We investigated the interactions of atomic and molecular chlorine with plasma-conditioned stainless steel surfaces through both experiments and modelling. The recombination of Cl during adsorption and desorption of Cl2 was characterized using a rotating-substrate technique in which portions of the cylindrical substrate surface are periodically exposed to an inductively coupled chlorine plasma and then to an Auger electron spectrometer in separate, differentially pumped chambers. After several hours of exposure to the Cl2 plasma, the stainless steel substrate became coated with a Si-oxychloride-based layer (Fe : Si : O : Cl ≈ 1 : 13 : 13 : 3) due to chlorine adsorption and the erosion of the silica discharge tube. Desorption of Cl2 from this surface was monitored through measurements of pressure rises in the Auger chamber as a function of substrate rotation frequency. Significant adsorption and desorption of Cl2 was observed with the plasma off, similar to that observed previously on plasma-conditioned anodized aluminium surfaces, but with much faster desorption rates that are most likely attributable to the smoother and non-porous stainless steel surface morphology. When the plasma was turned on, a much larger pressure rise was observed due to Langmuir–Hinshelwood recombination of Cl atoms. Recombination coefficients, γCl, ranged from 0.004 to 0.03 and increased with Cl-to-Cl2 number density ratio. This behaviour was observed previously for anodized aluminium surfaces, and was explained by the blocking of Cl recombination sites by adsorbed Cl2. Application of this variable recombination coefficient to the modelling of high-density chlorine plasmas gives a much better agreement with measured Cl2 percent dissociations compared with predictions obtained with a recombination coefficient that is independent of plasma conditions.

No abstract

In the dual damascene microelectronics integration scheme during the last stage of plasma etching of dielectrics down to underlying Cu layers, Cu is sputtered onto the reactor walls and is believed to cause a drift in etching rates. For photoresist etching in an O2-containing plasma, a drop in etching rate suggests that Cu could cause a decrease in the O-atom concentration in the plasma, due perhaps to an increase in the O recombination rate on the chamber walls. We therefore studied the effects of traces of Cu on O recombination on an oxygen plasma-conditioned surface, using the spinning wall technique. With this method, a cylindrical substrate, here coated in situ with sputter-deposited Si and then oxidized in an O2 plasma, is rotated past skimmers, allowing the surface to be periodically exposed to the plasma and an Auger electron spectrometer with a pressure gauge in a differentially pumped chamber. Between plasma exposures, the sample could also be dosed with Cu from an evaporation source in a differentially pumped chamber. With no Cu on the surface, a pressure rise was observed in the Auger chamber, due to desorption of recombined O2. These measurements were used to derive a Langmuir–Hinshelwood recombination coefficient of γO=0.043 for the steady-state oxidized Si, Cu-free surface. The surface was then coated with a small fraction of a monolayer (roughly ∼0.002 monolayers of Cu with a dose of ∼1.4×1013 cm−2 and an assumed sticking coefficient of 0.3) and γO was found to increase to 0.069. Further dosing with Cu did not produce any further increases in γO. The initial low γO value could not be recovered by coating the surface with sputter Si, apparently due to rapid outdiffusion of Cu through Si at room temperature. Cu catalyzed recombination of O is ascribed to a redox cycling between Cu+ and Cu2+ oxidation states.

Mass and Auger electron spectroscopy studies of the interactions of atomic and molecular chlorine on a plasma reactor wallHigh temperature reaction of nitric oxide with Si surfaces: Formation of Si nanopillars through nitride masking and oxygen etching Modification of a high vacuum, crossed molecular beam scattering system to perform angle-resolved, gassurface scattering studies under ultrahigh vacuum conditions Rev.The interplay between chlorine inductively coupled plasmas (ICP) and reactor walls coated with silicon etching products has been studied in situ by Auger electron spectroscopy and line-of-sight mass spectrometry using the spinning wall method. A bare silicon wafer mounted on a radio frequency powered electrode (À108 V dc self-bias) was etched in a 13.56 MHz, 400 W ICP. Etching products, along with some oxygen due to erosion of the discharge tube, deposit a Si-oxychloride layer on the plasma reactor walls, including the rotating substrate surface. Without Si-substrate bias, the layer that was previously deposited on the walls with Si-substrate bias reacts with Cl-atoms in the chlorine plasma, forming products that desorb, fragment in the plasma, stick on the spinning wall and sometimes react, and then desorb and are detected by the mass spectrometer. In addition to mass-to-charge (m/e) signals at 63, 98, 133, and 168, corresponding to SiCl x (x ¼ 1 -4), many Si-oxychloride fragments with m/e ¼ 107, 177, 196, 212, 231, 247, 275, 291, 294, 307, 329, 345, 361, and 392 were also observed from what appear to be major products desorbing from the spinning wall. It is shown that the evolution of etching products is a complex "recycling" process in which these species deposit and desorb from the walls many times, and repeatedly fragment in the plasma before being detected by the mass spectrometer. SiCl 3 sticks on the walls and appears to desorb for at least milliseconds after exposure to the chlorine plasma. Notably absent are signals at m/e ¼ 70 and 72, indicating little or no Langmuir-Hinshelwood recombination of Cl on this surface, in contrast to previous studies done in the absence of Si etching.

