Ar Glow Discharge Research Articles

Characterization and depth profiling of E′ defects in buried SiO2

Oxygen-vacancy defects (E′) generated at the surface of buried SiO2 (BOX) layers formed by O+ implantation during the separation by implantation of oxygen process have been studied by electron spin resonance at 4.3 K. The E′ generation tool used was exposure to a dc Ar glow discharge that produces E′ defects predominantly in a surface layer of ≊100-Å thick, reaching local volume densities at the surface up to 8×1019 cm−3. This glow discharge exposure, alternated with step-by-step etch back, allowed mapping of the defect generation sensitivity depth profile for the entire buried layer, revealing a fairly uniform sensitivity with a strong decline towards the BOX/substrate interface. Besides the usual E′γ signal, the E′δ center—a delocalized variant of the E′ center—has been newly observed in the BOX. Reoxidation (950 °C; 1.1 atm O2) of the BOX was observed to reduce the E′ sensitivity close to that of regular dry thermal oxide (≥29 times lower), while the E′δ signal could no longer be generated—again similar to conventional thermal oxide. These data suggest an adapted model for the E′δ defect, based on the existence of small Si clusters (≳5 atoms) in the BOX serving as defect precursors. Generally, the results imply that the buried oxide contains excess Si, exceedingly so near the BOX/substrate interface.

Defect generation sensitivity depth profile in buried SiO2 using Ar plasma exposure

The sensitivity to oxygen-vacancy defect (E′) generation of buried SiO2 (BOX) layers in separation by implantation of oxygen (SIMOX) structures is studied by electron spin resonance. The E′ generation tool used is exposure to a dc Ar glow discharge that produces E’s predominantly in a surface layer of ≊100 Å thick, reaching local volume densities at the surface up to 8×1019 cm−3. Reoxidation of the layers is observed to reduce this value close to that of regular dry thermal oxide (≥29 times lower). This glow discharge exposure, alternated with step-by-step etch back, allowed mapping of the defect generation sensitivity depth profile for the entire buried layer, revealing a fairly uniform sensitivity with a strong decline towards the BOX/substrate interface. The results provide evidence that the buried oxide contains excess Si, exceedingly so near the BOX/substrate interface.

Si atomic layer epitaxy based on Si 2H 6 and remote He plasma bombardment

Atomic layer epitaxy (ALE) of Si has been demonstrated by using remote He plasma low energy ion bombardment to desorb H from an H-passivated Si(100) surface at low temperaturea and subsequently chemisorbing Si 2H 6 on the surface in a self-limiting fashion. Si substrates were prepared using an RCA clean followed by a dilute HF dip to provide a clean, dihydride-terminated (1 × 1) surface, and were loaded into a remote plasma chemical vapor deposition system in which the substrate is downstream from an r.f. noble gas (He or Ar) glow discharge in order to minimize plasma damage. An in situ remote H plasma clean at 250°C for 45 min was used to remove surface O and C and to provide an alternating monohydride and dihydride termination, as evidenced by a (3 × 1) reflection high energy electron diffraction (RHEED) pattern. It was found necessary to desorb the H from the Si surface to create adsorption sites for Si- bearing species such as Si 2H 6. Remote He plasma bombardment for 1–3 min was investigated over a range of temperature (250°C−410°C), pressures (50–400 mTorr) and r.f. powers (6–30 W) in order to desorb the H and to convert the (3 × 1) RHEED pattern to a (2 × 1) pattern which is characteristic of either a monohydride termination or a bare Si surface. It was found that as He pressures and r.f. powers are raised the plasma potential and mean free paths are reduced, leading to lower He bombardment energies but higher fluxes. Optimal He bombardment parameters were determined to be 30 W at 100 mTorr process pressure at 400°C for 1–3 min. He was found to be more effective than Ar bombardment because of the closer match of the He and H masses compared with that between Ar and H. Monte Carlo TRIM simulations of He and Ar bombardment of H-terminated Si surfaces were performed 3o validate this hypothesis and to predict that approximately 3 surface H atoms were displaced by the incident He atoms, with no bulk Si atom displacement for He energies in the range 15–60 eV. The He bombardment cycles were followed by Si 2H 6 dosing over a range of partial pressures (from 10 −7 Torr to 1.67 mTorr), temperatures (250°C–400°C) and times (from 20s to 3 min) without plasma excitation, because it is believed that Si 2H 6 can chemisorb in a self-limiting fashion on a bare Si surface as two silyl (SiH 3) species, presumably leading to a H-terminated surface once again. The Si 2H 6 dosing pressures and times corresponded to saturation dosing (about 10 6 langmuirs). Alternate Si 2H 6 dosing and He low energy ion bombardment cycles (about 100–200) were performed to confirm the ALE mode of growth. It was found that the growth per cycle saturates with long Si 2H 6 dosing at a level which increases slightly with He bombardment time. At 400°C, for 2 min He bombardment at 100 mTorr and 30 W, the growth per cycle saturates at about 0.1 monolayers cycle −1, while for 3 min He bombardment the Si growth saturates at about 0.15 monolayers cycle −1. It was also confirmed that the growth is achieved only by using alternate He bombardment and Si 2H 6 dosing. He bombardment alone for a comparable time (3 min × 100 cycles) causes a negligible change in the Si film thickness (less than 5 Å). Similarly, thermal growth using Si 2H 6 under these conditions for (3 min × 100 cycles) causes negligible deposition (less than 5 Å).

