HIGH-GRADIENT INSULATOR CAVITY MODE FILTER
STATEMENT OF GOVERNMENT INTEREST The United States Government has rights in this invention pursuant
to Contract No. W-7405-ENG-48 between the United States Department of Energy and the University of California for the operation of Lawrence Livermore National Laboratory.
CROSS-REFERENCE TO RELATED APPLICATIONS This Application is a Continuation-ln-Part of United States Patent
Application serial number 08/889,587, filed 7/8/97, for "HIGH GRADIENT INSULATOR" and Provisional United States Patent Application serial
number 60/035,463, filed 1/14/97, for "HIGH GRADIENT INSULATOR
CAVITY MODE FILTER. Ail such applications are incorporated herein by
reference.
BACKGROUND OF THE INVENTION The present invention relates in general to high-power devices and,
in particular, to insulators and mode filters for such high power devices.
The invention relates especially to induction accelerators and more
particularly to the use of an insulator structure having alternating layers of
dielectric and metal in the accelerating gaps of the induction cell stages in
inductive accelerators.
Induction accelerators are a unique source for high-current, high-
brightness, charged particle beams. Induction accelerators are used to
drive very-high-power microwave sources such as free electron lasers and relativistic klystrons, as a driver for intense X-ray sources for radiographic
applications and as intense beam sources for material processing. Induction accelerators are also expensive, and the higher energy induction
accelerators have been limited to major national laboratories. For these
accelerators to become commercially viable, the basic unit of the accelerator, the induction module, must be made more compact, more
efficient and less costly.
A typical induction accelerator is described in United States Patent
4,730,166, entitled, "Electron Beam Accelerator with Magnetic Pulse
Compression and Accelerator Switching", and which patent is is
incorporated by reference herein. A typical induction accelerator includes serially arranged induction cells that each have a conductor shell around a
core of ferrite or other ferromagnetic material. When a pulsed voltage is
applied to such conductor, an equivalent voltage is developed across a
portion of the conductor shell at what is referred to as the accelertor gap.
Staging such accelerator cells along the trajectory of a charged particle
beam can impart an acceleration energy to the beam. A supporting
conductive structure is used to contain and support the ferromagnetic
material. Only a part of the conductive structure functions as the actual
conductor around the ferromagnetic material.
Because of an adverse interaction that can occur between air
molecules at atmospheric pressure and the electron beam which is to be
accelerated, induction accelerators use sealed and evacuated beam pipes
that thread through ferrite torroids. The evacuated beam pipe allows an unimpeded propagation of the charged particle beam. In the induction cell,
the regions surrounding the ferrite cores are typically operated at high
electric fields, so present designs use a high dielectric strength fluid, e.g.,
oil to surround the ferrite. The oil acts like a self-healing insulator and enhances the resistance to electrical breakdown over the accelerator gaps. The dielectric fluid and the vacuum needed within the evacuated pipe are
typically separated at the accelerating gap by an alumina insulator.
It is well known that a propagating charged particle beam, particularly
one propagating at speeds close to that of the speed of light, will radiate significant amounts of electromagnetic energy. Cavities, and cavities
coupled through a neck, tube or gap will express certain preferential
electromagnetic modes and/or frequencies of oscillation, e.g., resonance.
These preferential modes or frequencies are readily calculated and can be
estimated form the dimension and shape of the system, the constituent
materials and the way the electromagnetic energy is initially introduced. Such energy can excite various electromagnetic modes and/or frequencies
in the beam pipe, in the accelerating gap in the induction cell, and in the
ferrite regions. Resulting interactions may produce undesirable
electromagnetic modes and/or frequencies. Such interactions can adversely affect the controlled propagation of the beam through the
accelerator by causing the beam to oscillate transverse to the direction of
propagation. Such oscillations, referred to as beam break-up instabilities
(BBU), generally start as small effects at the beginning of the beam, but may be amplified in succeeding portions of the accelerator to reach significantly
high magnitudes.
