UCRL-JC-128812
PREPRINT
Multilayer Ultra High Gradient Insulator Technology
S. E. Sampayan
I’. A. Vitello
M. L. Krogh
J. M. Elizondo
This paper was prepared for submittal to the
International Symposium on Discharges and Electrical Insulation in Vacuum
Eindhoven, The Netherlands
August 17-21,1998
March 27,1998
This is a preprint of a paper intended for publication in a journal or proceedings. Since
1~
. . .
understanding that it will not be cited or reproduced without the permission of the
. .
DISCLAIMER
This document was prepared as an account of work sponsored by an agency of
the United States Government. Neither the United States Government nor the
University of California nor any of their employees, makes any warranty, express
or implied, or assumes any legal liability or responsibility for the accuracy,
completeness, or usefulness of any information, apparatus, product, or process
disclosed, or represents that its use would not infringe privately owned rights.
Reference herein to any specific commercial product, process, or service by trade
name, trademark, manufacturer, or otherwise, does not necessarily constitute or
imply its endorsement, recommendation, or favoring by the United States
Government or the University of California The views and opinions of authors
expressed herein do not necessarily state or reflect those of the United States
Government or the University of California, and shall not be used for advertising
or product endorsement purposes.
MULTILAYER ULTRA-HIGH GRADIENT INSULATOR TECHNOLOGY
S. E. Sampayan, P. A Vitello, Lawrence Livermore National Laboratory, P.0 Box 808,
Livermore, CA 9455 1, USA, M. L. Krogh and J. M. Elizondo,
AlliedSignal, FM&T, P. 0. Box 419159, Kansas City, MO 64141, USA
Abstract
- We are investigating a novel insulator con-
cept which involves the use of alternating layers of
conductors and insulators with periods less than 1 mm.
These structures perform many times better (about 1.5
to 4 times higher breakdown electric field) than con-
ventional insulators in long pulse, short pulse, and
alternating polarity applications. We present our ongo-
ing studies investigating the degradation of the break-
down electric field resulting from surface roughness,
the effect of gas pressure, and the performance of the
insulator structure under bi-polar stress. Further, we
present our initial modeling studies.
1. INTRODUCTION
A high gradient insulator consists of a series of very
thin (Cl mm) stacked laminations interleaved with
conductive planes. This insulator technology was
originally conceived and disclosed by Eoin Gray in the
early 1980’s [l] and follows from experimental obser-
vations that the threshold electric field for surface
flashover increases with deceased insulator length [2,
31. The concept laid dormant until the mid-to-late
1980’s: when this technology was pursued and success-
fully showed substantial increases in the breakdown
threshold over conventional, single substrate insulators
[4]. In more recent work we pursued verification of the
technology and showed an increase of 1.5 to 4.0 times
that over conventional insulator technology [5] In later
studies, we explored the properties of these structures
in the context of switching applications, investigating
their behavior under high fluence photon bombardment
WI.
2. EXPERIMENTS
Small sample testing (approximately 2.5 cm diameter
by 0.5 cm thick) was performed in a turbo-molecular
pumped stainless-steel chamber at approximately 10m6
T. High voltage was developed with a 10 J “mini-
Marx”. The Marx developed a pulsed voltage of ap-
proximately 1 to 10 l.ts (base-to-base) and up to 250 kV
amplitude across the sample. Diagnostics consisted of
an electric field sensor and a current viewing resistor.
Failure of the insulator was determined by a prompt
increase in Marx current and a prompt collapse in the
voltage across the sample.
Samples were fabricated by interleaving layers of 0.25
mm fused silica. The interleaved metallic lavers were
formed by depositing gold on each planar insulator
surface by a sputtering technique and then bonding the
stacked layers by heating while applying pressure.
Bond strength between the gold layer and substrate
using this technique was measured to exceed 10 kpsi.
To perform the breakdown experiments, the structure
was slightly compressed between highly polished bare
aluminum electrodes which establish the electric field
for the tests.
To obtain a particular data set, the insulators were
subjected to several low voltage conditioning pulses.
The voltage was then increased a small amount incre-
mentally until breakdown occurred. Voltage was then
reduced for several shots and then incrementally in-
creased again until a constant value was achieved. In
these experiments, however, we generally observed that
these insulators did not condition. Once a breakdown
occurred at a particular field, reducing the voltage
slightly and increasing it again did not cause an in-
crease in breakdown field To produce a given data set
we would apply up to 150-200 shots to a given struc-
ture and would attempt to determine if any damage to
the structure occurred which significantly altered the
breakdown characteristics At these applied energics,
we generally did not observe any degradation. These
data were then reduced to reliability plots by determin-
ing the total number of successful shots over the total
number of applied shots. In these data we define the
electric field as the applied voltage divided by the total
insulator length. We define reliability at a given elec-
tric field as the total number of successful shots over
the total number of shots.
Using this method, we observed flashover of the small
samples at approximately 175 kV/cm for these Iitsed
silica substrates (Fig. 1). The effect of pulse width from
l-10 l.ts on this breakdown threshold was well within
the statistical nature of our data. The trend in conven-
tional insulator technology (Fig 2) for 0” insulators
indicates a breakdown threshold of approximately 50
kV/cm. Thus, there was a net increase in the perform-
ance with these insulators over conventional technol-
ogy of approximately 3.5.
We are also studying various effects which can ad-
versely affect these new structures. For instance, to
ensure concentricity, a finish grinding operation was
performed on the outside diameter. This process is a
time consuming second operation and an alternate
on the multilayer high gradient structures. Similar re-
sults are obtained for bipolar pulses (Figure 6). This
result leads us to conclude that this lamination tech-
nique scales the performance of thin insulators to the
entire structure with each layer acting independently.
