A few days ago, I posted the following problems

Prove that \begin{equation} \int_0^{\pi/2}\ln^2(\cos x)\,dx=\frac{\pi}{2}\ln^2 2+\frac{\pi^3}{24}\\[20pt] -\int_0^{\pi/2}\ln^3(\cos x)\,dx=\frac{\pi}{2}\ln^3 2+\frac{\pi^3}{8}\ln 2 +\frac{3\pi}{4}\zeta(3) \end{equation}

and the OP receives some good answers even I then could answer it.

My next question is finding the closed-forms for

\begin{align} \int_0^{\pi/4}\ln^2(\sin x)\,dx\tag1\\[20pt] \int_0^{\pi/4}\ln^2(\cos x)\,dx\tag2\\[20pt] \int_0^1\frac{\ln t~\ln\big(1+t^2\big)}{1+t^2}dt\tag3 \end{align}

I have a strong feeling that the closed-forms exist because we have nice closed-forms for \begin{equation} \int_0^{\pi/4}\ln(\sin x)\ dx=-\frac12\left(C+\frac\pi2\ln2\right)\\ \text{and}\\ \int_0^{\pi/4}\ln(\cos x)\ dx=\frac12\left(C-\frac\pi2\ln2\right). \end{equation} The complete proofs can be found here.

As shown by Mr. Lucian in his answer below, the three integrals are closely related, so finding the closed-form one of them will also find the other closed-forms. How to find the closed-forms of the integrals? Could anyone here please help me to find the closed-form, only one of them, preferably with elementary ways (high school methods)? If possible, please avoiding contour integration and double summation. Any help would be greatly appreciated. Thank you.

  • $\begingroup$ What type of integration is this, I was taught only basic integrals, can you tell me waht topic which they come under so that I can read about them and learn about them. $\endgroup$
    – RE60K
    Sep 2, 2014 at 16:02
  • $\begingroup$ @Aditya Log-Trig integrals $\endgroup$ Sep 2, 2014 at 16:04
  • 1
    $\begingroup$ As I wrote in my answer on this thread, this question is equivalent to finding the closed form expression for $\text{Li}_3\Big(\frac{1+i}2\Big)$. $\endgroup$
    – Lucian
    Sep 4, 2014 at 23:55

7 Answers 7


Following the same approach as in this answer,

$$ \begin{align} &\int_{0}^{\pi/4} \log^{2} (2 \sin x) \ dx = \int_{0}^{\pi/4} \log^{2}(2) \ dx + 2 \log 2 \int_{0}^{\pi/4}\log(\sin x) \ dx + \int_{0}^{\pi /4}\log^{2}(\sin x) \ dx \\ &= \frac{\pi}{4} \log^{2}(2) - \log (2) \left(G + \frac{\pi}{2} \log (2) \right) + \int_{0}^{\pi/4} \log^{2}(\sin x) \ dx \\ &= \int_{0}^{\pi /4} \left(x- \frac{\pi}{2} \right)^{2} \ dx + \text{Re} \int_{0}^{\pi/4} \log^{2}(1-e^{2ix}) \ dx \\ &= \frac{7 \pi^{3}}{192} + \frac{1}{2} \text{Im} \int_{{\color{red}{1}}}^{i} \frac{\log^{2}(1-z)}{z} \ dz \\ &= \frac{7 \pi^{3}}{192} + \frac{1}{2} \text{Im} \left(\log^{2}(1-i) \log(i) + 2 \log(1-i) \text{Li}_{2}(1-i) - 2 \text{Li}_{3}(1-i) \right) \\ &= \frac{7 \pi^{3}}{192} + \frac{1}{2} \left(\frac{\pi}{8} \log^{2}(2) - \frac{\pi^{3}}{32} + \log(2) \ \text{Im} \ \text{Li}_{2}(1-i) - \frac{\pi}{2} \text{Re} \ \text{Li}_{2}(1-i)- 2 \ \text{Im} \ \text{Li}_{3}(1-i)\right) . \end{align}$$


$$ \begin{align}\int_{0}^{\pi/4} \log^{2}(\sin x) \ dx &= \frac{\pi^{3}}{48} + G \log(2)+ \frac{5 \pi}{16}\log^{2}(2) + \frac{\log(2)}{2} \text{Im} \ \text{Li}_{2}(1-i) - \frac{\pi}{4} \text{Re} \ \text{Li}_{2}(1-i) \\ &- \text{Im} \ \text{Li}_{3}(1-i) \approx 2.0290341368 . \end{align}$$

