Fluctuations of the rightmost particle in the catalytic branching Brownian motion (Q2135122)

Consider a catalytic branching Brownian motion with a single catalytic point at the origin. This process can informally be defined as a branching particle system in which each particle moves according to an i.i.d. Brownian motion, such as a particle at position \(x\) branches at rate \(\beta \delta_0(x)\), where \(\delta_0\) is the Dirac mass at \(0\). For all \(t \geq 0\), we write \(R_t\) for the position of the maximal particle at time \(t\). It is known [\textit{S. Bocharov} and \textit{S. C. Harris}, Electron. Commun. Probab. 21, Paper No. 70, 12 p. (2016; Zbl 1346.60128)] that \(R_t - \frac{\beta}{2} t\) converges in distribution to a random variable. The aim of the present article is to study the almost sure fluctuations of \(R_t - \frac{\beta}{2} t\), namely \[ \limsup_{t \to \infty} \frac{R_t - \frac{\beta}{2}t}{\log t} = \frac{1}{\beta}\ \text{ and }\ \liminf_{t \to \infty} \frac{R_t - \frac{\beta}{2}t}{\log \log t} = -\frac{1}{\beta}\text{ a.s.} \] The proofs techniques are analogue to the one used to prove similar results for spatially homogeneous branching processes [\textit{Y. Hu} and \textit{Z. Shi}, Ann. Probab. 37, No. 2, 742--789 (2009; Zbl 1169.60021)], using first and second moment methods to estimate precisely the probability of presence of a particle around position \(\frac{\beta}{2}t + O(\log t)\).