Calculation of macroparticle flow in filtered vacuum ARC plasma systems
Model of macroparticle behavior inside plasma guiding channels of vacuum arc systems and program for calculation of macroparticle flow in plasma filters (separators) used in vacuum arc film deposition are described. Several magnetic filters have been modeled to compare calculated and experimental da...
Збережено в:
| Дата: | 2012 |
|---|---|
| Автор: | |
| Формат: | Стаття |
| Мова: | English |
| Опубліковано: |
Національний науковий центр «Харківський фізико-технічний інститут» НАН України
2012
|
| Назва видання: | Вопросы атомной науки и техники |
| Теми: | |
| Онлайн доступ: | http://dspace.nbuv.gov.ua/handle/123456789/108630 |
| Теги: |
Додати тег
Немає тегів, Будьте першим, хто поставить тег для цього запису!
|
| Назва журналу: | Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| Цитувати: | Calculation of macroparticle flow in filtered vacuum ARC plasma systems / D.S. Aksyonov // Вопросы атомной науки и техники. — 2012. — № 2. — С. 108-113. — Бібліогр.: 4 назв. — англ. |
Репозитарії
Digital Library of Periodicals of National Academy of Sciences of Ukraine| id |
irk-123456789-108630 |
|---|---|
| record_format |
dspace |
| spelling |
irk-123456789-1086302016-11-13T03:02:19Z Calculation of macroparticle flow in filtered vacuum ARC plasma systems Aksyonov, D.S. Физика радиационных и ионно-плазменных технологий Model of macroparticle behavior inside plasma guiding channels of vacuum arc systems and program for calculation of macroparticle flow in plasma filters (separators) used in vacuum arc film deposition are described. Several magnetic filters have been modeled to compare calculated and experimental data. The capability of the program to optimize filters is demonstrated on example of straight filter. The high compliance degree of calculated and experimental data is shown. Представлены модель, алгоритм работы и возможности программы для расчёта и оптимизации плазменных фильтров (сепараторов), используемых в вакуумно-дуговом осаждении покрытий. Промоделировано несколько магнитных фильтров для сопоставления расчётных и экспериментальных данных. Продемонстрирована возможность оптимизации одного из фильтров с помощью разработанной программы. Показана высокая степень соответствия расчётных данных экспериментальным. Наведено модель, алгоритм роботи та можливості програми для розрахунку та оптимізації плазмових фільтрів (сепараторів), які використовуються у вакуумно-дуговому осадженні покриттів. Промодельовані декілька магнітних фільтрів для порівняння розрахункових та експериментальних даних. Продемонстрована можливість оптимізації одного з фільтрів за допомогою розробленої програми. Показано високий ступінь відповідності розрахункових даних експериментальним. 2012 Article Calculation of macroparticle flow in filtered vacuum ARC plasma systems / D.S. Aksyonov // Вопросы атомной науки и техники. — 2012. — № 2. — С. 108-113. — Бібліогр.: 4 назв. — англ. 1562-6016 http://dspace.nbuv.gov.ua/handle/123456789/108630 621.793 en Вопросы атомной науки и техники Національний науковий центр «Харківський фізико-технічний інститут» НАН України |
| institution |
Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| collection |
DSpace DC |
| language |
English |
| topic |
Физика радиационных и ионно-плазменных технологий Физика радиационных и ионно-плазменных технологий |
| spellingShingle |
Физика радиационных и ионно-плазменных технологий Физика радиационных и ионно-плазменных технологий Aksyonov, D.S. Calculation of macroparticle flow in filtered vacuum ARC plasma systems Вопросы атомной науки и техники |
| description |
Model of macroparticle behavior inside plasma guiding channels of vacuum arc systems and program for calculation of macroparticle flow in plasma filters (separators) used in vacuum arc film deposition are described. Several magnetic filters have been modeled to compare calculated and experimental data. The capability of the program to optimize filters is demonstrated on example of straight filter. The high compliance degree of calculated and experimental data is shown. |
| format |
Article |
| author |
Aksyonov, D.S. |
| author_facet |
Aksyonov, D.S. |
| author_sort |
Aksyonov, D.S. |
| title |
Calculation of macroparticle flow in filtered vacuum ARC plasma systems |
| title_short |
Calculation of macroparticle flow in filtered vacuum ARC plasma systems |
| title_full |
Calculation of macroparticle flow in filtered vacuum ARC plasma systems |
| title_fullStr |
Calculation of macroparticle flow in filtered vacuum ARC plasma systems |