The answer could be further simplified using the dilogarithm reflection formula $$\text{Li}_{2}(x) {\color{red}{+}} \text{Li}_{2}(1-x) = \frac{\pi^{2}}{6} - \log(x) \log(1-x) $$

and the fact that $$ \text{Li}_{2}(i) = - \frac{\pi^{2}}{48} + i G.$$


Specifically, $$\text{Li}_{2}(1-i) = \frac{\pi^{2}}{16} - i G - \frac{i \pi}{4} \log(2). $$

So $$\int_{0}^{\pi /4} \log^{2}(\sin x) \ dx = \frac{\pi^{3}}{192} + G\frac{ \log(2)}{2} + \frac{3 \pi}{16} \log^{2}(2) - \text{Im} \ \text{Li}_{3}(1-i).$$

  • $\begingroup$ Thank you Mr. Random variable, for now I'll upvote this answer and study it (ô‿ô) $\endgroup$ Sep 5, 2014 at 12:21
  • $\begingroup$ There were a couple of errors that I fixed. $\endgroup$ Sep 5, 2014 at 15:54
  • $\begingroup$ I added a bit more detail to the other answer. $\endgroup$ Sep 6, 2014 at 7:52
  • 2
    $\begingroup$ Ack. Apparently I accidentally downvoted, and now it's locked in. I'll fix this if you edit. :( $\endgroup$
    – David H
    Sep 6, 2014 at 10:57
  • 3
    $\begingroup$ @DavidH Why did you do that? I consider this answer as a strong candidate winner for my bounty! ᕙ(`皿´)╯ $\endgroup$ Sep 6, 2014 at 11:18

$$\int_0^\frac\pi4\Big(\ln\sin x\Big)^2~dx~=~\dfrac{23}{384}\cdot\pi^3~+~\dfrac9{32}\cdot\pi\cdot\ln^22~+~\underbrace{\beta(2)}_\text{Catalan}\cdot\dfrac{\ln2}2~-~\Im\bigg[\text{Li}_3\bigg(\dfrac{1+i}2\bigg)\bigg].$$

$$\int_0^\frac\pi4\Big(\ln\cos x\Big)^2~dx~=~\dfrac{-7}{384}\cdot\pi^3~+~\dfrac7{32}\cdot\pi\cdot\ln^22~-~\underbrace{\beta(2)}_\text{Catalan}\cdot\dfrac{\ln2}2~+~\Im\bigg[\text{Li}_3\bigg(\dfrac{1+i}2\bigg)\bigg].$$

$$S=\int_0^\frac\pi4\Big(\ln\sin x\Big)^2~dx~+~\int_0^\frac\pi4\Big(\ln\cos x\Big)^2~dx=I+J.$$

But, by a simple change of variable, $t=\dfrac\pi2-x,~J$ can be shown to equal $\displaystyle\int_\frac\pi4^\frac\pi2\Big(\ln\sin x\Big)^2~dx$,

in which case $I+J=\displaystyle\int_0^\frac\pi2\Big(\ln\sin x\Big)^2~dx=\dfrac{\pi^3}{24}+\dfrac\pi2\ln^22.~$ So we know their sum! Now all

that's left to do is to find out their difference, $D=I-J.~$ Then we'll have $I=\dfrac{S+D}2$ and


$$D=I-J=\int_0^\frac\pi4\Big(\ln\sin x\Big)^2~dx-\int_0^\frac\pi4\Big(\ln\cos x\Big)^2~dx=\int_0^\frac\pi4\Big(\ln^2\sin x-\ln^2\cos x\Big)~dx$$

$$=\int_0^\frac\pi4\Big(\ln\sin x-\ln\cos x\Big)~\Big(\ln\sin x+\ln\cos x\Big)~dx=\int_0^\frac\pi4\ln\frac{\sin x}{\cos x}~\ln\big(\sin x~\cos x\big)~dx=$$

$$=\int_0^\frac\pi4\ln\tan x\cdot\ln\frac{\sin2x}2~dx=\frac12\int_0^\frac\pi2\ln\tan\frac x2\cdot\ln\frac{\sin x}2~dx=\int_0^1\ln t\cdot\ln\frac t{1+t^2}\cdot\frac{dt}{1+t^2}$$

where the last expression was obtained by using the famous Weierstrass substitution, $t=\tan\dfrac x2$