Abstract:In low-pressure plasmas commonly used in materials processing, plasma-wall inter actions play a crucial role in the evolution of the plasma properties both over time and across large-area wafers. We have recently studied the heterogeneous recombination of O and Cl atoms on reactor walls in O 2 and Cl 2 plasmas through both experiments and modeling. The Langmuir-Hinshelwood (i.e., delayed) recombination was investigated using a "spinning-wall" technique in which a portion of the substrate surface is periodically exposed to an inductively coupled plasma and to a differentially pumped chamber where either Auger electron spectroscopy (AES) or line-of-sight mass spectrometry (MS) is used to detect surface and desorbing species. In this paper, a review of the various effects driving the O and Cl atoms recombination dynamics on anodized aluminum (AA) and stainless steel (SS) surfaces is presented. It is shown that recombination probabilities, γ, can vary following plasma exposure due to surface conditioning. In Cl 2 plasmas, γ was also found to depend on the Cl-toCl 2 number density ratio, a mechanism ascribed to a competition for adsorption sites between Cl and Cl 2 . We have also determined the recombination rates of Cl atoms in Cl 2 high-density plasmas sustained by electromagnetic surface waves by comparing the measured degrees of dissociation of Cl 2 to those predicted by an isothermal fluid model. For a reactor with large SS and quartz surfaces exposed to the plasma, γ values and their dependence on the Cl-toCl 2 number density ratio were consistent with those obtained from the rotating substrate technique. Similar values were obtained for plasmas sustained in a quartz discharge tube. It is expected that for plasmas sustained in or adjacent to a silica tube or plate, the Cl atoms recombination coefficient becomes independent of chamber wall material due to reactor seasoning, producing a silicon-oxychloride layer.

Spatially resolved trace rare gases optical emission spectroscopy was used to analyze the electron energy-distribution function (EEDF) in low-pressure argon plasma columns sustained by surface waves. At frequencies >1 GHz, in the microwave-sustained region, the EEDF departs from a Maxwellian, characterized by a depletion of low-energy electrons and a high-energy tail, whereas in the field-free zone, the EEDF is Maxwellian. Abnormal behavior of the EEDF results from the acceleration of low-energy electrons due to the conversion of surface waves into volume plasmons at the resonance point where the plasma frequency equals the wave frequency and their absorption by either collisional or Landau damping.

Electron energy distribution functions ͑EEDFs͒ were measured in a 50 mTorr oxygen plasma column sustained by propagating surface waves. Trace-rare-gas-optical-emission spectroscopy was used to derive EEDFs by selecting lines to extract "electron temperature" ͑T e ͒ corresponding to either lower energy electrons that excite high-lying levels through stepwise excitation via metastable states or higher energy electrons that excite emission directly from the ground state. Lower energy T e 's decreased from 8 to 5.5 eV with distance from the wave launcher, while T e Ϸ 6 eV for higher energy electrons and T e Ͼ 20 eV for a high-energy tail. Mechanisms for such EEDFs are discussed.

Chlorine atom recombination coefficients were measured on silicon oxy-chloride surfaces deposited in a chlorine inductively coupled plasma (ICP) with varying oxygen concentrations, using the spinning wall technique. A small cylinder embedded in the walls of the plasma reactor chamber was rapidly rotated, repetitively exposing its surface to the plasma chamber and a differentially pumped analysis chamber housing a quadruple mass spectrometer for line-of-sight desorbing species detection, or an Auger electron spectrometer for in situ surface analysis. The spinning wall frequency was varied from 800 to 30 000 rpm resulting in a detection time, t (the time a point on the surface takes to rotate from plasma chamber to the position facing the mass or Auger spectrometer), of ∼1–40 ms. Desorbing Cl2, due to Langmuir–Hinshelwood (LH) Cl atom recombination on the reactor wall surfaces, was detected by the mass spectrometer and also by a pressure rise in one of the differentially pumped chambers. LH Cl recombination coefficients were calculated by extrapolating time-resolved desorption decay curves to t = 0. A silicon-covered electrode immersed in the plasma was either powered at 13 MHz, creating a dc bias of −119 V, or allowed to electrically float with no bias power. After long exposure to a Cl2 ICP without substrate bias, slow etching of the Si wafer coats the chamber and spinning wall surfaces with an Si-chloride layer with a relatively small amount of oxygen (due to a slow erosion of the quartz discharge tube) with a stoichiometry of Si:O:Cl = 1:0.38:0.38. On this low-oxygen-coverage surface, any Cl2 desorption after LH recombination of Cl was below the detection limit. Adding 5% O2 to the Cl2 feed gas stopped etching of the Si wafer (with no rf bias) and increased the oxygen content of the wall deposits, while decreasing the Cl content (Si:O:Cl = 1:1.09:0.08). Cl2 desorption was detectable for Cl recombination on the spinning wall surface coated with this layer, and a recombination probability of γCl = 0.03 was obtained. After this surface was conditioned with a pure oxygen plasma for ∼60 min, γCl increased to 0.044 and the surface layer was slightly enriched in oxygen fraction (Si:O:Cl = 1:1.09:0.04). This behavior is attributed to a mechanism whereby Cl LH recombination occurs mainly on chlorinated oxygen sites on the silicon oxy-chloride surface, because of the weak Cl–O bond compared to the Cl–Si bond.