Electron Energy Distribution Function at Potential Drop between two Negative Ar Glow Discharges

AbstractThe electron energy distribution function was measured at the potential drops (double layer) formed between two negative Ar glows. The acceleration of the electrons by the electric field and the electron trapping due to the high plasma potential formed a double‐humped distribution function.

High sensitivity absorption spectroscopy in glow discharge plasmas

A highly sensitive absorption experiment for diagnosing glow discharge plasmas is described. This experiment is applicable from the VUV to the IR. A very stable Xe arc lamp is used as a source of continuum radiation. An echelle spectrometer equipped with a gated, image-intensified, charge-coupled device detector array is used to disperse and detect the continuum, with absorption features, after it has traversed the glow discharge. Digital subtraction is used to discriminate against the line emission from the glow discharge and detect only the continuum emission from the arc discharge. Estimates of the relative spectral radiances of glow and arc discharges suggests the subtraction technique is broadly applicable to glow discharge studies. A fractional absorption of 10−3 is detectable with a signal-to-noise ratio limited primarily by shot noise. A detection limit for excited Hg atoms of 7×109 cm−2 is demonstrated in a 400-mA Hg–Ar glow discharge. Further improvements in the experiment are proposed.

Photodesorption from copper, beryllium, and thin films

Ever increasing circulating currents in electron-positron colliders and light sources demand lower and lower photodesportion (PSD) from the surfaces of their vacuum chambers and their photon absorbers. This is particularly important in compact electron storage rings and B meson factories where photon power of several kW cm−1 is deposited on the surfaces. Given the above factors we have measured PSD from 1 m long bars of (a) solid copper and solid beryllium, and (b), TiN, Au, and C thin films deposited on solid copper bars. Each sample was exposed to about 1023 photons/m with a critical energy of 500 eV at the vacuum ultraviolet ring of the National Synchrotron Light Source. PSD was recorded for two conditions: after a 200 °C bakeout and after an Ar glow discharge cleaning. In addition, we also measured reflected photons, photoelectrons, and desorption as functions of normal, 75, 100, and 125 mrad incident photons.

