General Disclaimer

One or more of the Following Statements may affect this Document

This document has been reproduced from the best copy furnished by the

organizational source. It is being released in the interest of making available as

much information as possible.

This document may contain data, which exceeds the sheet parameters. It was

furnished in this condition by the organizational source and is the best copy

available.

This document may contain tone-on-tone or color graphs, charts and/or pictures,

which have been reproduced in black and white.

This document is paginated as submitted by the original source.

Portions of this document are not fully legible due to the historical nature of some

of the material. However, it is the best reproduction available from the original

submission.

Produced by the NASA Center for Aerospace Information (CASI)

lwNz"' DATEs

s)C.AF

i-%ECEIVED BYESA . SDS

(1EE; i37s

;Y

I TYi..• j , J S t^A

^ T '.SCIE IC NOTE

X

r

i

ESA SN-124

September 1976

CARBON DIOXIDE AS WORKING GASFOR LABORATORY PLASMAS

(,4A5A-CP- 14 Q 893) CARBON DIOXIDF, AS WOPKTNG N71-20885

G4S FOR LABORATOPY °LASMAS (TPxas Univ.)24 P HC A02/MF A^1

'inclasG3/75 21752

R, Kist*

Itwitul /fir 1'hYsiculisc •he Weltrawn lorwhung

Freihctrk, Wesl Gerniun_v

l*ESRO Frlloit at University of Tevas in 1y74i

d

EUROPEAN SPACE AGENCY

AGENCE SPATIALE EUROPEENNE

t i

ES

, 5

"",1

24

A~'

Cfle

e !)

pal

qle

I u

.rup

een

nc

KIS

T.

R.

(ln

sulu

l f.

Ph.)

s. \

\cll

n·l

m(o

rsch

un

g.

nC'

.bU

Tg.l

Le g

:u carbu",qu~ l.'

Omm~

<ou

rer

de

pbsm

.a.s

.u b

bo

utO

lfC

. ~p,cmhrf'

Iq1f

J .\

+ 2

0 p

a6f's

. 16

fig

u f'C

1.i,.

t-"

:lngt

:us-

Dc\

me"

",,,,

, o

nl ~

Ie

1 ... 11~ a

r.au

Je .

run

e ~"

'nde

H

r. d

'un

.Jl

:dY

'leur

a p

ute

nll

el d

e fR'lfU~ e

l d

'un

"p

e1:l

rmn

dre

d

e R

U'-W

.l

Jn~

UO

it \."

fh:r

tnh:

a

pb.. ..

m.1

tn

ulih~nt

t'IU

f a

tQur

d

u p

z

"'ar

hu

nlq

ue.

de

rU'H

e, d

e r.

1l!

,'n C

'f dl

.· n

.e·h

um

:lh

n d

r "n

mp

uef

I~'\

C.1

u'te

nstJ

qul.

~ d

r '-

e5

dl\!

"'f"

p'!

po

ur

1\.h

t..-n

lhJn

dl'

pla .

. m:1

.. en

!;3

bOr"

'lnu-

e. O

an) 1\

'n~m

"'I~

. o

n c

on

sUl ..

. qu~ ~ p

:r

o.: .. :u

h.'lm

qu\·

d.v

nn£'

d.~

" pl

us (

on

t-.

den!

o&te

s dt

" rU

mlJ

. id

e p

lu§

1J.lb

i,,'"

pre~~I')R\.

de p

; ne

'J!r

e en

m

...·n

le

tem

p:-

qu·u

n.:'

plu

) g

nn

de c

flm

po

san

te ~, ..

ellb

tnne

dl."

b p

(tp

ub

uo

n d

.. ~ul.naqu.:

.1\"'r

C'

un

e (t

'mpi

.:n.

tun·

l:k

ttro

Ulq

ue f

tl()

I!l.

.. he,

rS

'S'·

124

\0:n

cc ~P"

ltDl

e It!rn~enne

II: i~

T.

R,

(In'

1I1u

' I,

Ph~

), "

~ell

raum

fuc4

ioCh

ung,

Ire

-Ibu

rg.)

I e p

i c:

".t-

omlq

ue l

.'(lm

me

~iOU

f .. "'

" de

pb\

m:l

s.1I

u bho~tuu\'.

$ep

tem

bre

1 q

.. "

1\ .

.. 2

0 p

.ges

. 16

fig

urK

Jon

.n

gl ...

. s.

1 '"

Oc~

mew

",'"

on

' e

t,-,

f;Il

IC'~

ll

':lI

J~ I

fum

-o;

ool,h

: II

I.

,.ru

n aO

:lI)'5

eul

, pt

lfen

tk"l

de

freu

ugt!

CI

J'u

n ~pcl.lrnmclrC'

de

m;tW

" d.