The amount of growth of these transverse oscillations can be understood and characterized by well known techniques. A useful parameter in this analysis is the transverse interaction impedance, Z^ of the cavity (i.e., the impedance normal to the direction of propagation of the
beam). If the transverse impedance is sufficiently large at a given electromagnetic mode and/or frequency, oscillations can continue to grow.
And if the transverse impedance is sufficiently small at a given
electromagnetic mode and/or frequency, oscillations will damp at a large
rate and become inconsequential. Thus it is desirable to design an accelerator cell with as low a transverse impedance as is possible over the
frequency range of the unwanted oscillations to reduce or minimize BBU.
In current induction accelerator designs, the transverse impedance is
sufficiently high at multiple frequencies that problematic oscillations can
result. The large transverse impedances at some of these frequencies are
known to result from the accelerator gap in the conductive shell of the beam
pipe in the induction cell. Because the transverse impedance from the accelerator gap is inversely proportional to the square of the beam pipe
radius, a large beam pipe has been required to prevent BBU. However, if an amount of decoupling can be realized between the beam pipe and the
induction gap at the appropriate frequencies, the transverse impedance at
these frequencies can be reduced. This in turn will allow a smaller diameter
beam pipe to be used. This will facilitate a smaller and less costly overall accelerator.
Glass, ceramic, and other such materials are universally relied on as insulator materials in high voltage systems and in present accelerators. But
such materials allow the insulators to be damaged by avalanche and
flashovers that occur when the insulator has been subjected to a voltage
over-stressing. Fine cracks can develop that lower the insulator's
breakdown voltage to successive exposures to stress voltages. Conventional bulk material insulators also tend to be very large, and the systems that incorporate them must necessarily provide enough room to
accommodate them. If a compact insulator having a high breakdown
voltage can be utiltized, the size and cost of the accelerator can be reduced.
SUMMARY OF THE INVENTION
It is therefore an object of the present invention is to provide a more
compact, less costly and more efficient induction accelerator.
Another object of the present invention is to provide an improved
induction module for induction accelerators.
A further object of the present invention is to provide a more compact induction module.
Another object of the present invention is to enable the design of an
induction accelerator having a beam pipe of reduced diameter.
Another object of the present invention is to provide an induction
module having an accelerating gap insulator having an improved vacuum surface flashover capability, especially in the presence of a cathode and an electron beam.
A further object of the present invention is to provide an accelerating gap insulator having lower field stress on the vacuum surfaces.
Yet another object of the present invention is to provide an
accelerating gap presenting reduced (low) transverse interaction impedance at beam breakup frequencies to minimize beam breakup instabilities.
A still further object of the present invention is to provide a gap
insulating structure presenting reduced (low) transverse interaction
impedance to minimize beam breakup instabilities.
Still another object is of the present invention is to provide a gap
insulator having reduced (low) longitudinal impedance at the beam
modulation frequency and harmonics to minimize power loss. Still another object is to control beam-structure interaction
impedance's in cavities for microwave sources.
Briefly, these and other objects are provided by employing a
laminated high gradient insulator (HGI) structure as the gap insulator in the
induction cells of the induction accelerator. The HGI structure for the gap
insulator is a hollow cylinder composed of multiple thin annular layers of dielectric with a thin annular conductive layer between each. The maximum insulator characteristic, and especially resistance to surface flashover, is
achieved when the applied electric field traverses perpendicular to the laminate structure. HGI's have demonstrated much greater (factors of 1.5 to
4 times) vacuum surface flashover capability than insulators made from a uniform dielectric. The HGI insulator also provides a transverse interaction
impedance at the accelerator gap that inhibits beam instabilities. The HGI
furthermore can be bipolar in performance since its breakdown voltage and impedance characteristics are identical when the beam propagates in either
direction in the beam tube.
