dynamic-SASAMAL Code Research Articles

Computer Simulation of Plasma Immersion Ion Implantation and Deposition

By using a newly developed simulation program "PEGASUS", plasma behavior was analyzed for the plasma immersion ion implantation and deposition (PIII&D). For plasma analysis of low pressure gas which is used in PIII&D, the software uses a particle in cell (PIC) method for the analysis of electric and magnetic fields and the motion of charged particles. A Monte Carlo collision method is used for collisions of ions, electrons and neutrals in the plasma, and the dynamic-SASAMAL code is used for the ion-solid surface interactions. Spatial distributions of potential, electron density and ion density together with the ion flux distribution on the target surface were calculated for the case where a negative pulse voltage was applied to a trench shaped target immersed in a high density Ar plasma (1010 cm-3). The time evolution of sheath length obtained by the simulations for a flat plane part of the surface agreed with the analytical result obtained by the Child-Langmuir method. In a bipolar pulse PIII&D system, a positive and a negative pulse voltages are applied alternately to a workpiece without any other external plasma source. Simulation has been conducted for a target immersed in a very low density Ar plasma (107 cm-3) to compare the plasma generated by a negative and a positive pulse voltage applied to the target. When a negative pulse voltage is applied to the target, only a weak plasma is generated. In contrast to it, when a positive pulse voltage is applied, a two-order or more high density plasma is generated under the same condition. The plasma behavior around a trench shaped target is also presented.

Dynamic-sasamal: simulation software for high-dose ion implantation

A computer code, ‘Simulation of Atom Scattering in Amorphous MAterials based on Liquid model (dynamic-SASAMAL)’ has been developed to simulate the dose dependence of composition profile and sputtering yield of a non-crystalline target under ion implantation, which uses the binary collision model. Among many kinds of binary collision programs, the program TRIM and its derivatives have spread widely. In addition, there are some TRIM derivatives that can treat the dose dependence, i.e. TRIDYN and EVOLVE. However, these dynamic programs are not so obtainable as TRIM. Authors have studied many dose-dependent phenomena under ion implantation by the dynamic-SASAMAL code. Recently, we made the code easy to operate. At first, by using the periodic table, we are requested to input the atomic number of the implanted ion species and the target structure, the number of layers, and the maximum dose and the dose step. After the first calculation, if we need to continue the calculation by changing the energy or the ion species, we can. During the calculation, the depth profiles of atomic composition, vacancies and interstitial atoms, and the angle dependence of number and energy of sputtered atoms are shown on the display. We can set the concentration limit for a high-dose implantation of a reactive ion to a metal, such as 50% for N implantation to Ti. The outline of the software is presented together with the dose-dependent phenomena, which have been studied by using the dynamic-SASAMAL.

Dynamic MC simulation of nitrogen implanted Si/C and Zr/C bilayers

Efforts have been dedicated to the formation of carbon nitride, a super-hard material, by many researchers. However, the nitrogen to carbon ratio obtained by high dose nitrogen implantation into carbon is quite limited by nitrogen diffusion towards the surface during the implantation. In order to study the behavior of implanted nitrogen, dynamic Monte Carlo simulations were carried out using the dynamic-SASAMAL code for nitrogen implantation into carbon, Si/C and Zr/C bilayers with energies of 25, 50 and 150 keV at doses of 5 × 10 16 to 10 18 cm −2. The calculated results were compared with the experimental results obtained by NRA, RBS, and XPS for RT implantation with and without annealing, and for 700°C implantation. For the RT implantation at a low fluence up to 1 × 10 17 ions/cm −2, the measured NRA profile agreed well with the calculated result and almost all implanted nitrogen retained in the target. But diffusion began around 2 × 10 17 cm −2 and the maximum concentration did not reach the calculated value. And the nitrogen concentration saturated at 25 at.%, and retained doses were much lower than the calculated values in all systems. The atomic ion mixing and the sputtered losses obtained by the calculation and the experiment were agreed very well.