,,,,,,,I ,,,,,,,’ ,I
01
1
IO
Insulator Period, mm
Fig. 6 The effect of period length scaling under bi-
polar stress [S]
3. MODELING
Voltage breakdown for insulators surrounded by vac-
uum is believed to take place along the insulator sur-
face and not within the insulator or exterior to the insu-
lator in the vacuum. Breakdown occurs on the time
scale of nanoseconds, making this a difficult process to
study experimentally. The basic physical processes
involved are poorly understood Present theories of
surface breakdown are mainly descriptive, and are not
suited for the purpose of designing insulators. Due to
the computationally intensive nature of a self-
consistently model of surface breakdown, few attempts
have been made to model the complete process.
The two most widely accepted published models for
surface breakdown focus on the initial process occur-
ring either just below or just above the insulator surface
[9, lo] These models focus respectively on solid state
physics phenomena, and the propagation and emission
of electrons through the vacuum just exterior to the
insulator surface. Both models lead to surface heating,
and evaporation of gas from the insulator. This evapo-
rated gas is the medium where ultimately the voltage
breakdown occurs along very localized “streamer”
channels. Pressures in the evaporated gas close to the
insulator surface where the discharges form can reach
a significant fraction of atmospheric pressure. The
surface breakdown is likely to be closely related to high
pressure voltage breakdown processes [ 111.
We have compared processes involved in surface volt-
age breakdown with other well understood voltage
breakdown phenomena at high pressure. Our conclu-
sion is that an accurate theoretical model of surface
breakdown must also include the long time scale evo-
lution of the streamer discharge though the evaporated
gas. Further, the tip of a propagating streamer is
known to produce intense high energy radiation emis-
sion which we believe can lead to photo-conduction in
the insulator. As conduction in the insulator will
strongly modify the voltage which drives the streamer
discharge, accurate coupling of the streamer to the in-
sulator must also be included. Thus, our modeling ap-
proach is that the surface breakdown couples compet-
ing processes inside and external to the dielectric sur-
face, and that a detailed self-consistent model must be
built to accurately study this process
Our baseline code is INDUCT95 [12] This code is a
plasma discharge code that has been applied to plasma
reactors, flat panel plasma display discharges, streamer
discharges and other problems involving complex sys-
tems. It solves the standard time dependent fluid equa-
tions of the form
an,=
NC
Bt
-V .n,v, + c
R,,
,= I
(1)
and:
F=-V (niv, v,)+F--+,kq -$r,v,~(~)
I I
,=I
for the ions, where
Ri
is the source term for the
changes in the ion densities. And for the electrons
a1
c=
81
-v 1-, + 2 Re,
j=l
(3)
with the energy balance equation.
8f
e = -v Q - ere . E - P,“,,,,,i,
a
(4)
where r, is the electron flux and qneIarhc is the energy
loss rate due to inelastic collisions
Preliminary examples of results are shown in Figures 7
and 8 This simulation was performed with the updated
version of INDUCT95/HYBRID and uses a Monte-
Carlo treatment of electrons and a fluid treatment of
ions to follow the discharge development. Several ge-
ometry configurations and surface materials were in-
vestigated. The standard configuration studied had an
electrode separation of 1 mm with an applied voltage
of 50 kV/cm An insulator with a dielectric constant,
&= 15 was placed between the electrodes The dielectric
was considered to either have a flat surface (Figure 7)
or a step discontinuity (Figure 8) halfway between the
electrodes. Electrons were launched with random ini-
tial direction from the cathode triple point until they
either struck the insulator or the anode. Upon striking
the insulator, electrons generated secondary electrons.
The secondary emission yield is a strong function of
incident electron energy Secondary yield profiles were
fabrication means was pursued To simplify fabrica-
tion we attempted an ultrasonic machining process.
Although it was possible to fabricate the part in a sin-
gle operation, the surface was left slightly rougher.
Comparison of the breakdown characteristics of these
samples showed significantly more scatter and on av-
erage a slightly decreased breakdown threshold of ap-
proximately 25% (Fig 3)
i
00
0 50 100 150 200 250
Electric Field, kVlcm
Fig 1 Pulsed surface breakdown reliability ground
fnsed silica high gradient insulator
10' I@
I@
Pulse Wiih ns
Fig 2. Pulsed surface breakdown electric field as a
function of pulse width for single substrate,
straight wall insulators
The structures were also subjected to increased pres-
sures to determine susceptibility to breakdown (Fig. 4).
In these data, using the previously described procedure,
a fixed reliability was established at the various pres-
sures All data was then normalized to a mean break-
down electric field. Susceptibility to breakdown stays
relatively constant up until about the 10m3 T range, at
which point, the field at which breakdown occurs de-
creases rapidly. Also shown are data from previous
work by Smith [7]. From this comparison, it appears
that these new structures show a lower breakdown
electric field threshold than these previous data.
D 40 80
120 180
200
Electric Field, kV/cm
Fig 3 Pulsed surface breakdown reliability rough-
ened fused silica high gradient insulator
Fig. 4 Effect of gas pressure on the performance of
high gradient insulators
Fig. 5 The effect of insulator length scaling
(conventional insulators, [3]) compared to
individual layer thickness (multilayer insula-
tors) on the surface breakdown electric field.
Our data for varied layer thickness and materials were
compared to classical insulator length scaling (Figure
5). The trend is generally equivalent when the scale
length is defined as the individual lamination thickness
![Fig. 5 The effect of insulator length scaling (conventional insulators, [3]) compared to individual layer thickness (multilayer insulators) on the surface breakdown electric field.](/figures/fig-5-the-effect-of-insulator-length-scaling-conventional-9cu4t0vm.webp)