| title_full_unstemmed |
Calculation of macroparticle flow in filtered vacuum ARC plasma systems |
| title_sort |
calculation of macroparticle flow in filtered vacuum arc plasma systems |
| publisher |
Національний науковий центр «Харківський фізико-технічний інститут» НАН України |
| publishDate |
2012 |
| topic_facet |
Физика радиационных и ионно-плазменных технологий |
| url |
http://dspace.nbuv.gov.ua/handle/123456789/108630 |
| citation_txt |
Calculation of macroparticle flow in filtered vacuum ARC plasma systems / D.S. Aksyonov // Вопросы атомной науки и техники. — 2012. — № 2. — С. 108-113. — Бібліогр.: 4 назв. — англ. |
| series |
Вопросы атомной науки и техники |
| work_keys_str_mv |
AT aksyonovds calculationofmacroparticleflowinfilteredvacuumarcplasmasystems |
| first_indexed |
2025-07-07T21:50:52Z |
| last_indexed |
2025-07-07T21:50:52Z |
| _version_ |
1837026564105043968 |
| fulltext |
108 ISSN 1562-6016. ВАНТ. 2012. №2(78)
UDC 621.793
CALCULATION OF MACROPARTICLE FLOW IN FILTERED VACUUM
ARC PLASMA SYSTEMS
D.S. Aksyonov
National Science Center "Kharkov Institute of Physics and Technology",
Kharkov, Ukraine
E-mail: dsaksyonov@gmail.com
Model of macroparticle behavior inside plasma guiding channels of vacuum arc systems and program for
calculation of macroparticle flow in plasma filters (separators) used in vacuum arc film deposition are described.
Several magnetic filters have been modeled to compare calculated and experimental data. The capability of the
program to optimize filters is demonstrated on example of straight filter. The high compliance degree of calculated
and experimental data is shown.
INTRODUCTION
Vacuum arc deposition is one of the most promising
and commonly used techniques to produce coatings of
different purposes. During several decades the
technique plays an important role in such industry
branches as engineering, instrument making,
electronics, metallurgy, tools production and decorative
coatings. However, vacuum arc method has a significant
drawback – the presence of macroparticles among the
products of cathode erosion. Macroparticles are solid or
liquid fragments of cathode material. Their size varies
from tenths to several tens of micrometers. After contact
with a substrate, macroparticle may stick to it, bounce
from it or fall off from it some time after sticking, which
in any case leads to the formation of defective
coating [1,2]. Therefore the use of vacuum arc method
in some cases may be unacceptable or impractical.
There are some well-known ways to suppress the
emission of macroparticle phase or to decrease its
amount in deposited films. The most common approach
in practice is a spatial separation of plasma and
macroparticle trajectories. Magnetoelectric plasma
filters (separators) are successfully used for this
purpose. In such filters plasma streams are guided by
magnetic field along curved path from the area of
plasma formation (near-cathode space), to the coating
deposition area – a workpiece (substrate). At the same
time, there must not be a direct line-of-sight between the
cathode and the substrate. Macroparticles have low
negative (or neutral) charge [1,2] and their mass is much
greater then ion mass. As a result, macroparticles are
almost not influenced by magnetic and electric fields,
used for plasma transport. Therefore their movement
paths are representing straight lines. Such spatial
separation of the trajectories allows intercepting the
flow of macroparticles by plasma duct walls, making
the substrate unreachable for macroparticles.
As was shown by practical use of curvilinear filters,
an absence of a direct line-of-sight between a cathode
and a substrate is necessary but not sufficient condition
for removing macroparticles from the plasma. There are
some possible scenarios when macroparticle collides
with plasma duct walls. The macroparticle may scatter
on several smaller parts. Solid macroparticles may also
bounce from duct walls, which is especially relevant
when graphite cathode is used for films deposition. A
third option is also possible – macroparticle may stick to
the wall. In first two cases, nothing prevents
macroparticle from hitting the substrate after collision
with filter walls. Therefore filters are equipped with
baffles (fins), which are acting as additional obstacles
for macroparticles and as source of additional collisions
for them. During collision with baffles and filter walls
macroparticle gradually loses its kinetic energy. And
after a series of such collisions it eventually sticks to an
obstacle.
