This question has some relationship to this integral: Let $\mathrm{Li}_2$ be the dilogarithm. Then, numerically, $$ \int_0^1 \Bigl(\mathrm{Li}_2\bigl(\frac{x-1}{x}\bigr)\Bigr)^2 \mathrm{d}x = \frac{17}{180}\pi^4, $$ as can be checked by Mathematica code like

NIntegrate[ PolyLog[2, (x-1)/x]^2, {x,0,1}, WorkingPrecision -> 100]


N[ 17/180 * Pi^4, 100]   

both of which output the same digits

9.199747486544674627886031420599927173585383 ...

Is this a known identity? Some related results appear in Freitas and Laurenzi, but without this specific argument in the dilogarithm.


OK, this follows from Freitas' paper. In Table 3 there

$$\int_0^1 \mathrm{Li}_2^2(x) \mathrm{d}x = 6 - 2\zeta(2) - 4\zeta(3) + \frac{5}{2}\zeta(4),$$ and Formula (2.2.5) of Table of integrals and formulae for Feynman diagram calculations states $$\mathrm{Li}_2\Bigl(-\frac{y}{1-y}\Bigr) = - \mathrm{Li}_2(y) - \frac{1}{2}\ln^2(1-y),$$ and Mathematica can solve $$ \frac{1}{4}\int_0^1 \ln^4(1-x)\mathrm{d}x = 6 $$ as well as $$ \int_0^1 \mathrm{Li}_2(x) \ln^2(1-x) \mathrm{d}x = \frac{1}{15} \left(60 (\zeta (3)-3)+5 \pi ^2+\pi ^4\right) $$

Using these formulae, the claim follows.

  • $\begingroup$ Impressive! I wanted to come back to this integral later, and I probably will - as the cancellation of $\zeta(3)$ hints that there should be also an elementary derivation. (+1) $\endgroup$ – Start wearing purple May 14 '13 at 16:05

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.