$$=\int_0^1\frac{\ln t\cdot\Big[\ln t-\ln(1+t^2)\Big]}{1+t^2}dt~=~\int_0^1\frac{\ln^2t}{1+t^2}dt~-~\int_0^1\frac{\ln t~\ln\big(1+t^2\big)}{1+t^2}dt~=~\frac{\pi^3}{16}-K,$$

where $~K=2~\Im\bigg[\text{Li}_3\bigg(\dfrac{1+i}2\bigg)\bigg]-\dfrac{\pi^3}{64}-\dfrac\pi{16}\ln^22-\underbrace{\beta(2)}_\text{Catalan}\ln2.~$ It follows then that our two

definite integrals possess a closed form expression if and only if $~\text{Li}_3\bigg(\dfrac{1+i}2\bigg)$ has one as well. As

an aside, $~\Re\bigg[\text{Li}_3\bigg(\dfrac{1+i}2\bigg)\bigg]=\dfrac{\ln^32}{48}-\dfrac5{192}~\pi^2~\ln2+\dfrac{35}{64}~\zeta(3).~$ Also, $~K=\displaystyle\sum_{n=1}^\infty\frac{(-1)^n~H_n}{(2n+1)^2}$.

  • $\begingroup$ +1 Mr. Lucian, your method is easy to follow but I have a question, how to evaluate $$\int_0^1\frac{\ln t~\ln\big(1+t^2\big)}{1+t^2}dt$$ $\endgroup$ Sep 3, 2014 at 10:41
  • 1
    $\begingroup$ @Lucian Do you mind explaining how you compute $\displaystyle K=\sum^\infty_{n=1}\frac{(-1)^n H_n}{(2n+1)^2}$? Thanks. $\endgroup$ Sep 3, 2014 at 10:53
  • $\begingroup$ @SuperAbound: I didn't. Mathematica did. Basically, I struggled for hours to find an expression which my CAS could finally evaluate, since all other intermediary results available on this page were impenetrable, and, in the end, I actually did. Then I posted the result. As far as the integral is concerned, I tried the two main approaches: series expansion and the Feynman trick, both of which failed miserably, but your Euler sum gives me hope. $($How on earth did you even come up with it, anyway?$)$ My only idea would be to write $H_n$ as a sum, then switch the order of the two summations. $\endgroup$
    – Lucian
    Sep 3, 2014 at 15:03
  • $\begingroup$ @V-Moy: In my opinion, SuperAbound's Euler sum seems the way to go. Indeed, many such similar sums have been solved on this site, so I am rather hopeful. $\endgroup$
    – Lucian
    Sep 3, 2014 at 15:18
  • 1
    $\begingroup$ @Anastasiya-Romanova 秀 check this link math.stackexchange.com/q/3449505 $\endgroup$ Aug 29, 2021 at 19:11

By setting $x=\arctan t$ we have: $$\int_{0}^{\pi/4}\log^2(\cos x)\,dx = \frac{1}{4}\int_{0}^{1}\frac{\log^2(1+t^2)}{1+t^2}.$$ Attack plan: get the Taylor series of $\log^2(1+t^2)$ and integrate it termwise.

Since $$-\log(1-z)=\sum_{n=1}^{+\infty}\frac{z^n}{n}$$ it follows that $$[z^n]\log^2(1-z)=\sum_{k=1}^{n-1}\frac{1}{k(n-k)}=2\frac{H_{n-1}}{n},$$ $$\log^2(1+t^2)=\sum_{n=2}^{+\infty}2\frac{H_{n-1}}{n}(-1)^n t^{2n}.\tag{1}$$ If now we set $$\mathcal{J}_m = \int_{0}^{1}\frac{t^{2m}}{t^2+1}\,dt $$ we have $\mathcal{J}_0=\frac{\pi}{4}$ and $\mathcal{J}_{m+1}+\mathcal{J}_m = \frac{1}{2m+1}$, hence: $$\mathcal{J}_m = (\mathcal{J}_m+\mathcal{J}_{m-1})-(\mathcal{J}_{m-1}+\mathcal{J}_{m-2})+\ldots\pm(\mathcal{J}_1+\mathcal{J}_0)\mp\mathcal{J}_0,$$ $$\mathcal{J}_m = \sum_{j=0}^{m-1}\frac{(-1)^j}{(2m-2j-1)}+(-1)^m\frac{\pi}{4}=(-1)^m \sum_{j\geq m}\frac{(-1)^j}{2j+1}.\tag{2}$$ From $(1)$ and $(2)$ it follows that: $$\int_{0}^{\pi/4}\log^2(\cos x)\,dx=\frac{1}{2}\sum_{n=2}^{+\infty}\frac{H_{n-1}}{n}\sum_{r\geq n}\frac{(-1)^r}{2r+1},\tag{3}$$ and summation by parts gives:

$$\int_{0}^{\pi/4}\log^2(\cos x)\,dx=\frac{1}{4}\sum_{n=2}^{+\infty}(H_n^2-H_n^{(2)})\frac{(-1)^n}{2n+1}.\tag{4}$$

