eomhhmhmhhl ehmheehhh eemmmhhhheo · 2014-09-27 · zone contains a complex , turbulent mixture of...
TRANSCRIPT
7RD-A149 258 ANALYSIS OF THE DISSIPATION MECHANISM WITHIN
Ni/INSTATIONARY CATHODE SPOT.-CU) STUTTGART UNIV (GERMANY
U U CRS FR) INST FUER RAUMFAHNTANTRIEBE _Hi0 SCHRRDEET AL,NCRSIFEDVAUG 84 IRA-84-IB-6 AFOSR-TR-84-i16 F G 281/3
EhMhEEhhhEomhhmhmhhlEEmmmhhhhEo
,- . . . . . . L - . ' - . , , , . r, ' "
.- -. • . '- r -. - -.- '.-: -- " '- . - . -" ' . - ° " ' 1
I"
@3 6
IIII
1.0 , ~~I Ill IIg
V1.2--5 11111 .6
MICROCOPY RESOLUTION TEST CHART
NA' A VpAV AN: k, :
.- I
AFCSR-82-0298
Analysis of the Dissipationmechanism
within an Instationary Cathode Spot
of Metalvapor Arcs
M01.0 H.Q. Schrade, H.L. Kurtz, M. Auweter-Kurtz
CNI and M. Kirschner
IRAINSTITUT FUR RAUMFAHRTANTRIEBE
UNIVERSITAT STUTTGART
Ij 4s 0 i
PFAFFENWALD RING 31
* 7000 STUTTGART 80 (VAIHINGEN)
dlsr~!:~o~lnlq>8 4 1 2 _7 0 9 33SN~dXJ 1N3VYNH3AQOj IV U3JflUOid3I8
. . .. .. i
AFOSR-82-0298
Analysis of the Dissipationmechanism
within an Instationary Cathode Spot -.
of Metalvapor Arcs S
H.O. Schrade, H.L. Kurtz, M. Auweter-Kurtz
and M. Kirschner
0IRA 84i IB-6 ii
August 1984 nT 1.e
AInterim Scientific Report 0
Approved for public release; distribution unlimited
prepared for
Air Force Office of Scientific Research
Bolling AFB, D.C. 20332
and S
European Office of Aerospace Research
and Development
6 London, England0
REPORT DOCUMENTATION PAGE F: IN 7
rE P: r' C\ 7 AC E SSI:% NO V E'-Pi EN- % v i
AFOSR.TR. -4L & -2~7~~774 E &.' 5..bfm: F R F-6 rE
Analysis of the Dissipation-echanisnANiA
wi thin an Ins tati onary Cath,- e Spot 01A,1S : .c. lo f ME talvapcr A r cs 1ER00N GrEV rE
IRA - S. lB68 C~oN.PA;' Gr&N- %-~t
H .0. Schrade, H.L. Kurtz, M. Auweter-Kurtz AFOSR-S2-0-198and M. Kirschner
PC E %:L NN A,,: A:-1rESS PP :;Ak E-E%EN P .E. -A!-AZZ A V ':RK~
Institut fiir Raumfah-rtantriebe 61102FUniversit~t Stuttgart, Piaifenwaldring 317000 Stuttgart 80, West Germany - I_______
IRFORCE OFFICE CF SCIENTIFIC RESEIRCH,/N!jAAuut 17, 1B T LINO LFB D 233748I3-~Z
'A M. % A- :.7- %Avr' 5 A -.e S I~!rler f- C,:-<'~j U- :Sz.., .AiE S
Urciassif ied
__ __ __ __ __ __ __ __ __ __ __ __ __ __ _ e~ AS$ ;C A- DZA% ,A 'N
.A PprCvocd Fcr puiblic rice :z~uir r1it~
P B D.~'; A 7-M EN "-a sa eere - Blo, A 20. I di!f.e&nr from- Repor.)
10 S.PP-EmENAR N -ES__ _ _ _ _ _ _ __ _ _ _ __I
'9 KEY *ADPS (Cor:Mr. r -t' r* - f3 re~es arl rdc jden'!(3 t- blc-h num-tr,
4 Cathode spots, vapor pressure, energy dissipation mechanism,
d ten h ca t t r ar. sfe r.
2~ A B,17 R A -'~t or, , e C. side If r,.. essar) and fdewjfs- bv block numnber)
The paper deals with the dissipation mrechanism,- of microspots oil
lo~w prEssure arc cathcdes.
