Skip to main content

Full text of "An investigation of unipolar arcing at atmospheric pressure in Aluminum 2024 and aluminum coated glass slides."

See other formats


6002 



NAVAL POSTGRADUATE SCHOOL 

Monterey, California 




THESIS 



AN INVESTIGATION OF UNIPOLAR ARCING AT 
ATMOSPHERIC PRESSURE IN ALUMINUM 2024 
AND ALUMINUM COATED GLASS SLIDES 

by 

Steven Wayne Woodson 
June 1987 

Thesis Advisor: F. R. Schwirzke 

Approved for public release; distribution is unlimited. 



T 233796 




security Classification OP 'This "past 



REPORT DOCUMENTATION PAGE 



la REPORT SECURITY CLASSif 'CATION 

UNCLASSIFIED 



lb RESTRICTIVE MARKINGS 



2a SECURITY CLASSlflCATlON AUTHORITY 



2b DEClASSifiCATiON/ DOWNGRADING SCHEDULE 



3 DISTRIBUTION'/ AVAILABILITY Of REPORT 

Approved for public release; 
distribution is unlimited. 



J PERfORMiNG ORGANIZATION REPORT NUMBER(S) 



S MONITORING ORGANIZATION REPORT NUMBER(S) 



6a NAME Of PERfORMiNG ORGANIZATION 

Naval Postgraduate Schoo 



] 



6b Off ICE SYMBOL 
(If applicable) 

61 



7a NAME Of MONITORING ORGANIZATION 

Naval Postgraduate School 



6< AOORESS (Ofy, Sfefe, and /IP Code) 

Monterey, California 93943-5000 



7b ADDRESS (Cry State, sod HP Code) 

Monterey, California 93943-5000 



8a NAME Of fUNOiNG / SPONSORING 
ORGANIZATION 



8b OffiCE SYMBOL 
(if applicable) 



9 PROCUREMENT INSTRUMENT lOE N TiF iC A TiON NUMBER 



8c ADORE SS (Ctty, State, and /IP Code) 



10 SOURCE OP FUNDING NUMBERS 



PROGRAM 


PROJECT 


TASK 


WORK JNIT 


ELEMENT NO 


NO 


NO 


ACCESSION NO 



1 1 T;TlE (include Security Clarification) 

AN INVESTIGATION OF UNIPOLAR ARCING AT ATMOSPHERIC PRESSURE IN ALUMINUM 
2024 AND ALUMINUM COATED GLASS SLIDES 



Z PERSONA,. AUTHOR(S) 

Woodson, Steven W. 



3a r YPf Of REPORT 

Master's Thesis 



I 3b T'ME COVERED 
fROM TO 



14 DATE Of REPORT (Yt*r Month CUy) 

June 1987 



IS PAGE (OoN T 

53 



6 Supplementary notation 



” COSATI COOES 


c ElO 


GROUP 


Subgroup 















18 SUBJECT terms (Continue on reverie it necessary and identity by block number) 

Unipolar, Unipolar Arcing 



9 ABSTRACT (Continue on reverie it neceisary and identify by block number) 

An experimental investigation of unipolar arcing at atmospheric 
pressure was conducted using Aluminum 2024 and glass slides with a thin 
coating of pure aluminum. The plasma was produced on the surface using 
a neodymium- glass laser in the O-switched mode. It was found that the 
power density required for the onset of unipolar arcing was similar to 
that of samples irradiated in a vacuum, although the size and density 
of the resulting craters were significantly different. 



0 D S'R'3uTiON / availability Of ABSTRACT 
&] JNCL ASSif IE DAJNL'M iTE 0 □ SAME AS RPT Q DTiC USERS 



21 ABSTRACT SECURITY CLASSif ICA T I0N 

UNCLASSIFIED 



22a NAME Of RESPONSIBLE INDIVIDUAL 

F. R. Schwirzke 



22b TELEPHONE (irxlude ArttCoO*) 

646-2635 



22 c OffiLE SYMBOL 

61Sw 



DO FORM 1473.84 mar 



83 APR edition may be used until exhausted 
All other editions ate obsolete 



SECURITY CLASSIFICATION OP This PAGE 



UNCLASSIFIED 



Approved for public release; distribution is unlimited. 



An Investigation of Unipolar Arcing at 
Atmospheric Pressure in Aluminum 2024 
and Aluminum Coated Glass Slides 



by 



Steven Wayne Woodson 
Lieutenant, United States Navy 
B.O.E., University of Mississippi, 1979 



Submitted in partial fulfillment of the 
requirements for the degree of 



MASTER OF SCIENCE IN PHYSICS 
from the 

NAVAL POSTGRADUATE SCHOOL 
June 1987 



ABSTRACT 



An experimental investigation of unipolar arcing at 
atmospheric pressure was conducted using Aluminum 2024 and 
glass slides with a thin coating of pure aluminum. The 
plasma was produced on the surface using a neodymium-glass 
laser in the Q-switched mode. It was found that the power 
density required for the onset of unipolar arcing was similar 
to that of samples irradiated in a vacuum, although the size 
and density of the resulting craters were significantly 
dif f erent . 