Thus, the design of an effective filtering system is a
challenging task. In addition to providing the "no direct
line-of-sight" principle, a designer is forced to rely
mainly on his intuition because it is scarcely possible to
take into account all cases of macroparticle rebounds
(ricochet).
The process of filter designing can be significantly
simplified by using specially developed algorithms for
macroparticle trajectories calculation. Computer based
calculations were made earlier by other authors [1,3].
However, neither program description nor list of tasks it
can handle were not provided or referred to. Moreover,
the number of calculated filtering systems was very
limited. It can be seen from analysis of given results,
that used algorithm is able to calculate very small
amount of macroparticles (about 180) emitted only from
a single point. Any information about geometry
construction and calculation result output features the
program owns were also not been given. So the task of
creating a necessary tool for computer simulation and
analysis of filtering features of plasma separators still
remains actual.
For the purpose of solving this problem
Macroparticle Tracer (MPT) program have been
developed. It is able to assist designer to make a
qualitative evaluation of filtering abilities of various
systems, to significantly simplify the task of designing
of such systems and also to define the basic concepts of
baffle systems construction. The program is capable to
perform a real-time simulation, uses own user-friendly
interface and has tools for geometry construction,
visualization of problem and results. It also provides a
flexible adjustment of problem input and result output.
ISSN 1562-6016. ВАНТ. 2012. №2(78) 109
MACROPARTICLE BEHAVIOR MODEL
AND CALCULATION PROBLEM
Simulation of macroparticles rebounds from the
walls of plasma duct and baffles (hereinafter "system
borders" or simply "borders") a number of assumptions
have been made. Macroparticles (hereinafter "particles")
are solid spheres with radius tending to zero and
represented in calculation as dots. Collision of particle
with borders has partially elastic nature. After certain
number of collisions the energy of the particle becomes
insufficient for the subsequent rebound – such particle is
considered to be absorbed by filter. Borders are
completely smooth (no roughness) and represented as
broken lines. Segments of particle trajectory before and
after rebound are forming equal angles with respect to
the collided border (angle of incidence is equal to
reflection angle). Particles do not collide with each
other. Calculations are made in two-dimensional
approximation, i.e. in one plane. This plane should be
symmetry plane of the system. Particle emission from
the working surface of the cathode is equiprobable in
any direction. Collisions of particles with borders do not
lead to the particle fragmentation (scattering). An
influence of magnetic and electric fields, gravity and
other forces are not taken into account – particles are
always flying along straight paths (between collisions).
Used model is very similar to previously used [1,3].
Formulation of particle trajectory calculation
(tracing) problem is reduced to the following stages.
Definition of system borders: an input of coordinates of
all borders (defined as line segments) involved in
calculation. Definition of emitter (cathode working
surface) and emission direction. Setting tracing
parameters, such as: maximum (allowable) number of
particle impacts with borders, when the particle is
considered to be absorbed by filter due to complete lose
of particle energy; the number of emission points and
the way their coordinates are being calculated (for
example, fixed step between points or their total
number); the number of particles flying out from every
emission point (by step or total number); and also
emission base (starting) angle and its relation (to global
coordinate system or to local – with respect to emitter
orientation). It is also possible to define absorbing
borders, i.e. particle collided this borders will be
considered as absorbed by filter regardless of how many
collisions this particle has experienced before. Though
definition of such borders is not necessary for tracing
task setting.
Additional postprocessing tools are added to make
simulation results easier to analyze and to lower the
load on used computational system. For example, it is
possible to display trajectories of only those particles
that collided with specified border (or group of borders)
or these trajectories can be displayed in other color.
BASIC ALGORITHM
After initial coordinates and directions for all
involved in simulation particles have been specified, the
program enters calculation mode with the following
stages.
1. Calculation of equations coefficients of lines
from endpoints coordinates of line segments. These
lines describe borders.
2. Determination of equation coefficients of line,
which describe initial segment of particle trajectory.
3. Calculation of collision points coordinates for
every border.
4. Discarding collision points which do not belong
to line segments of corresponding borders.
5. Determination of the collision point which is
closest to emission point (or previous collision point).
