Hello, and thank you for your interest in my research!
I am currently investigating electrical breakdown in high-voltage,
dielectric-loaded systems. In general, I am interested in
the physics of plasmas and numerical simulation of such
systems.
If you have any particular questions, I can be reached at:
aldan at berkeley dot edu
Motivation:
Practically every high-voltage system is susceptible to insulator
breakdown in some form. We are interested in breakdown as an undesirable
effect, wherein many industries encounter breakdown as a limiting factor,
usually causing some sort of damage. For example:
| Application | Description | References |
| Microelectronics | This is a high-field case, where arcs can form between electrical contacts, damaging system components and possibly requiring refabrication of the whole system. | (Sandia) |
| HPM sources | High-Power Microwave (HPM) systems are loaded with both dielectric insulation and output windows. Surface effects include burning of the surface, which can introduce foreign species. | (TTU) |
Goal:
Characterize and order contributions to electrical breakdown and
provide the means to ultimately control breakdown.
Approach:
Starting from work done at the University of Michigan by Jordan,
et al., simulate multipactor breakdown results with
Particle-In-Cell (PIC) simulations and extend into space- and
surface-charge regimes. Include models for interactions with
gaseous species.
Current Focus:
Multipactor in dielectric-loaded, DC systems.
Control study with true-secondaries only, and extended study with scattered/reflected primaries.
Inclusion of proper triple-point emission model.
Extended test of angular dependence on breakdown with newer models.
Further Details:
The target system is an idealized model of a cylindrical capacitor
between two electrodes with azimuthal symmetry. The code I am
using is two-dimensional, which is sufficient given that the 2D-plane
is the shortest path to breakdown. The simulated system is semi-infinite
in a 2D plane such that the length of the electrodes is much longer
than the gap width. Additionally, a number of constraints are imposed
on the numerical representation of the system, paying particular
attention to the surface of the dielectric, where electron impacts occur.
An improved geometric model was developed by Taverniers in 2009 that
rotated the simulation system so that the dielectric surface sits on the
numerical grid, eliminating impact errors but introducing errors at the
electrode boundaries. Taverniers' model was slightly modified by Aldan
to include finer grid sizing in one direction which allows for lower
computational costs at high resolution while maintaining the same errors
as defined by Taverniers. I am also using an implementation of Vaughan's
secondary-emission model that more closely resembles the behavior of
secondary-emission on dielectrics fitted to empirical data from such
sources as Baker with a simple non-linear fit.
For simulations, I am using a PIC suite developed by Berkeley, called
XOOPIC,
an object-oriented implementation of PIC making modification tractable.
Diagnostics are based on the familiar Tcl/Tk libraries. Futher details
can be found at the
PTSG Software
website.
References:
| 1. | N.M. Jordan, Y.Y. Lau, D.M. French, R.M. Gilgenbach, and P.Pengavich, "Electric field and electron orbits near a triple point", Jour. Appl. Phys., Vol. 102, 033301, 2007. |
| 2. | J.R.M. Vaughan, "A new formula for secondary emission yield", IEEE Trans. Elec. Dev., Vol. 36, No. 9, pp. 1963-1967. |
| 3. | R. Vaughan, "Secondary emission formulas", IEEE Trans. Elec. Dev., Vol. 40, No. 4, pp. 830, 1993. |
| 4. | M.C. Baker, "Secondary electron emission from dielectrics", Master's Thesis, Texas Tech University. |
Acknowledgements:
This work is supported by an AFOSR grant on the Basic Physics of Distributed Plasma Discharges.
We extend our gratitude to the following individuals for discussion and input:
Dr. Ives, Dr. Read, Mr. Bui, and Mr. Collins of Calabazas Creek Research, Inc.