3 



TABLE OF CONTENTS 



//*' 

- » 



557 ^ 



I. INTRODUCTION 6 

II. BACKGROUND AND THEORY OF UNIPOLAR 

ARCING MODELS 10 

A. ROBSON-THONEMANN MODEL 10 

B. TAYLOR-SCHWIRZKE MODEL 13 

III. EXPERIMENTAL DESIGN AND PROCEDURES 20 

A. EXPERIMENTAL DESIGN 20 

B. EQUIPMENT 21 

1 . Laser 21 

2. Optical Microscope 23 

3. Scanning Electron Microscope 23 

4. Target Test Chamber 23 

C. TARGET PREPARATION 24 

1. Aluminum Coated Microscope Slides 24 

2. Aluminum Targets 25 

IV. EXPERIMENTAL RESULTS 26 

A. TYPE 2024 ALUMINUM TARGETS 26 

1. Description 26 

2. Target Damage 26 

3. Determination of Power Density Required 

for the Onset of Unipolar Arcing 30 

a. Experimental Results 30 

b. Analysis of Results 31 

c. Discussion of Results 35 



4 



B. ALUMINUM COATED GLASS SLIDES 



37 



1. Description 37 

2. Target Damage 38 

V. CONCLUSIONS 48 

LIST OF REFERENCES 50 

INITIAL DISTRIBUTION LIST 52 



5 



I. INTRODUCTION 



One of the most active areas of research that the 
scientific community is currently engaged in is that of power 
production. Fusion energy sources in particular have 
received great emphasis. In most fusion reactor design 
studies, the plasma is confined by strong magnetic fields. 
Some of the plasma will, however, diffuse perpendicular to 
the field lines and come into contact with the surrounding 
walls. This has led to extensive studies involving the 
interaction of a hot dense plasma with a surface. These 
studies have revealed that erosion due to several processes 
occurs. These processes include sputtering, unipolar arcing, 
pulse heating, and gas implantation [Ref. 1]. 

The process involved in sputtering is a momentum transfer 
process. When an energetic neutral atom or ion hits the 
solid surface it imparts its energy to a surface atom, and if 
this energy exceeds its binding energy, the particle can be 
released into the plasma [Ref. 2]. Heat pulses are a result 
of nonuniform energy deposition onto the wall from the plasma 
and may lead to localized evaporation and surface cracking. 
Gas implantation results when high energy hydrogen or helium 
plasma ions impact on the first wall. The ones that are not 
directly backsca ttered come to rest in the wall, generally at 



6 



an interstitial site [Ref. 1], Unipolar arcing is a process 
where an arc is established between the wall and the plasma 
with the wall acting as both the anode and cathode. 

These erosion processes lead to two major areas of 
concern. The first is that various components of the reactor 
will be damaged, resulting in a limited lifetime. The 
second, and perhaps more significant, is that these erosion 
processes introduce impurities into the plasma, and these 
impurities limit the available plasma parameters [Ref. 1: p. 
1047]. In an article by T. Taxima, he states that "the major 
energy losses from the present day’s tokamak plasmas are 
those to impurities" [Ref. 3], 

Evidence of unipolar arcing (i.e. arc tracks) has been 
observed in DITE, as well as PLT, ISX, Macrator and Pulsator 
Tokaraaks [Ref. 4], The arc tracks appear on the fixed 
limiter, probes inserted in the plasma and on parts of the 
torus vacuum structure. A study on impurity levels and 
erosion rates showed that unipolar arcing is the dominant 
mechanism for impurity production [Ref. 5], This is due to 
the fact that unipolar arcing is not a homogeneous energy 
deposition process, but rather one whereby the energy from 
the plasma is concentrated towards the cathode spots [Ref. 
6 ]. 

Plasma surface interactions are also important in newly 
evolving areas of technology. For example, weapon system 



7 



designs for both high energy lasers and particle beams 
include high speed plasma switches. The contacts on the 
switches will operate in a plasma environment and will be 
subject to these erosion processes. Also, in high energy 
laser weapons a plasma could be formed at the target which 
will result in surface damage. It is not known how severe 
this plasma surface damage will be compared to other damage 
mechanisms, but conceivably it could be significant. 

Although unipolar arcing was first observed in the late 
1950's, other problems dominated the fusion program and 
interest in arcing diminished. It was not until the late 
1970's that interest in unipolar arcing was revived, and 
since that time it has been an area receiving much study, 
although the process is still not fully understood. For this 
thesis, experimentation was done to observe plasma-surface 
damage which occurs at atmospheric pressure on a good 
conducting material (type 2024 Aluminum) and on a poor 
conductor (a glass slide) that was coated with a thin film of 
aluminum . 

The specific goals of this thesis were twofold. The 
first was to conduct a systematic experimental study to 
observe unipolar arcing at atmospheric pressure, determine 
the power density required for the onset of arcing (on type 
2024 Aluminum targets) and compare the observed results to 
those previously reported. Secondly, it -was a goal to try to 



8 



photograph the arcing events as they were occurring. For 



this, the aluminum coated glass slides were used 
camera placed behind the targets. 

The following section deals with the theory 
unipolar arcing phenomenon and gives details 
underlying model on unipolar arcing. 



with a 

of the 
of the 



9 



II. BACKGROUND AND THEORY OF UNIPOLAR ARCING MODELS 



A. ROBSON-THONEMANN MODEL 

In 1958 Robson and Thoneraann proposed the first model of 

an arcing phenomenon which "requires only one electrode and 

is maintained by the thermal energy of the plasma electrons" 

[Ref. 7]. They called this phenomenon "unipolar arcing." 

Two experiments were also conducted which verified their 

theory. In these, a strong plasma was generated by an 

electrodeless high-frequency discharge in a mercury cathode 

— 6 

tube which was evacuated to a gas pressure of less than 10 
mm Hg. An arc was then established by an externally applied 
anode voltage. As the plasma density increased, the arc 
current increased. The externally applied voltage was then 
turned off, but the arc continued to burn in the unipolar 
mode until the plasma density became too low to sustain it. 