J.n\

u

ne

encn

nll

: a

pb

"nu

en

uuh~nt

tou

r .I

lO

W

du

p

z

ar))

l,n

tqu

C',

de

ra7t

IIt.'

, ,k

r<1

r~l'

)n

C'I

dl:

I'h

chu

m .

1f.n

.I

e co

m p

arer

"':

~ C

.ll ..

dtr

tSU

qu

M d

e ct'

"'

dlH~n 1

21

pn

ur

J'n

bu

mll

nn

dt'

pl..l'i

m~s

en b

bor:

lWui

!".

lnn

< I

'enl

ic:m

hk,

on

L"t

IN-U

le q

ue

k-p

i

t:a.

rbon

tquc

don

nl"

de

JlIu

s fo

rtn

den

Sil

es d

t" pb

.sm

a a ~ p

lus

f:uh

le\ pr

e'\S

lun~

de

SJ.l

Mu

tre

en

mem

O."

tem

po;

qU'U

M p

lu'\ gJand~ C

tlmpn

gn

lr m

a",e

lhcn

nr

de

u p

np

ub

tll>

n el

ectr

on.lq

lh.'.

2\'«

" u

ne

tCnl

J".:r

.1lu

rt e

lecu

()n

• .;au

C' m

rnn

dre

.

1 88

t:S

A S

N-I

24

A

J!en

« S

pa.

ia'"

~ ..

KJS

T,

R.

(IR

Sli

tu,

f. P

ltys.

....

dto

.. m

fon

c .. u

••

'-ro

bu

.rz)

L

c D

U c

arb

on

tqu

t ro

mm

t mu

ttT

de

pb

sa

u •• b

bo

alO

ffC

. ~Iembre 1

916

IV

'" 2

0 p

ages

.. 16

rW

m:s

. Jon"'ll~

Iks ~fft 0

01 c

ia;

f",t

tS i

l'a

JCk

d' ...

. JO

tMIr

lit

. d

'uR

ana

lYW

'U1

~ p

t;le

,UId

..ac

fK~ r

l d

'UR

S

PC

(U

On

K'

t'

dr

MH

.'!II

t eb

as .

ne n

lttd

lt'

a p-

3nou

('

"R

Ull

ltgR

I It

tYf

..

10u.

J d

u J

.d

caro

on

Klu

r:. *

:a.

wle

. d

e r.

,p d

d

e 1

'Id

....

. afl

. dt

' ro

mp

Ud

k" aU"It"fl'Ctq~ d

lo (

C'''

W""

S g

al"

pour

I'o

btm

ho

fl c

it p

b.sm

;as

eft

ta

bo

n'o

tee.

[b""

l'~mhJ,-'. u

n L

O'b

Uh

: q

ue

k p.

.. C

lItlb

n.n.

qur-

do

• .,'III

! d

e pI

\G r

o,.c

cs c

It"IU

ui1:

de

pIuI

a:.

.1 de

(N¥

~ b

.t",

," pfl:»WQ~ d

e p

.. S

Wti

lh'

U

mC

lnC

tem

ps

qu

'u.

plus

gnR

lde

co ..

ptJQAt~ n

a.., ..

d.ht

·1,a

r d

e b

pt'p

uL.t

lOft

d.-

... lh

*lq

lM. ~"

n:

UiM

' l"'m~lure

declJoo.tquc~.

r.s \

SS

-I2

4

Ag

t'1't

oCe

Sp

aIG

Ie t

:.ur

opic

ent"

K

IST

. R.

. (In

stiI

YI

f. P

hys.

.... d

lnu

mrt

)l1

Ch

un

s. F

tew

bulJ

) L

r p

z QJbontqu~ r

om

lt1lr

IOW

ct d

r pb

..sa

as a

u bbnnIOtl~.

Sq

urm

brc

: 1

91

6

IV +

20 p~ 1

6 r'l

lues

.. l:.

n ut

;pus

.

1 ow

Des

~""t$ o

nt

etc:

ral

tes

11 ru

de d

·u .. !

Ond

e H

I.

d'w

n lI

rW}W

1Ir

•

pm-c

ftlll

ld d

oc (r

t."'

~8C'

c-I

d'u

n

spec

trom

et:r

e d

e ~

tUm

: u

ne

mcn

nlc

-a

pbM

Da

rn

ull

lIsu

.t

tou

r a

toU

T d

" p

i

c:u

bo

ruq

ue.

de

I'az

olt

... I

afJO

II (

"( cit

-l1

tdtt

8m .

... 'n

dt"

rom~.rn I

rs a

rxtU

lSllq

1lf

'S d

e ..

:~

dt"

(!t\

91

' po

w r

ob

tdll

ho

a c

k p

bsI

na

l. C

'a b

bo

olO

lft"

. ib

M r~~h ..