6. Calculation of equation coefficients of line,
describing trajectory segment of particle after rebound:
1 1mA + = ,
( )
( )
2 2
1 2 2
2
2
m n n n m n
m
m n n n n m
B B A A A B
B
A A B A B B+
− +
=
− +
,
( ) ( )
( )
2 2
1 2 2
2
2
n n m n m m n n
m
m n n n n m
C A A B B C A B
C
A A B A B B+
+ − +
=
− +
,
where Am, Bm, Cm − equation coefficients of line of
particle trajectory before rebound; An, Bn, Cn − equation
coefficients of line, which define border.
7. Determination of particle movement direction,
based on the definition of the half-planes (regarding the
border) calculated collision point and emission point (or
previous collision point) belongs to.
8. Repeat of steps 3 − 7 for each segment of the
trajectory line until the particle is absorbed by filter.
9. Repeat of steps 2 − 8 for each emitted particle.
10. Display the simulation results in numerical and
graphical forms with required postprocessing.
It should be noted, that line segments representing
borders and trajectories are treated as lines (through the
coefficients of line equation in general form) during
calculation process, while coordinates of their endpoints
are used instead for storage and display.
During the tracing process some additional
calculation may be needed to solve exceptional cases.
Description of the algorithms used for these additional
calculations is unreasonable in this work; therefore only
cases themselves are listed below. The most common
among them are: parallelism or coincidence of border
and trajectory lines; emitter enclosure by borders (for
example, in case of cup-shaped cathode); collision of
particle simultaneously with multiple borders in a single
point; particle fly out from simulated system if the last
one is not closed.
PROGRAM CAPABILITIES
A program for macroparticle tracing, as a tool for
filtering systems development, should provide rather
flexible way to define a calculation problem. Some
cases of simulation may demand exceptional functional,
which is not needed in most other cases of filtering
system designing. Therefore, when developing MPT, an
attempt was made to anticipate the needs of filter
designer and thus to create tools capable to satisfy these
needs.
One of the most important areas in particle tracing
task is the setting of emission. At the stage of emission
centers creation (the points on cathode working surface
macroparticles will "fly from" during simulation) it is
possible to set their total number or, for example,
110 ISSN 1562-6016. ВАНТ. 2012. №2(78)
specify their offset, i.e. distance between them. When
the emission centers have been set, an amount of
particles per emission center must be specified. The
amount can be defined the same ways as in case of
emission centers, i.e. by total number or by angular step.
Because it is not possible (expedient) to calculate every
possible direction of emitted particle, it is often handy to
have an opportunity to set up the orientation of the
particles bunch. This can be made by specifying base
(starting) emission angle. So every direction of particles
(in the bunch associated with emission center) will be
calculated with relation to this base angle. In addition,
base angle can be set in global coordinate system, or in
local – relative to emitter.
Allowable number of collisions (see above) with
system borders can be set to any number. The only
limitation is the performance of computer system used
for simulation. It is especially vital when comparing
different filters, which are well-designed and optimized.
Using small allowable number of impacts will give
barely noticeable difference between simulation results.
Described situation is also applicable to emission
settings (high or low). The higher the amount of emitted
particles, the more reliable results can be obtained.
Emission is also limited only by computer performance.
For detailed analysis of a separator it is often
important to know its filtering abilities at different
allowable number of collisions. For example, the same
filtering degree at different maximum allowable particle
impacts may indicate nonoptimality of the separator
construction (in particular – location and shape of used
baffles). Detailed examination may reveal "weak points"
of the filtering system, when minor modification to its
design leads to significant filtering abilities gain.
Mentioned above procedures are greatly simplified with
visualization of said dependencies by means of graphs.
Furthermore, graphs can provide additional data on
filters with the same final results, but having different
intermediate ones.
Week dependence of filtering degree from allowable
impacts number is typical for separators of open
architecture. In this case particle bouncing from vacuum
chamber walls is not considered. But the chamber
definitely has certain dimensions and construction, thus
its exclusion from simulation is rather questionable.
Such situation will be shown and discussed below.
One of the most useful features of the developed
program is to perform real-time calculation and results
display. Using this mode one can change separator
architecture and obtain results simultaneously. For
example, changing baffles orientation and see the result
changes during the rotation. This mode is also can be
used as "single-particle emission scanning". Both cases
allow designer to rapidly find a construction close to
optimal one.
