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.

Consider the function $f:\mathbb{R}^3\rightarrow \mathbb{R}$, $f(x) = |x|^{-1}$. It is locally integrable, and its distributional Fourier transform is $F(f)(k) = g(k) = 4\pi/|k|^2$. Intuitively, the long-range $|k|^{-2}$ decay of $g(k)$ comes from the singularity at $x=0$ in $f$.

Is this true? I.e., if $\chi$ is the characteristic function of an open, bounded set containing $0$, and consider $h = F(\chi f)$. Is it true that $h = O(|k|^{-2})$ as $|k|\rightarrow +\infty$?

At least, if the open bounded set in question is a ball of radius $R$, a direct computation gives $$ h(k) \propto \frac{sin^2(|k|R/2)}{|k|^2} = O(|k|^{-2})$$

In this case, the hypothesis is verified, since $R$ was arbitrary.

share|improve this question
add comment

1 Answer

Asking that $\chi$ be the characteristic function of an arbitrary open set creates difficulties, related to the fact that $F(\chi)$ decays slowly itself, that do not relate to the main phenomenon here. If instead we ask that $\chi=1$ in a neighborhood of zero, and is $C^\infty$ with compact support, then $h(k)$ is proportional to the convolution $$ h(k) \propto \int_{\mathbb R^3} \frac{F(\chi)(z)}{|k-z|^2}\,d^3z. $$ Because $F(\chi)$ is smooth and decays rapidly now, and $1/|k-z|^2$ is in $L^1_{\rm loc}$ and is bounded by, e.g., $4/|k|^2$ for $|k-z|\ge |k|/2$, one can perform the estimate $$ |h(k)|\le C \left( \frac4{|k|^2}\|F(\chi)\|_{L^1} + I(|k|) \sup_{|z-k|<|k|/2} |F(\chi)(z)|\right) = O(|k|^{-2}), $$ where $$I(k) = \int_{|z-k|<|k|/2} |z-k|^{-2}d^3z \propto |k|.$$ Note $$ \sup_{|z-k|<|k|/2} |F(\chi)(z)| \le \sup_{|z|\ge |k|/2}|F(\chi)(z)| $$ and this decays rapidly with $|k|$.

share|improve this answer
Thanks! I am also very interested in the case where $\chi$ is the characteristic function of, say, a box. This is because I am working with Fourier series of $1/r$ in the box: $F(\chi f)$ is analytic (Paley-Wiener theorem), and the Fourier coefficients are in fact $\beta F(\chi f)(\alpha k)$ for some constants $\alpha,beta$, depending on box size and normalization. –  Simen K. Oct 1 '13 at 12:54
add comment

Your Answer


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.