The basis of their theory lies in the idea that a 
floating potential is established between a plasma and the 
exposed metal plate. Within a plasma, the ion temperature 

(T^) does not necessarily equal the electron temperature 
(T e ). The electrons will have a much higher thermal velocity 
than will the ions due to their relatively small mass. When 
in the vicinity of a wall, many more electrons will initially 
contact the wall, resulting in a negative potential relative 
to the plasma. (One of the defining conditions for the 



10 



existence of a plasma is that it is "a quasineutral gas of 
charged and neutral particles" [Ref. 8], The idea behind 
quasineutrality is that if the plasma dimensions are large 
compared to a characteristic length called the Debye Length, 
then it will shield itself from electric potentials (or 
electric fields) in a distance on the order of the Debye 
Length. This results in the formation of a potential sheath 
(Fig. 2.1) which prevents all but the most energetic of the 
electrons to reach the wall while at the same time it 
functions to accelerate the ions within the sheath. The 
plasma potential will build up to the point where the ion 
loss level is equal to the electron loss level, thus 
maintaining an equilibrium situation where the net current is 
zero. For singly charged ions the sheath potential is given 



by : 



kT M. 




( E q n . 2.1) 



e 



where 



k = Boltzmann Constant 



e 



charge of the electron 



T 



e 



electron temperature 



M e = electron mass 



= ion mass 



11 



If the value of this floating potential exceeds a certain 
value, an arc will be initiated and sustained on the isolated 
plate. If a cathode spot is initiated on the plate, there 
will be a strong local emission of electrons from it. This 
results in a reduction in the floating potential to the value 
of V c « More electrons can then reach the plate, closing the 
current loop and maintaining the plasma's q ua s i n e u t r a 1 i t y . 
This circulating current is given by {Ref. 9]: 



I 

c 





(- eV c > 

(kT e ) 



-exp 



C-eVj ) 1 

TkT^r 



( Eqn 2.2) 



where 

n^= electron plasma density 
A = area of the exposed plate 
In order to maintain a stable cathode spot, there is a 
lower limit for I c * Above this value the arc will be 
maintained, and below it, it will not. This critical value 
of current is dependent on the material of the plate and is 
on the order of 10 Amperes. 



12 




/ i/sheath 



v =o 



plasma 



v = v. 



i 




sheath 



Figure 2.1 Plasma-Wall Potential Sheath. 

This model is widely referenced in many papers dealing 
with unipolar arcing and is generally accepted to explain the 
basic process of unipolar arcing. However, this model does 
not address many of the specifics of what is happening, and 
in order to gain a better understanding of this phenomenon, a 
more detailed model must be examined. 

B. TAYLOR-SCHWIRZKE MODEL 

An expanded version of the unipolar arcing phenomenon was 
introduced in 1980 by Schwirzke and Taylor [Ref. 10]. This 



13 



model expands on the basic ideas presented in the Thonemann 
model by elaborating "upon the electric fields which are set 
up in the plasma and drive the arc" [Ref. 11]. 

As discussed previously, a sheath potential is 
established between the plasma and the surface (as given by 
Equation 2.1). The length (perpendicular to the surface) 
over which this potential exists is proportional to the Debye 
Shielding Length (A^) which is given by: 

2 ^ 

\ d = (kT e /47T n e ; (Eqn. 2.3) 

In order for an arc to develop, Vf must increase sufficiently 
for an arc to ignite and be sustained. Another necessry 
condition, they argue, is that the density of ions above the 
cathode spot increases in order for a larger electron current 
to flow from the surface into the plasma (this is different 
from the previously discussed model in that it implies a 
constant plasma density). This increase is due to the 
ionization of neutral particles being released from the 
cathode spot (Fig. 2.2). Cathode spots can initiate from 
surface protrusions, inclusions, micro-whiskers, or other 
metallurgical inhomogeneities. 

This idea (of where the cathode spot begins) is 
consistent with a study done by Tien, Panayotou, Stevenson, 
and Gross on different materials which had been prepared 



14 



similarly, but upon which some samples were slightly etched 
to expose their characteristic grain structure. They found 
that the etched samples had evidence of a higher area 




/ 




/ 





cl 



wal I 



Figure 2.2 Schwirzke-Taylor Unipolar Arc Model. 



coverage of arcing than did the polished samples. They also 
reported that "the locations of the arc spots are not random 
in the slightly etched samples." They continue by saying, 
"It is interesting that, on the polished specimens, ...» arc 
spots are more randomly distributed." One of their main 
conclusions is that the cathode spots are preferentially 
initiated at raicrostructural inhomogeneities [Ref. 12], 
Figures 2.3 and 2.4 illustrate this. 



15 




© I ON 



$ ELECTRON 



Figure 2.3 Isolated Plate Without Cathode Spot. 




© ION $ ELECTRON 

Figure 2.4 Isolated Plate with Cathode Spot. 



16 



Using values which are characteristic of tokaraak edge 
plasmas, Schwirzke and Taylor calculate the Debye Length (A^) 
to be on the order of 10”^ cm. Making assumptions about the 
neutral particle density (n^ ) in the sheath (which is a 
function of the desorption and evaporation rates) and 
assuming a value for the ionization cross section, the mean 
free path length can be calculated. From this, the 
probability for ionization of a neutral particle can be 
found. Using the assumed plasma and surface characteristics, 
a probability of 2x10“^ was calculated. This leads to an 
increase in the plasma density by a factor of 4000. The 
Debye Length is inversely proportional to the square root of 
the electron plasma density, and decreases by a factor of 
about 60 . 

