In the realm of mathematics, two-dimensional (2D) spaces are akin to an infinitely thin sheet of paper, confining existence and movement to a single plane. In the physical world, we find materials that closely emulate this concept. These 2D materials are typically a thousand times thinner than a human hair and as thin as a few atoms. Their unique dimensionality allows for the ballistic transfer of heat and electricity within the plane. Despite their thinness, these materials exhibit robust mechanical properties, making them truly remarkable. Among these 2D materials, a promising class known as transition metal dichalcogenides (TMDs) stands out. These materials, akin to atomically thin sandwiches arranged in a honeycomb pattern, hold potential for a wide range of applications, from electronic transistors in smartphones to solar cells and batteries. However, their fabrication presents a significant challenge.

Figure 1: Interesting properties of 2D materials (1)

The creation of these 2D materials can be likened to an artist painting a masterpiece on a canvas, but with an unpredictable twist. Each brush stroke must create a perfectly aligned and continuous layer of paint. Any misplaced strokes are unacceptable. Even minor mishaps can dramatically alter the entire portrayal, transforming a tranquil landscape into a turbulent seascape. This is the challenge faced by scientists working with 2D materials, where microscopic structural defects can drastically alter material properties and affect the electron transport landscape. However, just as an unexpected change in portrayal might add depth to an artist’s work, certain defects can enhance a material’s properties. Nevertheless, scientists must understand and control these changes to exploit them effectively. Substitutional oxygen, a prominent defect in TMDs, can not only alter their properties but also accelerate their decay through oxidation, as seen in tungsten disulfide (WS2). Imagine the despair of an artist watching their masterpiece self-consume, changing and degrading before their eyes.

In this study, we aim to combine two different approaches to tackle the issue of defects in TMDs. Traditional experimental and computational approaches to studying 2D materials, while complementary and powerful, are often set up under idealistic conditions, with complexities common in practical settings often minimized. With recent advancements in operando nanometrology, we can now observe these materials in action under near-ambient conditions using experimental techniques such as Environmental Scanning Electron Microscopy (SEM) and ellipsometry. However, these techniques alone cannot provide a complete atomic-level understanding. High-performance computing can be rewarding in this regard. First principles quantum mechanics, such as density functional theory (DFT), have been employed with great success in computationally predicting material properties and energetics. However, the cost of simulations can lead to a disparity in reproducing realistic complexity, and simulations are often limited to small, undiversified systems.

Figure 2: Schematic of an operando SEM investigating individual 2D material islands (left). A neural network representation (middle). Concept of ML interatomic potential applied to 2D materials (right).

Machine learning (ML) models can help bridge this complexity gap. By developing ML interatomic potentials trained on DFT data, we can model larger and more complex systems with the accuracy of DFT but at a fraction of the cost. In my project, I will be using a combination of data from operando experimental techniques and ML interatomic potentials to model and understand the oxidation kinetics of the 2D material tungsten disulfide (WS2). Using a technique called the Nudged Elastic Band (NEB) method, we can calculate the energy barriers for different defect movements, providing insights into how oxidation occurs at the atomic level. This will help us understand and control the properties of these fascinating 2D materials, bringing us one step closer to harnessing their full potential and, metaphorically, saving the painter’s canvas!

References:
(1) Rojaee, R., & Shahbazian-Yassar, R. (2020). Two-dimensional materials to address the lithium battery challenges. ACS nano, 14(3), 2628-2658.

Jad Jaafar

NanoDTC PhD Student, c2022