MINFF: A Family of Classical Force Fields for Minerals with Tunable Angle Terms

The Earth is built from minerals, vast and varied assemblages of atoms that form the bedrock of our planet and the foundation of countless human technologies. From the clay in a riverbank to the silica in a computer chip, these materials control how water moves through soil, how pollutants are trapped, and how energy is stored. To understand them, scientists often turn to computer simulations, creating digital models where atoms interact according to the laws of physics. However, for these models to be useful, they must rely on a set of rules, known as a force field, that dictates how atoms attract, repel, and bond with one another. If these rules are slightly off, the digital mineral might crumble, swell incorrectly, or fail to hold water, rendering the simulation useless for predicting real-world behavior. For decades, researchers have relied on a few standard sets of rules to model these complex structures, but these older methods often struggle to capture the subtle balance between the rigid framework of a mineral and the flexible ways it bends and swells when wet.

A team of researchers at Umeå University in Sweden has introduced a new family of rules called MINFF, designed to fix these long-standing issues for a wide range of minerals, including clays, oxides, and porous frameworks used in filtration. The core of their innovation lies in how they handle the angles between atoms. In many previous models, the angles between metal and oxygen atoms were left completely free to move, which sometimes caused the digital structure to collapse into a shape that does not exist in nature. In other models, these angles were locked down so tightly that the mineral became too stiff to bend realistically. The new MINFF approach finds a middle ground. It uses a single, adjustable rule that controls the angles between every trio of metal and oxygen atoms. By turning a dial on this rule, researchers can make the mineral slightly more rigid or slightly more flexible, allowing the model to match the specific behavior of different materials without needing a completely new set of rules for each one.

The researchers tested this new system against more than thirty different minerals, ranging from simple rocks like quartz and corundum to complex clays like kaolinite and montmorillonite. They compared their results against two of the most widely used existing systems, known as CLAYFF and INTERFACE. The tests showed that the new MINFF rules produced digital structures that were significantly closer to the real, physical minerals measured in laboratories. Specifically, the new models matched the size and shape of the mineral crystals with an error of less than one percent, a level of precision that the older systems struggled to reach. Furthermore, the new models were much better at maintaining the correct internal geometry of the atoms, ensuring that the coordination of atoms around metal centers remained stable and chemically sensible, a metric the researchers measured using a stability index that dropped to very low values with their new method.

One of the most significant findings was how the new rules handled the interaction between minerals and water, a critical factor for understanding soil swelling and water filtration. When the researchers simulated water trapped between the layers of a swelling clay called montmorillonite, the new MINFF rules accurately reproduced the distance between the layers as water was added or removed. This is a difficult task because the clay must expand and contract smoothly without breaking apart. The older models often showed the clay layers behaving too rigidly or collapsing in ways that did not match experimental observations. The new system, particularly when tuned with a moderate angle rule, captured the soft, flexible nature of the clay layers while keeping the internal structure intact. This allowed the simulation to predict how much energy is required to wet the clay and how water molecules move within it, matching real-world data more closely than the previous standards.

The researchers also looked at how these minerals vibrate, which is a way of checking if the digital atoms are moving with the right stiffness. By analyzing the vibrational frequencies of kaolinite, a common clay mineral, they found that the new rules produced a spectrum of vibrations that aligned well with what is observed in real experiments. The older models tended to either blur these vibrations into a single broad noise or shift them to the wrong frequencies. The new approach, by carefully balancing the angle rules, resolved these vibrations into distinct peaks that matched the real material. This suggests that the new rules not only get the shape of the mineral right but also capture the dynamic energy of the atoms as they jiggle and shift, which is essential for understanding how minerals react to heat and pressure.

A key feature of this new system is its flexibility and accessibility. The researchers did not just create a single set of rules for one specific rock; they created a family of rules that can be adapted. They developed a general set that works well for a broad variety of minerals, and they also showed how to create a "tailored" set for a specific mineral if even higher precision is needed. This means that scientists can choose the level of detail they require for their specific problem. To make this accessible to others, the team released the computer code and tools needed to generate these models freely, allowing researchers around the world to apply these new rules to their own studies of soil, water, and industrial materials.

The work represents a significant step forward in the ability to simulate the natural world with high accuracy. By refining how the angles between atoms are treated, the researchers have provided a tool that is both more accurate and more versatile than what was previously available. The new system successfully bridges the gap between models that are too loose to hold their shape and those that are too stiff to bend, offering a realistic view of how minerals behave in the presence of water and other chemicals. This improved understanding could help in designing better materials for water purification, more efficient catalysts for chemical reactions, and safer methods for storing radioactive waste, all by ensuring that the digital models used to design them are grounded in the true physical behavior of the atoms.