a nice answer, so just let it be, but unfortunately you will probably not get the accept. – N3buchadnezzar Jan 25 '14 at 13:34
@N3buchadnezzar: I have added the solution for you that depends only on real variables. – TrialAndError Jan 25 '14 at 14:10

$\newcommand{\+}{^{\dagger}}% \newcommand{\angles}[1]{\left\langle #1 \right\rangle}% \newcommand{\braces}[1]{\left\lbrace #1 \right\rbrace}% \newcommand{\bracks}[1]{\left\lbrack #1 \right\rbrack}% \newcommand{\ceil}[1]{\,\left\lceil #1 \right\rceil\,}% \newcommand{\dd}{{\rm d}}% \newcommand{\down}{\downarrow}% \newcommand{\ds}[1]{\displaystyle{#1}}% \newcommand{\equalby}[1]{{#1 \atop {= \atop \vphantom{\huge A}}}}% \newcommand{\expo}[1]{\,{\rm e}^{#1}\,}% \newcommand{\fermi}{\,{\rm f}}% \newcommand{\floor}[1]{\,\left\lfloor #1 \right\rfloor\,}% \newcommand{\half}{{1 \over 2}}% \newcommand{\ic}{{\rm i}}% \newcommand{\iff}{\Longleftrightarrow} \newcommand{\imp}{\Longrightarrow}% \newcommand{\isdiv}{\,\left.\right\vert\,}% \newcommand{\ket}[1]{\left\vert #1\right\rangle}% \newcommand{\ol}[1]{\overline{#1}}% \newcommand{\pars}[1]{\left( #1 \right)}% \newcommand{\partiald}[3][]{\frac{\partial^{#1} #2}{\partial #3^{#1}}} \newcommand{\pp}{{\cal P}}% \newcommand{\root}[2][]{\,\sqrt[#1]{\,#2\,}\,}% \newcommand{\sech}{\,{\rm sech}}% \newcommand{\sgn}{\,{\rm sgn}}% \newcommand{\totald}[3][]{\frac{{\rm d}^{#1} #2}{{\rm d} #3^{#1}}} \newcommand{\ul}[1]{\underline{#1}}% \newcommand{\verts}[1]{\left\vert\, #1 \,\right\vert}$ $\ds{{1 \over 2\pi}\int_{0}^{2\pi}{R^{2} - r^{2} \over R^{2} - 2Rr\cos\pars{\phi-\theta} + r^{2}}\,\dd\theta = 1\,,\quad R > r > 0}$

The integral is obviously independent of $\phi$: Just derive the integral respect of $\phi$ and use $\ds{\partiald{}{\phi} = - \partiald{}{\theta}}$. Since the integrand is a periodic function of $\theta$ it turns out that the $\phi$-derivative of the integral is zero. Then, we just need to consider the integral: $$ {\cal I} \equiv {1 \over \pi}\int_{0}^{\pi}{1 - h^{2} \over 1 - 2h\cos\pars{\theta} + h^{2}}\,\dd\theta\,,\qquad 0< h \equiv {r \over R} < 1 $$

With Weierstrass substitution $\ds{t \equiv \tan\pars{\theta \over 2}}$: \begin{align} {\cal I} &= {1 - h^{2} \over \pi}\int_{0}^{\infty} {1 \over 1 - 2h\pars{1 - t^{2}}/\pars{1 + t^{2}} + h^{2}}\,{2\,\dd t \over 1 + t^{2}} \\[3mm]&= 2\,{1 - h^{2} \over \pi}\int_{0}^{\infty} {\dd t \over \pars{h^{2} + 2h + 1}t^{2} + h^{2} - 2h + 1} = {2 \over \pi}\,\int_{0}^{\infty} {\bracks{\pars{1 + h}/\pars{1 - h}}\,\dd t \over \bracks{\pars{1 + h}t/\pars{1 - h}}^{2} + 1} \\[3mm]&= {2 \over \pi}\ \overbrace{\int_{0}^{\infty}{1 \over t^{2} + 1}}^{\ds{=\ {\pi/2}}} =\color{#00f}{\Large 1} \end{align}

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.