Microwave Imaging Reflectometry (MIR) on NSTX

One of the current application machines is the National Spherical Torus Experiment (NSTX) located at the Princeton Plasma Physics Laboratory. Unlike interferometry and collective scattering, reflectometry is extremely sensitive to the electron density of the target plasma. This is due to the fact that the functions by reflecting off the appropriate plasma cutoff layer. Examples of NSTX characteristic profiles for low, intermediate, and high density conditions are provided in Figs. 3 and 4. The 3-D MIR system will reflect from the right hand X-mode cutoff (fR) in order to get the maximum plasma coverage at the highest frequencies (yielding a shorter wavelength and enhanced plasma resolution).

                                     Fig. 3. Characteristic frequency plots for low (left), and intermediate (right), density L mode NSTX discharges

                                         Fig. 4.  Characteristic frequency plots for intermediate (left), and high (right), density L mode NSTX discharges.

The MIR system for NSTX will be developed in one to two stages. The Stage 1 system targets low to intermediate density L-mode plasmas, with a frequency range of 38 to 52 GHz (see Fig. 3). The Stage 2 system adds a parallel system to target intermediate to high density L-mode plasmas with frequency coverage extending from 52 to 70 GHz (see Fig. 4). Both systems would employ the same window and plasma facing optics, using a dichroic plate beamsplitter to separate the two frequency ranges. On the longer term, the frequency coverage can be increased to 80 GHz should the physics results be deemed sufficiency important. Figure 5 shows the preliminary optical layout developed using the Code V optics code.

                                                                                    Fig. 5.  3-D view of the NSTX MIR receiver imaging optics.

The MIR receiver is comprised of a 2-D array of dual dipole antennas (see Fig. 6). These antennas have a broad radio frequency (RF) bandwidth of 20-25%. This is sufficient to achieve frequency coverage of 38 to 52 GHz as required in the Stage 1 implementation.


 Fig. 6. Schematic layout of 2-D MIR receiving array on NSTX

Unlike the TEXTOR system LINK TO THIS , the MIR system on NSTX will employ multiple illumination frequencies in order to simultaneously monitor density fluctuations on multiple cutoff surfaces. In its initial implementation, this will comprise two to eight frequencies whose separation may be varied remotely (see Fig. 7). The receiver array for the initial MIR implementation will be an 82 array, arranged with 8 channels poloidally and 2 channels toroidally. The plasma coverage could be doubled toroidally by employing a staggered 84 array geometry (see Fig. 6), although this would require increasing the width of the vacuum window from 14.5 cm to ~20 cm.

                             Fig. 7 Upconverting mixer approach to be employed to generate (a) two, or (b) eight simultaneous illumination frequencies

The signals from each imaging channel are amplified, and divided into 8 components (see Fig. 8). Each component is mixed with a distinct LO frequency, offset from its corresponding probing frequency by 70.0MHz. The downconverted signals are then demodulated using I-Q mixers to form in-phase (I) and quadrature (Q) output signals arising from density fluctuations at a distinct plasma location.

                                                              Fig. 8 Downconverting mixer approach to be employed to demodulate MIR signals.

The initial 8 by2 by22 (or possibly as high as 8 by 4 by 8) MIR system will be employed in the study of turbulent fluctuations in the low to intermediate density L-mode NSTX plasmas. Although similar in nature to the UCLA single channel 49 GHz quadrature reflectometer presently installed on Bay J, the multichannel quadrature data collected by the MIR system will image fluctuations across an extended plasma surface with higher resolution and expanded plasma coverage. Simultaneously probing two distinct frequencies, whose separation can be scanned over a 10 GHz bandwidth within a single discharge, allows the radial, poloidal and toroidal correlation properties of these fluctuations to be studied over an extended 3-D plasma volume.
Unlike the TEXTOR tokamak LINK , where an MIR system is presently installed and operating, the NSTX spherical tokamak has shaped plasmas. This presents unique challenges to the optical design, as the radius of curvature of the plasma cutoff layer changes not only as a function of frequency (which translates to a specific flux surface) but also as a function of triangularity and elongation. To first order, we deal with this problem with a careful optics design in which the illumination and receive beams are made roughly normal to the desired cutoff layer, thereby maximizing the return signal. Mechanical translation stages, attached to key optical elements in the illumination and receiver systems, will be remotely controlled to compensate for shot-to-shot changes in triangularity and elongation. As we move away from "proof-of-principle" experiments and into conducting detailed physics studies with the MIR diagnostic, it will becomes more and more important to add an electronically-controlled focusing element so as to adapt/configure our probing beam in real time to match flux surface movements as the plasma evolves in time and provide extended flexibility of the system to ensure that the probe beam is properly matched to the desired cutoff layer for any given operating mode. Here, we note that an electronically-controlled focusing capability would also find use on circular plasmas. For example, PPPL collaborator Dr. Mazzacato has shown that that the tilt tolerance must be better than 0.5 for the TEXTOR MIR system.
The UC Davis approach is to use a linear launch array with symmetric quadratic phase excitation to change the beamwidth, and hence to produce the desired focusing variation. When a quadratic phase curvature is applied to a PAA, the wavefront appears to be concave along the direction of propagation. The effect of changing the phase curvature is to make it appear that the beam emanates from some other waist location, which results in a change in the effective focal length of the PAA "lens". The concept is shown schematically in Fig. 9, while Fig. 10 illustrates how it would be applied to the NSTX MIR system.

                                                                         Fig. 9. Schematic illustration of the variable beam focusing concept.

                                                              Fig. 10. Schematic illustration of the variable beam focusing concept applied to NSTX.

Data collected with the initial non-PAA MIR system(s) will be studied and compared to detailed plasma computational runs, to determine how best to simultaneously match multiple cutoff surfaces (arising from multiple illumination frequencies), and then used to design the optimum PAA approach to achieve a best simultaneous match. Combined with laboratory tests of a scale model PAA system, and leveraged through our companion program on Innovative Plasma Imaging Diagnostics as well as a DARPA grant to develop microelectromechanical systems (MEMS) delay lines and PAAs, UC Davis will develop a wideband PAA "lens" or "lenses" to be incorporated into the final design.

Microwave and Millimeter Wave Beam Steering/Shaping Phased Antenna Arrays

                                                                  Last Modified: Aug., 2005                Send comments to luyang@ucdavis.edu