ICMCTF 2026 Session PP-ThP: Plasma and Vapor Deposition Processes Poster Session

As silicon-based devices continue to scale down, leakage current has emerged as a critical issue. To overcome this challenge, indium–gallium–zinc oxide (IGZO)–based oxide semiconductors, with their wide bandgap and superior leakage suppression, have been highlighted as promising next-generation channel materials. However, IGZO films face important limitations. Their amorphous structure, relatively low carrier concentration, and high Schottky barrier height (SBH) result in poor contact resistance, which has become a major obstacle for memory applications.

In this study, we propose a novel contact engineering approach that has not been reported previously. Specifically, Carrier boosting has previously been applied to the channel region to modulate the threshold voltage (Vth) of devices. We extend the technique to the source and drain regions. A boosting layer is selectively deposited using PVD method on the contact areas to enhance doping and improve interface properties. By locally increasing the doping concentration through this approach, the contact resistance can be effectively reduced. Finally, a high-performance metal electrode is deposited to ensure excellent electrical contact.

This integrated strategy is designed to reduce contact resistance while simultaneously improving the overall device performance.

Two-dimensional transition metal dichalcogenides (TMDs), particularly molybdenum disulfide (MoS₂), exhibit remarkable mechanical and electronic characteristics that make them promising for flexible and strain-engineered devices. However, establishing a direct connection between their growth mechanisms and mechanical performance remains a key challenge. In this work, MoS₂ thin films were synthesized by Chemical Vapor Deposition (CVD) on Si/SiO₂ substrates, revealing triangular domains dominated by a spiral growth pattern. This morphology originates from screw dislocations that promote a continuous step-flow mechanism during layer formation. The nucleation and propagation of these dislocations are strongly affected by the substrate’s surface topography, leading to distinct variations in domain size and layer stacking.

Raman spectroscopy confirmed the crystalline nature of the films, showing well-defined vibrational modes associated with MoS₂. The in-plane (E¹_2g) and out-of-plane (A_1g) modes appeared around 383 cm^-1 and 408 cm^-1, respectively, while additional features near 176 cm^-1, 228 cm^-1, and higher-order peaks above 418 cm^-1 indicated complex interlayer interactions and multilayer coupling. These spectral variations were used to assess the structural uniformity and thickness of the films.

Atomic Force Microscopy (AFM) analyses revealed the stepwise topology typical of spiral growth and enabled quantitative evaluation of the vertical layer spacing. Nanoindentation experiments demonstrated that the presence of screw dislocations enhances the mechanical resilience of the material, resulting in a combined elastic–elastoplastic response. The observed hardening behavior suggests that dislocation-mediated growth can be harnessed to modulate the toughness and strain accommodation of 2D materials.

Overall, this study provides fundamental insight into how growth dynamics influence the microstructural and mechanical characteristics of CVD-grown MoS₂. These results contribute to the development of robust and flexible TMD-based coatings for applications in nanoelectromechanical systems (NEMS), strain sensors, and energy-dissipative devices.

Computational simulations are rapidly transforming the way magnetron sputtering processes are designed, understood, and optimized. They offer a powerful means of increasing experimental efficiency, accelerating process development, and improving reproducibility—particularly when transitioning from laboratory-scale research to industrial-scale coating production. Despite major advances in plasma-assisted deposition, one fundamental challenge remains: precise control of particle sputtering from the target and their subsequent transport toward the substrate. These parameters govern the particle abundance and energy while arriving at the substrate, ultimately determining coating stoichiometry, phase structure, microstructure, and performance.

While compositional gradients and local variations can be highly beneficial for combinatorial thin-film research, they are detrimental in industrial environments, where uniformity in thickness, composition, and phase structure are essential for high-throughput and large-area coating.

In this contribution, we demonstrate the combined use of SDTrimSP and SiMTra simulation tools to predict industrial magnetron sputtering of metallic glasses based on Zr–Cu(–Ni/Al) systems and on W–Ni–B and W–Zr–B systems, representing examples with comparable and strongly contrasting atomic masses. The simulations provide detailed predictions of relative thickness profiles, elemental composition distributions, and the energy spectra of the arriving species under varying process conditions.

