Can we get a closed form for the following contour integral?. Let us assume that n is a non-negative integer,

$\frac{1}{2\pi i}\int^{c+i\infty}_{c-i\infty}\frac{\Gamma(n-s)\Gamma(s)\Gamma(k-s)}{\Gamma(1+n-s)}{}_1F_1(1+n-s,n+1,\frac{\alpha}{2b})\, b^s\,\mathrm{d}s$

and also how to choose the value of c in order solve the contour integral.


Clearly the confluent hypergeometric is independent on $s$, thus: $$ \begin{eqnarray} && \frac{1}{2\pi i}\int^{c+i\infty}_{c-i\infty}\frac{\Gamma(s+n)\Gamma(1-s+n)\Gamma(s+1)}{(-s+n)\Gamma(s)\Gamma(1+s-n)}{}_1F_1(n-1,n+1,\frac{c}{2b})\,\mathrm{d}s = \\ &&{}_1F_1(n-1,n+1,\frac{c}{2b}) \cdot \frac{1}{2\pi i}\int^{c+i\infty}_{c-i\infty}\frac{\Gamma(s+n)\Gamma(1-s+n)\Gamma(s+1)}{(-s+n)\Gamma(s)\Gamma(1+s-n)}\,\mathrm{d}s \end{eqnarray} $$

The integral multiplying the hypergeometric function is known as Barnes integral. Its value will depend on the value of $c$ relative to the patter of poles of integrand. Let $$ G_n(s) = \frac{\Gamma(s+n)\Gamma(1-s+n)\Gamma(s+1)}{(-s+n)\Gamma(s)\Gamma(1+s-n)} = \frac{\Gamma(s+n)\Gamma(-s+n)}{\Gamma(1+s-n)} \cdot \frac{\Gamma(s+1)}{\Gamma(s)} = \frac{\Gamma(s+n)\Gamma(-s+n)}{\Gamma(1+s-n)} \cdot s $$ I will further assume $n$ to be a non-negative integer, and $c$ to be a real number. Let's list zeros and poles of $G(s)$ explicitly, with $k \in \mathbb{Z}^+$: $$ p_k^{(1)} = -n - k \qquad p_k^{(2)} = n + k \qquad z^{(1)} = 0 \qquad z_k^{(2)} = n-1-k $$ Consider $n=0$, first. Then $G_0(s) = \Gamma(-s)$. Then if $c > 0$, the Barnes integral can be evaluated via residue theorem (you have to see that complete the integration contour into a closed loop with arc extending from $c+i\infty$ to $c-i \infty$ to the left is permitted): $$ \forall c > 0, \frac{1}{2 \pi i} \int_{c-i \infty}^{c + i \infty} \Gamma(-s) \mathrm{d} s = -\sum_{k=0}^\infty \operatorname{Res}_{s=k} \Gamma(-s) = \sum_{k=0}^\infty \frac{(-1)^k}{k!} = \frac{1}{\mathrm{e}} $$ If $n>0$, notice that for every pole $p_k^{(1)}$ there exists a zero $z^{(2)}_m$ that cancels it. So, again, assuming $c > -n$ we have: $$ \frac{1}{2 \pi i} \int_{c-i \infty}^{c + i \infty} G_n(s) \mathrm{d} s = -\sum_{k=0}^\infty \operatorname{Res}_{s=n+k} G_n(s) = \sum_{k=0}^\infty (-1)^{k} (k+n) \frac{(k+2n-1)!}{k!^2} $$ The sum is easily evaluated in terms of hypergeometric function: $$ \sum_{k=0}^\infty (-1)^{k} (k+n) \frac{(k+2n-1)!}{k!^2} = (2 n)! \left(\, _1F_1(2 n+1;1;-1)-\frac{1}{2} n \, _1F_1(2 n;1;-1)\right) $$

  • $\begingroup$ Thanks Sasha for your reply, I have just updated the integral. Now the function is dependent on s. $\endgroup$ – Sophie Nov 28 '11 at 16:18
  • $\begingroup$ I think now it looks more like an inverse mellin transform... $\endgroup$ – Sophie Nov 28 '11 at 16:24
  • $\begingroup$ Thanks for the explanation of the poles and zeros, I always struggle with residues... $\endgroup$ – Sophie Nov 28 '11 at 16:34

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.