Based cr. a nEw YFD stabili tv theorv and or, an exeILnal: un
em7pirical relation between spot current and crater radius orn purt.
copper and mnolybdenum cathodes, the time behavior of pressure and
DD ~~ 1473 EDION OF t 40V 65 IS OBSOLETE UCAS~E ___
Q-TV g~ CL jSSr' A'N Or ImIS PAG E a £rtrJ
SELRT Ty CL.ASS ,CAION OF TmIS PAGE.'14"IeV, Data nrervd9
temperature within the spot and the appertaining voltage drop
have been calculated. Emphasis has been placed on finding an
analytical solution for the time dependent heat flux into the
hemispherical crater surface as a function of the expansion
of the spot.
U:N(TL.'- SSI FIED
SECRe ITY CLA$SIFICATIOIN OF A - i E'A"' em :',&'f E
' - --V - - i - : - - / . . , . .. .
... * .I , " . . . . .. . . -" . . .. .. _ . _ . -
t7~ - ~-- --- ---- ---
-I-
PREFACE 0
This report deals with the time dependent energy
dissipation mechanisms which should occur in micro
spots on cold metal cathode surfaces of low pressure 0
arcs.
The Appendices A and B contain especially the
mathematical way of how the analytical solution of
the time dependent heat transfer equation and the
formula for the time dependent heat flux for a
growing spherical spot has been found.
0
This research has been sponsored in part by the Air
Force Office of Scientific Research/US Air Force
under Grant AFOSR 82-0298.
/
Ai
0 0'
-*E
* .
6 °
- .o
TABLE OF CONTENTS
Abstract 2
Survey of Pictures and Diagrams 3
List of Parameters 4
Introduction 7
Hemispherical Crater Model 8
Analysis 10
Results and Discussion 26
References 32
Appendix A 33
Appendix B 40
0
* 5°
0 0
-2-
ABSTRACT
The paper deals with the dissipation mechanism of
microspots on low pressure arc cathodes.
Based on a new MPD stability theory and on an experi-
mentally found empirical relation between spot cur-
rent and crater radius on pure copper and molybdenum
cathodes, the time behavior of pressure and tempera-
ture within the spot and the appertaining voltage
drop have been calculated. Emphasis has been placed -
on finding an analytical solution for the time depen-
dent heat flux into the hemispherical crater surface
as a function of the expansion of the spot.
*
* 0. ''
* S]
* 0j
0- I W."-. 7'
SURVEY OF PICTURES AND DIAGRAMS
Fig. 1: Hemispherical crater model 9
2: Zone I: Current carrying plasma channel 11
3: Zone II: Hemispherical crater 14
4: Vapor pressure curves 15
5: Zone II and III: crater and molten layer 17
6: Zone III and IV: Phase area liquid-solid 19
7: Zone IV: Solid metal cathode 21
8: Temperature distribution 23 "
9: Function y 24
10: Spot behavior (cold copper cathode) 28
11: Spot behavior (cold molybdenum cathode) 29 '
A 1: Temperature distribution around a ball with 33
the surface temperature Ts
~4
° .
LIST OF PARAMETERS
Ac Surface of the melting within the crater
A s Phase boundary area solid/fluid within the
crater
Az Crosc section area of the channel within the
plane of the cathode surface
Magnetic induction
c Evaporation constant
C 1 ,C 2 , Vapor pressure constants
C3
Cf Specific heat of the cathode material
EIntensity of electric field
ec Uit vector
ec z Unit vector
f Factor depending on current density
distritution 0
f(T) Function
h Enthalpy
I Current e
Current density
M Molar mass
rMassflow
m 0 Mass of an atom
P Pressure tensor
Pc:PD Vapor pressure above the melting
S
Pz Pressure within the area Az
p Ambient pressure
Qs Heat of fusion
Qv Heat of vaporization
q Heat flow
qc Heat flow into the area Ac
qf Heat flow out of the phase area As-tas9 Heat flow out of the melt into the phase
area As
R Universal gas constant
rs Crater radius of the phase area As
rso Radius of a sphere
s Velocity of the phase area As
ro Crater radius of the phase area Ac
S Radiation
Radiation into the melt
T Temperature
Tc Evaporation temperature = temperature of the
gas within the crater straight above the melt
Temperature at the surface of the sphere with
rs
Ts Melting temperature
T ,T* Cathode temperature at infinity
T*,Ts* Substituted temperature
t Time
Uk Voltage drop at the cathode
-6-
U Voltage drop in zone IIII .