MPT has its own visualization means of calculated
system geometry and calculation results. They
incorporate several types of viewport pan and scaling;
tuning such settings as color, lineweight, transparency
and linetypes (solid, dashed, etc.) for grids, borders,
trajectories, snaps and snap tracking lines.
Solving every tracing task starts from geometry
definition of the system simulated. In order to provide
ease of program use, speedup calculation task
formulation and provide independence from third-party
software, MPT features its own vector graphics editor.
The editor itself has standard for such editors tools, the
main ones are: drawing/modifying, rotating, copying,
moving, mirroring, scaling, array copying, lengthening,
trimming and extending of lines (borders) or group of
lines, and also measurement of angles and distances.
Each of listed tools has additional options, which can
modify behavior of the tools. For convenient and easy
construction of complex geometric objects the program
has some support functions: snap to various object parts
and tracking of the snaps. Viewport of the program
supports two independent tunable grids to simplify
navigation and drawing.
Results data can be obtained for each border both in
absolute values – as number of collided particles, and in
relative values – as a filtration degree for the border.
Filtration degree is a number of collided particles
normalized to emission and expressed as a percentage:
1 100Unfiltered particlesFiltration degree
Emitted particles
⎛ ⎞
= − ⋅⎜ ⎟
⎝ ⎠
.
Such approach allows evaluation of practical
significance of each baffle or plasma duct wall and
economic feasibility of their manufacturing. If
simulation revealed that one of the baffles have very
little collisions count (or have none at all), then
manufacturing of this baffle and installation of it in real
filter are meaningless, therefore it only increases final
cost and complexity of the filter.
Clarity of the results is sharply decreased in case of
filter simulation with high emission (lots of particles)
setting, because the most part of the calculated system
will be filled with a single color, corresponding to
trajectory display setting. That is, particle trajectories
become indistinguishable and moreover, displaying a
large number of lines significantly loads graphical core
of the computer used. To resolve this issue MPT
provides the following measures. It is possible to
display only trajectories which collided with the
substrate. In general, it can be any of the borders or a
group of them. But if there are enough computer
resources, mentioned trajectories can be displayed in
another color. This method unambiguously emphasizes
the trajectories of most interest without losing the
integrity of the picture. For the detailed separator
analyzing, it is often needed to view trajectories of
specific particles. For example, in the vicinity of some
baffles, in order to determine whether designer's idea
worked or not. MPT provides such ability: any
individual trajectory can be highlighted.
Display of calculated dependencies graphs is
realized in a separate from system geometry and
trajectories space. Functions of calculation and display
of graphs are incorporating sizing, scaling, positioning,
curves and axes visual settings adjusting, etc. A single
graph, as a simulation result, includes curves for all
borders of the system. The curves are tuned
independently from each other. The same applies to
graphs. But graphs also can inherit visual settings from
other graphs and be created using adjustable templates.
ISSN 1562-6016. ВАНТ. 2012. №2(78) 111
CALCULATION RESULTS
Using the developed program some previously
known filtering systems were simulated. Data on the
filtering abilities of the systems were obtained
experimentally by corresponding authors. These
systems include classical curvilinear filter in a quarter
torus form [3], straight filter with a disc in front of
cathode [3] and open architecture filter with 60 degrees
bend angle [4] in several variations described below.
Optimization of baffles of straight filter was also
performed, as an example.
Toroidal and straight filters simulation results are
presented in Fig. 1,a,b. Particle emission was set in such
way as to clearly demonstrate the difference of filtering
abilities of the separators. As it can be seen from the
figure, toroidal filter shows much better results as
compared to straight one. Filtering degree of toroidal
filter is approximately 12 times higher. This difference
is qualitatively agrees with experimentally obtained
data [3] (Table. 1).
Optimization of straight filter was made by means of
insignificant variation of its baffles angles. This allows
do decrease the amount of unfiltered macroparticles
approximately in 65 times. Having such baffle design
the filter should be capable to provide 5.5 times higher
filtering degree than toroidal one (Fig. 1c and Table. 1)
if macroparticles are not capable to bounce more than
11 times.
