From guesswork to predictive control: Decoding metal-organic precursor chemistry

Metal-organic (MO) precursors are the chemical building blocks at the heart of atomically precise complex oxide materials. Yet in vapor-phase deposition techniques like MOCVD, ALD, and hybrid-MBE, they have long been treated as a “black box”—their reactions poorly understood and often dismissed as “just another knob to tweak.”

A new study, recently published in npj Computational Materials changes that. By combining computationally intensive quantum mechanics with the efficiency of ReaxFF, and metadynamics, researchers mapped the complete reaction landscape of titanium isopropoxide (TTIP), a common precursor for complex oxide growth. The team revealed hidden steps, potential roadblocks, and byproduct pathways, transforming MO precursor chemistry into a more predictable and controllable process.

“Metal-organic precursors are the workhorses of complex oxide growth,” said lead author Nadire Nayir, head of PDI’s Computational Materials Science group. “Understanding their reaction pathways allows precise element incorporation, lowers evaporation temperatures, and improves control over material composition and stoichiometry.

“Yet the real challenge, lies in the reactions’ complexity. Molecules branch into multiple paths—some yield useful products, others end in metastable byproducts or dead-ends. These can slow or even trap the process. For decades, chemists struggled to predict which pathways would succeed.”

Nayir underscored the dedication of the team’s talented and self-driven Ph.D. students—Benazir Yalcin Fazlioglu (co-advised by Roman Engel-Herbert and Adri van Duin) and Cem Sanga (advised by Nayir)—for tackling this challenge. The team’s efforts led to the development of a multiphysics framework that—unlike previous models—bridges thermodynamic driving forces and kinetic constraints, enabling reliable predictions in complex systems beyond the reach of equilibrium models.

“This strategy lets us understand and eventually control reactions that were previously opaque,” Nayir said. “As Harald Schäfer noted 50 years ago, ‘without knowledge of reaction pathways, one cannot control or exploit them.’ Now, we can anticipate reaction outcomes and refine our models in real time.”

Collaboration was key: Simulations were led at PDI with contributions from Penn State and Istanbul Technical University. “One of the most exciting parts of this project was the constant dialog with experimentalists, which was vital in shaping and refining our model,” she added, crediting Roman Engel-Herbert, PDI’s director and leading the h-MBE experimental efforts, for his invaluable discussions and guidance.

Engel-Herbert emphasized the impact of this collaboration. “Before this work, the process was somewhat of a black box. Working closely with the simulation team allowed us to think about our experiments differently. Now, we can see the reaction landscape, including metastable intermediates and dead-end pathways, which helps us design smarter synthesis strategies.

“This project highlights the power of dialog between theory and experiment, enabling us to see problems through each other’s eyes.”

The project also nurtured young talent. Through PDI’s outreach efforts, Physics undergraduate Irem Erpay from Istanbul Technical University made meaningful contributions to the research, showing that high-impact science isn’t limited to Ph.D.-level work.

By opening the black box of precursor chemistry, the team is laying the groundwork for more efficient, predictable, and scalable nanomaterials manufacturing. “This is just the tip of the iceberg,” Nayir said. “Our ultimate goal is to move from trial-and-error chemistry to predictive synthesis—faster material development, less waste, and precise atomic control—a major step toward efficient and reliable thin-film manufacturing.”