NERSC Helps Uncover the Mechanism Behind the ‘Dolomite Problem’

Science Background

Dolomite, or calcium magnesium carbonate (CaMg(CO3)2), is a major component of carbonate rocks but has historically been an outlier among minerals: while crystals tend to grow in supersaturated solutions at ambient temperatures, dolomite has not been readily produced in the lab – the so-called “dolomite problem.”  ’

Science Breakthrough

Using supercomputers at NERSC, researchers at the University of Michigan led by Assistant Professor and past Lawrence Berkeley National Laboratory (Berkeley Lab) Staff Scientist Wenhao Sun have revealed the key mechanism that inhibits dolomite growth and explained how it forms in nature. Using atomistic simulations, Ph.D. student Joonsoo Kim showed that in a supersaturated solution, dolomite initially precipitates a disordered surface, which inhibits further crystal growth – but also that mild undersaturation of the solution dissolved those disordered areas, allowing them to be rebuilt in a more ordered fashion. By cycling a solution between supersaturation and slight undersaturation, they were able to speed dolomite growth by up to seven orders of magnitude. Validation experiments using in situ liquid cell electron microscopy confirmed their theory. Their work was published in Science in November 2023.

Science Breakdown

To study crystal growth at nanometer length scales and geological timescales, the team developed a Kinetic Monte Carlo (KMC) simulation to observe the disordering of the dolomite surface and its re-ordering. This simulation required quantum-mechanically accurate formation energies of random configurations of disordered dolomite surfaces. However, using Density Functional Theory (DFT) to calculate every step of the KMC simulation was deemed prohibitively expensive.

Instead, they used a cluster expansion model to calculate the energies of random disordered dolomite surface structures while maintaining quantum-mechanical accuracy, a process that took two microseconds on a desktop computer rather than 5,000 CPU hours or more on a supercomputer.

To train the cluster expansion algorithm, the researchers used NERSC’s Cori supercomputer to produce a training set comprising approximately 600 DFT-calculated energies of different configurations of disordered dolomite. They then used the trained cluster expansion algorithm to execute the dolomite surface ordering process via KMC simulation, allowing them to computationally validate their key hypothesis: that dolomite growth, in fact, requires dissolution and the elimination of the Ca/Mg disordering that inhibits the growth of the crystals.

Research Lead

Joonsoo Kim, Department of Materials Science and Engineering, University of Michigan

Co-authors

Yuki Kimura, Brian Puchala, Tomoya Yamazaki, Udo Becker, Wenhao Sun

Publication

Joonsoo Kim et al., “Dissolution enables dolomite crystal growth near ambient conditions.” Science 382, 915-920 (2023). DOI:10.1126/science.adi3690

Funding

The U.S. Department of Energy Office of Science, Office of Basic Energy Sciences, partially funded this work.

User Facilities

National Energy Research Scientific Computing Center (NERSC)