(lB

CU

tlA.U

C q

ltC li

t pi

acb

on

lQu

c do

tIne

<It

ph

il f

ort

es c

IeR

sIlU

de

pI;a

5nu

a d.

... pl

lU fd

lio

pf~ Jo

:o p

i M

'Ulr

c tt

l

mE1

M'

It'!m

ps q

U'U

M ".

.as I"

rtd

c r

om

po

un

lr m

nw

dhe

nn

t cit

-LII

po.

lpub

llO

fl t

.:kdl

l1fU

qUC

, iI

"n.

un

c te

"mp

er..

lulr

dec

uo

ruq

w motodr~

1 88

CARBON DIOXIDE AS WORKING GAS

FOR LABORATORY PLASMAS

hYR. Kist

biwitut .Jiir Ph_i •si(-alische Welirmon fi4schung

freihtiriz, West Gerntcttt_t,

( fhe vowk described in this rt ,port was perlim p ie •d during the vithor's tonlre of all LSlk'o

External Ve/lowship at the (.-itive•rsit.t , of Texas in /974).

EUROPEAN SPALE r GENCY

AGENCE SPATIALE EUROPEENNE114, avenue Charles de Gaulle, 92522-Neuilly, France

761520

ill

Ix ^

i`

TABLE OF CONTENTS

I, INTRODUCTION 1

2• GLNERAL CONSIDERATIONS 2

3. CONSIDERATIONS CONCERNING CO 2 AS WORKING GAS 3 is3.1 Vacuum physical considerations 3

};i3.2 Breakdown voltage and ionisation arobability 4

3.3 Relevant reactions of CO2 5 I

4. ML'ASUREMENTS 6

i,1r

4.1 Cryogenic pumping of CO, 6 I

4.2 Breakdown voltage 8 ' ^14.3 Electron density 94.4 Electron-energy distribution 94.5 The final laboratory plasma system 13

a 1

4.6 Plasma diagnostic probe measurements 14

4.7 Mass-spectrometer measurements 17 !{

S. CONCLUDING REMARKS 19 I`

ACKNOWLEDGEMENTS 20 i

J '

LIST OF FIGURES

IV

Schematics of the Vacuum system,Paschen curves for various gases (after M, KNOLL, F. OLLENDORFF, and

It. ROMI , Gasenlladungslabellen, Springer-Verlag, Berlin (1935)),

Ionisation probability for various gases (F.P. Lossing, AA Tickner, and

W.A, Bryce, J. Chem. Phys. 19, 1254(1951)).

Effects of sorption pump (SP) oil tank pressure p, .

CO2 -pressure in the tank for minimum and maximum setting of file iris dia-

phragm applying cryogenic pumping at LN2 temperature,

Breakdown voltage UBD as a function of pressure p, for CO 2 and N2.

Current/voltage characteristics obtained with the RPA in CO 2 (a), N2 (b), A

(c), and l-Ic (d).(a) Saturation current Isal taken from RPA curves versus cold cathode discharge

voltage UDC (with neutral gas pressure approximately constant) for CO 2 , N2

and A. (b) Saturation current /sul taken from RPA curves versus neutral gas

pressure p, (constant discharge voltage UDC) for CO2 , N 2 and A.

Plasma frequency JN in the tank as a function of discharge voltage UDC for

CO 2 and N2 at three different neutral gas pressures p, .

General three-component feature of the I/U characteristic as seen with the RPA

in CO2 and N2.

(a) Temperature components 7G and 70 versus pressure p for CO2 and N 2 at

constant discharge voltage UDC. (b) Temperature components 7e and 70

versus discharge voltage UDC for CO2 mid N 2 at constant pressure p,. 7. for

N2 is not included, sincethe 0-component could not be identified clearly

enough from the 11U characteristics.

Figure I.

Figure 2.

Figure 3.

Figure 4,

Figure 5.

Figure G.

Figure 7.

Figure 8.

Figure 9.

Figure 10.

Figure I I

Figure 12. Laboratory plasma system with plasma source and diagnostic probes.

Figure 13. Cylindrical RF-probe and principle of RF impedance measurement.

Figure 14. Langntuir characteristics for various values of the bias voltage U, of the plasma-

source heating circuit.

Figure 15. Plasma impedance magnitude of the cylindrical two-electrode system.Figure 16. (a) Mass spectra with 25 V acceleration voltage applied to dic electron beam of

the ion source. (b) Mass spectra with 90 V acceleration voltage applied to the

electron beam of the ion source.