UPDATE: the question is now set in an answer to another question. This site (many thanks to @gammatester) is devoted to the evaluation of sums like the one appearing in the RHS of $(4)$. Through Euler-Landen's identity (see the line below $(608)$ in the linked site) it is not too much difficult to see that the RHS of $(4)$ depends on $\operatorname{Li}_3\left(\frac{1+i}{2}\right)$ as stated in the @Lucian's answer.

  • 1
    $\begingroup$ +1 for your answer before I sleep Mr. Jack D'Aurizio (ô‿ô) $\endgroup$ Sep 2, 2014 at 18:00
  • 1
    $\begingroup$ another approach: using the following identity: $$\displaystyle \frac{\ln^2(1-x)}{1-x}=\sum_{n=1}^{\infty}\left(H_n^2-H_n^{(2)}\right)x^n$$ replacing $x$ with $-x^2$ and integrating both sides from $0$ to $1$ gives: $$\int_{0}^{1}\frac{\log^2(1+x^2)}{1+x^2}\ dx=\sum_{n=1}^{\infty}(H_n^2-H_n^{(2)})\frac{(-1)^n}{2n+1}$$ $\endgroup$ May 2, 2019 at 1:35

The strategy in this post will be included in another paper.

A solution (in large steps) by Cornel Ioan Valean

In my opinion, this is a very magical & powerful way that manages to circumvent the necessity of using the already famous method proposed by Random Variable which I think most posts on MSE use it for such integrals. It's time for a new way to come in place and join the existing one!

In this post, we magically prove that $$\int_0^1\frac{\log x\log(1+x^2)}{1+x^2}\textrm{d}x=-\frac{\pi}{16} \log ^2(2) - \log (2)G-\frac{\pi ^3}{64}+2\Im\biggr \{\operatorname{Li}_3\left(\frac{1+i}{2}\right)\biggr \},$$ by wisely combining a result from the book, (Almost) Impossible Integrals, Sums, and Series, namely the special Fourier series (see eq. 3.284, page 244, and eq. 3.288, page 247), \begin{equation} \begin{aligned} \small \sum_{n=1}^{\infty} (-1)^{n-1}\left(\psi\left(\frac{n+1}{2}\right)-\psi\left(\frac{n}{2}\right)-\frac{1}{n}\right)\sin(2nx)&\small=\sum_{n=1}^{\infty} (-1)^{n-1}\left(\int_0^1 t^{n-1}\frac{1-t}{1+t} \textrm{d}t\right)\sin(2nx)\\ &=-\cot(x)\log(\cos(x)), \end{aligned} \end{equation} where $\displaystyle 0< x<\frac{\pi}{2}$, and the Cornel's integral,

$$\int_0^{\pi/2} x\frac{\log(\cos x)}{\sin x}\textrm{d}x=2\log(2)G-\frac{\pi}{8}\log^2(2)-\frac{5}{32}\pi^3+4\Im\left\{\text{Li}_3\left(\frac{1+i}{2}\right)\right\},$$ already calculated in this post How can you approach $\int_0^{\pi/2} x\frac{\ln(\cos x)}{\sin x}dx$.

Proof: We differentiate both sides of the Fourier series that leads to $$2 \sum_{n=1}^{\infty} (-1)^{n-1}\left(\int_0^1 t^{n-1}\frac{1-t}{1+t} \textrm{d}t\right)n\cos(2nx)=1+\frac{\log(\cos(x))}{\sin^2(x)},$$ and if we multiply both side by $x \sin(x)$ and integrate from $x=0$ to $x=\pi/2$, we arrive at $$\int_0^{\pi/2} x\sin(x)\textrm{d}x+\int_0^{\pi/2}x\frac{\log(\cos(x))}{\sin(x)}\textrm{d}x$$