Depth Profiling of Oxygen Vacancy Defect Generation in Buried SiO2

ABSTRACTOxygen-vacancy defects (E') generated at the surface of buried SiO2 layers formed by O+ implantation during the separation by implantation of oxygen (SIMOX) process have been studied by electron spin resonance (ESR) at 4.2 K. The E' damage was generated during exposure to a dc Ar glow discharge that produces defects predominantly in a surface layer of ≈100 Å thick, reaching local volume densities at the surface up to 8 × 1019 cm−3. This glow discharge exposure, alternated with step-by-step etch back, allowed mapping of a defect generation sensitivity depth profile of the buried oxide (BOX) layer, revealing a fairly uniform sensitivity with a strong decline towards the BOX/substrate interface. Besides the usual E' γ signal, the E'δj center — a delocalized variant of the E' center — has been newly observed in the BOX. Reoxidation of the BOX was observed to reduce the E' sensitivity close to that of regular dry thermal oxide (≥ 29 times lower), while the E'γ signal could no longer be generated, again similar to conventional thermal oxide. These data suggest a revision of the model for the E' defect. In general, the results strongly suggest that the buried oxide contains excess Si, exceedingly so near the BOX/substrate interface.

Two-dimensional profile of emissive species in a magnetized SiH/sub 4//Ar glow discharge for the scanning-plasma CVD method

In order to control reactive glow discharge plasmas, the authors have investigated a crossed magnetic field B perpendicular to the discharge electric field E. It was found that profiles of SiH* emission, had two peaks in both the negative glow region and the cathode sheath region. As a result of the modification of the plasma structure by the crossed magnetic field, subsequent modification of spatial profiles of optical emission from SiH and Ar were observed. The H/sub alpha / emission in the cathode sheath region, were not much influenced by the crossed magnetic field. By applying a modulated magnetic field in AC discharge plasmas (scanning plasma method), time-averaged uniform profiles of both plasmas and SiH radicals could be realized near the substrate sustained outside the discharge region. The minimum inhomogeneities of both profiles were less than 2.5% for the magnetic flux density of several tenths of a gauss, in contrast with 15% for the condition of no magnetic field.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

A tuned Langmuir probe for measurements in rf glow discharges

Measurements of charged-particle concentrations and the electron energy distribution function (EEDF) have been made in Ar and SF6 glow discharges using a tuned Langmuir probe technique. A simple passive circuit connected to the probe when properly tuned increases the impedance between the probe and ground, thereby forcing the probe to follow the instantaneous plasma potential. In this manner, rf-induced distortion of the probe characteristic is mitigated. At 13.56 MHz the electron collection characteristic of a detuned probe is distorted by rf interference; the ion collection characteristic is unaffected. The EEDF is highly non-Maxwellian in argon discharges, but quite Maxwellian in SF6 discharges. The mean electron energy increases with decreasing pressure and increasing power in argon discharges, but is independent of pressure and power in SF6 discharges. The measured distribution functions and charged particle concentrations are in good agreement with calculations.

Studies on ion formation in a glow discharge mass spectrometry ion source

The relative populations of singly- and mutiply-charged argon, matrix and argide ions in an Ar-glow discharge have been measured in different matrices using the VG-9000 glow discharge mass spectrometer. The measured relative densitites of multiply-charged argon ions (from +3 to +6) are in good agreement with the Lovett rate model for electron impact ionization/three-body recombination with an electron temperature of about 7 eV (R.J. Lovett, Spectrochim. Acta, 37B, 985 (1982)). An overpopulation of singly- and doubly-charged argon ions relative to this relationship has been observed in all cases. The overpopulation of singly-charged argon ions is caused by a process of quenching of metastables in a doubly excited system. So-called “secondary electrons” with an average energy about 7 eV produced in this reaction can be an agent creating multiply-charged argon ions. Sputtered sample/cathode atoms are ionized mainly by Penning ionization (PI). A dominance of PI has been found in the production of singly- and multiply-charged ionic species for which ionization energies are lower than the energy of argon metastable atoms. The remaining species appear to be ionized by the secondary electrons.

