Abstract
Sorbent-assisted atmospheric water harvesting (AWH) has great potential for decentralised sourcing of drinking water in many arid regions of the world. Metal–organic frameworks (MOFs) offer great potential for this application because of their structural diversity and the possibility of their ‘for purpose’ design and functionalisation. The total number of synthesised MOF structures is difficult to estimate – the CoRE MOF 2019 database contains ~14,000 MOF structures, while the CSD MOF subset contains a striking number of ~100,000 structures. Considering the number of existing and theoretically possible structures, to significantly reduce the cost in time and resources for the best materials selection the use of computational techniques validated by experiment is crucial.Despite being the current state-of-the-art method, grand canonical Monte Carlo (GCMC) simulations for water adsorption have been reported as inefficient and prone to errors. The source of these problems lies in the mechanism of water adsorption in MOFs, which is overwhelmingly dominated by adsorption through the formation of water clusters that fill entire pores. In the course of the classical GCMC simulation, water clusters are difficult to form and break because of the large energy barriers between states. For this reason, we advocate the use of the grand canonical transition matrix Monte Carlo (GC-TMMC) method, which allows to calculate the free energy of the system, from which the adsorption isotherm can be determined. As the computational cost of the transition matrix methods is high, we propose a new interpolation scheme using the NVT + ghost swap method along with GC-TMMC simulations. Our approach makes it possible to efficiently determine the full adsorption isotherm and to distinguish between stable and metastable states, while dramatically reducing the time and number of required simulations (Figure 1). It makes also possible the temperature extrapolation of the free energy profile of the system. Thus, on the basis of only a few simulations, we can predict the adsorption properties of the system over a wide range of temperatures and pressures. This not only offers the potential to improve high-throughput screening studies in sorbent-assisted AWH using MOFs but also enhances the ability to optimise other processes requiring temperature/pressure swing.