A longstanding challenge in reactive magnetron sputtering is the quantitative prediction of the deposition rate, which is primarily determined by the partial metal sputtering yield from the oxide layer formed on the target surface during poisoning. The first step in addressing this issue is to determine the total sputtering yield of the oxide. This has been accomplished by refining a published semi-empirical model. This model has been applied to fit an extensive set of oxide sputtering yield data from the literature, comprising 65 datasets for 21 different materials. The fitting process establishes a relationship between the surface binding energies of metal and oxygen atoms and the cohesive energy of the oxide. The calculated partial sputtering yield of metal from a poisoned target is then compared with previously published experimental data on the metal sputtering yield during reactive magnetron sputtering. While both yields are linearly correlated, the magnetron-based sputtering yields are approximately eight times lower than the model predictions. This reduction in yield is attributed to the formation of an oxygen-rich surface layer, a hypothesis supported by binary collision approximation Monte Carlo simulations. However, these simulations do not fully capture the mechanism, as a more detailed description of the surface oxygen origin is needed. Despite this limitation, the experimental correlation provides a practical strategy for predicting deposition rates during reactive magnetron sputtering in fully poisoned mode. As demonstrated, the oxide sputtering yield can be calculated using standard data sources, and the empirical correlation between the sputtering yields enables a reliable estimate of the metal partial sputtering yield in poisoned mode, thus allowing for an accurate estimation of the deposition rate.

D. Depla, Note on the low deposition rate during reactive magnetron sputtering, Vacuum 228 (2024) 113546D. Depla, J. Van Bever, Calculation of oxide sputter yields Vacuum 222 (2024) 112994

This PhD research focuses on developing advanced thin-film coatings for substrates with complex geometries, aiming to achieve uniform properties and enhanced resistance under demanding operational environments. The study emphasizes optimizing deposition parameters to ensure consistent film characteristics—critical for the performance and durability of optical components in optronic systems. These systems incorporate diverse optical elements, including windows, lenses, filters, and dichroic plates, all requiring precise functionalization through thin-film treatments to meet stringent optical, mechanical, and chemical specifications.

Front optics in optronic devices play a key role in detection, observation, and identification. Operating under harsh and variable conditions—such as corrosive, erosive, and chemically aggressive environments—these components require coatings that maintain high optical transmission while exhibiting robust mechanical and chemical stability.

Typically, coatings are deposited by vapor-phase techniques such as Electron Beam Physical Vapor Deposition (EB-PVD) and Ion Beam Assisted Deposition (IBAD), enabling the formation of dense, uniform multilayer interference stacks. However, substrates with complex geometries—characterized by large diameters or high curvatures—pose significant challenges for achieving uniform coating deposition. Variations in thickness, density, and mechanical properties across the surface can lead to performance degradation, including chemical attack, delamination, and loss of optical quality, particularly under critical conditions such as saline fog exposure.

This project aims to elucidate the underlying growth mechanisms and physical phenomena at the material scale, focusing initially on single-layer coatings to establish a solid foundation of knowledge. Comprehensive characterizations—including nanoindentation, ellipsometry, and strain measurements—are employed to assess mechanical and optical properties and to study the influence of deposition parameters. The insights gained will guide the design of novel multilayer architectures incorporating new materials and interfaces to enhance thermomechanical performance.

Ultimately, this research supports the evolving specifications of optronic devices by delivering coatings with improved robustness and consistent functional properties, thereby advancing the performance and reliability of front optics in demanding operational environments.

Ion acceleration is one of the main process tools in the field of ionized physical vapor deposition (IPVD) for thin-film microstructure manipulation. However, the acceleration of film-forming ions onto insulating substrates has been limited, if not impossible, using conventional approaches. Recently, the demonstration of synchronized floating potential HiPIMS (SFP-HiPIMS) has opened new avenues for controlled metal-ion acceleration on insulating substrates [1].

In this presentation, we report on systematic studies of two industrially relevant materials – AlN and AlScN thin films – grown on a range of insulating substrates using SFP-HiPIMS. The substrates include single-crystalline silicon (001), Z-cut quartz, c-cut sapphire, and amorphous borosilicate glass. The concept of SFP-HiPIMS is based on synchronizing the arrival of film-forming ions at the substrate surface with the build-up of the negative floating potential. Since the sputtered species in HiPIMS are time-separated and the build-up of the negative floating potential is transient, achieving this requires precise synchronization between the HiPIMS pulse-on time and the floating potential-on time (defined as the time offset). Such synchronization allows the attraction of film-forming ions to the substrate while avoiding Ar⁺ process gas ion bombardment and incorporation into the growing film. Although the SFP-HiPIMS can be implemented using at least two HiPIMS pulses, we demonstrate its feasibility not only for a multiple-magnetron configuration but also for a single-magnetron setup. We evaluate the microstructural quality of AlN and AlScN thin films grown by conventional HiPIMS and SFP-HiPIMS under different magnetron configurations and time offsets in terms of rocking-curve full-width at a half maximum (FWHM), Ar content, and surface roughness.

