Take the 2-minute tour ×
MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

I ask about an idea to prove this formula:

$Γ(1/2-iβ)=((\sqrt{π})/(\sqrt{\coshπβ}))\exp(-i(2ϑ(β)+βln2π+\arctan(\tanh(1/2)πβ)))$

where $ϑ(β)$ is the Riemann Siegel function.

share|improve this question
    
I'm curious about this unusual phenomenon (yet expected, by the "Reversal" badge) of +15 points difference between question and answer! Some hints? –  Pietro Majer Jan 28 at 17:41
    
@PietroMajer: Myabe the question is very trivial and you can see that the answer is not trivial!!. –  China-Hong Kong Jan 28 at 17:47
    
Yes, but if a question originates a non trivial and interesting answer, it shouldn't be that bad. Anyway, I respect your sportsman-like attitude. –  Pietro Majer Jan 28 at 18:04

2 Answers 2

up vote 5 down vote accepted

I know two proof, the first uses $$\cos\frac{\pi s}{2}=\frac{1}{\sqrt{2}}\sqrt{\cosh(\pi t)}\,e^{-i\arctan(\tanh\frac{\pi t}{2})}.\qquad (1)$$ and $$\Gamma(\frac12+i\frac t 2)=|\Gamma(\frac14+i\frac t2)|\,e^{i(\vartheta(t)+\frac t 2\log\pi)},\qquad (2)$$

Since $$\Gamma(z)\Gamma(z+1/2)=2^{1-2z}\sqrt{\pi}\Gamma(2z);\quad \Gamma(z)\Gamma(1-z)=\frac{\pi}{\sin\pi z}$$ we get $$\Gamma(2z)=2^{2z-1}\pi^{-1/2} \Gamma(z)\frac{\pi}{\cos\pi z}\frac{1}{\Gamma(1/2-z)}.$$ We put now $z=\frac14+i\frac t 2$, $t$ real $$ \Gamma(\frac12+it)=2^{-\frac12+it}\frac{\pi^{\frac12}}{\cos\pi(\frac14+i\frac{t}{2})} \frac{\Gamma(\frac14+i\frac{t}{2})}{\Gamma(\frac14-i\frac{t}{2})}. $$ From (1) and (2) we get $$\Gamma(1/2+it)=2^{it}\frac{\pi^{1/2}}{\sqrt{\cosh \pi t}\; e^{-i \arctan\tanh(\pi t/2)}}\pi^{it} e^{2i\vartheta(t)}. $$ so that $$ \Gamma(1/2+it)=\sqrt{\frac{\pi}{\cosh\pi t}}\exp\bigl\{i(2\vartheta(t)+t\log(2\pi)+\arctan\tanh(\pi t/2))\bigr\} $$

The other proof I know uses the functional equation of the zeta function.

share|improve this answer
    
@Juan: Note that all the identities concerning arguments holds only modulo factors of 2π if the argument is being restricted to (−π,π] . How you can deal with arguments figured in this proof. This is a problem since these arguments depend on the variable t . –  China-Hong Kong Dec 16 '12 at 9:38
    
@RH The argument of $\Gamma(s)$ is well defined and harmonic on the plane with a cut along the negative real axis. –  juan Dec 16 '12 at 15:25

I extended juan's proof to a formula for all complex numbers.

Consequently, we get the Euler chi function $\chi(z):=\zeta(1-z)/\zeta(z)$.

(substitution)

$\tan^{-1}(\tanh(\pi t/2))\longrightarrow\frac{\mathrm{gd}(\pi t)}2$, where $\mathrm{gd}(z)$ is the Gudermannian function.

$\sqrt{\frac{\pi}{\cosh(\pi t)}}\longrightarrow\frac{\sqrt{\pi}}{\cosh(\pi t)}\left(\sinh(\frac{\pi t}2)\sin(\frac{\mathrm{gd}(\pi t)}{2})+\cosh(\frac{\pi t}{2})\cos(\frac{\mathrm{gd}(\pi t)}2)\right)$

$t\longrightarrow z$

(simplification) If we covert the above result(trigonometric functions) to exponentials, then we get the following simplified formula.

$\Gamma(\frac12+iz)=\frac{\sqrt{\pi}(1+i)(2\pi)^{iz}}{e^{\pi z}+i}e^{\frac{\pi z}2+2i\vartheta(z)}/;z\in\mathbb{C}$

Therefore,

$\Gamma(z)=\frac{(2\pi)^z}{1+e^{i\pi z}}e^{i(2\vartheta(\frac{i}2-iz)+\frac{\pi z}2)}=\frac{(2\pi)^z}{2\cos(\frac{\pi z}2)}e^{2i\vartheta(\frac{i}2-iz)}=\frac{(2\pi)^z}{2\cos(\frac{\pi z}2)}\chi(z)$

share|improve this answer
1  
@ Sangkyu Kim: But the gamma function has no closed form. –  China-Hong Kong Jul 14 '14 at 8:04
    
Thanks@Shpigle : I understood what you mean. –  Sangkyu Kim Jul 16 '14 at 8:21
    
@ Sangkyu Kim: It has no an explicit formula. –  China-Hong Kong Jul 16 '14 at 8:24

Your Answer

 
discard

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.