UII I Voltaqe drop in zone III
V Volume
v Velocity
V c Exhaust velocity of cathode material on the
crater surface
v z Velocity component in z-direction within the
area Az
x,x I Mathematical variables
y Function
,- f Matter constant of cathode material
* x) Function
9-crit Stability functions
Heat conductivity of cathode material
..0 Magnetic permeability of free space
f Thermo diffusion constant of cathode material
Density *1c Gas density above the area Ac
Density of cathode material
Electrical conductivity of cathode material
Time
Potential function
U x) Error function
e Work function
* . S
INTRODUCTION
According to todays knowledge the attachment region
of a low pressure arc or a so-called metal vapor arc
consists of one or more bright "spots" of very high
current-density ( 1012 A/m 2 and even higher). A vapor
jet (containing even droplets) and a current-carrying
plasmachannel emanate from each of these microspots.
The discharge channel which electrically connects the
cathode with the anode or with the main inter-
electrode plasma region is therefore exposed to a
very strong axial flow.
I4
These microspots are highly nonstationary, they
propagate and disappear and, depending on the
electrode material (oxidized surface layer, surface
chemistry), surface roughness, overall temperature,
and type and pressure of gas, these microspots can
cluster together to macrospots of different sizes
witn different erosion rates.
This report presents an analytical investigation of •
the instationary energy dissipation mechanisms within
an individual microspot on a pure copper and
molybdenum cathode. It is based on the knowledge of
an experimentally found, empirical relation between
spDt current and spot size and assumes a strong
enough axii ow in order to maintain a stable
straight discharge channel.
* -5
HEMISPHERICAL CRATER MODEL
The following analysis considers a simplified model which,
*however, according t o experimental findings 3)4% i s
quite reasonable. At first the arc attachment region is
divided into five different zones (Fig. 1). 0
Zone I is the current carrying plasma channel which
electrically connects the low pressure plasma of the
interelectrode space with the crater zone II . This latter
zone contains a complex , turbulent mixture of evaporated
material like atoms, ions and electrons under fairly high
pressure. Adjacent to the high pressure zone II is the
molten layer zone III, which is connected to the solid
cathode zone IV through a hemispherical phase transition
area . Zone V is the gas and vapor region outside the
* current carrying channel and is characterized by the fact
that n o, or only an insignificant, current density is
* present here.
Cathode material is evaporated out of the crater which
forms a gas and vapor jet, causing an axial flow within
the current carrying plasma channel (zone I) and a
diverging flow in the non-, or only weakly, electrically
conducting ambient zone V. As often observed in experi-
ments,this jet can contain droplets which indicate that
* the surface of the melt within the crater is boiling and,
as will be shown later,that due to an explosively
increasing pressure in the crater (zone II), molten
m ma ter ial1 is pressed andJ e jec t ed out around t he crater
rim. The behavior is indicated by the protruding rims of
t he craters l e ft on the surface of cold metal cathodes
that have been struck by an arc.
, 4l
-9-
CURRENT DIRECTIONI "
EVAPORATED
CATHODE MATERIAL
,' DROPLETS
ZONE I CURRENT CARRYING PLASMA CHANNEL "
ZONE II HEM SPHERICAL CRATER
ZONE III MOLTEN METAL LAYERS _ I
ZONE IV SOLID METAL CATHODE
ZONE V NON CURRENT CARRYING GAS AND VAPOR REGION
S Fia. 1: Hemispherical crater model
- 10 -
In our analysis we assume a hemispherical crater, neglec-
ting this crater rim and the thrown out droplets. However, -
since these droplets have a minor effect on the energy
balance, and since the crater radius follows as a function
* of the spot current from empirical relations of experimen-
tal observations, one indirectly accounts for the material0
* loss due to this rim and droplet effect in zone III any-
* way.
Now we will investigate the zones separately and match the
0 boundary conditions in between. Thus we obtain an integral0
view of the energy dissipation and at the same time the
voltage drop Uk between the plasma channel (zone I) and
the solid cathode (zone IV). It will be shown that this
C voltage drop f its very well with the cathode drop of a
- , -spotty arc discharge. .1-
First we look at t he current carrying plasma channel,
-I.+
zone~~~~~ ~ ~ ~ ~ I (Fg 2) omrivsiaiosi r tblt
and f 0.8 w
which means that the axial impulse transport within the
discharge channel rve'aer has to be larger than
Z Jd 4* 0 4 "J
hpea of J
ta0bevtosoeidrcl acut o hemtra ,
CHANNEL AXIS
GAS AND PLASMA FLOW
xa
DISTRBUTIO
\ ii
THE ELECTRICALLY CONDUCTIVE CHANNEL is 1 KINK 1 UNSTABLE
AND BENDS MORE AND MORE IF:
WHERE /
AND JA',
Fi.2 Zone I: current carrying plasma channel
1% V I. .- V* - S
- 12 -
For a pure copper cathode, we know from experiments 7) ".i
that / 0.12 (m/A) . Using this we can neglect
the second term within the curved brackets as long as the
constant outer magnetic field is small enough (Bo _ 1 T)
so the criterion above is reduced to r O.