CARBON DIOXIDE AS WORKING GAS

FOR I AHORATORI' PLASMAS

ABSTRACT

31caturcment y ►► ith a Rl probe. retarding poteuttai anal , t • ser and »lass spectrometer in alahnrator,r plasma taut, hair heen perowmed using the gases ('(1 2 , N 2 . A and M , in order to

compare their properties as working gases jor lahorah)r. ► plasma production 17te overall result

is that CO, kadv to Itigher plasma clettsities at hmver neutral-gas pressures as ► yell as to a larger

Maxwelhan component uJ the electron population. ivitile the electron temperature is lower than

that when .V2 , A and He are used.

1. INTRODUCTION

One of the tasks involved in a laboratory plasma-diagnostics protect at thc University of

Texas at Dallas WTD1 was to eduip a cylindrical vacumn chamber, 70 cm long and SU cm indiameter. with a plasma source providing a collisionless plasina with a Maxwellian, low-teri

perature velocity distribution of the electrons. To each end flange of the tank was attached abell jar, as sketched in Figure I. The general concept is tv produce a primary plasma

TANK

BELL JAR

P G P I

P2z i

j^ a5

l t¢ure 1. Schematics of the ractnon tYstem.

2 GRNUALCONSIDERATIONS

(discharge) in a region of high Pressure (bell jar) and let it expand through all into a

region of low Pressure (tank). The present scientific note is concerned with the preparatoryphase of the project in which different gases (CO 2 , N 2 A, Ile) were compared in order to learn

About their properties as working gases for laboratory plasmas.

2. GENERAL CONSIDERATIONS

'I'hc pressure p, am:' 1) 2 in the tank and bell jar, respectively, are governed by the followingequations (see Pig. I):

2Q

PI = -S (1)

AP = P2 Pa = G

(2)

P2 = Q (S + d) (3)

Q is the gas stream in Torr I V introduced into each bell jar, G is the gas kinetic conductivityIn 19" of the orifice and S is the pumping speed in i V applied to the tank. For S»G we get

p2 ^ Q/G r Ap, i.e. p, «P2 . The first problem is to produce values for p, and P2 such that

P2 is sufficiently high to run a discharge and p, is sufficiently low to allow fora collisionlessplasma in the tank. 'Collision less' means that there are practically no collisions betweenelectrons and neutral gas particles. The collision frequency I'coli then is governed by collisions

with the tank walls according to

vcoli ^ Pre (4)

where D is essentially the diameter of the tank and Ve is the mean velocity of the electrons.

The plasma density Ne in the lank is governed by the continuity equation

aNc, _ P

at —

The production teran P describes the quantity of electrons that enter the tank per unit timethrough the orifice. It is a complicated function of Yhc discharge parameters such as working

gas, working pressure P 2 , geometry, material and voltages of electrodes (cathode, anode, grids)etc.

The loss tern L describes the quantity of electrons that get lost per unit time primarily clue

(5)

i

i

CONSIDERATIONS CONCERNING COr AS WORKING GAS 3I

to recombination at the wall. In a very general way G call be described in terms of a mean

lifetime re that is spent by an electron in the tank before gelling lost:

NL e

Te

e

In a stationary-state situation ((BNe/3!) = 0) the electron density in the lank will then be

controlled by the production term P and the mean lifetime according to

Ne—P, re(6)

The lifetime re will be closely related to the lifetime rr of the ions, since, owing to their lowmass and the strong Coulomb forces, the electrons will always be tied to the ions ill such a

way that the plasma remains quasineutral. If ions get los4 quickly (low r j), electrons will

immediately get lost at it corresponding rate (low Tc), Thu,; it is to be expected that it and

hence rc will decrease when the pressure p is decreased, since at low pressure (below -^ IV

pascal*) ions will travel directly to the wall, whereas at higher pressures p, they collide withneutrals, which leads to a long totai ion path before the wall is reached. Collisions of electronswith the wall do not necessarily mean electron loss, since many electrons are reflected back bythe potential within the ion sheath. As long its high plasma bulk velocities (wake phenomena)

and/or magnetic fields (Lorentz forces) are not involved, the plasma potential with respect to

the wall always attains a value such that the mean lifetime of the electrons matches that of the

ions, Then the plasma body is quasineutral, and potential differences are confined to surface

sheaths, the thickness of which is of the order of several Debye lengths.

ALrording to equation (6) a decrease in the charged-particle lifetime must be compensated far

s; y the p,ovision of a sufficiently high production rate P to maintain a required plasma density.

ii3, CONSIDERATIONS CONCERNING CO 2 AS WORKING GAS

3.1 VACUUM PHYSICAL CONSIDERATIONS

The demand for a large pressure difference p 2 — p t, t)2 typically being I pascal and p, lessthan 10" pascal, Ind to a required pumping speed S in the order of a few 10 1 1 s' , a much

higher value than the available turbomolecular point) could provide (only 260 1 s' ). In order to

* I pascal ;^: 10' 2 'rorr = 10 µm}ig

4 CONSIht RA I IOtis ( ON( I RNING CO, AS w, )RkING GAS

reach the required pumping speed at relatively low cost, cryogenic pumping with liquid

nitrogen seemed to he appropriate.

