I want to prove, the Laurent expansion of gamma function. \begin{align} \Gamma(z) = \frac1z-\gamma+\frac12\left(\gamma^2+\frac {\pi^2}6\right)z-\frac16\left(\gamma^3+\frac {\gamma\pi^2}2+2 \zeta(3)\right)z^2+O(z^3). \end{align}

First, my guess of obtaing above expansion, is starting from the definitions of gamma function \begin{align} \Gamma(z) &= \int_0^{\infty} dt e^{-t} t^{z-1} \\ & = \int_1^\infty dt e^{-t}t^{z-1} + \int_0^1dt e^{-t} t^{z-1} \\ & = \int_1^\infty dte^{-t}t^{z-1} + \int_0^1 dt t^{z-1} \sum_{n=0}^{\infty}\frac{(-1)^n}{n!}t^n \\ & = \int_1^\infty dt e^{-t}t^{z-1} + \sum_{n=0}^{\infty} \frac{(-1)^n}{n!}\frac{1}{z+n} \end{align} This only gives the gamma function as a function of $\frac{1}{z}$...

Or should I start with \begin{align} \Gamma(z) = \lim_{n \rightarrow \infty} \frac{n! n^z}{z(z+1) \cdots(z+n)} \end{align}

  • $\begingroup$ I have deleted equations related with errors. $\endgroup$ – phy_math May 18 '15 at 8:02

Let $\Gamma(z)$ be represented by the integral

$$\Gamma(z)=\int_0^\infty x^{z-1}e^{-x}\,dx\tag1$$

for $\text{Re}(z)>0$. Integration by parts the integral in $(1)$ with $u=e^{-x}$ and $v=\frac{x^z}{z}$ reveals

$$\Gamma(z)=\frac1z\int_0^\infty x^ze^{-x}\,dx\tag2$$

Next, we expand $x^z$ in a power series of $z$ to obtain

$$\begin{align} \Gamma(z)&=\frac1z\sum_{n=0}^\infty \frac{z^n}{n!}\int_0^\infty e^{-x}\log^n(x)\,dx\\\\ &=\frac1z+\underbrace{\int_0^\infty \log(x)e^{-x}\,dx}_{=-\gamma}+\frac12 z\underbrace{\int_0^\infty \log^2(x)e^{-x}\,dx}_{\gamma^2+\frac{\pi^2}6}+\frac16z^2 \underbrace{\int_0^\infty \log^3(x)e^{-x}\,dx}_{-\gamma^3-\gamma\pi^2/2-2\zeta(3)}+O(z^3) \end{align}$$

as was to be shown.


The coefficients $\int_0^\infty \log^n(x)e^{-x}\,dx$ for $n=2,3$ can be found by using the relationship, $\Gamma'(x)=\Gamma(x)\psi(x)$, between the logarithmic derivative of the Gamma function and the Digamma function, along with values of $\psi'(1)=\zeta(2)$ and $\psi''(1)=\zeta(3)$.

  • $\begingroup$ Please let me know how I can improve my answer. I really want to give you the best answer I can. $\endgroup$ – Mark Viola Aug 26 '18 at 15:12
  • $\begingroup$ Would the down voter care to comment. $\endgroup$ – Mark Viola Aug 30 '18 at 17:39

An idea for you to develop:

The Weierstrass Formula tells us that

$$\frac1{\Gamma(z)}=ze^{\gamma z}\prod_{n=1}^\infty\left(1+\frac zn\right)e^{-z/n}$$

Now take logarithms on both sides to get a more or less well known relation:

$$-\log\Gamma(z)=\log z+\gamma z+\sum_{n=1}^\infty\left[\log\left(1+\frac zn\right)-\frac zn\right]$$

Now differentiate the above to get the logarithmic derivative of the Gamma function:

$$\frac{\Gamma'(z)}{\Gamma(z)}=-\frac1z-\gamma-\sum_{n=1}^\infty\frac1n\left[\frac n{z+n}-1\right]=-\frac1z-\gamma+\sum_{n=1}^\infty\frac z{n(z+n)}$$

and etc. You can try to integrate the $\;-\dfrac1z\;$ term into the series, too.


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.