日本大学生産工学部　生産工学部研究報告A51-2

6/38

─ 4 ─dicates that saturation of the number of N atoms in the surface layer is one of the important conditions for the sputtering of the PPP substrate, i. e., the CN layer formed on the surface acts as a source of detaching particles. Generally, nitrogen at-oms can not make a chain structure like as carbon atoms, therefore, N atom injections into the saturated CN layer give rise to destruction of chemical bonding network, which is a cause of the etching the surface. At this injection energy, 25 eV, it seems that incident hydrogen atoms as NH3 molecule do not further the etching in contrast with general expectations for RIE where hydrogen ions/radicals play important role in the etching processes15–17). It can be attributed to the bonding energy of carbon-hydrogen bond is larger than that of nitro-gen-hydrogen bond23). Then, hydrogen atoms included in a NH3 molecule are taken away by the substrate carbon network when it reaches the substrate surface, and separated nitrogen atom is desorbed from the surface when the injection energy is not high enough. Therfore, the PPP surface is hydrogenated stronger than nitridation for the NH3 injections at 25 eV. It can be seen in Fig.4 which shows the densities of particle number of each atom, Nα (α is atomic species) as a function of the height measured from the lowest point of the substrate at 73.4 × 1015cm−2 of injection dose of (a) N, (b) N2, and (c) NH3 in-jections. It is indicated that the CN layer is formed on the PPP surface for all three cases. Thickness of these CN layers is about 2.5 nm. Moreover, in the case of N atom injection Fig.4(a), the CN layer has the region of the ratio of the parti-cle number, NN/NC, is almost unity, which is larger than the other two cases. In this case, from the numerical data of the present simulation, the region of such thickness is about 1.6 nm to 2.8 nm in its height for N injections and 2.0 nm to 2.5 nm for N2 injections. And, “almost unity” means that the value of NN/NC is in the range of approximately 0.9 to 1.2. This re-sults support the experimental facts10, 14) that the etching is strengthened by the formation of the CN layre with NN/NC ~ 1. Consequently, we can see from Figs.2-4 that, (1) injected N atom shows higher eciency for the CN layer formation and the etching of the substrate rather than N2 or NH3 molecule in-jections, (2) in these cases, the CN layer does not act as a pas-sivation layer, (3) the sputtering yield of the substrate C atoms shows rapidly increase when the amount of N atoms in the substrate saturates, (4) the ratio NN/NC reaches unity when the saturation is achieved.In Fig.5, we replot the net erosion yields indicated in Fig.3 as sputtering yields of the substrate C atom for N, N2, and NH3 injections. In present cases, there are no injected C atoms, then, the sputtering yield is identical to the net erosion yield. The results support our opinions mentioned above, because the sputtering yield of C atom for N atom injections is explic-itly larger than the other two cases.Moreover, Fig.6 shows number densities of main three types of bonds (NCN=x, x is 1, 1.5, or 2) among carbon-nitrogen bonds per unit volume contained in the substrate at 48.0 × 1015cm−2 of injection dose versus the height of the substrate Fig.4: The densities of particle number of each atom (Nα, α =C, N, or H) per unit volume vs. the height of the substrate, for (a) N, (b) N2, and (c) NH3 injections.(a)00.91.82.73.64.504080NCNHheight [nm]mn[seitisnedrebmun−3]NNN injections (25ev)(b)00.91.82.73.64.504080NCNHheight [nm]mn[seitisnedrebmun−3]NNN2 injections (25ev)(c)00.91.82.73.64.504080NCNHheight [nm]mn[seitisnedrebmun−3]NNNH3 injections (25ev)Fig.5: Sputtering yields of the substrate C atom for N, N2, and NH3 injections at 25 eV as indicated in the gure. These are extractions of YC’s from Fig.3.09.619.228.838.44857.667.2−0.100.10.2dose[1015 / cm2]sdleiygnirettupsN injections (25eV)N2 injections (25eV)NH3 injections (25eV)