This vacuum physical requirement called for a working gas with a vapour pressure less than

about IU 3 pascal at 77 K. ('0 2 appeared to be a suitable candidate, since its vapour pressureis 5 x I V 'Corr.

I{III-AKD(MN VOI.TAGF AND IONISA I ION PKONABI I ITY

III to get a high production terns P the working gas shouid have low breakdownvoltage and high ionisation probability. I i-irc 2 shows the Paschen curve for ('02 incomparison with several other gases. The distar,x d between cathode and anode for our systemis typically 10 2 nnm. With P2 in the order of I pascal, the parameterp . d for our discharge will

he somewhere around I Since the slope of the Paschen curve in this region is very steep, even

-I displacement with respect to another curve (compare for example ('02 with air) leads

to an appreciable decrease ill breakdown voltage.

As to the ionisation probability, Figure 3 shows that CO 2 can be expected to provide a

higher discharge current for a given pressure than N 2 and A. As an example. electrons within

'ge avalanche with an energy of 15 eV would produce about 30 times more

n CO 2 than in N 2 and A.

ness of energy states (electronic, vibrational, rotational) of the CO 2 molecule

w the discharge electrons to randomise their kinetic energy more quickly, i.e. over a

u.ion length, than would be possible for electrons in N 2 , A or Ile.

6 10 14 18 22 26 30

pd(MM-MM Hg)

Paschen crrri, r„r r,rri„tr^ gaws (arlier ,Al. KA'OLL. l-. 0LLEVD0RI-'1', and

R ROMPS, Springer-('erlag. Berlin (/935)1.

W 19000

14000Z 10003g 600

YW noCr

0 0 2

100e

A

-71, q

^

_a(:I

2 W X

x ~^ti

W q S

^

qC

N2a

iaaa

a

b

a 1,C ,^ jt-zulXa

U

zO

0110 11 12 13 14 15 16 17

ELECTRON ENERGY (VOLTS)

I'd

1

CONSIDERATIONS CONUHNING CO, AS WORKING GAS

S^i

1%igure 3, Ionisation probability for various gases W.P. Lossing, A,W, Tlckner, andW.A. Boce, J. Chrm. PAys. 19, 12540951)).

it

i!ii

3.3 RELEVANT REACTIONS OF CO2

(a) Thermal dissociation: At 2000 K, 1,8 per cent of the CO 2 dissociates according to

2CO 2 42CO+02

The neutral gas temperature should stay tar below 2000 K; it might approach about1000 K in the vicinity of a heated cathode. Thus only a very low thermal dissociation rate is ro

be expected in the discharge,

(b) Reactions in the CO 2 discharge involving charged particles are as follows:

(1) Dissociative excitation

CO 2 +e - CO* +0'+e

J

CO 2 +c- C*+O'+0"+e

6 MEASURLMRNTS

(2) lonisalion

CO, + c -* COi 4-20

(3) Dissociative recombination

CO=+e^ CO* +0'

pa +e-+0*+0'

'4) Charge exchange: CO + + CO 2 - CO + CO (very fast)

0 + COz + —* Ot + CO —+ 0 + + CO=

Little is known about the rate coefficients of these reactions. r

(c) The consituents of tltc CO, plasma ro be expected are listed below:

Neutrals furs

CO2 C0=

co CO,Oz s020 0+

C

The diagnostic investigations that were to be performed later in the plasma were concerned

with the motion of plasma electrons (electromagnetic and clectroacouslic plasma waves) so

that the presence of different kinds of neutrals and ions was of no consequence as tar as the

Interpretation of the measurements was concerned.

4. MEASUREMENTS

4.t CRYOGENIC PUMPING OF CO2

A preliminary investigation was made in order to determine the effect of a cryogenic type ofpump on the tank pressure p, for various gases, ror this purpose, a LNz Goole.' sorption

pump was temporarily connected to the tank. Figure 4 shows its effect on pn for lie, N 2 and

COV The effect is strongest for CO 2 and zero for tle. Although !it this case two effects were

mix9d, i.e. the absorption effect of the sieve material and the low temperature of the trapping

f

NFASURI NINTS 7

pia volvo to SP olosod

I p y vawo to SP opon

.— 00=

1 >^

1 5 10 15 p1,1poscol

Figure 4. Effects of sorpilon pump (SP) on the lank pressure pt.

6,0-3 6q-3 1 0-2 20-2 pl/Mascot

r6ire S. ('02-pressure In the tank for mininium and maximulil selling of the irds dia-phragm applying cryogenic ptmlping at Leis temperature.

t.tAll

^s

D

8 MPASURIi IM'S

surface, it was to be expected that, with CO 2 as working gas, operation of it cryogenic wallwould lead to it considerably higher pumping speed twin if the turbomolecular pump weree,Icruted alone.

