Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

I have considered the following definitions for the Fourier integral pairs:

$\Phi(\omega) = \frac{1}{(2\pi)^3} \int_{\mathbb{R}^3} e^{ix\omega \cos\theta}R(x)x^2 dx \sin\theta d\theta d\phi$

$R(x) = \int_{\mathbb{R}^3} e^{-ix\omega \cos\theta}\Phi(\omega)\omega^2 d\omega \sin\theta d\theta d\phi$

If one takes the derivative from both sides of the first equation with respect to $\omega$, we will have:

$\Phi^{\prime}(\omega) = \frac{1}{(2\pi)^3} \int_{\mathbb{R}^3} (ix \cos\theta) e^{ix\omega \cos\theta}R(x)x^2 dx \sin\theta d\theta d\phi$

If now we evaluate the value of the derivative at $\omega = 0$ we have simply $\Phi^{\prime}(\omega=0) = 0$.

The paris of the functions $R(x) = \frac{1}{(1+\pi^2 x^2/4)^2}$ and $\Phi(\omega) = \frac{1}{\pi^4} e^{\frac{-2\omega}{\pi}}$ satisfy the pair of Fourier integrals, but $\Phi^{\prime}(0)$ is not equal to $0$. Does someone have an idea about why this problem happens?

share|improve this question

1 Answer 1

up vote 1 down vote accepted

Differentiation under integral sign is a delicate operation, subject to some restrictive, and often hard-to-check, conditions. (Witness the mess on that Wikipedia article.) When you approach it without checking those assumptions, don't be surprised by a nonsensical result.

Here is the one-dimensional version of what you encountered. Let $\Phi(x)=e^{-|x|}$. Since $\Phi$ is even, its Fourier integral representation involves only cosine function $\cos t x $. When differentiated, the cosine turns to $-t \sin t x$ which is $0$ at $x=0$. So, the differentiation under integral sign results in the integral being zero when $x=0$. But $\Phi'(0)$ does not exist.

If you dig deeper, you'll find that in this and similar examples the formal differentiation results in the average of one-sided derivatives, in the same way how the Fourier series converges to $\frac12(f(x^+)+f(x^-))$ at a jump discontinuity. In higher dimensions, the property of $\Phi$ being even corresponds to rotational symmetry, and the same averaging effect occurs for any function that depends on the distance to the origin only.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.