The electric field in the normal direction is given by: 

E n = V f/A d ( Eqn 2.4) 

Therefore a decrease in the Debye Length will cause a 
corresponding increase in the electric field which 
accelerates the ions towards the surface. The increased 
number of ions bombarding the surface leads to an increase in 
surface temperature which causes more desorption and 
evaporation, and the process continues. 



17 



The increase of the plasma density above the cathode spot 



also leads to a radial electrical field given by: 



E 



r 




dn e 

dr 



(Eqn 2.5) 



This causes a lowered potential in a ring-like area which 
allows electrons to return to the surface and maintain the 
plasma's quasineutrality. The lowered potential ( AV) is: 



AV 




n. 



£n 



n 



eo 



(Eqn 2.6) 



Equating the sheath potential to AV yields: 



kT M. 

e n 1 

2e n 2W 



kT n 

e „ e 
Jin 



(Eqn 2.7) 



n 



eo 



With zero sheath potential, the electron saturation current 
to a surface over A is: 



I g = en e V e A (Eqn 2.8) 

4 

Solving for area (A) and using commonly occurring values for 
the variables in the equation, a radius on the order of 10 
microns is obtained. This value for the return current area 
is of the same order as the radius of the outer crater rim, 



18 



suggesting that the location of the outer crater rim is an 
indication of how far the return current area has expanded 
during the burn time of the arc. 

The Robson and Thonemann model stated that the area was 
the whole surface area that was exposed to the plasma. It 
was once thought that if the exposed surface was subdivided 
by insulating strips, arcing would be prevented, however, for 
areas of the magnitude predicted from the Schwirzke and 
Taylor model, this is clearly impractical. 

While the Robson and Thonemann model seems to account for 
the basic mechanism of unipolar arcing, this (Taylor and 
Schwirzke) model predicts the occurrence of the cathode spot 
hole which is due to high rates of desorption and 
evaporation. It also suggests relatively small values for 
the area of the return current. 



19 



III. EXPERIMENTAL DESIGN AND PROCEDURES 



A. EXPERIMENTAL DESIGN 

Target samples were irradiated by a neodymium glass laser 
to produce a plasma at the surface. The initial aim of this 
thesis was to then attempt to photograph the resultant 
unipolar arcing. Several difficulties, however, arose to 
prevent this. Primarily there was the problem of obtaining a 
camera that would satisfactorily accomplish the task. The 
primary stipulation on the camera was to be able to obtain 
magnifications on the order of at least 100X. In order to 
accomplish this, with available equipment, the camera had to 
be approximately an inch from the target. This immediately 
precluded the ability to use the evacuated target test 
chamber, since the viewing port is approximately eight (8) 
inches from the target. Even when the target was mounted 
outside the chamber no acceptable manner of positioning the 
camera could be found that would yield a direct enough 
picture, be out of the primary laser beam, and where the 
camera’s optics could be protected against the resulting 
plasma and reflected radiation. 

The next approach was to use a thin target (less than 5 
micrometers thick since that is the approximate depth of the 
ensuing crater) and attempt to photograph arcing from behind 



20 



the target. In a setup of this manner, an attenuating filter 
could be placed between the target and the camera which would 
protect the camera from the laser radiation and there would 
be virtually no limit to how close the camera could be 
positioned to the target. With this in mind, a vacuum 
deposition of aluminum was placed onto glass microscope 
slides. Two thicknesses of the coatings were used (thin 
coatings of approximately . 2 y m and a thick coating of 
approximately 1 ym). 

The process of unipolar arcing in atmospheric pressure 
has not been studied. A second procedure involving 
irradiating aluminum targets in atmosphere with a neodymium 
glass laser was done in order to determine the onset of 
arcing. All samples were examined under an optical and a 
scanning microscope to determine the surface conditions 
before and after the plasma surface interaction. Figure 3.1 
is a schematic of the experimental setup. 

B. EQUIPMENT 

1 . Laser 

A KORAD K-1500 Q-switched neodymium glass laser was 
used to irradiate the test samples producing a hot-dense 
plasma over the surface. The laser outputs a wavelength of 
1.06 micrometers. Nominal output energy ranges from .2-15J, 
depending on the applied voltage to the oscillator and 



21 




BEAM SPLITTER 




METER 



/ 

/ 



/ 




Figure 3.1 Experimental Setup. 



22 



amplifier flash lamps. It was found to be easier and more 
controllable to use constant voltages to the amplifier and 
oscillator and reduce the energy by placing transmission 
filters in the path of the beam. The output energy was 
measured using a Laser Precision RK-3200 Series Pyroelectric 
Energy Meter calibrated to an accuracy of about 20%. Nominal 
pulse width is 25 nanoseconds, and the unfocused beam has a 
cross sectional area of 4.04 ± .2 cm2. 

2 . Optical Microscope 

The optical microscope that was used is a Bausch and 
Lomb stereoscopic light microscope. Magnification ranges 
from 20X to 800X. The target samples were observed before 
and after laser illumination. 

3 . Scanning Electron Microscope 

The Scanning Electron Microscope used was a Cambridge 
Stereoscan S4-10. The images obtained were formed from 
secondary electrons. It has the capability of magnifications 
in the range from 20X to 100,000X. 

4 . Target Test Chamber 

The aluminum 2024 target samples were mounted in a 
cubic target test chamber (maintained at atmospheric 
pressure) with an internal volume of 12.9 ± 0.3 liters. The 
chamber is composed of unbaked aluminum with a target holder 
probe running into the chamber that can be rotated for 
alignment and to expose fresh . target samples. The targets 



23 



are aligned at 30 degrees to the laser beam. A focusing lens 
was mounted on the chamber to obtain the desirable laser beam 
spot size. 

