In this thesis we have expanded the applicability of POMSimulator from isopolyoxometalates to heteropolyoxometalates. To tackle the increasing complexity derived from the expansion to heteropolyoxometalates (HPA) systems, we have developed two main methodologies. First, we propose a general equation for the scaling of formation constants, universal for all polyoxometalates and independent of the Density Functional Theory (DFT) method, based on Multi-Linear Regression (MLR) models. Second, we have developed a statistical workflow based on clustering techniques to manage the large number of speciation models generated by POMSimulator.

Using these data-driven approaches we have successfully simulated the aqueous speciation of two HPA systems such as the phosphomolybdate (PMo) and the arsenomolybdate (AsMo) in agreement with experimental results. We have predicted the formation of the well-known PMo Keggin PMo₁₂ anion in contrast with the absence of the equivalent structure in the AsMo system. We have also confirmed the importance of the As₂Mo₆ and AsMo₉ species in the arsenomolybdate system.

We have reported the first speciation phase diagrams for HPA systems giving a general overview on the chemistry of heteropolyoxometalates. The current development of POMSimulator opens the door for further studies in the field and builds up the synergy with experimental studies to improve the understanding of the complex self-assembly of polyoxometalates.

If you would like to follow the PhD thesis defense via zoom, please, register here!