I am trying to solve the following integral (which can be seen also as a Hankel Transform):

$\int_{0}^{\infty} \frac{1}{\left( k-1 \right)^2} J_0(k r) \, dk$

where $J_0(k r)$ is the zeroth order Bessel function of the first kind, and $r$ is a positive real number.

My idea was to use integration by parts considering that $\frac{1}{\left( k-1 \right)^2} = \frac{d}{dk} \left(\frac{-1}{\left( k-1 \right)} \right)$ that is to say:

$\int_{0}^{\infty} \frac{1}{\left( k-1 \right)^2} J_0(k r) dk= \left[\frac{-J_0(k r)}{\left( k-1 \right)} \right]_{0}^{\infty}+r \, \int_{0}^{\infty} \frac{1}{\left( k-1 \right)} J_1(k r) dk$

however i cannot find any solution for the integral $\, \int_{0}^{\infty} \frac{1}{\left( k-1 \right)} J_1(k r) dk$

also, as this problem comes from a physical problem, it is important that the solution (which is a function of $r$) has to go to zero at $r \rightarrow \infty$ (however it can diverge at $r=0$), do you know if this can easily checked?

I checked Proudnikov Integrals and Series Vol.2 but I couldn't find anything helpful.

do you have any hint?

  • $\begingroup$ Don't you think that the integral diverges at $k=1$, or are you looking for a principal part integral? $\endgroup$ – Paul Enta Mar 16 '17 at 23:02
  • $\begingroup$ I know that the integrando diverges for $k=1$. But maybe the integral can still have a closed form. For example the integral of $\frac{J_0 (k r)}{k-1}$ has a closed form in terms of Struve H and Bessel Y functions (from proudnikov vol2 pag176) $\endgroup$ – SSC Napoli Mar 16 '17 at 23:07

One can decompose $$I=\int_0^\infty \frac{J_1(kr)}{k-1}\,dk=-\int_0^\infty J_1(kr)\,dk+\int_0^\infty \frac{kJ_1(kr)}{k-1}\,dk$$. Which can be evaluated as$$I=-\frac{1}{r}-\int_0^\infty \frac{d}{dr} \frac{J_0(kr)}{k-1}\,dk$$ Though validity of the inversion of the order integration / derivation is unclear, using the expression you quote: $$\int_0^\infty \frac{J_0(kr)}{k-1}\,dk=\varphi(r) $$ you get$$I=-\frac{1}{r}-\varphi'(r)$$

  • $\begingroup$ Thanks a lot for your answer. Is it possible to exchange the derivation sign in the integration? I should have a look at Leibniz rule. Also from the physical meaning of the problem it is important that the solution goes to zero as r goes to infinity . Whereas from your result it is clear that it will go to 1 $\endgroup$ – SSC Napoli Mar 17 '17 at 9:41

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.