Artificial Chemistry seeks to explore how life-like systems can emerge from a pre-biotic environment. This thesis begins with the background of this research area and a re-implementation of an existing Artificial Chemistry as a case study. From this basis, ingredients and properties of Artificial Chemistries are identified. This leads to a novel form of molecular representation – sub-symbolic. A group of novel Artificial Chemistries called RBN-World is developed using Random Boolean Networks as a sub-symbolic molecular representation. It is shown that RBN-World has several properties of interest, and variants of RBN-World and elemental subsets with those properties are identified from many alternatives. This thesis concludes by comparing RBNWorld to the case study and properties discussed earlier, and identifies avenues for future work.