Behavior of implanted nitrogen in Zr/carbon bilayer

The nitrogen to carbon ratio obtained by high dose nitrogen implantation into carbon is rather low, since nitrogen diffuses towards the surface during implantation. In order to study the behavior of implanted nitrogen, 15N ions were implanted into Zr/carbon bilayer with energies of 50 and 150 keV to doses up to 10 18 cm −2 at RT and at 700 oC. After implantation, some specimens were annealed at 700 oC for 1 h. The depth profile of the implanted nitrogen was measured by NRA, and the other components were measured by RBS. At a low fluence up to 1 × 10 17 cm −2, the measured nitrogen depth profile agrees with the theoretical profile obtained by dynamic Monte-Carlo simulations using the dynamic-SASAMAL code, but at a higher fluence, the measured profile spreads to both sides. As a result, much higher fluence than expected by the simulation is necessary to reach saturation. At 50 keV with high fluence, severe atomic mixing occurs and the implanted nitrogen diffuses more remarkably towards the upper Zr layer. Since the depth profiles do not change during the post-implantation annealing, the diffusion is radiation stimulated. The diffusion was more remarkable for the hot implantation at 700 oC. The surface topography observed by AFM showed no blisters even for 50 keV, 1 × 10 18 cm −2 implantation at RT. This is considered to be a result of carbon mixing into the Zr layer.

Nitride layers formed by nitrogen implantation with an energy scanning mode

We reported that a very smooth surface was obtained by room temperature nitrogen implantation into Zr with the energy increasing mode from 35 to 65 keV even with a high dose of 7×10 17 ions cm −2 in contrast to the implantation with the energy decreasing mode and 50 keV mono-energy implantation by which many blisters were observed on the surface. In order to understand the mechanism that causes radiation damage, a dynamic Monte Carlo simulation with the dynamic-SASAMAL code was performed for such an energy scanning mode. The movement of the nitrogen at the stage where the excess nitrogen over stoichiometric concentration just starts to migrate towards the surface was carefully analyzed and the results for the energy increasing mode from 15 to 85 keV with a uniform distribution were compared with that of the energy decreasing mode and 50 keV mono-energy implantation. It was found that, when nitrogen concentration exceeds the stoichiometry, the nitrogen that is enforced to migrate towards the surface is not the newly implanted nitrogen but is the previously implanted one which is moved by the newly implanted nitrogen through binary collisions. The moment given through the collision is towards the surface in the energy increasing mode and it will enhance the migration. In contrast to this, the moment is towards the deeper side in the energy decreasing mode and the mono-energetic implantation, so it will disturb the migration and cause nitrogen gas bubble formation which induces the surface radiation damage with many blisters. Finally, it is proposed that a thick nitride layer without surface damage can be obtained by the energy increasing mode implantation with an adequate energy distribution.

Saturated thickness of nitride layers formed by high fluence nitrogen implantation into metals

The thickness of the nitride layer formed by a high fluence nitrogen implantation to a metal surface is limited mainly because of the sputtering effects and the migration of the implanted nitrogen towards the surface. We have investigated nitrogen depth profiles implanted to several kinds of metals theoretically by Monte Carlo simulation using the dynamic-SASAMAL code and experimentally by RBS and NRA using the 15N(p,αγ) 12C reaction. In this paper, the saturation fluence and the saturated nitride thickness were calculated for 1–200 keV nitrogen implantations into several kinds of metals (Al, Ti, Fe, Zr, and Hf). In order to find the method to overcome the limited thickness and the severe radiation damage problem at the high fluence implantations necessary for the saturation, the nitrogen depth profiles for an energy scanning implantation with a uniform and a non-uniform energy distribution was calculated and compared with those of mono-energetic implantations.

High-fluence nitrogen implantation into metals

Nitride ceramics are formed on metal surfaces by high-fluence nitrogen implantation and the mechanical and chemical properties of the surfaces are modified. However, the thickness of the nitride layer is limited by the sputtering effect and the migration of implanted nitrogen towards the surface. In this work, we studied these effects for nitrogen implantation into several kinds of metals (Al, Si, Ti, V, Fe, Co, Ni, Zr, Nb, Mo, Ta and W) by Monte Carlo simulations using the dynamic sasamal code in comparison with the experimental results obtained by NRA. The dynamic sasamal code takes into account target alterations which occur under high-fluence implantations, so that dynamic changes in composition, deposited energy, retained probability of the implanted ion, sputtering effects, etc. with fluence can be calculated. For nitrogen implantations into metals, the saturated nitrogen concentration is assumed to equal that of the saturated nitride phase and nitrogen atoms above stoichiometry are assumed to diffuse towards the surface. The depth profiles of nitrogen concentration and the retention of nitrogen obtained by the code agree very well with the experimental results for 50 keV nitrogen implantations into metals with fluences of 10 17–10 18 ions cm −2. The calculated sputtering yields were compared with the semi-empirical values.

