0
$\begingroup$

I am reading the book "Special functions: an introduction to classical functions of mathematical physics" by Nico M. Temme and I'm having trouble understanding how to find a constant (in page 62).

I understand why

$$ \log\Gamma(z+1)=(z+\frac{1}{2})\log z - z + \int_{0}^{\infty}g(t)e^{-zt}dt + C, $$

where $g$ is bounded for $t\gt0$ and $C$ is a constant of integration, which has to be determined. I know that Legendre's duplication formula can be written in the form

$$ \log \frac{2^{2z}\Gamma(z+1)\Gamma(z+\frac{1}{2})}{\Gamma(2z+1)} = \frac{1}{2}\log \pi. $$

Now, the book asserts the following: Letting $z\to\infty$ and using Legendre's duplication formula we obtain $C=\frac{1}{2}\log 2\pi.$ But I don't understand how to use the duplication formula, when I let $z\to\infty$ the integral vanishes (since $g$ is bounded). Thus

$$ C=\lim_{z\to\infty} \log\Gamma(z+1)-(z+\frac{1}{2})\log z + z. $$

Now, using Stirling's approximation, the above limit equals $\frac{1}{2}\log 2\pi$, but I don't want to use Stirling's approximation, I want to know how to use the duplication formula.

$\endgroup$
  • $\begingroup$ It has nothing obvious that $\log\Gamma(z+1)=(z+\frac{1}{2})\log z - z + \int_{-\infty}^{\infty}g(t)e^{-zt}dt + C$ even less that $g(t) = 0$ for $t< 0$. And it depends on the tools you want to use : complex analysis, Fourier analysis, some theorems like duplication, reflection, Stirling $\endgroup$ – reuns Mar 24 at 19:10
  • $\begingroup$ @reuns I only want to use the duplication formula $\endgroup$ – Sebitas Mar 24 at 19:14
  • $\begingroup$ Once you know that $\log\Gamma(z+1)=(z+\frac{1}{2})\log z - z + \int_{0}^{\infty}g(t)e^{-zt}dt + C$ then the duplication formula gives $C = \lim_{z \to \infty}\log 2^{2z} + \log\Gamma(z+1)+\log \Gamma(z+\frac{1}{2})-\log\Gamma(2z+1)$. But again $\log\Gamma(z+1)=(z+\frac{1}{2})\log z - z + \int_{0}^{\infty}g(t)e^{-zt}dt + C$ has nothing obvious and you should probably look again at the steps they used to find it. $\endgroup$ – reuns Mar 24 at 19:35
  • $\begingroup$ I know it is not obvious, but after some trouble I was able to prove that $\log\Gamma(z+1)=(z+\frac{1}{2})\log z-z + \int_{0}^{\infty} g(t)e^{-zt}dt + C$ but from this I really don't see how the duplication formula gives $C=\lim_{z\to\infty} \log 2^{2z} + \log \Gamma(z+1) + \log \Gamma(z+\frac{1}{2})-\log\Gamma(2z+1).$ Could you please elaborate that line? $\endgroup$ – Sebitas Mar 24 at 20:24
1
$\begingroup$

First, define the two functions $$ f(z) := (z + \frac12)\log(z)-z,\quad F(z) := f(z) + f(z - \frac12) - f(2z). $$ Second, find that the power series expansion at infinity of $\ F(z) \ $ is $$ F(z) = -2\log(2)\ z - \frac12 \log(2) - \frac18 z^{-1} + O(z^{-2}). $$ The rest is a simple exercise.

$\endgroup$
  • $\begingroup$ Finally I understand why you defined such functions, thank you very much! $\endgroup$ – Sebitas Mar 24 at 21:21

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.