SuPERPORES project

Our society currently relies mostly on fossil fuels, harmful to the environment, for its energy consumption. Due to the damaging nature of fossil fuels consumption, there is an urging need to move towards more sustainable forms of energy production. In the future, we will likely depend more on renewable energies such as solar or wind energies. On top of the challenges associated with the efficiency of harvesting such energies, there are issues related to the storage of the energy produced. Indeed, the sun does not shine at night and the wind does not blow at all times. As a consequence, developing powerful and reliable energy storage systems is a necessary step to move towards renewable energies. In addition to large-scale storage, such systems can be used to reduce greenhouse gas emissions in applications such as regenerative energy braking or start-stop systems.

In this context, supercapacitors are of great interest as they exhibit very high rates of charge/discharge, long cycle lifes, and they are made of cheap and light materials. These attractive properties arise from the electrostatic nature of the charge storage which results from ion adsorption in the electrode pores. In 2006, it was demonstrated that ions can enter pores of sub-nanometer sizes leading to a huge increase of capacitance. This was an important breakthrough as the energy density of supercapacitors, relatively low compared to batteries, is what currently limits their application.

While this finding has generated a great deal of technological activity to refine potential devices and fundamental research to examine the underlying molecular phenomena, no major improvements of the energy density were observed. The progress towards more powerful supercapacitors is limited by our incomplete understanding of the relation between their performance, in particular their capacitance and charging rate, and the complex structure of the porous carbon electrodes. To make progress we need a better fundamental understanding of the ion transport and electrolyte structure in the pores but we are lacking the experimental and theoretical methods to do so.

The aim of SuPERPORES is to carry out a systematic multi-scale simulation study of supercapacitors. The combination of molecular and mesoscopic simulations will allow us to calculate the capacitive and transport properties of a wide range of systems. In particular, molecular simulations will be used to model ordered three-dimensional porous carbons. This will allow us to vary geometric descriptors, e.g. pore size and ion size, in a systematic way and obtain relevant microscopic information for the subsequent computational screening of porous carbons, achieved through very efficient lattice simulations. We will then be able to formulate design principles for a new, and much improved, generation of supercapacitors. The simulations will also provide other macroscopic properties, e.g. adsorption isotherms and pair distribution functions, which will be used to propose a new method to determine accurately the structure of disordered porous carbons.

At this stage of the project, a number of activities have been undertaken. In order to do simulations of ions confined in three-dimensional porous carbons, we need structures for such materials. This is actually one of the main issues when simulating supercapacitors. Indeed, due to the disordered nature of porous carbons used in energy storage systems, it is a challenge to determine their structure in a precise manner. In the context of this project, the three-dimensional models can be of two types: atomistic structures (for molecular simulations) and pore network models (for mesoscopic simulations). We have worked in collaboration with teams from the University of Cambridge and from Université Picardie Jules Verne on two approaches to obtain such models. The most common method to generate atomistic models is to conduct quench molecular dynamics simulations, i.e. simulations where a liquid is quenched to low temperatures thus leading to amorphous structures. Most of the time, this quench is simulated using interaction potentials parameterised on experimental data. These interaction potentials sometimes exaggerate certain characteristics (e.g. bonds or angles too rigid) leading to inaccuracies in the final structures. With the progress of machine learning, it is now possible to generate interaction potentials with a precision equivalent to Density Functional Theory (DFT) methods and with a computing cost equivalent to usual interaction potentials. Following such methodology, we have generated porous carbon structures of various densities and characterised their structures. These structures are now being used in molecular simulations of confined ions. For the pore network models, we have followed a strategy to extract pore networks directly from tomography images. While this approach has only been used in the context of Li-air batteries so far, it can also be implemented to obtain realistic carbon structures for mesoscopic simulations of supercapacitors.

Regarding the core of the project, one of the main objectives of the SuPERPORES project is to use molecular simulations to extract key trends for the relationship between geometrical descriptors and structural, dynamic and capacitive properties of the electrode/electrolyte interface. As a step towards this objective, simulations of pure ionic liquids and an organic electrolyte in contact with a set of ordered porous carbons have been conducted. These carbon structures have different pore sizes and topologies. To study more systematically the relationship between structure and properties, simulations with different ion sizes and charges have been conducted. For all simulations, we have calculated diffusion coefficients, ionic densities, coordination numbers and free energy profiles in order to characterise the carbon/electrolyte interface. The most promising carbons showing the highest diffusion coefficients are now used in simulations of model supercapacitors. Regarding mesoscopic simulations, an existing program suitable for the calculation of NMR spectra of adsorbed ions has been adapted to the calculation of capacitive properties. In particular, it is now possible to calculate quantities of adsorbed ions for various potentials and porous carbons with various pore size distributions. The model has been applied to the porous carbons used in the molecular simulations and the results from the two methods are being compared. These results will also be compared with more classical continuum models such as Poisson-Boltzmann and its derivatives. With respect to the development of a new method to determine accurately the structure of disordered carbons, we are in the process of investigating the suitability of Nuclear Magnetic Resonance (NMR) parameters as a constraint, in addition to the usual pair distribution function analysis, to improve the accuracy of the models. As a first step, we are building a database of chemical shifts for molecules or solids having similarities with the porous carbons of interest. We will then work on a machine learning strategy to predict the chemical shifts for more complex structures in reasonable times.

The work undertaken so far goes beyond the state of the art in several ways. The molecular simulations are done on a large set of original porous carbon structures in a systematic fashion. Investigations of the ion size and ion charge effects have not been explored in the past. The mesoscopic model developed, which is used to determine capacitive properties of realistic porous carbons in a very computationally efficient fashion is original because it is intermediate between molecular simulations and classical theories such as the Poisson-Boltzmann or the Stern models. Regarding magnetic properties of porous carbons, previous works have mainly focused on ideal aromatic molecules. Here we enlarge the set of aromatic molecules studied, include defects (e.g. holes in the molecules) and do calculations on periodic amorphous solids. Until the end of the project, we aim at i) disentangling the correlations between the ion size, the pore size, the curvature and the capacitance, ii) proposing new design rules for electrode materials with optimised performances, and iii) providing the community with a program to determine porous carbon structures from experimental results which would allow one to use information from multiple sources including pair distribution functions and NMR.

This project receives funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No 714581).