Utilizing the momentum equation for the crater zone II,
the vapor pressure Pc above the molten layer can be rela-
ted to the axial mass flow (Fig. 3). Neglecting the ambi-
ent pressure p , the criterion above can be expressed as
OF r A ~ (-3
which means that in the case of our hemispherical crater
model the vapor pressure above the melt has to be about
3 times the average magnetic pressure 1I
If this condition does not hold, a small disturbance, ]
which means a small curvature of the channel axis, willbend the channel more and more until it touches the
electrode surface at another point and a new spot is
created. Fig. 4 shows the vapor pressure curves for copper
and molybdenum with another ordinate, the product of
current and current density, Ij, with dimension A 2 /cm 2 .
For Ij above these curves the current carrying plasma
channel is unstable; correspondingly if Ij is less than
the value given by these curves we have a stable discharge
channel.
With the assumption that the mean evaporation rate fromthe fluid surface Ac within the crater corresponds to the I
vapor temperature Tc, which is the temperature of the gas
within the crater straight above the melt and is
77 (5)fr A - -
- 1 3 - - -
the criterion (4) becomes
with 0
c- 'S[ L ]7
M is the atomic or molecular mass of the evaporating
cathode material.
The minimal evaporating rates for one spot at a pure •
copper and a pure molybdenum cathode yield to
(6a)'foS
Pio ;r J, .60 bL34 ( 6b)
with I in amperes and T in Kelvin.
These results are now used for the analysis of the dissi- I
pation mechanism together with the justified assumption
that the voltage fraction of the spot is just high enough
to obtain a stable spot.
First the energy equation for the crater zone II is formu-
lated, which means that the power converted within zone II
given by the electrical power input*- -
" " 0V
Pao dAz
FLUID 4
CATHODE
CRATER S
4
- -_ - .0FORCE BALANCE: 4 ,D+ TdA =J'x ~
UNIT VECTOR
0Az Ac 1
P AND T ARE HOMOGENEOUS ON THE INNER SURFACE OF THE CRATER -
dIc e d A,& 022 1Z{ f- 7.
,V 2 aAz p -Ac - g - 0,227,
STABILITY CRITERION: S
Fig. 3: Zone II: hemispherical crater
1,[A 2/cmrn2 E 3 P CBARJ
C,: 2,2El 0C,=: -1,21
C -- 40 341.3
I E2
1E9
IEI
COPPER 1E8
* MOLYBDENUM
C4:: 9 429E10C2:=-1.148C3= -7 6 42 6.1
1 E- 1
1 E2
1E5
IE-3 *
2000. 3000. 4000. 5000. 5000. 7000.
T EK)
Fiz. 4: vapor pressure curves for copper and molybdenum
PC* C2 e(C 3/7)
accordi4ng to 10) and extrapolated for pressures
,LO above 1 bar
- 16 -
must be equal to the difference of energy which leaves and
which enters the crater zone per unit time. _
Within this calculation the purely convective part
and the parts due to heat transfer and radiation have to
be separated. The convective energy flow can be expressed 0
as inh, where ; is the evaporated mass through the crater
area Ac. Because of the continuity equation, m is equal to
the massflow through the circular crater orifice area Az .
-"h is the enthalpy difference of the plasma-vapor mixture
between the orifice area Az and the region directly above
the melting zone III. Because the mean pressure and
temperature at both areas are known, this enthalpy
difference can be calculated. In this evaluation the 0
resultant heat conduction and radiation loss through the
area Az can be neglected with respect to the convective
effects. The radiation and heat transfers (which include
ion bombardment) *hrouarh the area Act however, contribute 0
importantly to the heating up of the melting zone and to
the evaporation of the cathode material.
0 The energy balance of zone III can be written in a similar _
way. The electrical energy (I U III) within this zone
equals the difference of outflowing and inflowing energy
per second.
*.-0
* S
* 0
- 17 -
- CURRENT CARRYING
V PLASMA CHANNEL
A
A c : •7 A CATHODE
.. . A s( qs
ENERGY BALANCE 11I: q h (7-d ) + . ' ' " A f 'Z (
Ar_#A z Ac4,Z P
7h Ah + 2rr 7FS, 4 Sc/ -f v, gvdA4~z
ENERGY BALANCE IIl:; 7 4, 2 7(~4~ 2 712 + Sc.c
* 0
S U M I I A N D I I I : U = L 4' ,
F2 .ar an ml
0 Fig. 5: Zone II and III: crater and molten layer0
- 18 -
Two other terms have to be added in the energy balance of
zone III: mQv because of the transition of evaporated
material to zone II. and I e because of the work function
of the emerging electrons (Fig. 5). The energy of the
particles which flow over the crater rim and that of the
ejected droplets are neglected. It is a main fraction in
respect to the mass but a minor fraction in respect to the
energy in this zone.
