Exponentially small heteroclinic breakdown in the generic Hopf-zero singularity (Q2377294)

Following the notation introduced in the paper, a Hopf-zero (or central) singularity is a vector field \(X^*: \mathbb{R}^3 \to \mathbb{R}^3\) with the origin as a critical point and whose linear part near the origin, after a linear change of variables, is of the form \[ DX^{*}(0,0,0) = \left( \begin{matrix} 0 & \alpha^* & 0 \\ -\alpha^* & 0 & 0 \\ 0 & 0 & 0 \end{matrix} \right), \] for some \(\alpha^* \neq 0\). The paper deals with generic analytic families \(X_{\mu, \nu}\) of vector fields on \(\mathbb{R}^3\) depending on two parameters \((\mu, \nu) \in \mathbb{R}^2\) such that \(X_{0,0} \, = \, X^*\). After some changes of variables (using normal form theory) and some rescalings, the considered families of vector fields can be presented as \[ \begin{matrix} \displaystyle \frac{d\bar{x}}{dt} & = & \bar{x}(\nu - \beta_1\bar{z}) + \bar{y}(\alpha_0+\alpha_1 \nu+ \alpha_2 \mu+ \alpha_3 \bar{z}) \, + \, O_3(\bar{x}, \bar{y}, \bar{z}, \mu, \nu), \\ \displaystyle \frac{d\bar{y}}{dt} & = & -\bar{x}(\alpha_0+\alpha_1 \nu+ \alpha_2 \mu+ \alpha_3 \bar{z}) + \bar{y}(\nu-\beta_1 \bar{z}) \, + \, O_3(\bar{x}, \bar{y}, \bar{z}, \mu, \nu), \\ \displaystyle \frac{d\bar{z}}{dt} & = & -\mu+\bar{z}^2+\gamma_2(\bar{x}^2+\bar{y}^2)+\gamma_3\mu^2+\gamma_4 \nu^2+\gamma_5\mu \nu \, + \, O_3(\bar{x}, \bar{y}, \bar{z}, \mu, \nu). \end{matrix} \tag{1} \] The coefficient \(\alpha_0\) corresponds to \(\alpha^*\) after a rescaling and the coefficients \(\beta_1\) and \(\gamma_2\) depend exclusively on the vector field \(X^*\). \newline We quote the main result of the paper. \noindent { Theorem} Consider the system ({1}) with \(\mu, \beta_1>0\) and \(|\nu|<\beta_1\sqrt{\mu}\), which has two critical points \(S_{\pm}(\mu, \nu)\) of saddle-focus type. Then there exists a constant \(C^*\) depending on the full jet of \(X^*\) such that the distance \(\bar{d}^{u,s}\) between the one-dimensional stable manifold of \(S_{-}(\mu, \nu)\) and the one-dimensional unstable manifold of \(S_{+}(\mu, \nu)\) when they meet the plane \(\bar{z}=0\) is given asymptotically, as \(\mu \to 0\), by: \[ \bar{d}^{u,s} = \mu^{-\frac{\beta_1}{2}} e^{-\frac{\alpha_0\pi}{2\sqrt{\mu}}} e^{\frac{\pi}{2}(\alpha_0h_0-\frac{\alpha_1\nu}{\sqrt{\mu}}+\alpha_3)} \left(C^* + O\left(\frac{1}{\log(1/\mu)}\right) \right), \] where \(-h_0\) is the coefficient of \(\bar{z}^3\) in the third equation of system ({1}). As a corollary, if \(C^* \neq 0\) and \(\mu\) is sufficiently small, the breakdown of the heteroclinic connection is proved. We remark that the formula for the distance \(\bar{d}^{u,s}\) is explicit and, thus, one can numerically check whether the one-dimensional heteroclinic connection is broken. We observe that the formula for \(\bar{d}^{u,s}\) is an exponentially small function in the perturbation parameter. The given result is a first step towards providing conditions to ensure the existence of a Šil'nikov bifurcation.