ALD/ALE 2026 Session EM-MoP: Emerging Materials and Processes Poster Session

In this study, high performance 2D MoS₂ photodetectors incorporating multi-heterojunction structures were fabricated through Atomic Layer Etching (ALE). By precisely controlling the thickness of the MoS₂ thin films using an ALE technique, MoS₂ heterojunctions composed of monolayer and multilayer regions were successfully formed. This engineered bandgap modulation facilitated the generation and separation of photo-induced carriers, thereby significantly enhancing the device's photoresponsivity.

Furthermore, the formation of multi-junction interfaces increased the effective contact area with the electrodes, securing superior electrical characteristics. Unlike conventional approaches that rely on hybrid heterostructures, this work distinguishes itself by simultaneously optimizing electrical and optoelectronic performance within a single material platform. These results demonstrate that ALE is a scalable and effective technique for manipulating 2D materials, paving a new way for next-generation optoelectronic applications.

Precise control of doping type and concentration in transition metal dichalcogenides (TMDs) is essential for realizing TMD-based electronic devices, including p-n diodes, field-effect transistors (FETs), and tunnel FETs. However, conventional ion implantation techniques are not suitable for two-dimensional materials due to severe lattice damage and structural degradation.

This study presents an approach for achieving controlled n-type doping and structural phase transition in CVD-grown MoS₂ through atomic-scale plasma-assisted defect engineering. Although radical-based surface doping is generally limited by shallow penetration depth, Argon plasma treatment was applied to the MoS₂ surface, creating sulfur vacancies that effectively introduced n-type doping characteristics. The density of these sulfur vacancies and thereby the doping concentration was precisely tuned by varying the plasma exposure duration.

In addition to doping, the defect formation induced a phase transformation from the semiconducting 2H phase to the metallic 1T phase. Raman spectroscopy confirmed the emergence of new peaks associated with the 1T phase. Raman spectroscopy confirmed the emergence of new peaks associated with the 1T phase, while X-ray photoelectron spectroscopy (XPS) provided evidence of both chemical and structural changes during the process. These plasma – induced modifications resulted in substantial improvements in device performance, such as increased field-effect mobility and higher on/off current ratios. The ability to simultaneously control both carrier type and structural phase makes this method particularly promising for the design of high-performance two-dimensional (2D) electronic devices.

Hafnium-based hybrid photoresists have emerged as promising candidates for extreme ultraviolet (EUV) lithography due to their high EUV absorption, superior etch resistance, and potential for sub-10 nm patterning.^1,2 Beyond their conventional single-tone behavior, our studies have revealed that Hf-based inorganic–organic hybrid materials can exhibit dual-tone responses under EUV exposure, enabling both positive- and negative-tone patterning within the same resist platform. This unique capability opens new opportunities for process simplification and enhanced patterning flexibility in advanced lithographic applications.

In this work, we demonstrate dual-tone EUV patterning using Hf-based hybrid photoresists and systematically evaluate their lithographic performance under both EUV exposure and low-energy electron-beam lithography (100 V EBL). We show that the resist tone can be switched via post-deposition treatment, which induces chemical changes within the Hf-based resist and results in tone inversion from conventional negative to positive tone behavior. The Hf-based hybrid resist films were fabricated via molecular layer deposition (MLD) using TDMA-Hf as the inorganic precursor and 2,3-dimercapto-1-propanol (DMP) as the organic linker. The MLD-derived hybrid resist exhibits high sensitivity, with the critical doses of 17.6 and 18.5 mJ/cm² for negative- and positive-tone modes, respectively. The critical doses required for both negative- and positive-tone patterning are comparable, indicating minimal sensitivity penalty when switching between tone modes.

Mechanistic investigations were conducted using in-situ FTIR and XPS analysis. The results indicate that negative-tone behavior arises from exposure-induced crosslinking, which stabilizes the resist film after development. In contrast, post-treated films exhibit positive-tone behavior, where film hardening is followed by bond scission under EUV or e-beam exposure, leading to enhanced solubility in the developer.

These findings highlight the versatility of Hf-based hybrid photoresists as multifunctional EUV resist materials and underscore their potential for next-generation EUV lithography, where adaptable tone control and simplified process integration are increasingly critical.

This work is supported by the U.S. DOE Office of Science Accelerate Initiative Award 2023-BNL-NC033-Fund. This research is also partially supported by the National R&D program (2022M3H4A3052556) through the National Research Foundation of Korea (NRF), funded by the Ministry of Science and ICT in Korea.

[1] N. Mojarad et al., Nanoscale, 2015,7, 4031–4037.

[2] Y. Wang et al., J. Mater. Chem. A, 2025, 13, 29860.