For more complex analysis of toroidal and straight
filters emission was increased by 100 times. Calculation
was performed for different macroparticle energy
values, represented as number of allowable impacts (see
above). The results are shown in Fig. 2. Both filters
demonstrate good filtering of particles that bounce no
more than 4 times. Macroparticles bouncing 6 times still
nearly do not leave toroidal separator.
a
b
c
Fig. 1. Schematic representation of macroparticle tracing results for classical toroidal filter [3] (a), straight
filter [3] (b) and optimized straight filter (c). Number of emitted macroparticles is about 9900, distance between
emission centers is 0.6 mm, angular step – 1.8º and particles don't bounce from substrate.
Green – filtered particles, red – unfiltered
Calculation and experimental results
Chamber
present
Emitted
particles
Unfiltered
particles
Filtration
degree, %
Macroparticles
density [3,4], cm−2
Measured
ratio
Calculated
ratio
Torus [3] − 998706 4097 99.59 1·106 1 1
Straight [3] − 999000 48827 95.11 16·106 16 11.92
Straight (opt.) − 999000 752 99.92 − − 0.18
No 999000 110724 88.92 4.88 Open
(rectangles) [4] Yes 999000 222714 77.71
4.3·106 10.75
2.62
No 999000 22679 97.73 1 Open
(triangles) [4] Yes 999000 85124 91.48
0.4·106 1
1
No 999000 37932 96.20 1.67 Open (squares)
Yes 999000 198967 80.08
− −
2.34
No 999000 21749 97.82 0.96 Open (rotated
squares) Yes 999000 90852 90.91
− −
1.07
Allowable number of impacts is 10, distance between emission points is 20 μm for open type filters and 53 μm for
rest, angular step is 0.18 degrees. "Chamber present" means that the spiral duct and the substrate of the system were
enclosed in chamber with dimensions 800×800 mm. Measured and calculated ratios are for comparison of filters inside
of their own group. Closed filters are compared to the toroidal filter, open filters – to the filter with triangular cross-
section of solenoid turns
112 ISSN 1562-6016. ВАНТ. 2012. №2(78)
Meanwhile initial (not optimized) straight filter loses
filtering abilities to a considerable extent. As the energy
of emitted macroparticles continues to grow, filtering
degree of both filters continues to fall. Such a trend is
more pronounced in straight filter what makes toroidal
one more preferable in the areas where maximum
plasma filtering degree is needed.
The optimization of straight filter baffles makes it
able to compete with toroidal filter on condition that
particles can bounce from separator wall less than
12 times. An amount of particles, unfiltered by both
optimized straight filter and its unoptimized version,
undergoes increase for macroparticles with higher
energies. It should be stated, that in practice actual
number of particle rebounds is 9 or less [1, 3]. From this
point of view, straight filter with modified baffle system
has the best filtering performance among three
examined above separators. It is obvious, that baffles
optimization of the toroidal separator will increase its
filtering abilities, and they will be greater than the
straight one obtained due to optimization.
Architecture of the open type filters with 60 degrees
bending is reconstructed from the data specified in
work [4]. In order to understand determinant factors of
their design (according to filtering) – is it a cross-
section of plasma guiding solenoid turns or their
orientation, two more design modifications were
simulated. Cross-section of the solenoid was chosen to
be a square with 10 mm edge size. First modification
has these squares directed to the filter curvature center
by one of the edges, second modification – by one of the
corners. Hereinafter, second modification will be
referenced as rotated square filter.
In addition, all considered open type filters were
calculated twice. The first calculation was made to
determine the value of the conception used. Therefore
macroparticles left the spiral plasma duct was
considered as absorbed by filter. In practice, the filters
must have some kind of vacuum chamber, which they
are encapsulated in. Thus the second calculation takes
into account macroparticles rebounds from chamber
walls. It should be noted, that the simulation result are
strongly dependent from the accuracy of separator
design reproduction. Authors of the work [4] have not
provided any data about the filter enclosure. It means
that more significant difference between calculated and
experimental data should be expected. It was assumed
that vacuum chamber is 800×800 mm in size, and the
plasma source used was reconstructed from given
schematic figure.
Fig. 3 presents calculation results for filters of open
architecture with rectangular and triangular section of
spiral plasma duct turns. Numerical data is shown in
Table. As depicted in Fig. 3, the number of unfiltered
macroparticles reaches its maximum at 2 − 3 rebounds
value if vacuum chamber is not present (ignored).
Thereafter number of unfiltered particles remains nearly
the same. It is due to impossibility of macroparticles to
return back inside filter space by means of bouncing
from chamber walls. Being inside closed system, most
macroparticles will finally get to the substrate after a
certain number of rebounds. That's the reason why the
amount of particles monotonically increases if the
chamber is present. The last case is more likely from the
practical point of view.
