I am trying to prove that $$ \mathcal{L}[J_0(\sqrt {t^2+2t})] = \frac {e^{ \sqrt {s^2+1}}}{\sqrt{s^2+1}} $$

where $ \mathcal{L}[f(t)]$ is the Laplace transform of f(t) and $J_0(f(t))$ is the p-Bessel function where p = 0

My approach


The p-Bessel function is defined as: $$J_p(x) = \sum_{m=0}^\infty \frac{(-1)^m}{m!\Gamma(m+p+1)}\left(\frac{x}{2}\right)^{2m+p}$$ With $p = 0$

and the Laplace Transform of a function as: $$\mathcal{L}[f(t)](s) = \int_0^\infty f(t) e^{-st}dt$$


It is easy to prove that: $$\mathcal{L}\left[J_0(t)\right](s)= \frac{1}{\sqrt{s^2 + 1}}$$ Thus we need to find a way to make $e^{\sqrt {s^2+1}}$ appear in the numerator.

For $p=0$ and $ x = \sqrt {t^2+2t} $ the Bessel function is: $$J_0(\sqrt {t^2+2t} ) = \sum_{m=0}^\infty \frac{(-1)^m}{m! \Gamma(m+1)}\left(\frac{\sqrt {t^2+2t} }{2}\right)^{2m} = \sum_{m=0}^\infty \frac{(-1)^m}{m! \Gamma(m+1)}\left(\frac{t^2+2t}{4}\right)^{m} $$ and then I'm stuck. I am hoping that $e^{\sqrt {s^2+1}}$ will appear somehow through the Gamma function, but I have no idea how to proceed.


Intuitively a better approach seems to be this one: $$ \mathcal{L}[J_0(\sqrt {t^2+2t})] = \frac {e^{ \sqrt {s^2+1}}}{\sqrt{s^2+1}} \iff \\ \mathcal{L}^{-1}[\frac {e^{ \sqrt {s^2+1}}}{\sqrt{s^2+1}}] = J_0(\sqrt {t^2+2t}) $$

The problem seems much easier now, I'll keep this post updated.

  • $\begingroup$ Looks like: $\mathcal{L}_s^{-1}\left[\frac{e^{\sqrt{1+s^2}}}{\sqrt{1+s^2}}\right](t)=J_0\left(\sqrt{-1+t^2 }\right) \theta (1+t)$ $\endgroup$ Mar 11, 2020 at 20:09

2 Answers 2


Your Laplace transform is incorrect, as RHS does not tends to $0$ as $s\to\infty$. The correct version should be $$\mathcal{L}[J_0(\sqrt {t^2+2t})] = \frac {e^{ s-\sqrt {s^2+1}}}{\sqrt{s^2+1}}$$

It can be proved as follows: for any $z,h$, we have $$\tag{1}(z+h)^{-\nu /2}J_\nu (\sqrt{z+h}) = \sum_{m=0}^\infty \frac{(-h/2)^m}{m!}z^{-(\nu+m)/2}J_{\nu+m}(\sqrt{z})$$ then take $z=t^2, h=2t$ gives $$J_0 (\sqrt{t^2+2t}) = \sum_{m=0}^\infty \frac{(-1)^m}{m!}J_{m}(t)$$ Taking Laplace transform both sides, using $\mathcal{L}[J_\nu(t)] = \frac{(s+\sqrt{1+s^2})^{-\nu}}{\sqrt{1+s^2}}$ completes the proof.

For a proof of $(1)$: view it as a function of $h$, expanding at $h=0$ gives $$(z+h)^{-\nu /2}J_\nu (\sqrt{z+h}) = \sum_{m=0}^\infty \frac{h^m}{m!}\frac{d^m}{dz^m}(z^{-\nu/2}J_\nu(\sqrt{z}))$$ Basic properties of Bessel functions shows that $$\frac{d}{dz}(z^{-\nu/2}J_\nu(\sqrt{z}))=-\frac{1}{2}z^{-(\nu+1)/2}J_{\nu+1}(\sqrt{z})$$ then use induction.


If we complete the square under the square root and then make the substitution $u = t+1$, we get $$\begin{align} \int_{0}^{\infty} e^{-st} J_{0} (\sqrt{t^{2}+2t}) \, \mathrm dt &= \int_{0}^{\infty} e^{-st} J_{0} \left( \sqrt{(t+1)^{2}-1}\right) \, \mathrm dt \\ &= e^{s} \int_{1}^{\infty} e^{-su} J_{0} (\sqrt{u^{2}-1}) \, \mathrm du. \tag{1}\end{align}$$

Integral $(1)$ is the case $\beta=1$ of the integral formula $$\int_{1}^{\infty} e^{-st}J_{0} (\beta\sqrt{t^{2}-1}) \, \mathrm dt = \frac{e^{-\sqrt{s^{2}+\beta^{2}}}}{\sqrt{s^{2}+\beta^{2}}}, \quad \left(s\ge 0, \ \beta \in \mathbb{R} \right). $$

See this answer.

Therefore, $$\mathcal{L} \left[J_0(\sqrt {t^2+2t})\right](s) = \frac{e^{s-\sqrt{s^{2}+1}}}{\sqrt{s^{2}+1}}, \quad s \ge 0. $$

More generally, we have $$\mathcal{L} \left[J_0(\sqrt {t^2+2 rt})\right](s) = \frac{e^{r(s-\sqrt{s^{2}+1})}}{\sqrt{s^{2}+1}}, \quad s,r \ge 0, $$ since $$\int_{r}^{\infty} e^{-st}J_{0} (\beta \sqrt{t^{2}-r^{2}}) \, \mathrm dt = \frac{e^{-r\sqrt{s^{2}+\beta^{2}}}}{\sqrt{s^{2}+\beta^{2}}}.$$


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.