REBORN : Optimization and modeling of 2D h-BN epitaxy and its high-quality layered heterostructures using artificial intelligence.

h-BN Applications: vdW epitaxial h-BN can function in several ways (many of which first demonstrated by GT-CNRS IRL 2958). First, used as a release layer, it permits simplified lift-off and transfer of III-N optoelectronic devices (LEDs, solar cells, sensors, high power transistors,etc.) to arbitrary supports, permitting more flexible use and improving thermal performance. Second, used as an active layer, h-BN can be used as a p-doped material and electron blocking layer for for UV LEDS. Third, it can enable heterogenous integration of different semiconductor materials, leading to the highest pixel count three-color LED display so far reported. Finally, it can be used in supercapacitors and hydrogen ion separating units.

Existing obstacles and approach to overcoming them: (1) When h-BN is used as a release layer, the subsequent growth stack has a higher defect density than for current GaN on Sapphire processes. Given that each parameter is a function of time (growth temperature, pressure in the growth chamber, growth rate, III/V flow rate), new techniques are needed to find optimal growth conditions to improve the initial h-BN layer, the subsequent layers, and the 3D GaN-based materials. (2) When h-BN is used as an active layer, improvements to defect density, doping and interfaces with 3D AlGaN materials are needed for UV emission levels to become competitive. (3) Simulations of MOVPE growth dynamics are notoriously unreliable and the parameter space is too large to explore by simulation only. These obstacles point to the need for more advanced characterization of samples in different stages of growth (DIADEM), and the use of AI-guided optimization procedures to identify the right growth parameters to overcome obstacles to the production of higher quality h-BN. Especially relevant is the need to more efficiently explore higher temperature growth regimes, as higher temperature growth will soon become available in the GT- CNRS reactor. The regimes to be studied are (i) formation of the first BN layer on sapphire, (ii) formation of subsequent layers, and (iii) initial growth of 3D materials on h-BN layers.

Partnership: The partners in this project include state-of-the art platforms and leaders in 2-D hBN growth and prototyping: DIADEM partners (SOLEIL, ESRF, METS), CEA/IRIG/MEM Lab, who offers complementing set of techniques for material exploration and is able to provide comprehensive and concrete answers to materials or processes problems. GT-CNRS IRL 2958, who has European leadership role in h-BN epitaxy (publication track record culminating in 2023-2024 publications in Science and Nature). The project also includes ASDL Lab, a Georgia Tech Lab with researchers experienced in the US NSF Materials Genome project and use of AI for improved manufacturing of materials with desired properties. Given the track record of the consortium, the exciting recent developments in h-BN, and the potential impact for production of practical devices, REBORN is poised to produce excellent science with real-world impact.