Tracking Angstrom Resolution of Nano Crystals and Thin Solid Films via Transmission Electron Microscopy

To meet the expectations of down scaling and to use less precious material, Atomic Layer Deposition (ALD) currently sets the benchmark and ultimate limit for miniaturization in manufacturing. The growth of ALD films starts with a nucleation phase differing significantly from the later growth regime. In this phase, individual nm-sized islands start developing to later coalesce to a continuous film. This phase is yet poorly understood. As the demand for films below 10 nm thickness rises, knowledge of this phase becomes indispensable as such films are entirely determined by its specifics. The intend of Trånslate was to perform the first ever in situ atomistic characterization, of the early-stage evolution of three major ALD material system, noble metals, metal-oxides and metal-sulfides. Being prominent examples; Platinum (Pt), group IV oxides and zinc sulfides served as model systems for generalizable insights across all major ALD materials and mechanisms. We collaborated with the Nanoscale Prototyping Laboratory at Stanford University (SU). The project was conducted from April 1st 2018 until April 24th 2022. We started looking into Pt thin films gaining insight into its nucleation, nanostructuring and alloying for increasing its catalytic mass activity. We explored the interaction of Pt atoms with their substrate, created compressive strain through alloying and developed 3D nanostructuring strategies. We achieved a high mass activity to 0.8 A per mg Pt and tuned the catalyst-charge carrier interface of fuel cell electrodes. The nucleation of ALD metallic copper (Cu) thin films for CO2 reduction processes was investigated. We deposited Cu on 3D Nickel and Titanium mesh substrates for Oxygen Evolution Reaction (OER) devices. Substrate dependent nucleation and oxide formation was observed, and Ni based substrates performed better in OER than Ti based ones. Titania (TiO2), zirconia (ZrO2), and hafnia (HfO2) thin films proved valuable in the modification of Magnesium (Mg) and Titanium (Ti) based biomaterials. ALD TiO2 reduced the hydrogen evolution rate of a porous Mg alloys by 68% compared to conventionally produced counterparts. Further, ALD TiO2 and ZrO2 coatings reduced the impact of corrosion towards the mechanical failure (Ultimate Tensile Strength by 6% and 40%, strain at break by 70% and 76%, respectively). We related biocompatibility to corrosion resistance, where cells (L292) cultured on 100 nm thick TiO2 coated Mg alloy showed lower viability than on ZrO2 and HfO2 counterparts. We gained significant insight into the nucleation of zinc sulfide (ZnS), which we observed growing in situ on TiO2 substrates at the Stanford Radiation Lightsource. Sulfide and sulfate species form during the nucleation phase. A high-throughput screening method was developed to determine the most probable atomic configurations as the film structure evolves, which consisted of a supervised machine learning analysis of thousands of simulated X-ray spectra. Atomic-level insight was gained into changes in the coordination of surface species as they transitioned from nucleation toward crystalline ZnS. An ALD in situ Transmission Electron Microscopy (TEM) chip was created and validated that may facilitate in situ studies inside TEMs around the world. During this development, ultrathin (< 50 nm) free-standing alumina tunnel like shell membranes with of 475:1 etch depth to structure width and 9750:1 etch depth to membrane thicknesses aspects were created, which can have implications in the field of sensors. Two unique in situ instruments were developed; one in situ ALD-FTIR tool as well as an in-situ ALD-X-ray spectroscopy tool that will be available to future researchers and is documented in publications. Trånslate led to many future projects including an ERC Starting Grant and several follow up projects with Stanford University and the University of Michigan. 15 papers including work in Nature Materials, Nature Catalysis, ACS Chemistry of Materials and Advanced Materials were published. The ESIAM conference, an internationally recognized conference series with around 150 participants every second year series was established. The PI took a role in the Stanford seminar and lecture series “Affiliate Program”, where results are conveyed to industries from US, Austria, and Norway. The Pt nanoparticle optimization is ongoing through the ERC and projects with Stanford. ALD oxide coatings as boundary layers to the biological is its own research agenda pursued by collaborating researchers at NTNU. The work on sulfides is important for understanding of buffer layers in solar cells and led to a method for discerning the nucleation behavior via X-ray characterization, quantum simulations and machine learning. Many researchers are starting to pursue similar approaches. Our collaborators from the University of Michigan established a related research agenda.

Understanding the nucleation of Pt led to applications in catalysis, e.g. for polymer electrolyte fuel cells. The tuning of Pt particles led to high mass activity reducing one of the main cost drivers in fuel cells. This triggered scientific interest and will lead to follow up projects. Understanding oxides led to an improved corrosion resistance and biocompatibility of Mg and Ti alloys. A closed loop between the X-ray in situ experiments and the description and the nucleation of ZnS could be established through simulations and machine learning, which will be key to investigate the nucleation phase of other thin films. In situ TEM chips were created in a new workflow allowing nm thin, gas tight channels. We reduced cost drivers in fuel cells contributing to climate change mitigation. We worked on biomaterial coatings for their extended use in implants, developed a novel method to study and understand thin films and established a new MEMS process.

To meet the expectations of down scaling and to use less precious material, Atomic Layer Deposition currently sets the benchmark and ultimate limit for miniaturization in manufacturing. The growth of ALD films starts with a nucleation phase differing significantly from the later growth regime. In this phase, individual nm-sized islands start developing to later coalesce to a continuous film. This phase is yet poorly understood. As the demand for films below 10 nm thickness rises, knowledge of this phase becomes indispensable as such films are entirely determined by the specifics of this phase. We propose to perform the first ever in situ ALD Transmission Electron Microscopy experiment allowing atomic scale manufacturing and characterization in situ during film evolution at working conditions. The world leading TEM capabilities available at the Norwegian Centre for TEM will enable these experiments allowing a multitude of crystallographic, chemical and topological characterization within one experiment. Complementary, we will perform state-of-the-art plane wave quantum chemical simulations on nucleation mechanisms under study using experimental results as inputs.