References:

[1] Nature Communications (2025) 16:4719

Ti-based carbonitride thin films such as TiN, TiC, and TiCN have been used to enhance wear resistance and lifetime of cutting tool. Recent studies employed additional elements such as Si or B to form multi component thin films like TiSiCN and TiBCN to improve oxidation resistance and thermal stability. These films are considered to consist of Ti(C,N) nanocrystals dispersed in amorphous TSiCN or TiBCN, which effectively suppresses grain coarsening and also enhances oxidation resistance.

However, most of the TiSiCN and TiBCN thin films have been synthesized to date by physical vapor deposition methods such as magnetron sputtering or arc evaporation, which often result in poor step coverage and interfacial adhesion strength on complex-shaped substrates. Most of previous studies employed physical vapor deposition methods such as magnetron sputtering or arc evaporation, while plasma enhanced chemical vapor deposition (CVD) provides potential advantage on higher conformality, stronger interfacial adhesion and low temperature fabrication for complex-shaped tools and components.

In this study, TiSiCN and TiBCN thin films were deposited on Si and cemented carbide (WC–Co) substrates using direct current plasma CVD at around 600 ℃ where WC–Co substrates were pretreated with aqua regia to improve interfacial adhesion.

The precursor gases were TiCl₄, CH₄, N₂, tetramethyl silane (TMS), and BBr₃. The effects of deposition parameters on the film structure and physical properties were systematically investigated using X-ray diffraction, X-ray photoelectron spectroscopy, and nanoindentation.

Si content in the TiSiCN thin films increased with increasing TMS flow rate, while the B content in the TiBCN thin films also increased with increasing BBr₃ flow rate. TiSiCN thin films exhibited higher hardness as maximum value of HV 2585 than that of TiCN thin film. However, the hardness of TiSiCN film decreased according to increase of Si content in the film. The effects of addition of Si and B on grain refinement and structural densification will be discussed.

Plasma nitriding is a valuable and well-established technique for surface hardening of metals to improve their mechanical and tribological properties, such as hardness and wear resistance. Typically, plasma nitriding involves the use of a glow discharge of a mixture of nitrogen and hydrogen, where the metal component to be treated is the cathode and the chamber wall is the anode. The low-pressure plasma (15–1500 Pa) produced by the application of a DC potential (0.3–1.0 kV) contains nitrogen ions, which are accelerated towards the cathode and implanted in the surface of the metal. The treatment time, surface concentration of nitrogen, and temperature of the metal component determine the depth and gradient of the nitride layer, but various tens of microns are often formed. The same process cannot directly be used to nitride insulators since such materials cannot be used as an electrode in a DC plasma.

We have developed a variant of the normal plasma nitriding scheme where the discharge is produced by applying an RF potential and the piece to be nitrided is mounted on a magnetron cathode. This is, of course, the same as a RF magnetron sputtering system. It is known that the DC voltage bias generated on the surface of such an insulting target with a RF magnetron sputtering cathode depends on the following factors: the relative areas of the anode and cathode, the applied RF power, the gas pressure and composition, and the degree of matching of the impedance of the electrical supply to the impedance of the plasma. We have measured the DC bias potential and the rate of sputtering of a quartz target mounted on a 2” MAK magnetron cathode in a pure nitrogen gas discharge as a function of the area of the anode, the applied power, and the degree of matching indicated by the ratio RF(Reflected Power) / RF(Forward Power). Using conditions which produced a minimum sputter etching of the target, we produced three nitrided samples. We present the dimensions and composition of the nitride layers measured using XPS and RBS, the hardness and wear resistance and the optical properties measured using multi-wavelength ellipsometry of these layers.

Magnetron sputtering is a widely used Physical Vapor Deposition technique for thin film growth. A target (or cathode) is bombarded by ions from the plasma resulting in the sputtering of the former and emission of secondary electrons. Electrons are confined near the target by a magnetic field created by permanent magnets placed under the cathode. This confinement enhances the ionization efficiency of the working gas allowing to operate at relatively low pressure (0.4 – 4 Pa).