$$=2 \log (2)-1+2 \log (2)\underbrace{\int_0^1 \frac{\log (x)}{1+x^2}\textrm{d}x}_{\displaystyle \text{Trivial}}+\frac{1}{2}\underbrace{\int_0^1 \log (x) \log \left(1-x^2\right)\textrm{d}x}_{\displaystyle \text{Trivial}}$$ $$+\frac{1}{2}\underbrace{\int_0^1\frac{\log (x) \log \left(1-x^2\right)}{x^2}\textrm{d}x}_{\displaystyle \text{Trivial}}-2\underbrace{\int_0^1\frac{ \log (x) \log \left(1-x^4\right)}{1-x^4}\textrm{d}x}_{\displaystyle \text{Beta function in disguise}}$$ $$+2\underbrace{\int_0^1\frac{x^2 \log (x) \log \left(1-x^4\right)}{1-x^4}\textrm{d}x}_{\displaystyle \text{Beta function in disguise}}+2\color{blue}{\int_0 ^1 \frac{\log (x) \log(1+x^2)}{1+x^2}\textrm{d}x},$$ from which the desired result follows.

Note the following values of the Beta function forms in disguise:

$$\int_0^1 \frac{\log (x) \log \left(1-x^4\right)}{1-x^4} \textrm{d}x=\frac{1}{16}\int_0^1 \frac{\log(x)\log (1-x)}{ x^{3/4}(1-x) } \textrm{d}x$$ $$=\frac{7 }{4}\zeta (3)+\frac{\pi ^3}{32}-\frac{3}{16}\log (2)\pi ^2-\frac{\pi }{4}G-\frac{3}{2}\log(2)G,$$ and $$\int_0^1 \frac{x^2\log (x) \log \left(1-x^4\right)}{1-x^4} \textrm{d}x=\frac{1}{16}\int_0^1 \frac{\log(x)\log (1-x)}{x^{1/4}(1-x)} \textrm{d}x$$ $$=\frac{7}{4} \zeta (3)+\frac{3}{2} \log (2)G-\frac{1}{4} \pi G-\frac{3}{16}\log(2)\pi^2-\frac{\pi ^3}{32}.$$

A note: this method can also be adjusted to extract other very difficult integrals, which is possible by further exploiting and developing ideas like the ones in the paper A symmetry-related treatment of two fascinating sums of integrals by C.I. Valean.

End of story

  • 2
    $\begingroup$ (+1) hats off to Cornel. His magic never stop. $\endgroup$ Aug 20, 2020 at 22:47

we can prove, using the same strategy of Random Variable, the following equality:

$$\int_0^{\pi/4}\ln^2(\cos x)\ dx=\frac7{192}\pi^3+\frac5{16}\pi\ln^22-\frac12\ln2G-\text{Im}\operatorname{Li_3}(1+i)$$ proof :
\begin{align*} \ln(1+e^{2ix}) &= \ln (e^{-ix}+e^{ix}) + \ln(e^{ix}) \\ &= \ln(2\cos x) + ix \end{align*}

squaring both sides and integrating, we get

$$\int_0^{\pi/4}\ln^2(1+e^{2ix})\ dx=\int_0^{\pi/4}(\ln(2\cos x)+ix)^2\ dx$$ equating the real parts on both sides and rearranging the terms, we have:

$$ \int_0^{\pi/4}\ln^2(\cos x)\ dx=\int_0^{\pi/4}(x^2-\ln^22)\ dx-2\ln2\int_0^{\pi/4}\ln(\cos x)\ dx$$ $$+\text{Re}\int_0^{\pi/4}\ln^2(1+e^{2ix})\ dx$$ $$=\frac{\pi^3}{192}-\frac{\pi}{4}\ln^22-2\ln2\left(\frac12G-\frac{\pi}{4}\ln2\right)+\text{Re}\int_0^{\pi/4}\ln^2(1+e^{2ix})\ dx$$ $$=\frac{\pi^3}{192}+\frac{\pi}{4}\ln^22-\ln2G+\text{Re}\int_0^{\pi/4}\ln^2(1+e^{2ix})\ dx \tag{1}$$

Evaluating the last integral: \begin{align*} I&=\text{Re}\int_0^{\pi/4}\ln^2(1+e^{2ix})\ dx=\frac12\text{Im}\int_1^i\frac{\ln^2(1+x)}{x}\ dx\\ &=\frac12\text{Im}\left(\ln(-i)\ln^2(1+i)+2\ln(1+i)\operatorname{Li_2}(1+i)-2\operatorname{Li_3}(1+i)\right)\\ &=\frac{\pi^3}{32}+\frac{\pi}{16}\ln^22+\frac12\ln2G-\text{Im}\operatorname{Li_3}(1+i)\tag{2} \end{align*} Plugging $(2)$ in $(1)$ we get our result.