C. TARGET PREPARATION 

1 . Aluminum Coated Microscope Slides 

The microscope slides used in this portion of the 
experiment are of an ordinary soda-lime window glass type. 
The approximate composition is 72% Selica (Si02)* 15% Soda 
(Na 20 ), 9% Lime (CaO), 3% Magnesium Oxide (MgO), Aluminum 

Oxide (AI 2 O 3 ) and .03% Iron (Fe 203 )* 

The slides were thoroughly cleaned with soap and 
water and then with methyl alcohol. They were then coated 
with a thin film of pure (99.99%) aluminum. The deposition 
of aluminum on the slide surface was accomplished by an 
evaporation process under high vacuum using a Veeco Series 
401 system. The slides were then stored in a desicator until 
ready for use, at which point any surface dust was removed by 
using high pressure freon gas. 

The thickness of the aluminum coatings was 
approximated by comparing the mass of the slide before and 
after depositing the aluminum coating. This assumes a 
uniform thickness of aluminum on the coated area. 



24 



2 . Aluminum Targets 

The disk targets were made of type 2024 aluminum. 
They were machined on a lathe to a final diameter of 
approximately 1/4 inch. The samples were mounted in bakelite 
and rough sanded. They were then polished using an AB Duo 
Belt Wet Sander (400 grit) and fine ground using 600 grit wet 
paper. Following this, they were dry sanded and polished 
using three (3) slurries of .05 AI 2 O 3 to reduce surface 
roughness . 

The specimens were then removed from the bakelite, 
cleaned with methyl alcohol and u 1 1 r a s o n i c a 1 1 y cleaned in 
acetone. They were then stored in a desicator until ready 
for use in the target chamber. 



25 



IV. EXPERIMENTAL RESULTS 



A. TYPE 2024 ALUMINUM TARGETS 

1 , Description 

Numerous highly polished Aluminum 2024 targets were 
irradiated in the target chamber at atmospheric pressure. 
The power density was varied by the use of transmission 
filters and the energy was measured with the energy meter. 
The beam spot size was controlled by using a glass lens and 
was measured by irradiating an exposed Polaroid film at the 
target plane and then measuring the diameter of the spot. 
Using an oscilloscope, the 3 db pulse width was found to be 
20 nanoseconds. A hand held Polaroid camera was positioned 
above the targets to note any plasma formation. 

2 . Target Damage 

After being exposed to the laser radiation each 
sample was observed using an optical and scanning electron 
microscope. Target damage was in the form of unipolar arcing 
and surface melting. Figure 4.1 shows target damage in the 
region of maximum irradiation. In all of the field of view 
with the exception of the right central portion it is evident 
that melting has occurred. At greater magnification, Figure 
4.2 shows the boundary region of melting. The plasma flow 
direction away from the focal spot in the figure is from top 



26 




Figure 4.1 High Irradiance Region of 
Aluminum Target, 220X, SEM 




Figure 4 . 2 



Boundary Region for Melting 
on Aluminum Target, 1200X, SEM 



27 



to bottom. 



The small holes around which the flow was 



diverted are the craters where, simultaneously, unipolar 
arcing was occurring. Outside of the region of the melting, 
unipolar arcing was the only visible damage mechanism as 
evidenced by the characteristic crater and rim formation. 

At lower energy densities the only observable damage 
mechanism was unipolar arcing. Figure 4.3, at 220X 
magnification, using the scanning electron microscope, shows 
the region of maximum intensity on a sample irradiated at an 
order of magnitude less energy density than the preceding 
figures where melting was evident. The unipolar arc craters 
are not evenly distributed, but are bunched in locally 
concentrated areas. This is partially due to hot spots in 
the laser beam but also may be a result of surface 
in ho m o g e n e i t i e s where arcing is initiated (such as surface 
inhomogeneities like whiskers or at grain boundaries). At 

high magnification, Figure 4.4 clearly shows the unipolar arc 

£ 

craters. The crater density is on the order of 1.7 X 10° 
craters per square centimeter. 

Further analysis from Figure 4.5 at higher 
magnification shows a distinctive view of the unipolar arc 
craters. The craters with the largest diameter, 12 to 13 
microns, correspond to a long arc duration as compared to the 
ones with small diameters, 3 to 4 microns. The central dark 
regions are the cathode spots and range- from .08 microns to 



28 




Figure 4.3 Unipolar Arcing on the Aluminum 
Target, 220X, SEM 



Figure 4.4 



Unipolar Arc Craters on 
Aluminum Target, 540X, SEM 




29 






Figure 4.5 Unipolar Arc Craters on 

Aluminum Target, 1200X, SEM 

2.5 microns in diameter, again a function of the arc duration 
time. The hemispherical crater is formed by the outflow of 
molten metal where localized melting has occurred due to the 
radial plasma pressure. 

3 . Determination of Power Density Required for the 
Onset of Unipolar Arcing 

a. Experimental Results 

The incident energy upon various Aluminum 2024 
targets was systematically varied in order to determine the 



30 



minimum power density required for unipolar arcing to occur. 
Average power density, F, can be calculated by: 

F = E/AT 

where E = Incident Energy 
A = Exposed Area 
T = 3 db pulse width 

A summary of results is contained in Table I. 
b. Analysis of Results 

From these results an indication of the onset of 
arcing can be determined although there are some 
inconsistencies in the data. Targets la and 2 showed no 
evidence of arcing while targets 8 and 12 had arcing damage 
even though the calculated value for power density was lower. 
It is important to note that targets la through 5 were 
irradiated on the same day and that targets 6 through 13 were 
irradiated approximately three days later. 

