Development of Self-Assembled Co3O4 Nanocrystal Arrays for Electrocatalytic and Electronic Device Applications

Abstract

Colloidal nanocrystals are powerful building blocks for functional materials because their size, shape, exposed facets, composition, and surface chemistry can be programmed with nanoscale precision during synthesis. Translating this precision into device-level function, however, remains a central bottleneck: conventional deposition methods such as drop casting, spin coating, and dip- or spray-coating tend to produce randomly aggregated, non-uniform films with buried active surfaces, while highly ordered approaches such as Langmuir–Blodgett assembly suffer from transfer difficulties and limited scalability. The behavior expressed by an ensemble of nanocrystals is therefore governed not only by the individual particles but also by how they are organized into a working architecture. This dissertation develops Marangoni-driven assembly as a simple, scalable, and substrate-compatible route to organize colloidal Co3O4 nanocrystals into ordered arrays, and shows that the resulting architecture serves as a common platform that converts nanocrystal precision into both electrocatalytic and electronic device functionality. The assembly proceeds through evaporation-induced surface-tension gradients in a binary toluene/hexane solvent, in which preferential evaporation of hexane drives a thin liquid film to spread, form tear-like droplets, and recede, depositing nanocrystals as an ordered array. Rheometry shows that the PhD./MS 20222004 Jun Beom Hwang (황준범). Development of Self-Assembled Co3O4 Nanocrystal Arrays for Electrocatalytic and Electronic Device Applications (자가조립 Co3O4 나노결정 어레이의 개발과 전기촉매 및 전자소자 응용). Department of Materials Science and Engineering. College of Engineering. 2026. 151 p. Prof. Sanghan Lee solution viscosity stays nearly constant across compositions, so the flow is governed by the evaporation- induced surface-tension gradient rather than by viscosity. Building on this, the evaporation rate is identified as the key process parameter linking vial geometry, Marangoni flow, and nanocrystal transport, and a semi- empirical scaling relation (𝑅evap ∝ 𝐴open 2 /𝑉dead) is derived that predicts an optimal evaporation window of ~70–100 mm3 h–1 (≈1–2% of the solution volume per hour) for scale-up. Under the optimized condition (~91 mm3 h–1), Co3O4 nanocubes assemble into arrays with a uniform particle size (10.22 nm), a controlled interparticle nanogap (2.84 nm), and square-lattice translational and long-range orientational order, as quantified by automated image analysis together with radial-distribution and bond-orientational correlation functions. The process extends beyond planar Si to glass, three-dimensional porous scaffolds, and 4-inch wafer-scale substrates, and is adaptable to other nanocrystal building blocks, establishing it as a general assembly strategy. This ordered-array platform is then used to express facet-defined surface chemistry at the electrode level for electrocatalytic nitrate reduction to ammonia–an ambient route that upcycles a common water pollutant into a value-added chemical. Co3O4 nanocubes and nano-octahedra, exposing predominantly {100} and {111} facets through halide-assisted shape control, are assembled into identical array electrodes for a clean facet-level comparison. Density functional theory indicates that the {111} surface has a higher surface energy, stronger nitrate adsorption, and a lower oxygen-vacancy formation energy than the {100} surface, while XPS, Raman, X-ray absorption spectroscopy, and EPR consistently reveal a more Co3+-rich and vacancy-enriched, defect-active {111} surface. A comparison with a drop-cast electrode made from the same nanocubes confirms that the ordered architecture—not merely the intrinsic chemistry—is decisive: the array preserves accessible active facets, sustains higher current and double-layer-capacitance retention, and suppresses particle aggregation and cobalt dissolution. As a result, the Co3O4 {111} array delivers a high ammonia yield rate of 75.64 mg h–1 cm–2 at ~98% Faradaic efficiency under ampere-level partial current density, with stable operation over prolonged cycling; DMPO-assisted EPR and in situ electrochemical Raman spectroscopy further reveal a hydrogenation-active interface and earlier, more pronounced evolution of nitrogen-containing intermediates on the {111} array. The same ordered arrays are finally exploited as structurally discontinuous oxide platforms for electronic devices. Unlike a continuous Co3O4 thin film, the nanocube array confines charge transport to well-defined interparticle nanogaps, so that conduction is dominated by interface-limited processes rather than bulk percolation. Whereas the continuous film shows featureless conduction or irreversible failure under high bias, the ordered array exhibits stable volatile memristive switching at an ultralow operating current of ~10 nA with highly uniform set/reset voltages (coefficient of variation below 9%). Bias-dependent analysis reveals successive Ohmic, Schottky-emission, and Fowler–Nordheim tunneling regimes, and time-resolved measurements show dual relaxation dynamics that combine fast electronic recovery on the microsecond scale with slower ionic recovery—behavior well suited to energy-efficient, temporally dynamic devices. Taken together, this dissertation establishes Marangoni-driven assembly not as a mere film-forming step but as a structural design strategy, and demonstrates that coupling nanocrystal precision with electrode- level architectural control translates colloidal chemistry into scalable, high-performance functionality across both sustainable electrocatalysis and next-generation electronic devices.