The energy balance of the spotty cathode voltage drop is
the sum of the voltage drops of the crater zone II and the
liquid zone III (Fig. 5). This yields to
17r+
where (2 -r- s ) is the energy flow per second through the
phase area solid-fluid because of heat conduction.
The evaporated mass m cannot become larger than the molten
mass in the same time period; this leads to the require-
ment:
S(10)
From Fig. 6 the energy which causes the melting of thf
solid material can be calculated to
where Qs is the heat of fusion and qf the radial heat flow
directed from the phase area liquid-solid to the solid
zone IV.
V - SK _.__
- 19 -
• ;, .;.SOLID CATHODE•
SS
PHASE AREA r
2 Of
REQUEST: ,, 27r I25- --2., --
INTO THE ENERGY EQUATION
U r-; ) , t •
U A + 0t ) +
Fic. 6: Zone III and IV: phase area fluid-solid
200- 20 -..
With Eqs. (10) and (11) and the results from Fig. 5, the
heat flow in the liquid zone can be expressed with respect _
to the heat flow in the solid zone. Therefore the cathode
voltage drop follows as:
'rJ I ) # rJ f ) (12)
Before looking at the heat conduction qf within the solid
zone IV, the limitation of rb caused by the stability cri- •
terion (6) and the above requirement (10) will be added.
On a cold solid metal cathode the evaporation rate m with-
in a spot is limited by
To calculate the heat flow q , the heat flow equation S
within zone IV (Fig. 7) has to be solved. Therefore two
assumptions are made:
a) the heat transfer coefficient f respectively the ther-
mal diffusion coef fii'_ ent f within the solid is constant 0
and
b) the term fj 2 , which takes the ohmic heating within the
silid into account, can be approximated step by step by a
potential function with .f 1/r.
Hence it is
r
"
I - 21 -
qff
E SOLID CATHODEAREAS OF CONSTANT TEMPRAURE
HEAT CONDUCTION EQUATION OF ZONE IV:
T(r2 7rr 2fr a
C :CONST,. J z_
..pc Cf x (
INITIAL CONDITIONS:
0 (t 00
05 P
BOUNDARY CONDITIONS:
r r5 T'r T9 CONST.
r - ~T~r~ Tce, CONST,
Fic;. 7:- Zonec IV: solid metal cathode
- 22 -
in the region the potential function is
given as _
- r (rdE704s
S
Stin CAS I thr olw•eas
of g ¢c16)1
which is in fact no potential function. But within a small
r e go. I y' o r- -k the term (14) can be approximated by 0
where - is nearly realized This can be done within
all1 regions 1; 2 I23.. With this step by stepwhrich tio of f j2 by a potential function Bu the heatflow equation can be rearranged c
Replacing T by iii i i.!
where T* is te new independent variable bdn wh
the new differential equation follows as
T 7-
PT( 18
-. -- - -;-~'--*~---- -
-23-
SOLUTION T T (r,t) IN THE RANGE 0;
rs :s r< -.5-- o
T -Too F- ( T r. ) I " C ; :
r-~ r -5
S. 0
tz
X26 () t 2o x•TT-= _ _ rs_/~d _
= I
0t
6(t) fs -' x)iI2x]00
rrST
rs t, t t •
Fi.8 eprtuedSrbto
00.
450. . ......... .........
40... .. . .
.0 . ... . . . . . . . . . * . ... . . . . . . . . .
I20 . .. . . . .
10 . ... . . . . . . . . . . . . . . . . . . . . .
*0 . .... ... . . .
0.0 .5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
Fig. 9 Function Y.
* 0
- 25 -
which is solved for the given initial and boundary con-
ditions in Appendix A. Equation (17) leads to the solution - 6
for the temperature T(r,t) which is shown in Fig. 8. Using
this solution T(r,t) the heat conduction loss can be cal-
culated by forming suitable gradients on the solid side of
3 the phase area fluid-solid (r=rs; exactly: r=lim (rs+Lr). 0
_r-O
In Appendix B the heat flow into the solid zone IV is cal-
culated taking care of the fact that at time t:O the ]
initial radius is rso=O;
N T {-19)
The function 2 ( t- .. 4 ) is displayed in Fig. 9. Taking
the rises of current I(t) and spot radius rs(t) from ex-
@1 periments 2) Eqs. ? allows the calculation of the time •
dependent heat flow into the solid electrode. With this it
is possible to determine the cathode drop voltage Uk
(without inductive part), the minimal vapor pressure pc
0 and the spot temperature Tc as a function of time using 0
Eqs. (12),(4) and vapor pressure diagrams (Fig. 4).