From the first group (la through 5) it can be 

stated that the onset of arcing occurs between 13.1 and 20.3 

2 

M V / cm • The second group (6 through 13) indicates a region 

2 

of between 6.8 and 9.7 MW/ cm . Although these values differ, 
the data obtained in each group is self consistent. 

There are at least three possible explanations 
for the inconsistencies in the data. These are: 



31 



la 

lb 

2 

3 

4 

5 

6 

7 

8 

9 

10 

11 

12 

13 



I. SUMMARY OF RESULTS TO DETERMINE THE ONSET OF 
ARCING FOR ALUMINUM 2024 AT ATMOSPHERIC PRESSURE 



Calculated Power 
Density (MW/CM^) 


Light 


Arcing 


13.1 


no 


no 


103.0 


yes 


yes 


12.6 


no 


no 


26.4 


yes 


yes 


20.3 


yes 


yes 


19.4 


no 


no 


31.7 


yes 


yes 


21.2 


yes 


yes 


9.7 


no 


yes 


101.0 


yes 


yes 


17.6 


no 


yes 


29.0 


yes 


yes 


10.6 


no 


yes 


6.8 


no 


no 



32 



(1) The targets were somehow different (due to formation 
of aluminum oxide with time). 

(2) The diagnostic instrumentation was different. 

(3) The laser mode was different. 

Any one of the above or a combination of the 
above could have caused the results to differ from one 
another . 

All of the above targets were identically 
prepared and stored in a desicator until their use. It is 
conceivable, however, that between the time the first and 
second groups were irradiated the surface of the targets 
changed due to absorption of moisture or to surface oxidation 
or contamination. A change in the surface characteristics 
could have caused a different point for the onset of arcing. 

The hardware and experimental setup on the two 
days in question was identical. It was noted, however, that 
the energy meter had regions of deterioration on the detector 
surface. Although care was taken on both days to avoid those 
regions, it is possible that alignment on one of the days was 
such that enough energy was incident on those regions to 
account for the change in indicated energy. 

The next possible cause is the output mode of the 
laser. As previously stated, the intensity distribution of 
the output beam is not gaussian or flat top in shape. The 
high intensity region of the beam is a crescent shaped 



33 



portion in the lower part of the mode (as determined by 
actual target damage regions). The performance of the laser 
is temperature sensitive. Between the two runs the 
atmospheric conditions could have changed sufficiently to 
cause a slightly different spatial intensity distribution 
yielding a different peak to average ratio and causing 
inconsistent results. 

In a previous experiment at the Naval 
Postgraduate School, Beelby and Ulrich III discovered a 
similar inconsistency. They had done experimentation to 
determine the onset of arcing in a vacuum for various types 
of materials. It was determined that the onset of arcing for 
type 304 stainless steel was between 5.1 and 5.4 Mw/ cm^ • 
However, while determining the onset of arcing for aluminum 
2024 a sample of this type of stainless steel was irradiated 
and no plasma or arcing was present at 10.9 MW/ cm^. [Ref. 9] 

In addition to the above explanations for the 
inconsistency in the data there are also measurement errors. 
The first, and possibly most significant, is the fact that in 
calculating power density for the onset of arcing, a uniform 
spatial intensity distribution is assumed. This is certainly 
not the case and in making this assumption large errors may 
result. Furthermore, the calculation also assumes that the 
temporal content of the beam is a step function, which is 
also not the case. 



34 



A second source of error results from measuring 
spot size. The spot size seen using the exposed Polaroid 
film has an average diameter on the order of ,75cm. The 
response of the film to irradiation is not known exactly and 
the spot is slightly elliptical in shape. Assuming the 
measurement accuracy is 0.1 cm, a 3% error would result. 
Another assumption in calculating the power density was that 
each optical surface reflected 4% of the incident energy. 
The actual amount of reflected energy could not be measured. 
However, since the power meter was indirectly irradiated by 
the reflection from one of the optics, a slight error in the 
reflectivity could lead to relatively large measurement 
errors. Finally, the Energy Meter was calibrated to an 
accuracy of only 20%. 

In lieu of these error sources, calculated power 
densities can only be assumed to be accurate to an order of 
magnitude. 

c. Discussion of Results 

Previous studies of unipolar arcing have been 
conducted at the Naval Postgraduate School. The equipment 
and experimental setup in these studies were, for the most 
part, identical to those used for this thesis. In 
particular, the laser (with its ancillary equipment), target 
chamber, optics, and power meter were identical. The 



35 



variable of concern in this experiment was ambient pressure. 
Previous work has involved placing the samples in the target 
chamber, evacuating the chamber and then irradiating the 
targets. For this experiment the target chamber was allowed 
to remain at atmospheric pressure. This section will discuss 
and compare results from previous work with that done for 
this thesis. 

The damage mechanisms seen in this experiment 
were similar to those previously reported. Samples were 
observed to have evidence of melting and unipolar arcing. As 
the power density was decreased, unipolar arcing was the only 
damage mechanism present. 

Samples irradiated at atmospheric pressure had 
the characteristic hemispherical shaped craters associated 
with unipolar arcing. The density of craters found at 
atmospheric pressure (1.7 X 10^ craters per square 
centimeter) is an order of magnitude larger than previously 
reported values (3.0 X 10^ craters per square centimeter) 
[Ref. 9], There was also a difference found in the size of 
the crater rims. At atmospheric pressure the largest single 
craters found were approximately 13 microns in diameter while 
in a vacuum the longest burning arcs left craters up to 30-40 
microns in diameter. The cathode spots were similar in size 
and nature. 



