I was wondering if there is an explicit formulation for the series $$ \sum_{k=1}^\infty \frac{x^k}{k!\cdot k} $$ It is evident that the converges for any $x \in \mathbb{R}$. Any ideas on a formula?
-
True, corrected. – Bunder Jun 25 '13 at 10:19
3 Answers
Its derivative is $\sum_{k=1}^{\infty} \frac{x^{k-1}}{k!}$, which is $(e^x-1)/x$. So your function is the integral of my function, which you might or might not call 'closed form'.
- 50,853
-
-
When I was first-year undergrad, we were given so many examples of integrals that became arctans and logs that I thought 'closed form' mean exp,sin,cos,log, etc and no others. We didn't get Bessel, exponential integral etc until second year - but they appeared first as Taylor series so I thought you might be able to turn them into the functions I already knew. – Empy2 Jun 25 '13 at 13:05
You can have the closed form
$$\sum_{k=1}^{\infty}\frac{x^k}{k k!}= -\gamma-\ln(-x)-\Gamma(0, -x), $$
where $\Gamma(s,x)$ is the upper incomplete gamma function. Another possible form is
$$ \sum_{k=1}^\infty \frac{x^k}{k!\cdot k}=-\gamma-\ln \left( -x \right) -{\it Ei} \left( 1,-x \right), $$
where
$$ Ei(a, z) = \int_{1}^{\infty} \frac{e^{-tz}}{t^a}dt,\quad 0 < Re(z),$$
which is known as the exponential integral. The following relation between the exponential integral and the upper incomplete gamma function is useful
$$ Ei(a, z) = z^{a-1}\Gamma(1-a, z). $$
- 47,431
-
-
-
-
You see that you have no problem for $Re(z)>0$. See the exponential integral. Then they are other means which you can use to extend the domain known as analytic continuation. – Mhenni Benghorbal Jun 25 '13 at 10:31
-
Here is my attempt to unravel the mystery of the exponential integral. \begin{align*} e^x & = \sum_{k=0}^\infty \frac{x^k}{k!} \\ \frac 1xe^x & = \sum_{k=0}^\infty \frac{x^{k-1}}{k!} \\ \int_{1}^x \frac 1ye^y dy & = \sum_{k=1}^\infty \frac{x^k}{k! \cdot k} + \log x - \sum_{k=1}^\infty \frac{1}{k!k} \\ \int_{1}^x \frac 1ye^y dy - \log x + \sum_{k=1}^\infty \frac{1}{k!k} & = \sum_{k=1}^\infty \frac{x^k}{k!k} \end{align*} This derivation works for $x > 0$, but if $x < 0$, you only need to change the integral from $\int_1^x$ to $\int_{-1}^{x}$ and proceed from there.
- 10,303