Surface and Subsurface Reaction Mechanisms in Atomic Layer Deposition of Metals and Metal Oxides

Research output: ThesisDoctoral Thesis

Abstract

Nanotechnology has shown remarkable versatility and strength in response to large-scale challenges facing society today, despite many of its technical applications being on the atomic scale. From renewable energy devices to medical and sequencing technologies, to three-dimensional transistor architectures, advanced water purification, and novel organic-inorganic hybrid materials, nanotechnology has enabled powerful advances through precisely creating materials with specific chemistries and nanostructures. Thin films are particularly robust in their applications, and atomic layer deposition (ALD) is a thin film deposition technique with demonstrated strengths in precision, tunability, and structural control. The core principles of ALD that allow it to achieve these powerful results rely on self-limiting surface chemical reactions; however, despite the prevalence of ALD reports in the literature there are still many surface mechanisms that are poorly understood. ALD material properties can be highly sensitive to process conditions, impurities introduced from surface reactions, subtle changes in reaction rates, and many more phenomena, so it is critical to fully understand the surface reaction mechanisms at play in ALD to effectively implement processes and design chemistries for new materials. Especially as ALD is often idealized to behave in a simple self-limiting manner, the presence of more complex surface reactions that deviate from this behavior necessitates deeper study. As a result, this dissertation presents work to find, characterize, and model new reaction mechanisms in ALD that cause deviations from ideal behavior, then generalize that understanding and apply it to new chemical systems. The first half of the work focuses on activating surface species in the ALD of metal oxides. ALD of iron oxide using ozone is investigated as a case study, and we find that during the process ozone generates reactive oxygen species that migrate below the surface of the growing film where they are stored. The expansion of the ALD reactions beyond the surface of the film to a reservoir of active species in the subsurface region has a host of implications on the ALD process and the resulting material. The storage of reactive species results in high growth rates, and the physical movement of species through the film causes preferential crystalline rearrangement and film roughening. Further studies of nickel oxide ALD found related behaviors, indicating it grows by a similar mechanism of subsurface active species storage. In both cases, the oxygen species are reactive enough to activate surface combustion reactions, including in ALD of other materials grown on top of the reactive reservoirs. These mechanisms are consistent with reports of some other oxides, suggesting oxygen mobility and oxidizability of the metal center may be important factors in facilitating this reservoir mechanism. The second half of this thesis focuses instead on surface passivating species in metal ALD. As ALD hinges on self-limiting reactions resulting from the surface being passivated toward further reaction, the persistence of passivation is key to consistent and precise process function. A promising precursor for ruthenium ALD, Ru(DMBD)(CO)3, is studied due to its unique passivation mechanism of L-type ligands bonded to a zero-oxidation state metal center. Some reports have hypothesized this bonding results in excellent nucleation and growth properties, while others that it results in poor deposition control with better applications in continuous-deposition processes. By studying this ALD process, we can then gain insight into both surface passivation mechanisms and broader principles for process design of metals and zero-oxidation state compounds. We find the precursor undergoes a spontaneous decarbonylation reaction mechanism whereby the surface is initially passivated with carbonyl species that are lost with increasing temperature or time. Comparison of in situ characterization data and first principles kinetic modeling support these findings. Our results help explain inconsistencies in previous reports as well as observations of other zero-oxidation state precursors. Together, characterizing new surface mechanisms in ALD of both activating and passivating species gives a more complete picture of how a range of ALD processes can deviate from idealized and simplified self-limiting surface reactions. These findings span the breadth of both metal and metal oxide ALD, and we apply the insights to new ALD systems involving multicomponent and catalytically activated ALD processes. These grow the chemical toolbox of ALD and illustrate how a proper fundamental chemical understanding of ALD is important not only for effective implementation and control of existing processes but also for generalizing, expanding, and designing tools for ALD to further widen its horizons.
Original languageEnglish (US)
QualificationDoctor of Philosophy
Awarding Institution
  • Stanford University
Supervisors/Advisors
  • Bent, Stacey, Advisor, External person
Date of AwardMar 11 2022
Place of PublicationStanford, CA
Publisher
StatePublished - Mar 11 2022
Externally publishedYes

Fingerprint

Dive into the research topics of 'Surface and Subsurface Reaction Mechanisms in Atomic Layer Deposition of Metals and Metal Oxides'. Together they form a unique fingerprint.

Cite this