36 



The onset of arcing at atmospheric pressure for 
Aluminum 2024 was found to be between 6.8 and 20.3 MW /cm^ . 
In previous work done at the Naval Postgraduate School, 
Beelby and Ulrich III reported the onset of arcing for 
Aluminum 2024 in a vacuum to be between 4.6 and 11.2 MW/cm^ . 
Because of the accuracies of both of these measurements, the 
only conclusion which can be made is that the onset of arcing 
at atmospheric pressure occurs close to that of a target in a 
vacuum. That is, there does not appear to be a strong 
dependency on surrounding pressure. 

B. ALUMINUM COATED GLASS SLIDES 
1 . Description 

Numerous shots were made on the aluminum coated glass 
microscope slides. They were positioned outside of the 
target test chamber such that the incident beam made an angle 
of about 10 degrees relative to the surface normal. A narrow 
band filter (1.06pni) was placed immediately behind the coated 
slide and then a Polaroid camera was positioned to photograph 
the backside of the microscope slide. After the targets were 
irradiated, they were examined using an optical and scanning 
electron microscope (SEM). 

Examining the damage using the SEM proved difficult 
due to the lack of conductivity of the surface of the target. 
(The SEM requires an electrically conductive surface in order 



37 



to function properly.) In the figures produced by the SEM, 
the horizontal light colored bands are a result of the poor 
conductivity . 

2 . Target Damage 

After being irradiated, all of the targets had areas 
where the aluminum coating had been removed from the surface. 
The bright streaks in Figure 4.6 are the paths of molten 
aluminum particles occurring during the time the laser beam 
was incident on the target. The bright portion is indicative 
of a plasma being present. In Figure 4.7 a 90% attenuating 
broad band filter was placed before the camera. Again, the 
bright portion is evidence that a plasma was present and the 
crescent shape is where the aluminum was removed, 
corresponding to the high intensity portion of the beam. In 
the less intense regions of the beam the aluminum coating was 
not removed . 

Within the high intensity region there were brown 
colored elliptical shaped areas of even higher intensity. 
Figure 4.8 shows a portion of the cresent shaped high damage 
region and Figures 4.9 and 4.10 show extremely high intensity 
regions. Figure 4.11 is a magnified view of Figure 4.10. 
From this it can be seen that there are small circular areas 
of damage (top central region and the region in the center of 



38 




Figure 4.6 Polaroid Photograph of Aluminum 
Covered Glass Slide 




Figure 4.7 Same as above with a 90% 
Attenuating Filter 



39 





Figure 4.8 High Intensity Region of Aluminum 
Covered Glass Slide, 15X, SEM 





Figure 4.9 



Hot Spot within High Intensity Region of 
Aluminum Covered Glass Slide, 75X, SEM 



40 





'f 




I || 



• i|g |p 

' ; • • ; Y X ; :' ; &£>/ 

c> 

V- 

. 

" ,v '' v 



wp 

■ v wt ■ f 



Figure 4.10 Hot Spot within High Intensity Region of 
Aluminum Covered Glass Slide, 50X 
Optical Microscope 




Hot Spot within High Intensity Region 
of Aluminum Covered Glass Slide, 100X t 
Optical Microscope 



4.11 



F igur e 



41 




the photograph). Figure 4.12 is a magnified view of one of 
these regions. At 400X magnification on the optical 
microscope, Figure 4.13 shows that these areas are concentric 
rings. This interference pattern is seen frequently when 
studying unipolar arcing. The rings are a result of 
interference fringes from the laser and typically the high 
intensity bands have many unipolar arc craters while the low 
intensity regions have few or none. In this figure it is 
difficult to determine decisively if the unipolar arc craters 
are present, although in the central region where focus in 
the figure is the sharpest it appears that they are 
craterlike structures. The diameter of the crater is 
approximately 7.5 micrometers, corresponding to what would be 
expected if they were craters from unipolar arcing. Although 
clearly visible using an optical microscope, these regions 
did not appear using the scanning electron microscope. 

A feature that was evident in the "hot spots" within 
the high intensity regions, using the scanning electron 
microscope, was barnacle appearing structures. The 
structures are shown in increasing magnification in Figures 
4.14 through 4.16. The cores or "holes" in the center of the 
barnacles have diameters ranging from 0.3 to 1.9 microns, but 
are probably not unipolar arcs as they do not have the 
characteristic rims surrounding the cores. Similar 
structures have been observed previously with the conclusion 



42 




Figure 4.12 Magnified View of Figure 4.11, 

200X, Optical Microscope 




ure 4.13 Concentric Rings Occurring in Hot Spots 
on Aluminum Coated Glass Slides, 400X , 
Optical Microscope 



43 







Figure 4.14 Barnacle-like Structures Appearing 

in Hot Spots, 760X, SEM 




Figure 4.15 Same as above, 1500X, SEM 



44 




Figure 4.16 Barnacle-like Structures Appearing 

in Hot Spots, 7500X, SEM 




Figure 4.17 Boundary of Hot Spot Region, 760X, SEM 



45 



that since the samples were not outgassed, the barnacles may 
be a result of "expanding cavities" or "blisters" of trapped 
hydrogen due to the flash heating by the laser [Ref. 9]. 
Surface tension acting on the liquid metal can also 
contribute to the formation of the barnacle structures. 

At the boundary of the "hot spots" the barnacles are 
no longer present and the surface becomes much smoother (see 
Figure 4.17). In this region there is no evidence of 
unipolar arcing. 