note that we used: $$\ln(-i)=-\frac{\pi}{2}i$$ $$\ln(1+i)=\frac12\ln2+\frac{\pi}{4}i$$ $$\operatorname{Li_2}(1+i)=\frac{\pi^2}{16}+\left(\frac{\pi}{4}\ln2+G\right)i$$ which give us: $$\ln(-i)\ln^2(1+i)=\frac{\pi^2}{8}\ln2+\left(\frac{\pi^3}{32}-\frac{\pi}{8}\ln^22\right)i$$ $$\ln(1+i)\operatorname{Li_2}(1+i) =-\frac{\pi}{4}G-\frac{\pi^2}{32}\ln2+\left(\frac12\ln2G+\frac{\pi^3}{64}+\frac{\pi}{8}\ln^22\right)i$$


my approach to problem $(3)$: \begin{align} I&=\int_0^1\frac{\ln x\ln(1+x^2)}{1+x^2}\ dx=-2\int_0^{\pi/4}\ln(\tan x)\ln(\cos x)\ dx\\ &=-2\int_0^{\pi/4}\ln(\sin x)\ln(\cos x)\ dx+2\int_0^{\pi/4}\ln^2(\cos x)\ dx\\ &=-\int_0^{\pi/2}\ln(\sin x)\ln(\cos x)\ dx+2\int_0^{\pi/4}\ln^2(\cos x)\ dx\\ &=-\left(\frac{\pi}{2}\ln^22-\frac{\pi^3}{48}\right)+2\left(\frac7{192}\pi^3+\frac5{16}\pi\ln^22-\frac12\ln2~G-\text{Im}\operatorname{Li_3}(1+i)\right)\\ &=\frac3{32}\pi^3+\frac{\pi}8\ln^22-\ln2~G-2\text{Im}\operatorname{Li_3}(1+i) \end{align}

note that we evaluated the first integral using the derivative of beta function and as follows: \begin{align} J&=\int_0^{\pi/2}\ln(\sin x)\ln(\cos x)\ dx=\frac18\frac{\partial^2}{\partial{a}\partial{b}}\beta(a,b)\Bigg\rvert_{a\to1/2,~b\to1/2}\\ &=\frac18\beta(a,b)\left(\left(\psi(a)-\psi(a+b)\right)\left(\psi(b)-\psi(a+b)\right)-\psi^{(1)}(a+b)\right)\Bigg\rvert_{a\to1/2,~b\to1/2}\\ &=\frac18\beta(1/2,1/2)\left((\psi(1/2)-\psi(1))^2-\psi^{(1)}(1)\right)\\ &=\frac{\pi}8\left(4\ln^22-\zeta(2)\right)\\ &=\frac{\pi}2\ln^22-\frac{\pi^3}{48} \end{align}


Evaluation of $\int_0^1\frac{\ln(x)\ln(1+x^2)}{1+x^2}dx:$

First note that

$$\int_0^1\frac{\ln(x)\ln(1+x^2)}{1+x^2}dx=\sum_{n=1}^\infty (-1)^n H_n\int_0^1 \ln(x) x^{2n}dx$$ $$=\sum_{n=1}^\infty\frac{(-1)^nH_n}{(2n+1)^2}=\sum_{n=0}^\infty\frac{(-1)^nH_n}{(2n+1)^2}.$$

We have here



Multiply both sides by $\frac{(-1)^n}{(2n+1)^2}$ then $\sum_{n=0}^\infty$ we get













$$=\int_0^1\ln x\left(\int_0^1\frac{t}{(1+t^2)(1+tx)}dt\right)dx$$




$$\int_0^1\frac{\ln(x)\ln(1+x)}{1+x^2}dx=3\ \text{Im}\operatorname{Li}_3(1+i)-\frac{5\pi^3}{64}-\frac{3\pi}{16}\ln^2(2)-2\ln(2)\ G,$$

we get

$$\int_0^1\frac{\text{Li}_2(-t)}{1+t^2}dt=-3\ \text{Im}\operatorname{Li}_3(1+i)+\frac{7\pi^3}{96}+\frac{3\pi}{16}\ln^2(2)+\frac32\ln(2)\ G.\tag3$$

Substitute $(2)$ and $(3)$ in $(1)$, we get


  • 1
    $\begingroup$ (+1) Ingenious work. $\endgroup$ Aug 21, 2020 at 11:28
  • $\begingroup$ @user97357329 Thank you :) $\endgroup$ Aug 21, 2020 at 11:30

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.