Microfabricated Wire Arrays for Z-Pinch

From the source

1. Wire arrays for Z are made by hand. On the order of a hundred fine wires are manually strung over a machined mandrel, with weights on the ends to keep them taut and in their slots. The method is time-consuming, limited in the geometric complexity it can produce, and limited in material to what can be drawn into fine wire and then handled — typically tungsten and aluminium. Gold can be drawn to fine wire but is too fragile to handle in that form.Section 1 Introduction, with Figure 1, SAND2008-6088

2. Three reasons to want a fabricated array rather than a strung one, stated in the introduction. A wider selection of wire materials would let experimenters tailor the x-ray spectrum and open up plasma diagnostics. Geometrically structured targets would allow investigation of stabilising the z-pinch implosion — a spherical implosion can be approximated in cylindrical geometry by tailoring the mass density along the radial axis. And a z-pinch driven fusion reactor, if one were developed, would require the production of large numbers of wire arrays at low cost.Section 1 Introduction, SAND2008-6088

3. Of the four standard microfabrication routes for laying down metal, electroplating wins on two counts: the plating rate is relatively high, several microns per hour, and the film stress can be controlled through chemistry, temperature and bias. Chemical vapour deposition of tungsten was rejected despite tungsten being the incumbent wire material, because the deposited metal carries extremely high residual stress and the films are likely to peel during deposition and crumple the support after release. Evaporation and sputtering are both limited by low deposition rates to films too thin for this application.Section 2.1 with subsections 2.1.1 to 2.1.4, SAND2008-6088

4. The anode and cathode connections in Z are cantilevered from the edges of the vacuum vessel and are known to move fifty to one hundred microns further apart as the chamber is evacuated, so a wire soldered to both would be torn. The answer designed here is a planar thin-film spring fabricated in line with each wire, narrower than the 420 micron wire pitch. Finite-element modelling of a gold hairpin, at a Young’s modulus of 100 gigapascals and a yield strength of 140 megapascals, puts the elastic limit at 1.46 microns of displacement per cell with the stress concentrated on the inside of the hairpin rather than in the wire; springs can be stacked in series, and a serpentine variant lowers peak stress by about twenty per cent at the cost of much harder mask layout. The spring was designed but never fabricated before the project ended.Sections 2.2 and 2.6, with Figures 2 and 3, SAND2008-6088

5. A gallium arsenide film grown over an aluminium gallium arsenide layer carries a built-in strain gradient, so when the sacrificial layer beneath it is etched away the film peels off the substrate and rolls itself into a cylinder. By adjusting the layer thicknesses and composition, any radius of curvature from less than one millimetre to more than six centimetres can be obtained, which covers the full range of Z wire array applications, and there is a wide window of thickness over which the curvature barely changes, so the design tolerates growth non-uniformity. Releasing large films through isolated holes on a fifty micron pitch works where edge undercutting does not, and straight hydrofluoric acid etches about 180 microns in ninety minutes against about 12 for buffered oxide etchant, because the near-neutral buffered solution precipitates an aluminium-oxide particulate that clogs the etch channel.Section 2.3, with Figures 4 to 8, SAND2008-6088

6. The result that changed the project, and the gap that was left. Plated gold films turned out to be strong enough to handle directly, so the semiconductor support — the most expensive part of the original proposal — could be dropped entirely and the wires plated onto a sacrificial photoresist layer on any substrate, which the report says drops the cost per array enormously. Arrays 20 millimetres in diameter carrying 150 wires plated to 15 microns were made, with wire dimensional uniformity and roughness at least on the order of existing drawn tungsten wire, and a 300 millimetre line would yield twenty-four 40 millimetre arrays per wafer. What was never demonstrated is the last step: mounting an array into real Z hardware and firing it. The project was ended after two of its three years, because target development at Z had moved toward solid metal cylinders.Sections 2.4, 2.7, 2.8 and 4 Conclusions, with Figures 12, 13, 17 and 18, SAND2008-6088

The way in

How to cite it

Michael J. Cich, John F. Klem, Olga Blum Spahn, Gregory M. Peake, Adam M. Rowen, Christian L. Arrington, Dustin Heinz Romero, Thomas J. Nash (2008) Microfabricated Wire Arrays for Z-Pinch doi:10.2172/945909

Where it sits in the curriculum

• Fusion machines: pinches, focus devices and inertial drivers