The understanding of the plasma behavior is key to fully control the sputtered material and, therefore, the thin film growth. It involves complex phenomena such as ExB gradient, curvature and drift coupled with kinetic reactions and plasma-surface interactions. Hence, the motion of individual particles and the whole plasma is subject to intricate phenomena difficult to apprehend.

Recently the TMP-D&S (Theory and modelling of Plasma – Discharge & Surface) team of LPGP (Laboratoire de Physique des Gaz et des Plasmas) has developed a 2D Particle-In-Cell – Monte Carlo Collision exploiting the power of AMR (Adaptive Mesh Refinement). Because magnetron plasmas are highly inhomogeneous, using the smallest Debye length as the upper limit of the mesh size for the entire domain results in a waste of computational resources although it is a mandatory criterion to avoid numerical issues. Instead, AMR refines the mesh only where it is required.

The principle of the ARM approach will be presented with the advantages and its implementation. Further, the first results of the SAPIC (Saclay simple AMR Particle-in-Cell) code applied to a conventional magnetron show the effectiveness of the method and pave the way towards further numerical optimizations.

Gas aggregation cluster sources (GAS) have been emerging as a key technology for the production of clusters and nanoparticles (NPs) of precise size and composition. The resulting NP properties are significantly affected by the thermal balance during their growth in the aggregation zone. In this study, the characteristics of a novel controllable GAS setup are investigated, using a post (cylindrical) magnetron with a rotating magnetic circuit [1] equipped with a copper target in Ar and/or N₂ atmosphere, respectively. Energy fluxes are quantified by calorimetric measurements using passive thermal probes (PTP) [2], while the plasma parameters are assessed by Langmuir probes. These quantities are critical to develop a comprehensive understanding of the correlation between (external) process parameters (e.g., current, voltage, continuous or pulsing, gas pressure) and (internal) plasma parameters and their correlation with NP growth, transport and microstructure.

[1] D. Nikitin et al., Plasma Processes Polym. 18 (2021) 2100068.

[2] H. Kersten et al. Thin Solid Films 377–378 (2000) 585–591.

In recent semiconductor device fabrication processes, the epitaxial growth of Si and SiGe has become a critical technology. Beyond conventional selective epitaxial growth for source/drain formation, SiGe/Si superlattices are now employed as backbone structures in nanosheet-based transistors, extending the application of epitaxy to channel formation. Furthermore, the rapid adoption of 3D integration technologies—including vertically stacked device architectures and hybrid bonding—has significantly increased the demand for epitaxial processes that can be performed within a constrained thermal budget. These requirements have driven interest in high-order silanes (Si_nH_2n+2, n ≥ 2) as potential alternatives to conventional Si precursors, including SiHCl₃, SiH₂Cl₂, and SiH₄. High-order silanes contain Si–Si bonds with lower bond dissociation energies, enabling decomposition at lower temperatures and potentially achieving higher growth rates compared with conventional precursors.

In this study, we investigate the epitaxial growth of SiGe films using Si₂H₆, Si₃H₈, and Si₄H₁₀ under ultra-high vacuum chemical vapor deposition conditions. The effects of varying the flow rates of each Si precursor on growth characteristics were systematically examined. The Ge concentration and growth rate were analyzed under different growth conditions, and the crystallinity and surface morphology of the resulting films were evaluated. Our results demonstrate the feasibility of employing high-order silanes for low-temperature SiGe epitaxy, providing insight into their potential application in next-generation semiconductor device fabrication.

As semiconductor devices evolve toward vertically stacked architectures, selective epitaxial growth (SEG) of Si and SiGe for source/drain regions has become increasingly challenging due to strict thermal budget requirements. To overcome these limitations, new precursors capable of enabling efficient low-temperature reactions are needed. Although Cl₂ has recently been proposed as a potential alternative to traditional SEG etchant, its insufficient etch rate at extremely low temperatures (≤500℃) underscores the need for novel etchant gases. In this study, we report the first systematic investigation of GeBr₄ as a precursor for low-temperature SEG applications. The reaction characteristics of GeBr₄ were investigated on Si_1-xGe_x (:B) films and Si (:P) films using a UHV-CVD system.

