A graduate student in physics asked me if there was a closed form for the following series, which he had found as the second virial coefficient for a gas in an exponential potential: $$\sum_{k=0}^\infty \frac{x^k}{(k+1)^4 k!}$$ where $x < 0$. This is the generalized hypergeometric function $_4 F_4(1, 1, 1, 1; 2, 2, 2, 2; x)$ by definition. We were skeptical that it could be simplified, as neither the series nor any of its first few derivatives and antiderivatives is the Taylor series of a function that we could recognize. Even if there's not a truly closed form, we're still curious whether it has a formula with special functions other than the generalized hypergeometric.

  • 1
    $\begingroup$ the site suggested an analogous math.stackexchange.com/questions/2005837/… im not too optimistic of a more closed form then $\endgroup$ – user499752 Nov 8 '17 at 15:24
  • $\begingroup$ Even better, look for the commentary on the first answer math.stackexchange.com/questions/2006799/… $\endgroup$ – user499752 Nov 8 '17 at 15:27
  • $\begingroup$ The given series equals $$-\frac{1}{6}\int_{0}^{1}e^{xy}\log^3(y)\,dy$$ so its asymptotic behaviour is pretty simple to estimate, but a nice closed form... I won't bet on that. $\endgroup$ – Jack D'Aurizio Nov 8 '17 at 18:35
  • $\begingroup$ @JackD'Aurizio I don't immediately see where that integral comes from; can you give a justification? $\endgroup$ – Connor Harris Nov 8 '17 at 19:22
  • $\begingroup$ @ConnorHarris: $$\frac{1}{(k+1)^4}=-\frac{1}{6}\int_{0}^{1}y^k\log^3(y)\,dy,$$ hence by multiplying both sides by $\frac{x^k}{k!}$ and summing over $k\geq 0$ we get the previous integral representation. $\endgroup$ – Jack D'Aurizio Nov 8 '17 at 19:29

Let us denote \begin{equation} S_q(x) := \sum\limits_{k=0}^\infty \frac{x^k}{(k+1)^q k!} \end{equation} and $y=-x$.

Using JackD'Azurio's integral representation we get: \begin{eqnarray} S_4(x) &=& -\frac{1}{3!}\left. \frac{d^3}{d \theta^3} \int\limits_0^1 e^{x y} y^\theta dy \right|_{\theta=0} \\ &=&-\frac{1}{3!} \left. \frac{d^3}{d \theta^3} \frac{\Gamma(1+\theta) - \Gamma(1+\theta,-x)}{(-x)^{1+\theta}} \right|_{\theta=0} \\ &=& \frac{4 \zeta(3)+2 \gamma^3+ \gamma \pi^2+(6 \gamma^2+\pi^2)\log(y)+6 \gamma \log(y)^2+2 \log(y)^3+12 {\mathfrak f}(y)}{12 y} \end{eqnarray} where \begin{equation} {\mathfrak f}(y) := \int\limits_1^\infty \frac{e^{-y t}}{t} \frac{\log(t)^2}{2!} dt = G_{3,4}^{4,0}\left(y\left| \begin{array}{c} 1,1,1 \\ 0,0,0,0 \\ \end{array} \right.\right) \end{equation} where the function above is expressed through the Meijer-G function and $\gamma$ is the Euler's gamma constant.

| cite | improve this answer | |

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.