A copper shroud with copper tubes soldered on one side and covering about hair tl:: inner

wall of the tank was inserted Into tho chamber. It could be cooled down to LN 2 temperaturewithin about one hour. III

to the cryogenic wall, an iris diaphragm was installedbetween bell jar and lank to provide it diameter of the orifice between the dischargeand the measuring volume. As can be soon from Figure 5, it was possible to maintain apressure ratio p2/1)I or roughly two to three orders of magnitude by adjusting the irisdiaphragm and cooling the cryogenic wall down to liquid nitrogen temperature. Thus thevacuum requirements for producing it plasma were met by this device, using CO2as working gas.

4.2 BREAKDOWN VOLTAGE

Figure G shows measured values of the breakdown voltage (cold cathode discharge) as afunction of pressure for a fixed distance d between cathode and anode (both of circularshape). The breakdown at presrv°:r hclow about 2 pascal occurs at considerably lower voltages

In CO2 than m N2 . This is it' ,orw:ment with the Paschen curves hi Figure 2, it being expected

that air would behave in a similar fashion to N2.

UeD / V

II

Of

d = constant =100 nmCO 2 XXX

N 2 000

.,

w

0 1 2 3 L 5 6 7 8 9 10 11 p•,,/pascal

Figure 0. /Breakdown voltage Uljn as a function ref pressure p i Ji)r CO, and 1V2.

U

f

3

CO2

1Y

iE

AIGASURMUNTS;;

9

4.3 ELECTRON DENSITY

To investigate further the properties of CO2 as it gas for plasma production,

voltage current (U/J) characteristics were measured with :I potential analyser (RPA)

in plasmas of CO 2 , N 2 , A and He, RPA measurements were performed over it pressure interval0.5 to 10 pascal ( without cryogenic pumping) and it voltage range 600 V

6 UDC 6 3000 V. In these cases p2 is very close to p,. A representative set of these

characteristics is shown in the Figures 7 (a) through (d). The plot is ecmilogarithmic. Note the

different scale for the Iag J ordinates. Although the discharge parameters (pressure p,

discharge voltage UDC) are not the same for the four curves they claarly show that in the CO2plasma both the strongest Maxwellian component and the lowest electron temperature T. are

obtained. ThL Electron saturation current Jsnt (defined here as t1w current at U =+ I V) of theRPA characteristics can be taken as it mcusure of the electron density obtained III tank,Figure 8 (a) shows Log V,,,, I) versus discharge voltage UDC for CO 2 , N; and A. Although thepressure for the CO2 curve is lower (0.6 pascal) than the one for the N 2 and A curves (1.0 and

1.4 pascal, respectively) the saturation current for CO 2 is considerably higher than that for theother two gases. This holds over the entire range of UDC.

Figure 8(b) shows Log (^sar) as a function of pressurep, for UDC= 1200 V. It can be seenthat the value of Jsat is much higher for CO 2 than for N 2 and A in the pressure range below

about 3.5 pascal. At higher pressures the density provided by a CO 2 discharge goes down. This

fact may be due to volume recombination and/or formation of negative ions (C07• No

detailed interpretation of this fact has been attempted since the pressures p, we deal with laterare well below 3.5 pascal.

E

w

j

Simultaneously with the RIIA-measurements, the parallel resonance frequency FP of a

cylindrical RF prnbe immersed in the plasma was measured. Fig is identified as the plasma

frequency FN (neglecting corrections for the collisions) which is related to the plasma density

I!!N. according to:

N. [m a 1 = 1.2456 X 1010 FN2 [Mlizl (7)

iiFigure 9 shows FN as ,I of the discharge voltage UDC comparing CO 2 with N; at

i pressures p, of about 0.6, 1.0 and 2.0 pascal, This diagram again shows the striking advantage

of CO 2 for providing It production term P for the steady-state plasma in the tank.

II

4.4 ELECTRON-ENERGY DISTRIDuTI.ONF

Returning to the set of RPA measurements, we find it way of presenting the data

that reveals facts again favouring CO 2 for piasma, prod uelion. In general, most of the RPA

l;

t

3

^I10 MEASUREMENTS

Log13/A) t 1'l iet

4

3

2

-6 U/V

^1 0 -1 -2 -3 -4

5 Tea43

CO2 , Pi =2.1 pmeat

2Uec•1200V

-7 \ Te = 1900 K

Tel) 6w 60000K

7a \— 7-

Log 131A) Log n /A )

U/V 6 ILe,6g8)

+1 0 -1 -2 -3 -4

4

A, Pt =14 pascal He, P1 = 9.6 pascal6 Uac.=3000K UDC =3000V

4 To-27000K 3

3non- Maxwelilan

2-, 2

\11^u /V

^T -g«1 0-5

7c 7d

Figm,e Z Currentivoltage characteristics obtained with the RPA in CO 2 (a) Nz (b), A(c), and Ile (d).