Atomic-scale control of both the active switching medium and the ionic source is a critical requirement for achieving reliable and scalable resistive memory technologies. While two-dimensional (2D) materials offer an intrinsically thin and well-defined switching layer, resistive switching in filamentary memories is still dominated by stochastic variations originating from uncontrolled metal-ion supply. Here, we demonstrate a resistive memory platform in which the resistive switching behavior is deterministically governed by the simultaneous atomic-layer-level control of a 2D WS₂ switching medium and the confined supply of Ag ions, enabled by ALD/ALE-compatible thickness engineering.

The WS₂ switching layer was synthesized with atomic-scale thickness control, allowing systematic modulation of the vertical transport length and defect density. In parallel, the Ag electrode was intentionally confined in both thickness and areal distribution, limiting the total Ag reservoir available for electrochemical metallization. By independently tuning the WS₂ thickness and the Ag supply, we reveal a clear transition in resistive switching behavior from Ag-dominated metallic conduction to vacancy-mediated switching governed by sulfur-vacancy (Vₛ) percolation. Electrical measurements combined with thickness-dependent statistics show that the high-resistance state is primarily dictated by the WS₂ thickness, whereas excessive Ag supply leads to unstable low-resistance states and increased variability.

Detailed electrical analysis and finite-element simulations indicate that a partially formed Ag tip, generated under confined Ag conditions, produces a highly localized electric-field enhancement at the WS₂ interface. This localized field drives controlled Vₛ migration along grain boundaries without forming a continuous metallic filament. The resulting sequential switching process—comprising partial Ag filament formation, transient space-charge-limited conduction, and eventual vacancy filament percolation—yields sub-percent switching variability and stable non-volatile memory operation. Importantly, ex-situ structural and compositional analyses confirm the absence of residual Ag within the WS₂ layer after switching, highlighting the non-metallic nature of the final conductive pathway.

By demonstrating that resistive switching characteristics can be deterministically programmed through atomic-scale thickness control of a 2D switching medium and precise limitation of metal-ion supply, this work establishes a materials-level design strategy directly aligned with ALD/ALE processing. The presented approach provides a scalable pathway toward highly uniform RRAM and neuromorphic memory devices, in which variability is suppressed not by circuit-level compensation but by atomically engineered material interfaces.

CuNb₂O₆ is an intrinsic p-type metal oxide semiconductor in which hole transport originates from copper-derived electronic states embedded within an orthogonal Cu–O–Nb octahedral framework. This structurally ordered lattice provides stable and directional charge transport pathways, rendering crystallinity control essential for optimizing functional performance.

Herein, CuNb₂O₆ powders were synthesized via a sol–gel route followed by systematic thermal annealing to regulate crystal formation. The annealing-induced evolution of crystal structure and electronic states was comprehensively investigated using X-ray diffraction (XRD) and X-ray photoelectron spectroscopy (XPS). Distinct temperature-dependent crystallization regimes were identified, accompanied by progressive stabilization of the Cu²⁺ and Nb⁵⁺ electronic environments and enhanced lattice ordering.

These results elucidate the coupled structural and electronic evolution of orthogonal CuNb₂O₆ during thermal processing and establish a crystallization window for achieving phase-pure and electronically optimized materials.

Biasing substrates during atomic layer deposition (ALD) in gaining popularity as a method to alter the resulting film. This is hypothesized to alter precursor–surface interactions and provide a means to tune material properties. We performed zinc oxide (ZnO) ALD using diethylzinc (DEZ) and water on silicon with native oxide substrates at 150 °C in a sample holder designed to create a static electrical field by biasing one plate of a parallel plate capacitor during deposition. ZnO films prepared in an electric field/on a biased sample holder were thinner, changed relative crystalline composition, and contained more carbon compared to samples grown under identical conditions without bias. Density functional theory (DFT) calculations showed enhanced electron migration between dissociatively adsorbed DEZ molecules and the ZnO (002) facet with increasing force from an electric field at the substrate surface, which strengthens the electronic interactions between the surface and the adsorbate. These models offer a compelling explanation for the inhibited growth, changes in the crystallinity, and increased carbon content of films grown in an electric field/on biased plates.