An impulsive collision model for the vibrational and rotational excitation of sputtered molecules: R De Jonge et al,Nucl Instrum Meth Phys Res,B30, 1988, 159–172

Atomic layer epitaxy (ALE) of Si has been demonstrated by using remote He plasma low energy ion bombardment to desorb H from an H-passivated Si(100) surface at low temperaturea and subsequently chemisorbing Si2H6 on the surface in a self-limiting fashion. Si substrates were prepared using an RCA clean followed by a dilute HF dip to provide a clean, dihydride-terminated (1 × 1) surface, and were loaded into a remote plasma chemical vapor deposition system in which the substrate is downstream from an r.f. noble gas (He or Ar) glow discharge in order to minimize plasma damage. An in situ remote H plasma clean at 250°C for 45 min was used to remove surface O and C and to provide an alternating monohydride and dihydride termination, as evidenced by a (3 × 1) reflection high energy electron diffraction (RHEED) pattern. It was found necessary to desorb the H from the Si surface to create adsorption sites for Si- bearing species such as Si2H6. Remote He plasma bombardment for 1–3 min was investigated over a range of temperature (250°C−410°C), pressures (50–400 mTorr) and r.f. powers (6–30 W) in order to desorb the H and to convert the (3 × 1) RHEED pattern to a (2 × 1) pattern which is characteristic of either a monohydride termination or a bare Si surface. It was found that as He pressures and r.f. powers are raised the plasma potential and mean free paths are reduced, leading to lower He bombardment energies but higher fluxes. Optimal He bombardment parameters were determined to be 30 W at 100 mTorr process pressure at 400°C for 1–3 min. He was found to be more effective than Ar bombardment because of the closer match of the He and H masses compared with that between Ar and H. Monte Carlo TRIM simulations of He and Ar bombardment of H-terminated Si surfaces were performed 3o validate this hypothesis and to predict that approximately 3 surface H atoms were displaced by the incident He atoms, with no bulk Si atom displacement for He energies in the range 15–60 eV. The He bombardment cycles were followed by Si2H6 dosing over a range of partial pressures (from 10−7 Torr to 1.67 mTorr), temperatures (250°C–400°C) and times (from 20s to 3 min) without plasma excitation, because it is believed that Si2H6 can chemisorb in a self-limiting fashion on a bare Si surface as two silyl (SiH3) species, presumably leading to a H-terminated surface once again. The Si2H6 dosing pressures and times corresponded to saturation dosing (about 106 langmuirs). Alternate Si2H6 dosing and He low energy ion bombardment cycles (about 100–200) were performed to confirm the ALE mode of growth. It was found that the growth per cycle saturates with long Si2H6 dosing at a level which increases slightly with He bombardment time. At 400°C, for 2 min He bombardment at 100 mTorr and 30 W, the growth per cycle saturates at about 0.1 monolayers cycle−1, while for 3 min He bombardment the Si growth saturates at about 0.15 monolayers cycle−1. It was also confirmed that the growth is achieved only by using alternate He bombardment and Si2H6 dosing. He bombardment alone for a comparable time (3 min × 100 cycles) causes a negligible change in the Si film thickness (less than 5 Å). Similarly, thermal growth using Si2H6 under these conditions for (3 min × 100 cycles) causes negligible deposition (less than 5 Å).

Theoretical emission spectra of NeAl + and ArAl + in the vacuum ultraviolet region

Two diatomic ions, NeAl + and ArAl +, were studied via ab initio MO CI calculations. The first excited singlet state 1Π of both ions forms an excimer, and its estimated radiative lifetime is about 1 ns. The simulated emission spectra from ν′ = 0 peaks at 57900 cm −1 for NeAl + and 53800 cm −1 for ArAl +. The broad emission spectrum observed in an Ar glow discharge using an Al cathode by Yamashita is most likely the excimer band from hot ArAl +.

Rare gas-oxygen effects on the rf sputter deposition of platinum

Thin Pt films were sputter deposited on Si substrates using Ar and Ne rf glow discharges containing 0%–6% O2. Significant differences in electrical resistivity, crystallography, and Pt suboxide formation were found to depend upon the type of rare gas used for a given nominal O2 partial pressure. Several reasons for this behavior are proposed.