^•

pp

^

F i1 ,^

44

0

_I

7

f .^.

1

.2.

O^

:1200V

IgasoslPI i CO2 1 . 0 6 PascalPI (N 2 1 ` 1 PascalPt I A I `14 Pascal

I

NIPASURLhIr•.NTS-, .

Log I ISAT/AI

Log I IsAi/AI

-5 -SA

u

Pi /Pascal

LMOD

0

5 6

so

86

Figure $ (a) Saturation current 1;a t,taken from RPA curves versos cold cathode discharge

voltage UDC (with neutral gas pressure approximately constant) for CO 2 , N2

and A. (b) Saturation current [sat taken from 2PA curves versus neutral gas

pressure p, (constant discharge voltage UDC) for CO2 , N2 and A.

characteristics measured in CO 2 and N2 plasmas have the general form shown in Figure 10.

The semiloyarithmic plot usually contains three components (a, p and y) of the electron

population represented by straight or nearly straight lines which can be attributed to electron

temperatures V TO, 1`, respectively. Figures I 1 (a) and (b) represent these temperatures (71only as table) as a function of botli p i and UDC. The scale for 7c is on the left, the scale for0.1 on the right side of the diagrams.

The iepresentation shows the following points:

(I) 7e (CO 2 ) stays low (1000 to 3000 K) over the investigated ranges of p^ and UDCrespectively.

(2):(e (N 2 ) rises strongly towards the 10 1 K range when the pressure decreases to I pascal.

For constant pressure (1 pascal) it is about 101 K and rises slightly with increasing

discharge voltage UDC.

(3) 7¢ (CO 2 ) increases with increasing pressure for constant UDC (1200 V) and decreases

with increasing UDC at constant pressure (0.6 pascal) towards a constant level (about

3 X 104 K).

1

MEASUREMENTS

/ CO

'I

6o FN /MHz

50

40-.

COZ(2pascal)

—^Te 7

I "

LOG I

x

10N

10,6 pas

s

cal) x

xx

UDC/v500 1000 1500 2600 2500 3000

Figure 9.

Plasma frequency fN fit lite lank as a function of discharge voltage UDC forCO2 and N2 at three different neutral gas pressures ph,

a

. I

U

Figure 10, General three-component feature of the 1lU characteristic as seen with the RPA

in CO2 and N2,

i

r

nWASlntenlrNTS

To/Kc..^-.._^.. '---r.To /K

13

Ta,^K a To 1c02 1 To /K 111101

^----

,u010`Too I l,4 2 1tap 103.

PH1m" T r IN 1 /0V A9101 • 5.4 30 9,01 Uoc/V 1gr NA&V

423,0

2219 5a^ 1500 13 gp^

0 101 2.0 17 gio 2000 121.4 15 2500 12

Te IN21 3000 127101 7101

7^ co')I^4 p̂

010 06100 PI ICO 2 ! -0 6 Pascal

1.6.1200v P1 I N 2 1° 1 Pascal

5101

I142ardol2!Sp

b

•

3,0 1 3101Tc ICO21 TQ IN21 ' Te"1CO21

1 210 0 2102'0 Pt /Pascal 210 Uac/V

1 2 3 4 5 1200 2000 3000

110 11b

rigure ll. (it) Temperature components Te and 7^ versus pressure p 1 for CO2 and N2 at

constant discharge voltage UDC (b) Temperature components 7' and 70

versus discharge voltage UDC for CO 2 and N2 at constant pressur. p l . 7 for

N2 is not included, since the 0-component could not be identified clearlyenough from the h/U characteristics.

(4) 7e (N2 ) is low compared with T.- (CO 2 ) and rises slowly with increasing pressure.

4.5 THE FINAL LABORATORY PLASMA SYSTEM

The body of information collected by the measurements described above clearly favoured

the choice of CO 2 as a suitable working gas for producing a laboratory plasma with a strong

Maxwellian electron component of low temperature (a few thousand K). A source which was

then developed was of the back diffusion type, making use of a heated cathode in order tofurther increase the production rate P. The final system is shown in Figure 12. The brush

electrode was used temporarily for cold cathode discharge experiments and noted later as the

anode for the back diffusion source. 'file paddle proved very effective in baffling high-energyelectrons. The drawing also shows schematically the position and geometry of the diagnosticprobes as well as the LN 2 -cooled copper shroud for cryogenic pumping. A screen was present

in the chamber. This was occasionahhv used to influence the plasma Potential by having a bias

voltage applied to it, but in general it was kept at floating potential. Typical neutral-pressure

values for this laboratory plasma system in operation were a few pascal in the discharge region(bell jar) and 10-2 pascal in the tank.

