Is there a closed form solution for the following integral

$$\int_{0}^{\pi/4} e^{-(n^2\sec^2x)/2}\,dx$$ for $n>0$ ?

  • 7
    $\begingroup$ Why do you think that there should be one? $\endgroup$ – mrtaurho May 12 at 19:37

It's not that hard. We have the error function: $$\int e^{-x^2/2}dx=\sqrt{\frac{\pi}{2}}\operatorname{erf}\left(\frac{x}{\sqrt 2}\right)+C$$

Differentiate under the integral sign after putting $\tan x=t$

$$f(n)=\int_{0}^{\pi/4} \exp\left({-\frac{n^2\sec^2x}{2}}\right)dx=\int_0^1 \frac{\exp\left({-\frac{n^2(t^2+1)}{2}}\right)}{t^2+1}dt$$

$$f'(n)=-n \int_0^1 \exp\left({-\frac{n^2(t^2+1)}{2}}\right)dt=-\sqrt{\frac{\pi}{2}}e^{-n^2/2}\operatorname{erf}\left(\frac{n}{\sqrt 2}\right)$$ Now integrate back using $f(0)=\frac{\pi}{4}$ $$f(n)=\frac{\pi}{4}-\sqrt{\frac{\pi}{2}}\int_0^ne^{-x^2/2}\operatorname{erf}\left(\frac{x}{\sqrt 2}\right)dx=\frac{\pi}{4}\left(1-{\operatorname{erf}}^2\left(\frac{n}{\sqrt 2}\right)\right) $$

  • 1
    $\begingroup$ Also thought about applying Feynman but couldn't bring it to an end. However, well done (+1) $\endgroup$ – mrtaurho May 12 at 22:09

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.