Piezoelectric aluminum nitride (AlN) and doped-variant aluminum-scandium nitride (Al_(1-x)Sc_xN) thin films are commercially essential for RF filters for wireless communication and have additional applications as sensors for piezoelectric ultrasound transducers (PMUTs), and energy harvesters. Recently, there have been a flurry of reports of ferroelectric sputtered Al_(1-x)Sc_xN since its published discovery by Fichtner in 2019. Ferroelectric Al_(1-x)Sc_xN has enormous potential as a memory thin film due to its extremely large, switched polarization (>>100 µC/cm²) and Curie temperature > 1000° C, making it suitable for operation in extreme environments. Despite this excitement, there are few reports showing ferroelectricity in nominally undoped aluminum nitride, typically deposited using sputtering, due to the large electric fields (>6 MV/cm) required for switching. Here, we show ferroelectric AlN thin films grown by plasma-enhanced atomic layer deposition (PEALD) grown using ultra-high purity conditions. This is the first demonstration of ferroelectric PEALD AlN and only the second report of any ferroelectric nitride grown by PEALD. The PEALD AlN thin films were deposited at a substrate temperature of 300° C and exhibit room-temperature ferroelectric switching which is enabled by their enormous breakdown fields > 9 MV/cm. Trimethylaluminum (TMA) and alternatively N₂ and N₂/H₂ plasmas were implemented as co-precursors. The PEALD AlN films exhibited the wurtzite phase and grew in the c-axis (0002)-orientation on {111}-oriented Pt bottom electrodes deposited on (001) silicon with a 500 nm-thick thermal oxide.

Despite extensive research on two-dimensional (2D) materials, almost all experimentally synthesized 2D systems derive from van der Waals crystals. Beyond this class, only a limited number of compounds have so far been theoretically predicted to be stable in the 2D limit [1]. In this context, iron–sulfur compounds have recently emerged as promising candidates. Density functional theory predicts that both hexagonal FeS₂ and tetragonal FeS phases can exist as stable monolayers, exhibiting strain-tunable magnetic properties [2-4]. Although the Mermin-Wagner theorem precludes long-range magnetic order in ideal 2D isotropic systems, magnetic anisotropy can lift this constraint and allow stable ordering, enabling tunable magnetism at the atomic scale [5], which is essential for spintronics [2-4], as well as other emerging phenomena such as topological effects [6], multiferroicity [7], and proximity effects in heterostructures [8]. However, compared to their bulk counterparts, Fe–S systems generally display a rich phase diagram, characterized by multiple stoichiometries and atomic arrangements [9]. A similar complexity may also persist in the 2D limit, highlighting the need for a systematic experimental investigation into which 2D FeS_x phases can actually form. For this purpose, we systematically grow and characterize iron sulfide monolayers on Au(111) via in-situ co-deposition of Fe and S. Low-energy electron diffraction (LEED), scanning tunneling microscopy (STM), and X-ray photoelectron spectroscopy (XPS) reveal two primary phases: one with a hexagonal atomic arrangement consistent with FeS₂, forming a moiré superstructure, and one with a square arrangement. By tuning sulfur exposure, we observe additional phases with varying stoichiometry and atomic arrangement, all belonging to the same 5×5 superstructure family. These results confirm the existence of a complex 2D phase diagram for FeS_x monolayers, establishing them as a versatile and tunable platform for exploring 2D magnetism in non-van der Waals systems.

[1] Hongze, Gao et al., ACS Nano, 18, 16343 (2024)

[2] Yang, Ke et al., Physical Review B, 109, 014431 (2024)

[3] Wang, Duo et al., Physical Review B, 104, 245410 (2021)

[4] Lin, Haicheng et al., Nano Res., 11, 4722 (2018)

[5] Huang, Bevin et al., Nature, 546, 270–273 (2017)

[6] Kvashnin, Y. O. et al., Physical Review B, 102, 115162 (2020)

[7] Zhang, Junting et al., Phys. Rev. Letters, 125, 017601 (2020)

[8] Gong, Cheng et al., Science, 363, 4450 (2019)

[9] Espano, Jeremy R. B. et al., J. Am. Chem. Soc., 145, 18948 (2023)

In semiconductor device manufacturing processes, substrate surface roughness should be minimized because it leads to increased electrical resistance and added parasitic capacitance. Critical for device scaling, this is even more so when considering two-dimensional materials. Conventional methods for reducing substrate roughness include atomic layer deposition (ALD) and atomic layer etching (ALE), but a method that could reduce only the substrate roughness via surface chemical reactions without changing film thickness would represent a major breakthrough. We developed a numerical analysis method to evaluate the roughness-reduction effects of novel materials. The planarization capability of new precursors on various substrate materials was ranked for different topographic structures. The goal of the study at this stage is to prove the feasibility of chemical planarization of common interconnect metals.

