Chemical Vapor Deposition (CVD) epitaxy is the method of choice to produce high-quality layers in the semiconductor industry. The demand for increasing performances at contained costs motivates the design of modelling strategies able to predict growth rates and structure morphology. In this talk I will discuss the application of MulSKIPS [1], an open-source super-Lattice Kinetic Monte Carlo (KMCsL) code developed by our group and already demonstrated for PVD and LA processes, to the simulation of CVD growth of Si-based materials. Here, a crucial step addresses the description of the homogeneous and heterogeneous chemical reactions taking place in the reactor in terms of atomistic Monte Carlo events. Moreover, unlike usual LKMC codes, the super-lattice nature of MulSKIPS allows to simulate the spontaneous formation and evolution of defects as a function of the process parameters. Taking as an example the SiH2Cl2+H2+HCl reactor chemistry, simulation results will be shown for different initial substrate geometries and compared against experimental data.