Introduction

Polytetrafluorethylene is one of the most versatile materials, which is not only used by NASA in their rockets and space suits but also used in our day-to-day non-stick cooking utensils. From electrical appliances to musical instruments to healthcare, construction, and energy, it is used almost everywhere. But to everyone’s surprise, such a widely used material was, in fact, invented by accident while working with refrigerants. Today, it is commonly known by its brand name – Teflon. So, how long will we stay dependent on such coincidences or accidents?

With advances in computing, scientists have started using data and computational models to discover materials suited for specific applications. The process of finding new materials with desirable traits that can be used in a range of applications, such as storing energy, electronics, and medicine, is known as computational materials discovery. It requires creating and implementing new methods and techniques for identifying, synthesizing, and characterizing materials with specific properties.

What is the Traditional way of Material Discovery?

The discovery of new materials begins with defining and identifying the need for a specific material with specific properties, such as high strength, flexibility, or electrical conductivity. To identify and develop new materials that meet these requirements, researchers use a combination of property predictors, high-throughput screening, and exploratory synthesis and characterization techniques. For example, a material used in aerospace applications may need to be lightweight, strong, and able to withstand extreme temperatures.

Traditional computers do ‘material modeling’, where they use computational models to predict the properties of the material based on the structure of its atoms and molecules. Properties such as electronic structure, mechanical properties, and thermal behavior also provide valuable insights into the properties of the identified or given material. Researchers use computational methods to simulate the behavior, including density functional theory (DFT), molecular dynamics simulations, and Monte Carlo simulations. Alongside using atomic and molecular structure, parameters like temperature and pressure are also used for the prediction.

For instance, DFT can be used to calculate the electronic structure of a material, which can provide information on its electrical conductivity, optical properties, and other electronic properties. The mechanical behavior of materials, such as their strength, elasticity, and deformation under stress, can be studied using molecular dynamics simulations. Monte Carlo simulations can be used to study the thermodynamic properties of materials, such as their heat capacity and entropy. After the analysis, the most promising candidates are shortlisted, and further tests and optimizations are performed to refine them and their properties. This may involve modifying the chemical composition, changing the processing conditions, or adding different additives.

What is it lacking?

Although traditional computers are good enough for numerous tasks, there are a few shortcomings when it comes to the material discovery process.

• Simulations are often limited by the accuracy of the underlying models and approximations used, leading to incorrect predictions or deviations from the desirable characteristics. Additionally, materials modeling and high-throughput screening are computationally intensive and require high-performance computing resources that may not always be available.
• Traditional computers are usually limited where analysis of materials is done in idealized environments, including at absolute zero pressure and temperature. Real materials are frequently exposed to a variety of environmental factors, which might have an impact on their properties in ways that classical simulations might not be able to replicate.
• The complicated behavior of materials, such as the formation of unique features at the nanoscale, cannot always be captured by traditional computers and may call for more specialized methods like quantum computing that simulate the “true” quantum behavior of atoms and molecules.

How is Quantum Computing better suited?

Quantum computing is believed to be well-suited for material discovery due to its ability to efficiently simulate the behavior of complex systems. In classical computing, the simulation of a large number of atoms or molecules can quickly become computationally infeasible, but quantum computing can handle these large-scale simulations much more efficiently.

One of the main advantages of quantum computing is its ability to represent and manipulate quantum mechanical systems, which are the fundamental building blocks of materials. Quantum mechanical systems can be simulated using quantum algorithms such as the quantum version of molecular dynamics. This can enable the discovery of new materials and chemical compounds with specific properties, such as high thermal conductivity or superconductivity, that would be difficult or impossible to predict using classical computing methods. Additionally, the ability of quantum computing to perform optimization problems makes it an attractive option for finding the best structure and material properties from a given set of candidates.

Our Startup Partner – QpiAI Offerings

QpiAI provides products and solutions for accelerating materials discovery pipelines through their simulation platform QpiAI-Sim. The platform is equipped with tools that use AI and quantum computing to accelerate the discovery of next-generation materials and molecules by efficiently designing and navigating huge chemical spaces. For example, their force field calculations using QpiAI-Sim-Force can scale high-fidelity simulations and provide 100,000x speed-ups compared to traditional CCSD methods and 1,000x compared to DFT. Their QIO solver QpiAI-Opt provides up to 100x faster combinatorial optimization performance compared to traditional commercial solvers on million-scale problems with near-optimal solutions. It can build scalable algorithms for active chemical space exploration in materials discovery. Similarly, QpiAI-Sim-Gen uses generative modeling for the inverse design of functional materials with desired properties. Such methods can be applied across domains such as pharmaceuticals, energy, conversion and storage, aerospace, automotive, carbon capture, and many others. Many of the use cases that QpiAI has already developed include quantum models for automated drug discovery, organic LEDs, organic binders, and MOFs where exhaustive chemical exploration using traditional methods is not possible.

Conclusion

In the foreseeable future, Quantum computing will undoubtedly become more crucial to material discovery. It would become feasible to replicate materials at an even more granular and precise level as technology develops, leading to the discovery of new and improved materials with a wide range of applications. With its ability to perform complex simulations, it will help researchers to identify new materials with improved properties and characteristics and also to gain a deeper understanding of existing materials. While the technology is still maturing, early applications are on the horizon in the next 2-5 years. Organizations must start preparing themselves today for quantum disruption and adopt these advanced technologies in their workflows.

Thanks to QPiAI Team for their valuable input to the blog.