[en] We report optical emission spectroscopy (OES) measurements in cylindrical inductively coupled plasma (ICP) discharges, focusing on E-to-H mode transition in rare gases (Ar, Kr, Xe). The discharge configuration is designed primarily as a radical source for treatment in a high vacuum system, but little is known about its characteristics. Two discharge modes are observed in Ar and Kr on increasing the power injected into the discharge (E-mode at low injected power and H-mode at high injected power). In this paper, we use OES measurements and calculations to infer rare gas discharge characteristics and mode transitions for this geometry in rare gases. It is shown that emission line intensity ratios from the OES signal of a level that is strongly coupled to a metastable state with one whose principal production mechanism is by direct excitation can be used to infer information about the electronic parameters of the discharge

[en] The authors report measurements and modeling of wafer heating mechanisms in an Ar/O₂ inductively coupled plasma (ICP). The authors employed a commercially available on-wafer sensor system (PlasmaTemp developed by KLA-Tencor) consisting of an on-board electronics module housing battery power and data storage with 30 temperature sensors embedded onto the wafer at different radial positions. This system allows for real time, in situ wafer temperature measurements. Wafer heating mechanisms were investigated by combining temperature measurements from the PlasmaTemp sensor wafer with a three-dimensional heat transfer model of the wafer and a model of the ICP. Comparisons between pure Ar and Ar/O₂ discharges demonstrated that two additional wafer heating mechanisms can be important in molecular gas plasmas compared to atomic gas discharges. The two mechanisms are heating from the gas phase and O-atom surface recombination. These mechanisms were shown to contribute as much as 60% to wafer heating under conditions of low bias power. This study demonstrated how the ''on-wafer'' temperature sensor not only yields a temperature profile distribution across the wafer, but can be used to help determine plasma characteristics, such as ion flux profiles or plasma processing temperatures

[en] Molecular dynamics (MD) simulations have been carried out to examine fundamental etch limitations. Beams of Ar⁺, Ar⁺/F and CF_x⁺ (x = 2,3) with 2 nm diameter cylindrical confinement were utilized to mimic 'perfect' masks for small feature etching in silicon. The holes formed during etch exhibit sidewall damage and passivation as a result of ion-induced mixing. The MD results predict a minimum hole diameter of ∼5 nm after post-etch cleaning of the sidewall.

[en] A cylindrical inductively coupled plasma source (ICPS) is optimized for nitrogen atom production. How to operate in the bright mode (H-mode) in pure nitrogen at relatively low applied power is explained. Actinometry is developed for a quantitative determination of the nitrogen dissociation fraction in the ICPS. These results are compared with those obtained outside the ICPS by modulated beam mass spectrometry. A qualitative agreement is obtained between the nitrogen dissociative fraction inside and outside the ICPS. The higher magnitude of the dissociation fraction given by mass spectrometry compared with that derived from actinometry is explained by the contribution of dissociative recombination of N₂⁺ near the exit hole of the ICPS. A dissociation fraction of up to 0.7 is observed by mass spectrometry. Some results are also reported for argon-nitrogen gas mixtures

[en] The degradation of porous low-k materials, like SiOCH, under plasma processing continues to be a problem in the next generation of integrated-circuit fabrication. Due to the exposure of the film to many species during plasma treatment, such as photons, ions, radicals, etc, it is difficult to identify the mechanisms responsible for plasma-induced damage. Using a vacuum beam apparatus with a calibrated Xe vacuum ultraviolet (VUV) lamp, we show that 147 nm VUV photons and molecular O₂ alone can damage these low-k materials. Using Fourier-transform infrared (FTIR) spectroscopy, we show that VUV/O₂ exposure causes a loss of methylated species, resulting in a hydrophilic, SiO_x-like layer that is susceptible to H₂O absorption, leading to an increased dielectric constant. The effect of VUV radiation on chemical modification of porous SiOCH films in the vacuum beam apparatus and in Ar and O₂ plasma exposure was found to be a significant contributor to dielectric damage. Measurements of dielectric constant change using a mercury probe are consistent with chemical modification inferred from FTIR analysis. Furthermore, the extent of chemical modification appears to be limited by the penetration depth of the VUV photons, which is dependent on wavelength of radiation. The creation of a SiO_x-like layer near the surface of the material, which grows deeper as more methyl is extracted, introduces a dynamic change of VUV absorption throughout the material over time. As a result, the rate of methyl loss is continuously changing during the exposure. We present a model that attempts to capture this dynamic behaviour and compare the model predictions to experimental data through a fitting parameter that represents the effective photo-induced methyl removal. While this model accurately simulates the methyl loss through VUV exposure by the Xe lamp and Ar plasma, the methyl loss from VUV photons in O₂ plasma are only accurately depicted at longer exposure times. We conclude that other species, such as oxygen radicals or ions, may play a major role in chemical modification at short times near the surface of the material, while VUV photons contribute to the majority of the damage in the bulk.

