日本大学生産工学部　生産工学部研究報告A51-2

4/38

─ 2 ─etching of organic polymer surface using MD simulations. Previously, we have performed MD simulations of organic polymer etching by hydrocarbon beams15–17) using empirical inter-atomic potential functions proposed by Brenner18) with van-der-Waals interacction19). In these previous works, the polyparaphenylene (PPP) has been adopted as an organic polymer lm and it has been shown that injected C atoms tend to accumulate on the pristine surface and strong etching occur when hydrogen rich clusters (e. g., CH4 molecules) are inject-ed. And, we have observed that sputtered clusters show vari-ous compositions of carbon and hydrogen atoms. Therefore, a specic cluster including carbon atoms has relatively low sputtering yield (≤ 0.05) in spite of the sputtering yield of car-bon atom has visible value (~ 0.5 on average). The yield show sharp changes depending on injection dose and various dier-ent behaviours as dierent injection paths17).Moreover, we have already performed preliminary MD simulations of the PPP substrate etching process due to hydro-gen-nitrogen beam injections and attempted to give some etching mechanism of organic polymer substrate20) using the inter-atomic potential functions for (H, C, N) systems which have been developed by Yamada and Hamaguchi21). In this precedent simulations, we have mainly examined changes of chemical bonds among C, H, and N atoms due to N, N2, and NH3 injections to clarify the etching characteristics depending on the incident species.In the present paper, following and extending the preceding work20), we have analyzed etching characteristics in more de-tail such as etching products, sputtering yields, formation of CN layer by ion/radical injections, and described two opposite properties of the CN layer; one is generator of the sputtered species, for example, CN, HCN, and C2N2, the other is a pas-sivation for hydrogen radical injections. To focus on proper-ties of organic polymer lms arisen from phenyl ring22), we again employ PPP as a model substrate for organic polymer etching simulations using inter-atomic potential functions of (H, C, N) systems21). This potential function includes van-der-Waals potential which is similar to the function developed by Sinnott19).2　Procedures of the MD simulationsIn the following, to investigate the etching characteristics of PPP substrate, N, N2, and NH3 atoms/molecules are injected into PPP surface. The energy range of injected particles is mainly 25-50 eV and the beam direction is normal to the sub-strate surface.The simulation procedures used in this paper are essentially the same as those used in the previous works15–17, 20). In the fol-lowing, a summary of the numerical methods is given briey. The velocity Verlet algorithm is employed to integrate the equation of motion for each atom. A typical time step em-ployed here is 0.125 femtoseconds (fs), and the period of one injection event is 1.2 picoseconds (ps). As mentioned above, as a simplest model of organic polymer, a poly(1,4-phenylene)(i. e., PPP) is used as an organic polymer substrate. Figure 1 shows the initial PPP substrate thermally equilibrated at 300 K. The horizontal cross section (in the x-y plane) of the simu-lation box is 2.2 × 1.9 nm2. Periodic boundary conditions are applied in the x and y directions. The typical initial substrate is composed of four monolayers and one monolayer contains four chains of ve phenyl rings. Therefore, the initial substrate has 480 carbon (C) atoms and 320 hydrogen (H) atoms in to-tal. To avoid drift fo the whole substrate by energetic particle injection, the atoms in the lowest monolayer are rigidly xed. In our MD simulations, if any atom (injected or recoiled) passes through the lowest monolayer of the substrate, we con-sider that the substrate thickness is not enough, discard this event, add another polymer layer (which contains 80 H atoms and 120 C atoms) to the substrate from the bottom, equilibrate the new substrate at the room temperature, and restert the in-jection anew. And, the simulation cycle of a single injection was repeated 3000 times, which corresponds to the dose of about 73.4×1015cm−2. To simplify the arguments, we consid-ered that ions and charge neutral species for the same atomic species show essentially same reactivities. Therefore, all at-oms are assumed to be charge neutral and highly energetic impinging species are simply interpreted as beam ions. A typi-cal time scale in which a MD simulation can practically fol-low the dynamics of a many particle atomic system is a few ps after an injection of each beam species. Therefore, although thermal desorption process of surface species13) and substrate temperature dependencies of etching byproduct species6, 9) are also important for organic polymer etching, it cannot be taken into account in the present work since such thermal process is too slow to be simulated directly by MD simulation within the Fig.1: Horizontal view in (a) z-x plane and in (b) z-y plane of the initial PPP model substrate. White and green spheres represent carbon and hydrogen atoms, respectively.(a)(b)