Reactions between a precursor-absorbed SiO 2 surface and energetic ion species, i.e. Ar + or O + ions, in a plasma enhanced atomic layer deposition (PEALD) process were investigated using an in situ X-ray photoelectron spectrometer system and a molecular dynamics (MD) simulation. Both the experiment and simulation results showed that N and/or C atoms originating from the precursor molecules remained on the surface as amorphous carbon and/or cyanide after 50 eV ion bombardment. The precursor-originated atoms as well as the SiO 2 film were removed by the incident ions when its energy increased to 100 eV, and exceedingly small amount remained on the surface. The MD simulation results showed that chemical effects had a more obvious effect on the removal of C atoms at lower incident energies, while purely physical effects dominated at high O + ion incident energies. These results indicated the importance of ion energy in PEALD processes in terms of film quality.
1 April 2025 in Japanese Journal of Applied Physics
This article reports an ionization by Ar+ impact (i.e. β process) found in simulations of capacitively coupled Ar plasmas in a gaseous electronics conference reference cell. The simulation, which was performed using a combination of our in-house GPU-parallelized particle-in-cell, Monte Carlo collision model, and a comprehensive ion-neutral collision model, has revealed for the first time that ionization by Ar+ occurs in the vicinity of the powered electrode surface. Although its impact on the plasma density is so small that the ionization rate can be negligible in comparison to other ion reaction rates for e.g. short- and long-range charge exchanges, its effect becomes significant by raising the voltage applied to the powered electrode due to an increase in energetic Ar+ flux.
16 July 2024 in Journal of Vacuum Science & Technology A
The ion-neutral reactive collision model DNT+, which generates comprehensive ion-neutral collision cross section (CS) data sets for atoms and nonpolar molecules, has been extended to polar molecules. The extension is based on the average dipole orientation (ADO) theory, which adds the dipole moment to Langevin–Hassé CS. Furthermore, the ADO CS for short-range reactive collisions is covered with a rigid core to incorporate long-range elastic and charge-exchange collisions. The modified version of DNT+, i.e., DNT+DM, is applied to gas-phase H2O+–H2O and low-energy CF3+–CO collisions for its validation. The cross sections (CSs) for those collisions using DNT+DM show good agreement with literature data, proving that DNT+DM is valid to some extent. Help with ion swarm analyses and measurements is needed to make the predicted CSs more accurate.
The origin of intrinsic stress in amorphous and structurally disordered carbon films remains a fundamental challenge. Numerous previous studies have focused on deposition parameters and their relationship with stress, resulting in the hypothesis that a high sp 3 -C fraction is a key factor governing compressive stress. However, the precise role of geometrical parameters in stress generation is still unknown. Here, we show that intrinsic stress is mainly dominated by two geometrical parameters: standardized bond length and density, followed by sp 3 -C fraction. We found that the high sp 3 -C fraction independently contributes to tensile stress, contrary to previous experimental results. A data-driven assisted investigation of various amorphous carbon models prepared using first-principles calculations quantitatively revealed the effect of each parameter by isolating their contributions. As a result, a generalized formula is proposed to correlate the intrinsic stress and fundamental structural parameters.
16 March 2026 in Japanese Journal of Applied Physics
We developed a single/dual-frequency (DF) sheath model for arbitrary gas mixture plasmas by modifying the previous model for noble gas plasmas and by adopting ion collision cross-sections generated using DNT+DM in order to simulate fluxes and energy-angular distribution functions (EADFs) of fast neutrals produced in the sheath and transported to the wafer surface. In the 46th International Symposium on Dry Process (DPS2025), an application of the new sheath model to an Ar/H 2 capacitively coupled plasma (CCP) was demonstrated after proving its validity by comparing with data measured in Ar rf and DF CCPs. In this paper, we apply the same model to a DF CCP in a mixture of Ar and CF 4 gases and present the EADFs of various fast neutrals produced due to ion-neutral reactions in the sheath. Through the simulation, it is found that the EADFs of fast fragment neutrals are apparently characterized by two distributions which originate from ion dissociation and gas molecule dissociation. Furthermore, ion dissociation mechanisms in the sheath are investigated and discussed in detail.
16 June 2026 in Journal of Vacuum Science & Technology A
Continued downscaling of three-dimensional semiconductor devices demands dielectric deposition methods capable of void-free filling in highly confined features. Flowable chemical vapor deposition addresses this challenge through a transient liquidlike infiltration stage followed by curing, yet the molecular origins of precursor transport and wetting in the precure regime remain poorly understood. Here, classical molecular dynamics simulations are used to investigate tetraethoxysilane (TEOS) and TEOS-based siloxane liquids as representative model precursors for flowable oxide deposition. Focusing on the precure, transport-dominated state, we quantify how viscosity, surface tension, wetting behavior, and region-resolved substrate interactions govern infiltration in hydroxylated SiO2 trench geometries. Increasing silane substitution systematically raises viscosity and surface tension, reduces flowability, and delays gap filling. Contact-angle analysis and force decomposition further show that substrate interactions promote sidewall-dominated infiltration, whereas weaker driving in the trench interior contributes to transient center voiding. These results establish molecularly grounded precure transport and wetting boundary conditions for subsequent cure modeling. This work establishes a quantitative, precursor-dependent framework that directly links molecular-scale interactions to gap-fill kinetics in flowable systems.