Mr. Angelo Wong of UC Berkeley
Dr. Lau of the University of Michigan
Dr. Booske of the University of Wisconsin
Dr. Leopold of Rafael Laboratories
Presentations:
|
GEC 2011
(PowerPoint) |
(As of October 2011). The template for a new triple-point current source has been implemented and is currently being tested. The model is based on Schachter's theory for the electric field inducing a Fowler-Nordheim-like emission in the high-field region of the triple-point. This is a particularly important development when for the extension of this study into multiple electrodes, which means the system will have multiple potential seed points, i.e. triple points. The template for finite-thickness multiple electrodes has also been developed while minimizing grid errors. Previous case studies included only infinitesimally-thick electrodes. An important difference between the newer seed current models and previous seed current models used in this study is that the source will be distributed. It is observed that distributed sources significantly alter the system behavior, largely due to a distributed dielectric-surface charge resulting from secondary emission and/or electron absorption. The characteristics of the seed current affect both the time to breakdown and the breakdown voltage at a specified dielectric angle. Future work includes improving the triple-point model, a more thorough study of the effect of the seed current on breakdown characteristics, and the inclusion of gas from mTorr to atmospheric pressures. RF will also be considered. |
|
ICOPS 2011
(PDF) |
(As of June 2011). A bug in the legacy code of XOOPIC was found to be incorrectly initializing transverse velocities from the dielectric surface. After correction, positive charging is observed on the dielectric surface. At dielectric angles from 10° and below, multiplicative breakdown is observed within a few hundred picoseconds. On the approach to steady-state, secondary lifetimes are shortened, leading to an overall shift in secondary energies to the first-crossover energy. Output anode current saturates at the imposed input current. Above 17°, multiplicative breakdown is not observed due to space charge effects, although a current does develop in steady state that we are characterizing as dark current for the moment. Scattered/reflected secondaries (SRP) were studied briefly, showing a slight increase in breakdown strength. --- There are videos in this presentation, but as they are large, I've opted to remove them from the presentation. Videos are located here. Overall behavior in videos: The upper right window shows the primary phase space and the lower right window shows secondary phase space. Both are highly zoomed in to a window of about 2mm(x)×72μm(y); don't be fooled by the aspect ratio Additionally, particles are traveling from left to right. The lower left window shows the charge on the dielectric surface; particles are traveling from right to left for ease of viewing the surface charge. 1000V applied, 1A imposed current, 1ps timesteps. In both the 10° and 17° cases, initial behavior is similar with primaries emitted from the cathode (left-hand anode). Secondaries are generated as primaries are impacting with energies between first and second crossover. Positive charging is seen from the intial impact site. After a few picoseconds, a negatively-charged region right beside the initial impact site is seen in both cases resulting from the slow secondaries impacting quickly with low energy. Higher-energy secondaries get away in both cases, but in the 10° case, sufficient energy is imparted to these secondaries that further positive charging can occur. At 17° the electric field angle is sufficiently normal to the dielectric surface that secondary lifetimes as well as impact energies are reduced, leading to overall negative surface charging. |
|
Qualifying Presentation
(PDF) |
(As of April 2011) Update: As delineated in the ICOPS report, a bug in legacy code was setting tranverse velocities to zero. Discussion below is still valid in low-energy regimes, and the "breakdown current" from the discussion below is something we are calling "dark current" for the moment. This may be characterized differently once we understand the influence of the seed current, as we are setting an imposed current akin to a thermionic emitter, which allows for viable simulation with high statistics but wouldn't necessarily be seen in practice. --- Recent progress in the form of my Qualifying Exam presentation. Brief outline on dielectric charging as a necessary condition for a cascade. Negative surface charging shown to be predominant in simulation. Resolution of space-charge cloud is an issue in simulation, and fairly challenging to overcome. A simple reduction factor in y is used to ensure particle lifetime is resolved. Higher resolution modifies the details of the surface-charge geometry, although bulk properties are mostly unaffected, interestingly enough. In particular, time to initial breakdown current is roughly the same, and despite some variations, even the time to saturation is of the similar magnitude. Looking at true secondaries only. |
|
APS-DPP 2010 Poster
(PDF) |
(As of November 2010) Covers general results using the original numerical model developed by Taverniers. Things to note in the simulations are that only true-secondaries are studied (i.e. some physics are missing) and that the space charge cloud is under-resolved. General behavior suggests negative charging in most cases. Breakdown voltage vs. angle shown for true secondaries only. Preliminary representation of the breakdown voltage vs. angle shown. Breakdown current at the anode observed through dielectric angles from 0° through 40° using a Teflon insulation. This is in contrast to theoretical results from Jordan, which suggests a maximum angle of 20° for the configuration studied. Higher resolutions will be implemented to ensure particle lifetime is properly calculated; however, this is computationally expensive. |
Under Construction....