Retention of nitrogen implanted into metals

Retention of nitrogen implanted into various kinds of metals (Al, V, Ti, Fe, Co, Ni, Zr, Nb, Mo, Ta, and W) was calculated by the dynamic-SASAMAL code assuming planer surface binding energy and diffusion toward the surface for nitrogen over the concentration of saturated nitride phase. As the value of surface binding energy ( E s), the sum of the heat of sublimation of the metal ( H s) and the heat of formation of the nitride ( H f) times the nitrogen concentration ( C) at the surface layer was used ( E s = H s + CH f). The effects of values of H s, H f, and the displacement threshold energy on the retention were studied. The agreements between the calculated results and the experimental results obtained by NRA using 15N(p, αγ) 12C reaction were satisfactory for depth profiles and retention of nitrogen implanted into metals of Al, Ti, Ni, and Zr at a wide range of fluence from 6 × 10 16 to 6 × 10 17 ions cm −2.

Computer simulation of dynamic change of surface composition induced by high-fluence nitrogen implantations

Abstract Dynamic SASAMAL code takes into account target alterations which occur under high fluence implantations, so that dynamic changes of the atomic composition, the retention probability of implanted ions, etc., with the fluence can be calculated. The model calculations have been applied for implantations of nitrogen ions of energy 50 keV into zirconium with fluences of 10 17 –10 18 ion cm -2 , assuming (1) no special diffusion and (2) upward diffusion for the extra atoms over stoichiometry. Also, the saturation concentration of nitrogen is assumed to be equal to the saturated nitride phase. Depth profiles of the atomic composition and the retention rate of implanted ions obtained by the calculations were compared with the results of nuclear reaction analysis, and the results obtained with the assumption of “upward diffusion” agreed very well with the experimental results. The depth profiles of the energy deposited by the implanted ions on the target atoms through electronic and nuclear interactions also were calculated. These are discussed in comparison with the results of glancing-angle X-ray diffraction.

Dynamic behavior of implanted nitrogen in zirconium at various stages of implantations

Abstract Depth profiles of nitrogen implanted into Zr with an energy of 50 keV were calculated by dynamic SASAMAL code with three different assumptions for the diffusion of excess atoms over stoichiometry, i.e., ‘no diffusion', ‘both-sides-diffusion’ and ‘upward-diffusion'. To distinguish nitrogens implanted certain stage of implantations, alternate implantations of 15N and 14N were used. The results were compared with the experimental results by the resonance nuclear reaction analysis, NRA. For 15N implantation with fluences from 1 × 1017 to 1 × 1018 ions/cm2, the calculated results with ‘upward-diffusion’ agreed very well with the NRA results for all fluences. For the depth profile of pre-implanted 15N (1 × 1017 ions/cm2), which was changed by the subsequent 14N implantation with fluences of 1 ∼ 10 × 1017 ions/cm2, the agreement with the NRA results was satisfactory until the 14N fluence did not exceed 5 × 1017 ions/cm2, but for higher fluences, the retained probabilities of 15N obtained by the ‘upward-diffusion’ code were too low compared with the experimental value obtained by NRA. For the depth profiles of 15N (1 × 1017 ions/cm2) implanted following after implantations of 14N with fluences of 1 ∼ 10 × 1017 ions/cm2, the agreement with the NRA results was quite good for all 14N fluences. It is concluded that the approximation of ‘upward-diffution’ is proper satisfactorily for the treatment of atoms implanted at the final stage of implantations, but a problem is left for the treatment of atoms implanted at the early stage of implantations.

Composition and structure of zirconium implanted with nitrogen at high fluence

High-fluence implantations of 100 keV N 2 + ion into a zirconium layer deposited by an electron beam were carried out at room temperature and at liquid-nitrogen temperature with and without nitrogen gas of a pressure of 1 × 10 −4 Torr. Depth profiles of atom concentrations were determined by Rutherford backscattering spectrometry, Auger electron spectroscopy, and resonance nuclear reaction analysis. The experimental results were compared with the theoretical results obtained by Monte Carlo simulations of the dynamic-SASAMAL code. The depth profiles measured by RBS agreed well with the calculated profiles for all fluences from 1 × 10 17 to 5 × 10 17 ions/cm 2, but differed from those obtained by AES indicating the existence of easily released nitrogen in the Zr specimen implanted with nitrogen at high fluence. Glancing angle X-ray diffraction was used to confirm zirconium nitride formation.