The above objects and other objects, advantages, and features of the present invention will become apparent from the following detailed
description of the invention when considered in conjunction with the
accompanying drawings wherein like reference characters represent like or
similar parts throughout the several views and wherein:
BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a schematic partial cross-sectional view (taken along the
longitudinal center line of the accelerator) illustrating an accelerator
induction cell having a high gradient insulator disposed in the acceleration
gap according to the present invention;
Fig. 2 is a schematic of an experimental test cavity which represents a simplified induction module which experimentally illustrated the operation
of the HGI as the gap insulator in an induction module;
Fig. 3A and 3B are plots of the transverse impedance versus
frequency of the experimental structure of Fig. 2 including measurements
with a solid insulator and with several variations in the HGI insulator-to- metal ratio;
Fig. 4A, B and 4C are a plots of the transverse impedance versus
frequency resulting from a computer simulation of the operation of the
experimental structure of Fig. 2;
Fig. 4D is a schematic drawing of the a portion of the module design
of Fig. 1 illustrating the equipotential lines in the vicinity of the module gap;
Fig. 5 is a partially cut away exploded assembly view of a hollow
cylindrical HGI made of a fused stack of metalized flat annular dielectric rings suitable for use in the accelerator gap;
Figs. 6A and 6B illustrate an alternate non laminate structure for an
HGI embodiment; and
Figs. 7A, 7B and 7C illustrate alternative gap designs and gap
insulator designs using the HGI structure in prior art accelerator module designs.
DETAILED DESCRIPTION OF THE INVENTION
Referring now to the drawings, and in particular to Fig. 1 , which
shows a preferred embodiment of an induction module or induction cell 10
according to the present invention, is illustrated. The induction cell 10 is symmetrical with respect to the longitudinal centerline 12. Only the half of
the cell structure above the centerline 12 is shown to simplify the drawing as is conventional in the art. The illustrated embodiment is designed for the
Relativistic-Klystron Two-Beam Accelerator program of the U.S. Department of Energy at Lawrence Livermore National Laboratory and the Lawrence
Berkeley National Laboratory.
The induction cell 10 has a supporting cylindrical conductive shell formed by a hollow cylinder 14 forming the outer edge of the induction cell,
annular end plates 16 and 18, and interior annular magnet housing
members 20 and 22. The inner walls of the annular magnet housing members 20 and 22 form the surface of the beam pipe. The interior annular
magnet housing members 20 and 22 have similar annular channels in which
two annular low-field permanent magnet quadruples 23 and 24, respectively, are disposed.
An annular accelerator gap 26 in the wall of the beam pipe is
provided between magnet housing members 20 and 22. In operation, a
pulsed high voltage accelerating electric field is applied across the gap 26.
A HGI 28 is disposed across the gap 26. The HGI 28 incorporates larger
conductors 30 and 32 on each end that are used for mounting the insulator
in the gap 26. These two conductors are formed with a 0.5 cm radius of
curvature and a 90 degree bend on the inner radius and welded to the beam
line at a point of low surface electric field stress to create a vacuum seal.
On their outer radius, the conductors 30 and 32 have a 180 degree bend terminating in an electrical slip connection that rides on the magnet housing
members 20 and 22.
Five annular induction cores 34 of ferromagnetic material (Metglas in
the illustrated embodiment) are disposed within the supporting conductive shell. The induction cores are maintained away from the supporting
conductive shell by insulating supports 36 to allow the cores to be surrounded by high strength dielectric insulating fluid (e.g., oil) in the
channels 38 between the supporting structure including the HGI 28 and the
induction cores 34. An annular microwave absorber 40 is disposed behind
the accelerator gap 26 behind the HGI 28.
The operation of the structure of Fig. 1 is illustrated by the
experimental test cavity shown in Fig. 2. The insulators were comprised of
thin disks (0.05 mm) of polycarbonate or stainless steel. Figs. 3A and 3B
show experimental impedance measurements of the test cavity with an inner
diameter of 14.6 cm and an outside diameter of 19.7 cm and different ratios
of dielectric to conducting material. Fig. 3A illustrates the impedance
spectrum of the cavity for different dielectric and conductor ratios. Fig. 3B
illustrates the effect on the primary resonance due to varying the ratio of dielectric to conducting material. The smaller peaks flanking the primary
resonance at 1.35 GHz are related to the geometry of the transmission line
and are not specific to the test cavity. Addition of microwave absorbers at
the ends of transmission line pipe reduced these impedance's while not
affecting the primary 350 MHz resonance's. The TE/TM cutoff frequencies
for traveling wave modes in the transmission pipe are 1.2/1.57 GHz,
respectively. The 350 MHz resonance is not effected by the structure of the
insulator. The distribution of the layers was varied to produce different
ratios of insulator to conductor and different number of alternating layers, or
periods, for a specific ratio. The number of layers , or periods of dielectric (polycarbonate) were 9 for the 2:1 ratio, 7 for the 1 :1 ratio and 9 for the 1 :2
ratio.