Numerical evaluation of roughness-reduction effects followed two approaches: molecular dynamics (MD) and direct evaluation of the reaction-pathway thermodynamics. First, machine-learned potentials (MLP) based MD simulations described how novel precursor molecules interact on various metal surfaces with different roughness features. The analysis produced a picture in which certain reactive precursors, upon impinging on rough metal substrates, bond with surface metal atoms promoting their mobility, diffusion and surface flattening. By comparing the MD simulation temperature conditions at which metal surfaces become flattenable, we can rank the relative ease of planarization among different metals including ruthenium, cobalt and gold. Our second approach employed reaction-pathway (thermodynamics) to evaluate the activation energy for step-edge collapse for various metal surfaces. Smaller activation energies imply an easier to planarize surface. The same three metals in the MD study were evaluated and found to scale consistently. We also found that adsorption of the novel precursors reduces the activation energy for step-edge collapse, clarifying a mechanism in which precursor adsorption promotes diffusion of surface metal atoms and thereby enables surface planarization.

This study demonstrates molecular layer deposition (MLD) of alucone and hafnicone thin films using a previously unexplored organic precursor, salicylaldehyde, paired with trimethylaluminum (TMA) and tetrakis(dimethylamido)hafnium(IV) (TDMAHf), respectively. MLD of alucone and hafnicone at deposition temperatures between 120 and 190°C was confirmed with in-situ quartz crystal microbalance (QCM) and chemical analyses. The alucone chemistry was observed to be prone to self-termination, which could be mitigated with two methods: using an exposure process, and using H₂O as a coreactant. Conversely, the hafnicone chemistry showed strong resistance to self-termination owing to the high functionality and large ligand-to-ligand steric hindrance of TDMAHf. The potential of the alucone and hafnicone films as low dielectric constant (k) materials was evaluated, with the alucone and hafnicone films exhibiting k value of 4.2 and 4.7, respectively. The alucone and hafnicone films showed good thermal stability, with ~25% and 20% reduction in thickness upon baking in vacuum at 500°C for 15 min, respectively.

The effect of chemically inert plasmas on the surface of thin film materials is of interest because it allows for an additional means to deliver energy to a material beyond increasing the material’s temperature. This could be of use in cases where energy needs to be delivered to specific layers or interfaces in a heterostructure without subjecting the entire material stack to higher temperatures or to overcome energetic barriers, such as those associated with material crystallization (or amorphization) or the formation of metastable phases. However, plasmas make for a complex environment due to the various energy contributions from ions, electrons, excited species, and photons. Despite this complexity, an increased understanding of energy transfer at the plasmas-surface interface opens additional avenues in material engineering. This may be particularly true for plasma enhanced atomic layer deposition (PEALD) processes where deposited thin films are cyclically exposed to plasmas and thus could benefit from an increased understanding of the interaction between plasmas and material surfaces.

In this work we look at the impact of chemically inert remote Ar plasmas on crystalline VO₂ (c-VO₂) thin films deposited by ALD. Thermal ALD of VO₂ was carried out at 150 °C using TEMAV and O₃ and subsequently crystallized in a dedicated thermal annealing vacuum system using our typical process¹. We demonstrate that under certain plasma conditions we can remove the Raman signature of the c-VO₂ films without any change to the surface morphology of the material, and this capability was greater when operating at lower pressures during the plasma exposure. Additionally, by changing the Ar plasma conditions used, we are also capable of recovering the crystalline nature of the previously damaged films; this was accomplished by treating the sample to subsequent plasma exposures of descending power. This presentation will discuss the details of these experiments and, with the aid of in situ plasma diagnostics, attempt to elucidate the mechanisms responsible for the behavior observed.

¹ J. Phys. Chem. C 2017, 121, 19341−19347

As Si-based devices approach their physical scaling limits, monolithic three-dimensional (M3D) integration has emerged as a promising strategy for continued performance enhancement. In this architecture, transistors are vertically stacked within the back-end-of-line (BEOL) layers, imposing a strict process temperature limit below 400 °C to prevent degradation of front-end CMOS circuits. This restriction necessitates new channel materials that can be processed at low temperatures while maintaining high performance. Although substantial advances have been achieved in n-type oxides such as In₂O₃ and IGZO, the development of BEOL-compatible p-type channels remains limited.

Tellurium (Te) has attracted attention as a promising p-type semiconductor owing to its intrinsically high hole mobility and low melting point. However, direct ALD growth typically yields discontinuous, island-like films due to poor wettability and weak interchain van der Waals bonding. Previous approaches, such as increasing precursor pressure or introducing TeO₂ adhesion layers, have shown limited success because of poor conformality and interfacial oxygen residues. Here, we present a transformation-based strategy in which ALD-deposited TeO₂ is subsequently reduced to elemental Te. This method enables the formation of ultrathin, continuous, and oxygen-free Te layers with excellent conformality. Comprehensive structural and electrical analyses confirm complete phase conversion, smooth morphology, and stable p-type conduction, demonstrating a viable route for BEOL-compatible p-type channel integration in next-generation M3D electronics.