Battery Power, Products & Technology Magazine and Conference

Battery Power

Quantum Simulations Based Lithium-Ion Battery Materials Database and Design Engine Software: Accelerating the Pace of Li-Ion Battery Materials Development

NanoeXa has developed and is commercializing quantum simulation (QS) based Li-ion battery (LIB) materials software, which includes system level database and design engine software that determines the structure and relevant properties of LIB electrode materials. The design engine software, when combined with rapid synthesis, characterization and performance testing processes, enables faster and lower cost development of higher performance, safer and lower cost battery materials. NanoeXa’s multi-level materials to system level database and design engine software enables simulation of battery material properties at the system level for both coin cells and cylindrical cells.

The electronic structures and properties of common LIB electrode materials such as transition metal oxides or phosphates are complex and demand high accuracy QS. Due to the cubic scaling of the time relative to the number of atoms in the material models, the regular QS for big models on a first-principles basis are a formidable task even for single crystal materials containing very simple elemental structures. Modeling of the solid solution or of the composite materials including the transition metal oxides or phosphates for battery applications inevitably requires large super cells that could contain from several tens to hundreds of atoms in thousands of combinatorial mixtures.

Figure 1. NanoeXa QS Software based design and optimization cycle of Li-ion battery materials and systems.

The commonly used principle of battery material modeling by QS is based on the trial and error strategy to search a candidate structure located on a total energy minimum in a multi-dimensional space spanned by several lattice constants and 3N coordinate variables for N-atoms per unit cell. However, deducing the final coordinates of a possible structure starting with a random arrangement of atoms is a computationally demanding task, and currently impossible for complex materials. The efficiency of a high-throughput search of new materials using QS, thus, is significantly dependent on both the starting points and the efficiency of reaching the final structures on the multi-dimensional landscape.

NanoeXa is focused on fast and pragmatic solutions to the real world commercial needs rather than on the further development of the theoretical multi-scale modeling solutions popular in the academia and big laboratories. Alternatives of pure QS database mining and high throughput screening also take long and are not linked to system level design and testing cycles on computers, which are an important key to the rapid applications development and the final commercialization of the designed material.

The main bottleneck in the use of QS for high-throughput search, design and optimization of advanced battery materials is the lack of efficient and fast search and design algorithms that can avoid time consuming utilization of very large computing resources for the complex composite and solid-solution battery materials in the multi-dimensional landscape of performance, safety, longevity and cost. With the development of such search and design algorithms working on an appropriate QS database, a fast and high-throughput search and design engine using QS is feasible for the accelerated development of new low cost LIB materials for high power and capacity along with better safety and longer cycle life.

Figure 2. (Clockwise from Top Left) GUI, and examples of simulated structures, specific capacity, and XRD spectra predicted by the NanoeXa QS software.

NanoeXa has developed and has patents pending for such a database as well as search and design algorithms for fast and high-throughput development of solid-solution/composite electrode materials for LIBs using QS. The QS based approach is based on two facts: advances in parallel computing platforms have made feasible highly accurate QS on smaller building block materials structures (not necessarily unit cells) for building a database using moderate computing resources on a regular basis, and the materials structures of the complex solid-solution and composite materials for LIB applications are generally derivative of the constitutive parent building block materials.

Thus, the ability to reliably model and build an extensive database of building block material structures is the basis on which NanoeXa’s proprietary search and design algorithms predict the battery related behavior of the designed structures at the intrinsic materials level, such as the specific capacity, voltage, volume change on charge and discharge, and relative safety and cycling characteristics. Whereas the big multi-scale modeling efforts that attempt to link quantum simulated materials behavior with macro level battery cell design and testing have not succeeded thus far, due to too many levels and variables involved and too many CPU hours required for each trial run at the component level.

While we start with DFT (density function theory) level QS to build our proprietary database for predicting intrinsic materials level battery properties, our proprietary middle level physics-based modeling algorithms enable us to predict the system level discharge behavior of the quantum designed electrode materials against standard electrolyte and standard counter electrodes. If needed, we also have the ability to add custom electrolyte and/or custom counter electrode material properties to our database and test against those as well.

The GUI implements a query and answer methodology that can be used by engineers as well as materials scientists to first design and test target materials for the intrinsic voltage, capacity, discharge performance, safety and cost characteristics. The results are then fed to system level software that enables the prototyping and testing, in computers, of the newly designed LIB electrode materials against standard or customized electrolytes and counter electrode materials in full cell configurations. This enables us to screen a very large set of possible materials compositions and reduce a very large number of candidates to the few very best materials compositions and morphologies, which are then synthesized, characterized, and tested in the lab, validated through higher fidelity full QS, and then added to our ever-expanding proprietary database.

Based on the ever-expanding database and search and design algorithms, our QS software is a CAD tool to guide materials synthesis and performance testing, which is implemented for LIB materials and will be extended in the future to other green energy applications such as solar, fuel cells, and ultra capacitors. For the validation of the software, NanoeXa has successfully completed a number of in-house and external short term battery materials structural characterization and performance prediction projects. We are currently engaged in the design and development of new battery materials in-house and in partnerships with strategically selected global partnerships.

Our software has three layers. At the bottom is the QS data base which delivers fast predictions of the intrinsic battery materials properties information. The middle level involves the prediction of coin cell level discharge curves against standard or customized electrolytes and counter electrodes, and at the top level is a system level design and testing tool that enables the full simulation of the cylindrical or flat prismatic cell design, testing, and properties comparison against target specifications.

The current speeds of design and testing cycles are limited by the available computing resources. These speeds are bound to improve as the database, design algorithms, and application areas expand. Our ongoing software development is focused on extending the simulation capabilities of QS to increase the efficiency and reduce cost across the LIB value chain from materials to cell to module to pack level simulations and to interface with battery management system design and validation software. In addition we will support SAE and UL standards for verification and validation testing and acceptance as these standards are approved and posted. While new applications will require extension, development and optimization of new database formats and algorithms, we believe our current base platform enables considerable saving in both time and money to address new applications versus the development of equivalent performance software de novo.