[en] Damage incurred during plasma processing, leading to increases in dielectric constant k, is a persistent problem with porous ultra-low-k dielectric films, such as SiCOH. Although most of the proposed mechanisms of plasma-induced damage focus on the role of ion bombardment and radical attack, we show that plasma-generated vacuum ultraviolet (VUV) photons can play a role in creating damage leading to increases in the dielectric constant of this material. Using a vacuum beam apparatus with a calibrated VUV lamp, we show that 147 nm VUV photons impacting SiCOH results in post-exposure adsorption and reaction with water vapour from the atmosphere to form silanol bonds, thereby raising the dielectric constant. Furthermore, the level of damage increases synergistically under simultaneous exposure to VUV photons and O₂. The vacuum beam photon fluences are representative of typical plasma processes, as measured in a separate plasma tool. Fourier-transform infrared (FTIR) spectroscopy (ex situ) and mass spectrometry (in situ) imply that O₂ reacts with methyl radicals formed from scissioned Si-C bonds to create CO₂ and H₂O, the latter combining with Si dangling bonds to generate more SiOH groups than with photon exposure alone. In addition, sample near-surface diffusivity, manipulated through ion bombardment and sample heating, can be seen to affect this process. These results demonstrate that VUV photo-generated surface reactions can be potent contributors to ultra-low-k dielectric SiCOH film plasma-induced damage, and suggest that they could play analogous roles in other plasma-surface interactions.

[en] A commercially manufactured PlasmaVolt sensor wafer was studied in an inductively coupled plasma reactor in an effort to validate sensor measurements. A pure Ar plasma at various powers (25-420 W), for a range of pressures (10-80 mT), and bias voltages (0-250 V) was utilized. A numerical sheath simulation was simultaneously developed in order to interpret experimental results. It was found that PlasmaVolt sensor measurements are proportional to the rf-current through the sheath. Under conditions such that the sheath impedance is dominantly capacitive, sensor measurements follow a scaling law derived from the inhomogeneous sheath model of Lieberman and Lichtenberg, [Principles of Plasma Discharges and Materials Processing (Wiley, New York, 2005)]. Under these conditions, sensor measurements are proportional to the square root of the plasma density at the plasma-sheath interface, the one-fourth root of the electron temperature, and the one-fourth root of the rf bias voltage. When the sheath impedance becomes increasingly resistive, the sensor measurements deviate from the scaling law and tend to be directly proportional to the plasma density. The measurements and numerical sheath simulation demonstrate the scaling behavior as a function of changing sheath impedance for various plasma conditions.

[en] The transient temperature profile across a commercial wafer temperature sensor device in an inductively coupled Ar plasma is reported. The measured temperatures are compared to model predictions, based on a coupled plasma-wafer model. The radial temperature profile is the result of the radial profile in the ion energy flux. The ion energy flux profile is obtained by combining the Langmuir probe measurement, the ion wall flux probe measurement, and a plasma model. A methodology to estimate the ion flux profile using the sensor measurements has been validated by combining the plasma measurements, the wafer temperature measurements, and the plasma-wafer model. It is shown that with minimal heat transfer between the wafer and the chuck, the initial transient wafer temperature profile after plasma ignition can be used to estimate the ion energy flux profile across the wafer

[en] A key challenge in modeling atmospheric pressure plasma chemistry is to incorporate a wide range of time and length scales. Electron and ion dynamics often evolve over shorter time scales and smaller length scales than reactive neutral species, so to speed calculations, models often decouple charged and neutral species dynamics using various iteration schemes. However, improper decoupling can lead to inaccuracies since neutral species can affect charged species and vice versa. In this communication, we demonstrate that certain ion–neutral reactions can significantly alter reactive species densities in models of surface micro-discharges in humid air. (fast track communication)

[en] Absolute total dissociation cross sections, σ_t,diss, by electron-impact are reported for CHF₃ and C₃F₈ from 10 to 300 eV using the chemical gettering technique described by Winters and Inokuti (1982 Phys. Rev. A 25 1420). Data are concentrated in the near-threshold region (10-30 eV). The thresholds for dissociation of CHF₃ and C₃F₈ are determined to be 10.4 eV and 11.9 eV, respectively. Ionization thresholds occur at 16 eV for CHF₃ and 16.2 eV for C₃F₈. Neutral dissociation cross sections of both CHF₃ and C₃F₈ are obtained by subtracting the ionization cross sections, σ_t,ion, from the total dissociation cross sections, σ_t,diss