With a solid insulator disposed in the gap, the test structure shows a
large transverse impedance at approximately 1.35 GHz. Insertion of a HGI structure, where the conductive film is a thin metallic plate of equal thickness to the dielectric, affords a reduction in this impedance by a factor
of 3 to 4. A similar effect was observed in extensive computer simulation as shown in Figs. 4A, 4B and 4C wherein Fig 4A illustrates the calculated
impedance spectrum of the cavity and Fig. 4B illustrates effect on the
primary resonance. The effect of increasing the number of periods while
maintaining the same ratio (2:1 )of dielectric to conductor is shown in Fig. 4C
which illustrates the convergence of the impedance spectrum with increased
number of layers, or periods of insulating material. The impedance converged rapidly with increasing number of periods. Thus it is observed in
computer simulation and also experimentally observed that insertion of the
HGI into the test accelerator cell benefits the operation of the accelerator cell.
The Poisson simulation was used to determine electric field stresses
in the gap 26. An equal potential plot for the gap design is shown in Fig.
4D. The electric field across the insulator in a uniform 100kV/cm. On the oil side the of the insulator the field is 135kV/cm and at the vacuum entrance to
the gap 26 the highest field is 120kV/cm.
In comparison with the solid dielectric gap insulator of the prior art,
the use of the improved high voltage holdoff of the HGI should allow the use of a shorter accelerating gap or, alternatively, a higher voltage gradient
while the lower frequency-selective impedance and mode suppression
should allow the use of smaller diameter beam tubes for a particular accelerator design. The frequency sensitive transverse impedance of the accelerator module with the HGI insulator suggests that the HGI structure
may be used to enhance beam propagation or inhibit unwanted beam
oscillations in accelerator or amplification devices in general through
appropriate control of the transverse impedance in the amplifying cavity.
Figs. 7A-7C illustrate three prior art accelerator gap designs that may
be enhanced by the use of the HGI. Each of the designs utilizes an
insulator that has its vacuum side angled relative to the accelerator beam line to maximize the breakdown electric field. In addition, the gap in each design is curved or angled to offset the gap insulator from a straight line
path to the beam pipe. The HGI could be substituted for each of these
insulators and could include the curved vacuum facing surfaces utilized in
Figs. 7B and 7C. This should allow each of the design to be more compact
while providing equivalent high voltage holdoff and the frequency-selective transverse impedance provided by the HGI. Of course the HGI could be
angled in the design of Fig. 1 to increase the breakdown voltage but this
would also eliminate the bi-directional capability of the HGI when oriented parallel to the beam line, in Fig. 7A-7C the HGI is shown with the interfaces
of the parallel layers being oriented at an angle to the beam line; however,
the HGI may be fabricated to allow the surface of the insulator to be
oriented at an angle to the beam line but the interfaces of the parallel layer oriented normal to the beam line as shown at 28c in Fig. 7C.
Design and Fabrication of a High Gradient Insulator
The partially cut-away exploded assembly view of the HGI of Fig. 5 illustrates an HGI used to fill an accelerator gap 26. An HGI 28 preferably comprises a fused stack of insulator layers, represented in part by layers
100, 102 and 104 having a period that does not exceed non millimeter. For
example, the fused stack of insulator layers 100, 102, 104 can be fabricated
from bulk dielectric materials such as fused silica, quartz glass, alumina,
ceramic, sapphire, etc. Given enough numbers to the stack, any
longitudinal length may be constructed. A series of conductive layers,
represented in part by layers 106, 108 and 110, are disposed between the
insulator layers 100, 102, 104. The structure is equivalent to a capacitive
voltage divider, and may even retain a charge that is divided among the
many layers from the cathode to the anode. The conductive layers may be
fabricated as separate metal foils or plates, or as one or more evaporated
metals diffused into the respective surfaces of the insulator layers.