* .
*• i
*• 0
- 26 -
RESULTS AND DISCUSSION
It is known from experimental investigations 1),2),11)
that during the first 3 to 5 ns of a pulsed discharge with
a current rise of about 1 to 5 *109 A/s, a small circular
melting point occurs on the pure metal electrodes which
afterwards forms a crater; then the crater rim expands
with a velocity of 104 to 5"104 cm/s. After about 5 to 40
ns the final crater size is reached which means that the
current transportation is incured by a new spot.
On pure metal cathodes the crater radius rs increases with
the current I S . Experiments lead to the empirical
approach: (20)
with pure copper: ro=1,7 2:m and I*=83 A
with pure molybdenum: ro=1.2 ;im and I*=33 A 2).
It is now possible to analyze the time behavior of the
dissipation mechanism of a cathode spot and to calculate
the purely ohmic part of the cathode voltage as a function
of time if one takes the spot current I s as a given quan-
tity which linearly rises with time.
Figures 10 and 11 show the results of these calculations
for copper and molybdenum. In both cases the characte-
ristic traces of the curves are similar to each other for
the vapor pressure pc, the surface temperature Tc and for
the ohmic part of the voltage Uk.
During the initial phase the pressure rises explosively
and sooner or later, depending on electrode material and
current rise (dI/dt), it reaches a maximum. With copper
- 27 -
(dI/dT 2.67 109 A/s) this maximum of about 160 bar is
reached after about 30 ns; with molybdenum (dI/dt
8.109 A/s) it needs only about 3.7 ns to get to the maxi-
mum of about 55 bar. This very significant, explosive
pressure rise on the inner crater surface gives an expla-nation for the fluid metal which is ejected and which 1
forms a crown like rim.
Initially the temperature rises steeply, later on it rises
more slowly up to nearly 5000 K for copper, and nearly S
7000 K for molybdenum. Nevertheless the temperatures do
not reach the value which would be needed for a pure
thermionic emission even in the case of molybdenum.
To explain the very high electron emission on the cathode
G. Ecker postulated a combination of thermionic and field
emissions 12). This seems to be possible because new
experiments 13) show field amplifying "Taylor cones" at S
formerly plane fluid surfaces. With this explanation the
emission of the charged particles is not uniformly
distributed on the inner fluid crater surface but will
occur at one or more "Taylor cones" which emerge and S
vanish at the fluid surface. Therefore only the time
average of the emission is uniform over the inner crater
surface.
A complex, turbulent high pressure plasma exists in the
crater-zone where the Debye-length reaches quite possiblyonly 10 - 8 to 10-7m and therefore the electrical field at "
the tips of the "Taylor cones" could be some 109 V/m.
With that, the relatively high emissions of 108 A/cm 2
(and more) can be explained. A further possibility would
be the evaporation of negative ions out of the tip of such
Taylor cones, similar to those observed with positive ions S
- 28 -
UKIrVolti Fig. 10: Spot behavior (cold, pure copper cathode)
20K 5000 K-
20 T0
15 15-a
3500 K
10 100 bar THEORY
UK0
1 5 10 20 30 [ns)
J0
rSSrsI[A] 100
8,, gO rs : .:HA) I -
50
EXPERIMENT
* A- (GUILE-JUETTNER)
, 320
5 0 2 0 30 In [ s] ..
-29-
Uk[Volt] Fig. 11: Spot behavior (cold, pure molybdenum cathode)
[Volt]
20
15
10 -- 550OK'
10 bart [ns]
.01I
2 3 4SK 1 2 3 4s
[A] EXFERIMENTr (GUI LE-JUETTNER)
50
K40 r
30S
20
-10
1101
12 3 4 5
- 30 -
on anode tips 14),15). The life span of such negative ions
within the crater plasma is expected to be short; they de--
compose rapidly into free electrons and atoms; through
further impact the atoms are ionized into electrons and
positive ions. These considerations are only speculative,
but they could yield a plausible explanation of the0
* experimentally found high current densities of about 108
* A/cm2 and greater.
The purely ohmic part of the cathode voltage Uk above the
spot , first drops ste ply from relatively high v alu es ,
* passes through a minimum and then increases slightly. It
* is of interest, that the high voltages at the start are
* caused mainly by the high heat-conductivity losses into
* the electrode material at the beginning. In the first mo-
* ment after the ignition of the spot discharge, the tempe-
rature decay at the phase boundary layer solid-liquid is
very steep (see Fig. 8) and therefore the heat flux into 0
* the solid cathode is very high. The actual ignition is not
described by this analysis . The ignition voltage however
* has to be surely greater than or at least equal to the
* purely ohmic ot. voltage Uk, calculated here, to init-
iate a spot discharge which leads to a craterlike cavity.