Crystalline and amorphous undoped Si_1-xGe_x films (0≤x≤0.3) were prepared to evaluate etch selectivity, defined as the etch-rate ratio between amorphous and crystalline films. Experiments were conducted using H₂ and N₂ as the carrier gases within the temperature range of 375 to 500℃. The behavior of B-doped SiGe and P-doped Si films was compared with those of undoped films. Scanning electron microscopy and spectroscopic ellipsometry were employed to measure film thickness. Atomic force microscopy and high resolution-transmission electron microscopy were utilized to examine the surface morphology and microstructure of etched films, respectively. We performed surface analysis after the reaction with GeBr₄ using time-of-flight secondary ion mass spectrometry (ToF-SIMS). Additional density functional theory simulations were performed to elucidate the origin of the observed behavior of GeBr₄.

Molybdenum (Mo) has emerged as a promising candidate for next-generation contact and interconnect applications due to its favorable electrical properties and scalability. Similarly, molybdenum nitride (MoN), which can function as a seed layer, diffusion barrier, or electrode material, has attracted significant attention for advanced microelectronic applications owing to its high electrical conductivity and chemical stability. Although MoN deposition by atomic layer deposition (ALD) has been widely explored using halide and organometallic precursors, MoO₂Cl₂ has not been reported for MoN thin-film growth, and its ALD growth characteristics for MoN remain unexplored.

In this work, MoN thin films were deposited by thermal ALD at 600 ℃ using MoO₂Cl₂ as the metal precursor and NH₃ as the co-reactant. Film thickness was systematically varied by adjusting the number of ALD cycles to investigate the evolution of structural, chemical, and electrical properties as a function of thickness. The deposited films were characterized by X-ray diffraction, X-ray photoelectron spectroscopy, and transmission electron microscopy to analyze phase formation, bonding states, and microstructural development, along with electrical resistivity and surface roughness measurements. The thickness dependence of the electrical properties and microstructural evolution of the MoN films were systematically analyzed. This study establishes MoO₂Cl₂ as a viable precursor for thermal ALD of MoN, providing quantitative insights into thickness-dependent electrical properties governed by microstructural evolution and a fundamental baseline for future process optimization, including plasma-assisted or hybrid ALD approaches, to improve early-stage electrical performance.

Acknowledgment

This work was supported by the Technology Innovation Programs (RS-2023-00235609 & RS-2025-02306457) and the K-CHIPS program (25065-15FC) funded by the Ministry of Trade, Industry & Energy (MOTIE, Korea).

Metal nitrides have been studied extensively as they have properties useful in applications such as barrier coatings against Cu interdiffusion in electronics as well as hydrogen catalysis [1]. Tungsten nitride (WN) is an attractive material for these applications when deposited as a thin layer. Likewise, boron nitride (BN) thin films have found use in electronics as an insulative coating with high thermal stability and oxidative resistance. Thin films of WN and BN have been generated in multiple ways including sputtering, physical vapor deposition, chemical vapor deposition, and atomic layer deposition [1,2].

We are investigating WN and BN as novel thin film materials for potential use in X-ray mirrors. Thin layers of WN and BN could be tuned for Bragg reflections in the X-ray spectrum, which would prove useful in astronomical devices. Such a Bragg reflector would require hundreds of conformal layer pairs, with each layer under a nanometer thick and very smooth, with rms roughness ~1 Å and no interlayer diffusion [3].

Plasma-enhanced atomic layer deposition (PEALD) is attractive for this application as the use of plasma allows for lower temperature depositions, which are typically associated with amorphous structuring and lower generated roughness. An experimental PEALD reactor was used for deposition onto 1” Si wafers. Precursors bis(tert-butylimino)bis(dimethylamino)tungsten (BTBMW) and a nitrogen/hydrogen mixture for were studied for WN deposition, while tris(ethylmethylamino) borane (TEMAB) and the nitrogen/hydrogen mixture were studied for BN growth.

The PEALD reactor demonstrated WN growth rates of ~0.55 Å/PEALD cycle and BN growth rates of ~0.4 Å/PEALD cycle at relatively low temperatures of 200 °C. Very low rms roughness of 1 Å was observed for ultrathin (< 5 nm) deposited films. Varying PEALD process parameters such as the deposition temperature, plasma exposure times, and plasma power were found to have interesting effects on the film growth rate. Composition, roughness, and other optical properties for deposited films will also be presented.

[1] M. J. Sowa, Y. Yemane. et al. JVST A 34, 051516 (2016)

[2] Park, H., Kim, T., Cho, S. et al.Sci Rep 7, 40091 (2017)

[3] B. Salmaso, D. Spiga. et al. Proceedings Vol. 8147, Optics for EUV, X-Ray, and Gamma-Ray Astronomy V, 814710 (2011)