We have observed frequency shifts of the 2͞1 tearing mode in a tokamak that are consistent with injecting momentum with low frequency rf waves. The 2͞1 frequency increased for current drive antenna phasing, decreased for anticurrent drive phasing, and was linear with rf power and driven current. The change in the toroidal velocities derived from the frequency shifts and the calculated rf momentum is consistent. The frequency change also increased with magnetic field, independent of antenna phasing. This is the first demonstration of modifying the 2͞1 frequency through rf momentum input. [S0031-9007(96)00568-6] PACS numbers: 52.55. Fa, Both theory and experiment suggest that plasma rotation can significantly modify the stability and transport characteristics of tokamak plasmas. Simultaneous improvement of the plasma confinement and stability has been identified in reactor studies as the most significant means for improving the economic attractiveness of the tokamak concept [1,2].Plasma rotation's impact on plasma stability has been examined recently. In tokamak plasmas surrounded by a resistive shell, toroidal plasma rotation has been found to increase tearing mode stability [3]. Theoretical analysis suggests that external kink modes are also stabilized by toroidal rotation [4]. Experimental evidence of this stabilization has been reported by Strait et al. where the stabilization was achieved with rotation velocities a fraction of the Alfvén speed [5]. External control of the momentum input can be utilized to prevent locked modes which cause degradation of confinement and often are precursors to plasma disruptions [6,7]. In addition, sheared plasma rotation has been predicted to cause linear coupling of ballooning modes. This increases damping of these modes which may allow tokamak plasmas to reach the second stability regime [8,9].Models developed to describe the transport in experimentally observed enhanced confinement modes have identified the importance of plasma rotation and sheared rotation in the formation of transport barriers [10]. These models are consistent with the observation of shear flows in H-mode plasmas [11] and negative central shear mode plasmas in the DIII-D tokamak [12]. Experiments also have shown degradation of VH mode with decreasing local toroidal plasma flow [13]. In addition, the power threshold for transport barrier formation is predicted to decrease with increased shear [14].External control of the momentum input has been demonstrated with neutral beam injection. While the net momentum transfer is efficient, the deposition profile is broad. In addition, calculations suggest that plasma rotation driven by radio-frequency (rf) waves can be more efficient than neutral beams if the parallel index of refraction N k $ 30 [15]. Several experiments have reported changes in the plasma toroidal rotation velocity with the application of ion cyclotron heating [16][17][18]. The rotation is generated by the spatial diffusion or loss of fast, resonant ions which leads to a modification of the radial...

Real-time etch rate of thin transparent films is determined within seconds by an in-situ two-color laser interferometer. The use of two colors improves the accuracy of the calculated rates, provides an absolute measure of film thickness for endpoint prediction, and differentiates between etching and deposition. The tool state parameters, rf power to the antenna and the wafer stage, gas pressure and flow rates, are computer controlled and monitored. Real-time etch rate characterization is obtained by monitoring the etch dependence on varying tool state parameters. The density of the etch radicals, chlorine and fluorine, is obtained with xenon and argon actinometry using optical emission and mass spectra. An etch rate model, based on the input power to the wafer stage and the relative density of the etch radicals, is used to develop a model-based real-time control algorithm. This algorithm has been used to control the etch rate of unpatterned polysilicon and SiO2.

A novel electro-optical probe ͑EOP͒ has been developed to characterize the spatial uniformity of various plasma parameters. Spatially resolved electron density and energy distribution function, neutral and charged particle densities, as well as ion flow velocity are determined by the EOP. The design of the EOP combines a Mach probe, back-to-back charge collectors, and a collimated optical fiber. The light collection angle of the optical fiber is limited by recessing the fiber in a ceramic tube. The line-of-sight integration length of the plasma emission is bounded by the charge collector disk. A spatial resolution of 2.4 cm is obtained by the present design of the EOP. The ion flow velocity perpendicular to the charge collector surface is determined by the ratio of the ion saturation currents of the two counter facing charge collectors. Localized actinometry, that combines spatially resolved optical emission spectra and electron energy distribution functions, is used to determine the density of atomic chlorine and fluorine radicals. The spatial distribution is obtained by scanning the EOP across the plasma volume.

Real-time etch-rate selectivity of thin transparent films is determined within seconds by an in situ two-colour laser interferometer. Two-point laser reflectrometry is used to determine the etch rates and etch selectivity of unpatterned polysilicon and SiO 2 films. The state of a magnetically confined inductively coupled plasma tool used in the experiments, including rf power to the antenna and the wafer stage, gas pressure and flow rates, is computer controlled and monitored. The dependence of the etch selectivity on tool state parameters is characterized in real time by varying the tool state and monitoring the etch rate of the films. The etch rate measurements have been combined with density characterization of the etch radicals in the discharge. The time-dependent density of atomic chlorine during the process is obtained by temporal optical-emission actinometry combined with in-line mass spectrometry from a pressure calibrated residual gas analyser. The etch selectivity of polysilicon to SiO 2 in a chlorine discharge, determined by a simultaneous etch of both films, differs from the etch-rate ratios of the films when they are etched separately. It is suggested that this difference is due to the presence of oxygen in the polysilicon etch process when the two films are etched simultaneously.

The loop voltage response in the low-frequency current drive experiments is analyzed in order to extract information about the current drive profile and efficiency. 0 1995 American Znstitute of Physics.after the RF turn-on, are flat within experimental uncertainty in the bulk of the plasma (Fig. 4). The average Z,,= 1.8kO.5 in the Ohmic phase is consistent with 'the measured plasma resistance of 13.5 & that implies Z&= 1.9 for the temperature profile obtained from the-Thomson scattering.' Upon the ~~ RF turn-on, Z& rises to 2.5t0.5.The Z,, profile flatness, together with the constancy-of the electron temperature profile, allows us to represent the corresponding Spitzer conductivity as a separable function of space and time: u(r,t) = [(t)&(r), with ir normalized in such a way that l= 1 in the Ohmic phase. The toroidal electric field diffusion in such a case has been analyzed in the companion paper.3 Characterizing current drive by the source function S = JCdl&, where J, is the driven current density, the edge loop voltage change, for a constant total plasma current, can be expressed as 'L..