The conductive layers are used during fabrication to hard seal
insulator layers one to another. Pressure and heat welding, diffusion
bonding, brazing and soldering techniques may be used. The reverse should also be possible, e.g., fabricating flat metal conductor rings with oxides or other dielectric materials on their mating surfaces, and then
chemically bonding the metal oxides or other dielectric materials together.
For HGI structures fused by metal-to-metal bonding, each layer may
comprise a eutectic two part alloy with corresponding constituents that are deposited or diffused into respective faces of the insulator layers. Each conductive layer may alternatively comprise a multi-layer nanostructure foil
initiator that is ignited to bond metals that are already deposited or diffused
into respective faces of the insulator layers.
The maximum insulator characteristic, and especially resistance to surface flashover, is achieved when the applied electric field traverses
perpendicular to the laminate structure. Since the metalizations on each stacked dielectric ring run through from the inside to the outside, electric fields, and especially direct current, applied in line with the laminate
structure will find a low impedance metal conductor through the bulk of the
HGI 28.
In the radial direction, energetic radiation, especially microwave
energy, is low-pass filtered when it passes between the inside and outside
of HGI 28.
At very high frequencies, the inductive reactances of the individual radial paths on the ring conductors becomes significant. Similarly, the
capacitor formed by each pair of ring conductors and intervening dielectric
exhibits significant capacitive reactance. Together, the electrical equivalent is a pi-filter with parallel capacitor inputs and outputs and a series
inductance.
Fig. 6A and 6B illustrate an alternative embodiment of an HGI 28a which does not employ a laminated structure. A dielectric wall segment
120 has both inside parallel metal rings 122 and outside parallel metal rings
124 that are, in at least one embodiment, flush at their tops to the respective surrounding dielectric material surfaces. Therefore, unlike the metal rings of
HGI 28, the metal rings of the HGI 28a do not extend all the way through.
The vacuum integrity HGI 28a would be easier to maintain, but electrical currents through the walls would not be supported. In alternative
embodiments, the depth of penetration of rings 122 and 124 into the dielectric wall 120, and/or their extension from the surface, may be adjusted
to suit particular applications.
In both HGI 28 and HGI 28a, it may be advantageous to apply to the
inner and/or outer surfaces a field emission suppression coating, a
secondary suppression material, or a thin resistive, e.g., semi-conductive,
coating. Conventional deposition means and materials may be used.
Therefore, semiconductor fabrication methods can be used to form systems
of electrically isolated rings on the inside and outside surfaces of a hollow
cylinder of dielectric. Such rings preferably have a spacing period of 100- 1000 nm, and so the width of each conductor on the surface needs to be
under 50 nm. The depth of penetration needs to be no more than 1000 nm
into the surface of the dielectric. The conductive layers are used during fabrication to hard seal
insulator layers one to the another. Pressure and heat welding, diffusion bonding, brazing, and soldering techniques may be used. The reverse
should be possible too, e.g., fabricating flat metal conductor rings with oxides or other dielectric materials on their mating surfaces, and then
chemically bonding the metal oxides or other dielectric materials together.
For HGI structures fused by metal-to-metal bonding, each conductive layer may comprise a eutectic two-part alloy with corresponding constituents that
are deposited or diffused into respective faces of the insulator layers. Each
conductive layer may alternatively comprise a multi-layer nanostructure foil
initiator that is ignited to bond metals that are already deposited or diffused into respective faces of the insulator layers.
Although particular embodiments of the present invention have been
described and illustrated, such is not intended to limit the invention.
Modifications and changes will no doubt become apparent to those skilled in
the art, and it is intended that the invention only be limited by the scope of
the appended claims.