I've been asked to find $\int{artanh(x)} dx$

The first thing I did was I said:

let $y = artanh(x)$

$\therefore \dfrac{dy}{dx} = \dfrac{1}{1-x^2}$

$\therefore dx = (1-x^2)dy$

Also, from $y = artanh(x)$ I know that:

$tanh(y) = x$

$\therefore 1-x^2 = 1-tanh^2(y)$

From this is can therefore say:

$\int{artanh(x)} dx = \int{y(1-tanh^2y)}dy$

$= \int{y- y(tanh^2y)}dy$

$= \int{y}dy - \int{y(tanh^2y)}dy$

$= \frac{1}{2}y^2 - \int{y(tanh^2y)}dy$

To solve $\int{y(tanh^2y)}dy$ I tried integrating by parts, so I said:

let $u = y$ and let $\dfrac{dv}{dy} = tanh^2y$

$\therefore \dfrac{du}{dy} = 1$ and $v = \int{1 - sech^2y}dy$

$\therefore v = y - tanhy$

If I substitute these values into the equation:

$uv - \int{v}{\dfrac{du}{dx}}$

I get:

$\int{ytanh^2y}dy = ytanh^2y - \int{y - tanhy}dy$

I can therefore say that:

$\int{artanh(x)} dx = \frac{1}{2}y^2 - ytanh^2y + \int{y - tanhy}dy$

$= \frac{1}{2}y^2 -ytanh^2y + \frac{1}{2}y^2 -lncoshy + C$

$ = y^2 -ytanh^2y -lncoshy + C$

But $y = artanhx$

$\therefore y^2 = artanh^2x $

and $tanh^2y = (tanh(artanhx))^2 = x^2$

and $coshy = cosh(artanhx)$

To simplify this part I used substitution:

let $u = artanhx$

I'm therefore trying to find $cosh(u)$

from $u = artanhx$ I know that:

$tanhu = x$

$\therefore tanh^2u = x^2$

$\therefore 1- sech^2u = x^2 $

$\therefore cosh^2u = \dfrac{1}{1-x^2}$

$\therefore coshu = \dfrac{1}{\sqrt{1-x^2}}$

From this I can therefore say that:

$\int{artanh(x)} dx = artanh^2x - x^2(artanhx) - ln(\dfrac{1}{\sqrt{1-x^2}}) + C$

$\therefore \int{artanh(x)} dx = artanhx(artanhx - x^2) + \frac{1}{2}ln(1-x^2) + C$

However this isn't the answer, the answer is:

$\int{artanh(x)} dx = x(artanhx) + \frac{1}{2}ln(1-x^2) + C$

I can't see where I've gone wrong. I understand I've probably made this more difficult than it needs to be and therefore probably made a silly mistake somewhere. I've probably missed something that makes the whole question a lot simpler but I can't see what that would be.

Thank you :)



It seems that you used a very complex approach of your problem.

Why don't you integrate by parts using $u=\tanh ^{-1}(x)$ and $v'=dx$.

I am sure that you can take from here and arrive to the result.

  • $\begingroup$ Oh man, I can't believe I missed that! Thank you :D $\endgroup$ – Elise Mar 1 '14 at 9:15
  • $\begingroup$ @Elise. The number of things I missed and miss is close to infinity ! So, don't worry. $\endgroup$ – Claude Leibovici Mar 1 '14 at 9:17

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.