Also, the time when the channel becomes unstable is not
* included in this evaluation - the assumption was that the
stability criterion (1) or (14) is just satisfied at any
time. This means that the product of current and current
density Ij would reach its maximum at the same time as the
* vapor pressure P.. However, it is improbable that with a
further increase of temperature, current and voltage, the
emission ~r the average current density j at the liquid
surface of t he crater would decrease. An increase of
- 31 -
,j would however lead to an immediate falling short of the
stability criterion (4) and therefore cause the kinking of
the discharge h al ine 1 5),16) . A s a consequence o f t h is
plausible discrepancy, we have to expect the extinction of
reaches its maximum. In our calculated examples this maxi-
mum lies at ca. 30 ns and at a current of 80 A for copper
and at ca. ~4 ns and 33 A for molybdenum.
- 32 -
REFERENCES
1) B. JUttner "Beitr. Plasmaphysik" Vol. 19 p. 25 (1979)
2) A.E. Guile and B. JU ttner "IEEE Transactions on
Plasma Science" Vol. PS - 8 No3 p. 259 (1980)
3) J.E. Daalder "Diss. Thesis" Univ. Eindhoven (1978) 0
4) J.M. Lafferty "Vacuum Arcs" John Wiley and Sons (1980)
5) M. Auweter-Ming and H.0. Schrade "J. Nucl. Materials
93 and 914" p.799 (1980)
6) H.O. Schrade,M. Auweter-Ming and H. Kurtz "Proc. XV
Int. Conf. on Phen. in Ion. Gases" P. 531 (Minsk 1981)
7) J.P. Datlev, A.E. Guile, B. Jattner "Beitr. Plasma-
phys." Vol. 21, No 2, p. 135 (1980)
* 8) J.E. Daalder "J. Phys. D", Vol 9, p. 2379 (1977)
9) D.T. Tuma et. al. "J. Appl. Phys. 49" p. 3821 (1978)
10) Hultgren, Orv., Anderson and Kelly "Selected Values of
Thermodynamic Properties and Alloys" John Wiley and
Sons Inc. (1963) 0
11) E. Hantzsche, B. Jittner, V.F. Puchkarov, W. Rohrbeck
and H. Wolff "J. Phys. D." Vol. 9, p. 1771 (1976)
12) G. Ecker "Ergebnisse d. exakten Naturwiss. 23" (1961)
13) G.N. Fursey "Local Elementary Processes at the For-
mation and Operation of Vacuum Arcs"*)
14) R. Clampitt and D.K. Jefferies "Proc. of the third
Int. Conf. on Ion Beam Analysis" Washington, D.C.
(1977)
15) J. Mitterauer "Liquid Metal Field Ion and Electron
sources"*)
16) H.O. Schrade, M. Auweter-Ming and H.L. Kurtz "Analy-
* sis of the Nonstationary Low Pressure Arc Cathode •
Proc. VIIth Int. Conf. on Gasdischarges and their
Applications" London (Sept. 1982)
17) L.D. Landau, E.M. Lifschitz "Lehrbuch der theoret.
* Physik VI; Hydrodynamik" Akademie Verlag Berlin S
p. 230 (1966)
*) presented at the Vth Workshop on Electrode Phen. Berlin
1982; organized by Zentralinstitut fUr Elektronenphysik,
* Academy of Science of GDR 0
Li __
*] - V -,, .-; , - - . . *- * *' .. . . .- r --
I 0
- 33 -
APPENDIX A
Before solving the heat flow problem as demonstrated in
Fig. 7, the following problem must be considered:
During t<O, the temperature within the homogeneous medium
surrounding a sphere with radius rso is set to be T . At
t 0, the temperature at the surface of the sphere jumps
suddenly from T. to T o = T s and for t>0 it increases
K steadily, corresponding to To(t.O) Ts+'Tso(t). Beginning
with t = 0, a homogeneous, radial current flows through
the outer medium, which leads to additional ohmic heat
sources. The heat flow equation describing the time
dependent spacial temperature behavior is given by
X& ( r 7v, (A 1
A (, A
-- I
II
T. W
Fig. Al: Temperature distribution around a ball with the
surface temperature Ts .