In the lower intensity region, Figure 4.18 shows what 
appear to be craterlike formations. The diameters of these 
craters are .5 to 1.5 microns. These structures are likely 
to be unipolar arc craters with the rims not visible due to 
poor conductivity of the glass slide under the SEM. 



46 




Figure 4.18 



Craters in Low Intensity Region of 
Aluminum Covered Glass, 1500X, SEM 



47 




V. CONCLUSIONS 



The study undertaken in this thesis examined the unipolar 
arcing phenomenon at atmospheric pressure. On the aluminum 
targets it was found that anytime there was a plasma present, 
unipolar arcing occurred. The power density required for the 
onset of arcing at atmospheric pressure was found to be 
similar to that in a vacuum, although the crater density was 
found to be an order of magnitude higher, while the size of 
the resulting craters was more than a factor of 2 smaller. 
This suggests that although the power density required to 
initiate arcing was similar at atmospheric pressure, more arc 
initiation sites are present. Consequently, in order to 
maintain a balance they burn less intensely. 

The extremely high intensity regions on the aluminum 
coated glass slides when viewed under the SEM showed no 
evidence of unipolar arcing, but arc craters in these regions 
may have been obscured by the appearance of the barnacle-like 
structures. Using the optical microscope, these high 
intensity regions showed interference rings and craterlike 
structures which possibly were formed by unipolar arcing. In 
the regions of less intense radiation the barnacles were no 
longer seen and craters resembling those from unipolar arcing 
were apparent. 



48 



A previous study had suggested that the use of better 
vacuum systems would result in minimizing damage from 
unipolar arcing. This study suggests that while the number 
of arc sites may be minimized, the sites that did initiate 
and sustain an arc would burn more violently. The number of 
damaged spots from unipolar arcing would be fewer but the 
damage to the surface in these regions would be more drastic. 



49 



LIST OF REFERENCES 



1. Behrisch, R., "Surface Erosion from Plasma Materials 

Interaction," Journal of Nuclear Materials , v. 85-86, p. 
1047-1061, 1979. 

2. Keville, M. T. and Lautrup, R. W., "An Investigation of 
Unipolar Arcing Damage on Stainless Steel and Titanium 
Carbide Coated Surfaces," M.S. Thesis, Naval 
Postgraduate School, Monterey, California, 1980. 

3. Tazima, T., "First Wall Design Considerations of JT-60 
and Related Experiments," Journal of Nuclear Materials , 
v. 76-77, p. 594-599, 1978. 

4. Ryan, F. T. and Shedd, S. T., "A Study of the Unipolar 
Arcing Damage Mechanism on Selected Conductors and 
Semiconductors," M.S. Thesis, Naval Postgraduate School, 
Monterey, California, June 1981. 

5. Goodall, D.H.J., Conlon, T. W., Sofield, C. and 

McCraken, G. M. "Investigations of Arcing in the DITE 
Tokaraak," Journal of Nuclear M aterials , v. 76-77, p. 
492-498, 1978. 

6. Schwirzke, F., "Basic Mechanisms that Lead to Laser 
Target Damage," Naval Postgraduate School Paper, Oct 
1981. 

7. Robson, A. E. and Thoneraann, P. C., "An Arc Maintained 
on An Isolated Metal Plate Exposed to A Plasma," 
Institute of Electrical Engineers , v. 106, pt. A, supp. 
2, p. 508-512, April 1959. 

8. Chen, F., Introduction to Plasma Physics and Controlled 
Fusion , v. 1, Plenum Press, New York, N.Y., 1984. 

9. Beelby, M. H. and Ulrich III, H. G., "A Study of the 
Breakdown Mechanism of AISI 304 Stainless Steel, Type 
2024 Aluminum and Various Titanium Coatings," M.S. 
Thesis, Naval Postgraduate School, Monterey, California, 
1981. 

10. Schwirzke, F. and Taylor, R. J., "Surface Damage by 
Sheath Effects and Unipolar Arcs," Journal of Nuclear 
Materials , v. 93-94, p. 780-784, 1980. 



50 



11. Schwirzke, F., "Laser Induced Unipolar Arcing," Laser 

Interaction and Related Plasma Phenomena , v. 6, 1984. 

12. Tien, J. K., Panayotou, N. F., Stevenson, R. D., and 
Gross, R. A., "Unipolar Arc Damage of Materials in a 
Hot, Dense Deuterium Plasma," J__o _u _r _n a__l _o f__ N_ u_ c__l e_a_ jr 
Materials, v. 76-77, p. 481-488, 1978. 



51 



INITIAL DISTRIBUTION LIST 



No . 



1. Defense Technical Information Center 
Cameron Station 

Alexandria, Virginia 22304-6145 

2. Library, Code 0142 
Naval Postgraduate School 
Monterey, California 93943-5002 

3. Professor K. E. Woehler, Code 61Wh 
Department of Physics 

Naval Postgraduate School 
Monterey, California 93943 

4. Professor F. Schwirzke, Code 61Sw 
Department of Physics 

Naval Postgraduate School 
Monterey, California 93943 

5. LT Steven W. Woodson, USN 
202 Rossford 

White Sands Missile Range, New Mexico 88002 



Copies 

2 



2 



1 



2 



2 



52 



1 



I 



■l 

I 






Ji 




• ; rr ♦ ry 
O ^ < 






I 



DUDLEY KNOX LIVPAPY 
NAVAL POSTGHAD T T /• TE SCHOOL 
MOI7TEBEY, CAUffOfLBIA 93943-6002 




Thesis 

W843757 Woodsoh 
c 1 1 An investigation o£ 

unipolar arcing at at- 
mospheric pressure in 
Aluminum 2024 and alumi- 
num coated glass slides.