Quality of plasma filtering provided by the discussed
in this work separators both calculated and measured is
given in Table. The separators can be rated by their
particle filtering degree in a following order. Optimized
straight separator offers the best filtration. After it, in
descending order are toroidal, initial straight, and open
type ones. Inside the group of open architecture filters
sequence is as follows: with triangular, square and
rectangular cross-section of solenoid turns. Measured
ratio of unfiltered particles for toroidal and straight
filters equals 16, while calculated using MPT ratio is
about 12. Simulation data on open type filters allows
one to unambiguously assert that the choice of
triangular section of plasma duct solenoid turns is
optimal. This fact is also confirmed by authors of
experimental studies of the filters [4]. As it has been
stated above, due to inaccuracies of open filters design
reconstruction, a substantial difference between
calculated and experimentally obtained data is observed.
The use of rectangular and square cross-sections in
open type filter design provides nearly the same filtering
quality. The same situation is observed if calculation
data on triangular and rotated square filters are
compared. It is evident that the key role in such filters
plays the orientation of solenoid turns, not their cross-
section alone.
Fig. 2. Dependence of the number of macroparticles at
filter output on the allowable number of impacts.
Distance between emission centers is 53 μm, angular
step – 0.18º and particles don't bounce from substrate
Fig. 3. Dependence of the number of macroparticles at
filter output on the allowable number of impacts.
Distance between emission centers is 22 μm, angular
step – 0.18º and particles don't bounce from substrate
ISSN 1562-6016. ВАНТ. 2012. №2(78) 113
CONCLUSION
This work showed that the developed MPT program
is capable to do a qualitative comparison of filtering
abilities of different separators. It was demonstrated,
that using this program one can successfully create a
new filter designs and also optimize the known ones.
The results of the performed simulations are
substantially correlating with the experimentally
obtained data.
REFERENCES
1. I.I. Aksenov. A vacuum arc in erosion plasma
sources. Kharkov: NSC KIPT, 2005, 212 p. (in
Russian).
2. A. Anders. Cathodic Arcs: From Fractal Spots
to Energetic Condensation. New York, Springer, 2008,
542 p.
3. I.I. Aksenov, V.E. Strel’nitskij, V.V. Vasilyev,
D.Yu. Zaleskij. Efficiency of magnetic plasma filters
// Surf. Coat. Technol. 2003, v. 163−164, p. 118−127.
4. .Y.H. Liu, J.L. Zhang, D.P. Liu, T.C. Ma,
G. Benstetter. A triangular section magnetic solenoid
filter for removal of macro- and nano-particles from
pulsed graphite cathodic vacuum arc plasmas // Surf.
Coat. Technol. 2005, v. 200, p. 2243−2248.
Статья поступила в редакцию 10.01.2012 г.
РАСЧЁТ ПОТОКА МАКРОЧАСТИЦ В СИСТЕМАХ ФИЛЬТРОВАННОЙ
ВАКУУМНО-ДУГОВОЙ ПЛАЗМЫ
Д.С. Аксёнов
Представлены модель, алгоритм работы и возможности программы для расчёта и оптимизации
плазменных фильтров (сепараторов), используемых в вакуумно-дуговом осаждении покрытий.
Промоделировано несколько магнитных фильтров для сопоставления расчётных и экспериментальных
данных. Продемонстрирована возможность оптимизации одного из фильтров с помощью разработанной
программы. Показана высокая степень соответствия расчётных данных экспериментальным.
РОЗРАХУНОК ПОТОКУ МАКРОЧАСТИНОК У СИСТЕМАХ ФІЛЬТРОВАНОЇ
ВАКУУМНО-ДУГОВОЇ ПЛАЗМИ
Д.С. Аксьонов
Наведено модель, алгоритм роботи та можливості програми для розрахунку та оптимізації плазмових
фільтрів (сепараторів), які використовуються у вакуумно-дуговому осадженні покриттів. Промодельовані
декілька магнітних фільтрів для порівняння розрахункових та експериментальних даних. Продемонстрована
можливість оптимізації одного з фільтрів за допомогою розробленої програми. Показано високий ступінь
відповідності розрахункових даних експериментальним.
|