* 01
- 3 4 - 0
with the conditions:
ir 77
7T(r -- )P vo 1 7,-.r'-'_
Solu t ion:
- The term j 2 with is represented within each
hemispherical shell by the potential function
e- su"tsi t ut in: . ' '-,,)-T'(i/ J4J' tA3~ S.t
0
~ r ~z Ir2 (pe/ 'd
th! h a o on t n eqiat ion becomes
ra Pr.
witn n the t4mc ir -erval the conditions are
- < 7 -o T* ,.) . -- , ) .
o ,. < , T'( ., -T + 7-., €).
W; 7j- 4)h-
S-
- 35 -
The solution of this problem is found with Landau-Lif-
schitz 17) as _
7-A 0
where
00
With this approach the temperature field is determined if
the temperature T*(rso, ) on the surface of the sphere is
known at ar.y time. The location of the temperature front,
T., at any time is known by experiment. Therefore we try
to reduce the time dependent temperature behavior on the
surface of the sphere to the behavior of the temperature
front T s . 0
Choosin7
#.,(r" -- IAo
r-r,,
the solution witn - (A7)
results in
76*1 7o-.~~ 7T-r 4 J, ;T (tifzf .
(A8)7t ,,s A )</,_
.- 36 -
kL
where l ftc/y is the error function.
With the substitution-_
_ _ _ - _ _ _ r -r .t " d u - c) ' / 'Y JqC (l -) l t a z (A9)
and using (A7) r*rc'yi = r,# (Aloa)
the following solution is obtained:
(All)
Olt-
As can be shown analytically with the linear approach
A11
(A8) or (All) are the solutions of the problem. In the
same way, at any time r4.+4# ( small), the function
0 f(-.) can be approached with S
This yields also the equation (A8) or (All).
01
37 -
Substituting f(T ) in Eq. A(11) with respect to A(13) we
obtain
T( -r , f) 7o a T - Jo ,0 f& [ [ )- ,r
or by using r/ * S/I
and calculating ( ) with (A10)
(A15)
where
. ¢(A16)
and
I lit'
-38-
k o and Mi have to be chosen at any time with respect
to the condition that the location of the temperature -
front at time t in the surrounding medium is the experi-
mentally given one. So we obtain from
and with (A15):
/ (A17)
with
w h 7 (A18) - 2
Following from (A17), because of t 0, rs = rso, and Sxs o, -2J '
that is, ko = 0 (A 19)
In the same way we obtain with (A17) and (A19)
-. - _ -_ _ - -(A20 ) _0,lr(As) 4.L"?
Using this together with (A19), Eq. (A15) yields the
temperature distribution, at any time, and at any place "
* 0
too
0 f n0) rOwe _7"_- -
*j - "S
- 39 -
* where
XISS
((k ~~ 4)] 4 ~px) - x
4 #(x)~AfeX
- '. .. ; =. , .- % . -. r , - - " 5" : . - - ; -- o . " - . . ','.
400- '40 -
APPENDIX B
Calculation of the heat flow 2-r qf
The heat flow at time t through the surface of a half
sphere with radius r( > rs ) into the colder solid body
leads to
T(, (Bl)
The heat flow is obtained as a function of t with the
passage to the limit:
2r (e '') Tp2~ - l/ -T (B2)
T(r,t) can be calculated with Eq. (A21):r
~ Tri #() (Kj L-4 r)
Wi th L
..,,,- "-) 5e" (.4-1" °" • . .
L L
U.)
.~ .f. r. A LO..K d.f,-,<)]- Ae,.
In addition is
F-, ' £, ,fd .-, g,.
J/ 1.~q~J~/L ~4.Zi-
* . *. -... I
41
because off- (potential function) and therefore
ir L (- /A) [-- -Or 761 "p- ,P-/"'' ( ) - O 2 -,,"'<e" O.] ° o. - ,_I
for r- rS -p x-* x S results in
opt (7-7-)f2 /,,:'t, 7-/_^/, r.
z',, N g- .O"
KAW
"~ ~( 7-,," (7j, -/.- ./^-v, .•,- .. I /r - .
i"T} K, e ?CLIo -e6X)O (,- t6, (s - 7-
S c,7..) - A,.'e ,e• t
- 9 7j. , - ..-
_ [,, .7,, t . (<,----.j4)/
* */- ,'- 'i... -
4 .
.16.
(72 -70- 7- -- i
ge c
+1( - Or / -- -
j e
-- - "
'2(d ) -A)[ 4 L - - .rS
Therefore the hea" 'l"w follows as
+ Z I €C,<, - i + S7,< ) - /
- , -A , ,, o .
0
S S•
jl (7 -- 7o,)'+,('': ){<"° ""/'(B3)
Th~~ref'orer th et l'llor Ewe
,'., % m ", t & ,.)[,++ .,. z r7-j -0-c 1
a FILMED 4L 2-85
.i DTIc -*.~