Experimental setup showing how the powder sublimates into the expansion chamber (1) with a vapor pressure up to by opening valves V1 and V2. Gaseous is then introduced into the reaction chamber (2) and reacts with the samples S1 and/or S2, located in mini chambers (3) separated from the reactor by mini gate valves GV1 and GV2, respectively.
Details of sample holder (S1 and S2), including the loading cap and the holder. The holder fits inside the body valve port and penetrates deep inside to minimize the volume (in yellow) between the sample and the gate of the valve when closed. The sample is compressed close to the gate by two O-rings. Only the area inside the O-ring of the surface facing the gate valve is exposed to when the gate valve is open. Sealing is achieved using Kalrez O-rings.
Time evolution of the consumption showing the effect of contamination during a virtual sample loading (i.e., no sample introduced but reactor opened). Curve (a): exposure at 2.25 Torr of an empty reactor after exposure to air. Curve (b): exposure at 2.99 Torr of a clean (i.e., after several exposures without opening) empty reactor before exposure to research grade Argon. Curve (c) : exposure at 2.38 Torr of an empty reactor after exposure to research grade Argon.
XPS data from Mo samples before cleaning (bottom: green curve), and after acetone/methanol cleaning (top, red curve). Acetone/methanol treatment removes part of the Mo oxide. The remaining oxide ( and 235 eV) and contaminants are fully removed during the first exposure to .
IR absorbance spectra of the most intense vibrational modes of , , and (doublet: ).
Real-time FT-IR absorbance spectra obtained during silicon exposure to 1.2 Torr . The number of molecules introduced into the reactor is . At , the mini-gate valve protecting the sample is opened, inducing silicon etching as indicated by the simultaneous consumption and production.
Evolution of the produced amounts of and (the red scales on the right) and initial etching rates (the green scales on the left) for Mo sample (top) and Si sample (bottom) as a function of exposure (at 0.34 Torr). The ten exposures are performed on freshly loaded samples under research grade Argon purge.
Time evolution of the ratios ( or Si) during the tenth cycle of a series of etching at 1.2 Torr of a (a) Si and (b) Mo samples. Straight black lines indicate the expected values from Eqs. (1) and (2) for Mo (33%) and Si (50%), respectively.
Depth profiles obtained on Si (plain line) and Mo (dashed line) after etching. Insets: left: Mo sample; right: Si sample, at the same scale. The edge of the circular o-ring is located at ±6 mm.
Calculated fluorine density obtained at the end of each cycle of a series of ten consecutive cycles of etching at 1.2 Torr. The fluorine density decreases exponentially with distance.
Evolution of the fluorine density during one etching cycle at 1.2 Torr. The thickness of the fluorosilyl layer first decreases, due to a fast etching process and a slow diffusion mechanism, and then increases at the end of the cycle, when etching is slower and more atoms are able to diffuse.
Time evolution of the ratio during a series of ten consecutive cycles of etching at 1.2 Torr. These simulated evolutions appear similar to the variations observed experimentally on silicon in Fig. 9. During the first cycle, the ratio rises above 50%.
Article metrics loading...
Full text loading...