Chemical characterization by Auger electron spectroscopy and voltammetry of platinum electrode surfaces prepared in the gas phase

The aim of this work was to find conditions of preparation of a clean platinum surface for electrochemical experiments. Auger electron spectroscopy shows that the main contaminants of the platinum surfaces treated in Ar glow discharge are nitrogen and oxygen while the carbon content is reduced to a low level. Voltammetry shows that after glow discharge a surface compound is irreversibly oxidized. This compound is stable in air and is chemically oxidized by oxygen in presence of water. This compound is probably NO formed and adsorbed from residual atmosphere during glow discharge. When present it inhibits electrochemical adsorption of hydrogen by blocking pairs of hydrogen adsorption sites. The cleanliness of the surface prepared in the present conditions is such that after desorption of NO the electrode immediately adsorbs 95% of the maximum quantity of hydrogen adsorbed by the same electrode electrochemically activated.

Reduction of Effective Sputtering Yield by Honeycomb Structures

The sputtering yield of Mo honeycomb walls was compared with the yield of fiat walls in a d.c. Ar glow discharge. Under the experimental condition where the mean energy is approximately 1 keV and the angle of incidence of primary ions strongly distributed around normal, the yield is reduced by a factor of 20 to 30. Before and after the bombardment, the surface topography of the honeycomb targets was observed by a scanning electron microscope.

Reduction of Effective Sputtering Yield by Honeycomb Structures

The sputtering yield of Mo honeycomb walls was compared with the yield of fiat walls in a d.c. Ar glow discharge. Under the experimental condition where the mean energy is approximately 1 keV and the angle of incidence of primary ions strongly distributed around normal, the yield is reduced by a factor of 20 to 30. Before and after the bombardment, the surface topography of the honeycomb targets was observed by a scanning electron microscope.

Electron Excitation and Deexcitation Coefficients for the 3P2, 3P1, 3P0, and 1P1 Levels of Argon

Electron excitation and deexcitation coefficients have been measured for the first excited levels of Ar (3P0, 3P1, 3P2, and 1P1) using line absorption techniques to obtain excited state population densities in an Ar glow discharge in the pressure range of 0.2 to 10.0 Torr. In addition, several excitation transfer coefficients for excited Ar atom – ground state Ar atom collisions have been derived from the population density measurements and the results are in agreement with previous measurements obtained using pulsed Ar afterglows. The measured electron excitation and deexcitation coefficients yield velocity averaged cross sections for the various collision induced transitions. There is order of magnitude agreement with deexcitation cross sections calculated using a modified Thompson cross section and a Maxwellian distribution for the electrons. The assumption of a Maxwellian distribution is inappropriate for the calculation of electron excitation coefficients since the average electron energy is much smaller than the threshold for excitation of the first excited levels in Ar.

On the direct measurement of diffusion at temperatures less than 0.5 Tm

The important microsectioning techniques developed in the last decade for the direct measurement if diffusion at low temperatures (<0.5 T m ) are briefly reviewed and critically discussed. A powerful variant of these techniques is material removal by r.f. backscattering in an Ar glow discharge which enables tracer profiling to be carried out in steps of about 30 Å or more directly and expeditiously and also independently of the nature of material. The relative merits of the various techniques described are discussed by examining the quality of the recently available low temperature diffusion data for Au.

Positive ion ratio measurements in Ar, Kr, and Xe glow discharges

Results are presented of an experimental study of the relative concentration of ions in the positive column of small-diameter noble-gas discharges. The ratio of the molecular ion concentration to the atomic ion concentration has been measured as a function of discharge pressure and current for dc discharges in argon, krypton, and xenon. Discharge pressures and currents were varied from 0.1 to 10 Torr and 15 to 40 mA. A simple theoretical model of the positive column was developed and the pressure dependence of the ion concentration ratio was calculated. By parametrically fitting the predicted ion ratio equation to the experimental data, estimates of several reaction rate coefficients could be derived. Where possible, the ion ratio data and the derived rate coefficients have been compared to previously published data.