Induction-Module Development

The induction accelerator is the type of driver preferred by the US heavy-ion fusion program.  In this device, a pulsed electric field along the accelerator axis is induced by increasing the magnetic flux in a ferromagnetic core encircling the beam pipe.  In effect, each core is a one-turn transformer, with the beam acting as the secondary winding.  The pulser used to drive the magnetic flux, the induction core, and the gap across which the electric field is produced are together called an induction module or, equivalently, an induction cell.  The induction-module research at LLNL and LBNL has three main objectives:  designing relatively conventional modules for near-term experiments like HCX, developing advanced pulsers and module geometries, and improving the performance of insulators and cores.

Pulser design involves a trade-off between flexibility and cost.  For a fusion driver, pulsers should ideally have high peak power, a 5-10 Hz repetition rate, agile waveform control, and a production cost less than $20/J.  The least expensive type of pulser consists of a passive pulse-forming network (PFN) switched by a sparkgap.  This type of pulser, however, has poor waveform control and a relatively short spark-gap lifetime (~105 shots).  Replacing the spark gap with a thyratron increase the lifetime (~106 - 107shots), but the cost is substantially higher, and there is still little wavefrorm flexibility.  Solid-state pulser are being studied, despite their even higher cost, because they offer high reliability and long liftetime, as well as allowing programmable waveforms.   Under a small-business research contract, Diversified Technologies has investigated the use of low-inductance Insulated-Gate Bipolar Transistors (IGBTs) for switching, followed by magnetic pulse compression, and has reported that such a hybrid system could meet the  performance requirements of a fusion driver.  Also, on a separate small-business contract, First Point Scientific is developing a solid-state pulser which will add corrections with a voltage up to 20 kV to waveforms generated by more traditional pulsers, allowing more accurate control of the beam longitudinal profile.  When complete, this pulser will be tested on HCX.

The work on advanced induction-module designs is aimed at increasing the maximum voltage gradient across the gap.  The principal limitation on cell voltage is electrical breakdown across the insulators,  so improved insulator materials are being investigated as a way to increase the cell voltage.   The requirements for insulators are stringent, including a surface-breakdown strength of 100 kV/cm, compatibility with ultra-high vacuum (10-8 Torr), and high mechanical strength.  Nonetheless, recent tests of Mycalex, layered insulators, and glass insulators show promise, although these materials may exceed the $10/kV production-cost target.  Another approach to increasing cell voltage is to design vacuum-insulated induction modules with long radial insulators.  The principal questions about this approach are whether a suffiently good vacuum can be achieved in an accelerator to hold the required voltage and whether heat generated by the cores can be dissipated.  The HIF-VNL is considering construction of an induction-cell test stand both to examine major issue of the vacuum-insulated cell concept and to test various core configurations, insulator materials, and cell geometries.

Research into core materials and fabrication techniques is being carried out in parallel with pulser design.  During the last five years, many amorphous and nanocrystalline iron-based alloys have been tested for their magnetic properties, uniformity, and ease of fabrication.  To minimize losses, cores made from these materials must be built up in thin (~2 x 10-5 m) layers, and each layer must be insultated to prevent current flow across them.  Cores are typically fabricated by casting or rolling the material into a thin ribbon, coating it with insulation, and winding the ribbon onto a spool.  At present, amorphous materials are the least expensive, but available insulating materials that can survive the annealling process are either too thick, degrade performance, or do not hold sufficient voltage.  Silicon steel is more expensive and has high losses, but it is an alternative for longer pulse durations due to its higher magnetic-flux swing.  Nanocrystalline steel is also more expensive than amorphous materials and produces a lower voltage for a given core size, but it is attractive for applications where low core losses are important.

For comments or questions contact or  Work described here was supported by the Office of Fusion Energy at the US Department of Energy under contracts  DE-AC03-76SF00098 and W-7405-ENG-48.  This document was last revised June, 2002.