Colloids are ubiquitously present in nature and inspire studies of material design. One reason for this is that the colloidal particles are available to observe under a microscope and can be highly customized on the component, shape, size and surface chemistry. The assembly process is a crucial bridge connecting these microscopic properties with macroscopic functions and applications. Colloidal clusters as an intermediate state between single particles and bulk systems are widely observed during the assembly behaviors, and the conformations of which hint frustration and disorder in the interested material synthesis. To further understand how clusters would influence material formation, we explore the thermodynamics and kinetics of small clusters during this early stage of colloidal assembly.One significant limitation in material design is the existence of competing free energy minima. Free energy calculation is a powerful tool to probe into the assembly process. However, application of such methods is often hindered by high energy barriers, complicated pathways and slow convergence of algorithms. Alongside fundamental work exploring cluster assembly, we additionally attempt to validate a parallel bias strategy to simplify the high-dimensional calculation problem with multiple parallel 1-dimensional simulations, which is beneficial when exploring complex systems in material science and biology. This method is likely to have a significant impact when collaborating with recently developed machine learning free energy methods.