Finding the right solid electrolyte or electrode material is one of the most persistent bottlenecks in modern energy technology R&D. Whether the goal is to design a faster-charging solid-state battery, a more efficient fuel cell, or a highly selective membrane for water electrolysis, the core challenge remains the same: finding a crystalline material through which ions can travel quickly and efficiently.
Traditionally, R&D planners have been forced to choose between two unsatisfactory options when screening new materials. On one hand, they can run highly accurate quantum-mechanical simulations that model the movement of every atom over time. While precise, these simulations require massive supercomputing resources and can take weeks for a single material, making them far too slow for high-throughput screening. On the other hand, they can use simplified geometric models that look for open spaces in the crystal structure. While fast, these simple models often fail because they ignore the complex electrostatic forces that can block an ion even in a seemingly open channel.
A new computational framework aims to resolve this dilemma by extracting detailed ion transport mechanisms from a single, static crystal structure in a matter of seconds. Developed as a calibration-free method, this approach could allow researchers to rapidly filter out unviable materials before committing expensive computational or experimental resources.
The computational bottleneck in materials discovery
In the search for next-generation energy materials, computational screening acts as a vital filter before physical synthesis and laboratory testing. However, the sheer variety of possible crystal structures means that researchers must evaluate thousands of candidate compounds to find a handful of promising options.
The current gold standard for modelling these materials is Ab Initio Molecular Dynamics (AIMD). This method simulates the physical movement of atoms by calculating the quantum-mechanical forces acting on them at tiny time steps. Because it captures the complex vibrations of the crystal lattice, AIMD provides an accurate picture of ion movement. Unfortunately, this level of detail requires substantial supercomputing time, making it impractical as a primary screening tool for large material databases.
To bypass this bottleneck, researchers often turn to simpler surrogate models that look purely at the geometry of the crystal, searching for continuous pathways of empty space. While these geometric assessments are fast, they frequently produce false positives. They reduce complex, multi-dimensional pathways to a single transport number, ignoring the electrostatic environments and energy barriers that dictate whether an ion can actually make a jump.
How ions navigate the crystal lattice
To understand why simple geometric models fall short, it is necessary to examine how ions move through a solid. Unlike in liquids, where molecules slide past one another freely, atoms in a solid crystal are locked into a rigid, repeating lattice.
An ion travelling through this lattice must move by hopping between stable sites. Between these positions lies a narrow bottleneck where the migrating ion is squeezed by the surrounding atoms of the host crystal. Passing through this bottleneck requires energy to overcome the repulsive electrostatic and physical forces exerted by the lattice.
Furthermore, ion transport is often highly directional, flowing easily along two-dimensional planes but blocked in other directions. A material might possess high geometric openness, yet still be a poor conductor if those open spaces do not connect into a continuous, low-barrier pathway across the entire material. Conversely, a seemingly tight bottleneck might stretch easily to let an ion pass. A predictive screening tool must therefore capture the energetic barriers of these specific pathways without simulating dynamic motion over time.
A rapid network-based screening tool
The newly proposed framework, known as Topo-Spectral Percolation Descriptors (TSPD), addresses this challenge by representing the crystal structure as a mathematical network. Because the method is detailed in a preprint that has not yet undergone formal peer review, its findings remain subject to evaluation, but the underlying mechanics offer a novel way to bridge the gap between speed and accuracy.
Instead of simulating atomic movement over time, TSPD requires only a single, static crystal structure. From this input, the algorithm maps out a periodic migration network where nodes represent stable ion sites and edges represent the pathways connecting them.
To determine whether an ion can traverse these pathways, the framework calculates the energy barriers for each edge using physics-based energetics. Because these calculations are derived from fundamental physical principles rather than being fitted to experimental data, the method is entirely calibration-free and can be applied immediately to entirely new, hypothetical crystal structures.
Once this energy-weighted network is constructed, the framework analyses it using two mathematical concepts. First, it evaluates the barrier-threshold topology. By gradually raising the energy threshold, the algorithm determines the exact point at which individual hopping pathways connect to form a continuous, material-spanning network—the percolation threshold that identifies the easiest route an ion can take. Second, it calculates the spectrum of the barrier-weighted graph Laplacian, a mathematical matrix describing the overall connectivity and diffusion characteristics of the network to extract the transport mechanism in seconds.
Validation, boundaries, and pre-publication status
To demonstrate the validity of this approach, the developers tested the TSPD framework against eight different materials, including battery cathodes and solid electrolytes. These test cases represented transport behaviours ranging from one-dimensional channels to three-dimensional networks.
The predictions made by the TSPD model were compared directly against high-fidelity AIMD simulations and experimental neutron diffraction data. In all eight cases, the static network analysis successfully reproduced the established transport pathways and mechanisms. Most importantly, the model correctly identified materials where geometric openness did not lead to viable ion transport, flagging the electrostatic barriers that simple geometric models miss.
However, the static nature of the TSPD model imposes certain physical boundaries. Because it relies on a static structure, the method is most effective for pathway-limited, direction-dependent transport. It cannot fully capture highly cooperative transport mechanisms where multiple ions push each other in a coordinated chain. For materials where transport is strongly cooperative or nearly isotropic, the static network analysis is not intended to replace dynamic simulations entirely. Instead, it acts as a complementary tool, allowing researchers to quickly map out the structural skeleton of the transport pathways before dedicating computational resources to intensive dynamic modelling.
Sources
- Topo-Spectral Percolation Descriptors for Mechanistic Ion Transport Pathways from Static Crystal Structures, Diptendu Roy et al., arXiv preprint, not yet peer reviewed, 2026-10-01
The R&D takeaway
For organisations funding or planning materials discovery for energy applications, the TSPD framework demonstrates that physics-informed network analysis can successfully bypass the need for computationally expensive dynamic simulations during early-stage screening. By filtering out non-viable candidates in seconds rather than days, R&D managers can significantly optimise their high-performance computing budgets. This strategic approach ensures that costly supercomputing resources are reserved exclusively for the most promising solid-state candidates, accelerating the overall pipeline from molecular design to physical prototyping.
The R&D Innovate desk