I L,

BELLJAR

i

14 WASUR1 MI KITS

/ LN Z - COOLEV COPPER SHROUD

NELL, fVALVE

GAS -+

BRUSH IRISELECTRODE / CATHODE

GRI f• PADDLE..y,,

LP

DISCHARGE APLASMA t

STAINLESS STEEL

I

TMP

Figure l '. LaboraforI, plasma sYstem with plasnna source and diagnostic prob",

16.. 1

4.6 PLASMA DIAGNOSTIC-PROBE MFASURFAI:NTS

RF-measurements were performed with a cylindrical two-electrode system IF,, I•: 2 1. 'ISshown in Figure 13, The principle of the 10 --measurement is also shown. A - , ept-frequencyRl generator provides a signal of constant amplitude within the fre(juv ".:y interval of typicallyI to 25 Mllr. Both the RF-reference voltage U R at " I and 'he test vo; • . t,. U7• at I?: are

measured and compared as to their complex ratio

UTIUR z I..+Y"

by mcans of a network analyser I hp 8407).

fhe signals provided by the network analyser are magnitude

1 1,71

a = = f.•2 + F2 I d li IiR

and phase = arctan (FIE) in degrees. Magnitude and phase together are a measure of thecomplex plasma impedance 7. = X + i l between F I and E 2 .

1

-- 16. n m —•.,

'M fi. T..

-7 .

-8

MI. x!:11P F ME vTS

Up

1nF E 1E2 1nF=

ti 1vK 10K

i

Uk i u TT DC I DC

Figure 1.1

C •yliudrical Rl • '-probe acrd principle of RF hnpedan.-e mea,^uret►n.•»t

U/VOLT+6 +4 +2 0 -2 •4 -6

-r-- i--'-t—,-4

-5

0-6OJ

7!!_

i1 \e

►ov

Figure 14

Langmid-characteristics for rarh,us value's of the bras voltage U' 'Ohe j an;;,:swirce heating cire uit.

U,

1.;

-8-6

-5

-4

-3

-2

-I

OV-U,

EN {^^ NX

U,

-PROBEF U,

i

Co100

^

Ii

16% ► I ASURI MI NIS

In f= igure 14 current/voltage characteristics of a spherical (diameter: 10 nm ► ) stainless-

ace[ I .mt;rnu ► r probe are shown. The parameter of this set of curves is the bias voltage V, ofthe plasma source heating circuit. It can he .ern that the velocity distribution and temperature

Te of the electron. is markedly influenced by 11 1 . In the prese ►► t case the distribution function

is Ma%wellian m good approximation for ll t %.dues of 2 V, s V and 4 V. I hr

corresponding T, values are 0.55, 0.53 and 0.52 eV, respectively. For each of these 1_anginuir

curves the magnitude a measured as a function of frequency was plotted ()it an X-Y recorder.

Figure 15 shows the corresponding set of curves, which reveals the lollowing essential features:

a)

above the parallel resonance fp, which is in our case (no magnetic field) equal to theplasma frequency l:v. an additional re%on;uce fz occurs and

h) fy is clearly most pronounced for the case of Maxwelhan distribution of the electrons

with low electron ten ► prrattire Te IU, = - 2 V, - 3 V. 4 V).

I his resonance l! can he understood in terms of resonantly excited elect roacoustic waves

(also called electron- pressure or I andau waves). These waves are set up by an KF source above

?8.474£XPNO 09

-10 Ca

5 10 15 20 25

f/MHz

I ;^,rrr / i Phimp ►a of the c_rlrnclmal rtt—clectrode sYstcnr.

1

b.,

jAIEASLIRP&HNI'S 17

the plasma frequency. Excitation of this longitudinal type of plusml wave, which is damped

with Increasing frequency by collisionless or Landau damping, is predominantly responsible forthe real part of the Impedance of an electrode system immersed in a plasma. For a singleelectrode this real part would decrease monotonically with Increasing frequency. For ittwo-electrode system (E, , I1 2 ), a% used in our experiment, however, it characteristic electrode

- distance ti can be defined (distance between inner and outer cylinder). In 11115 case theelectroacouslfc wave call produce a standing-wave pattern between Li, and 1s 2- "" I ' s is expected

to occur essentially at efgenfrequencfes of the system (electrodes/plasma) 1'or which the

wavelength )`en (or integer multiples of it) matches the distance d, '1 lie detailed discussion ofthese resonances is oulsidc the scope of the present paper,

4.7 MASS-SPECTROMETER MEASUREMENTS

As regards the chemical composition of the CO 2 -plasma, I1leaslll'ealentS With it

neutral mass spectrometer (a laboratory model of the Lunar Atmospheric Composi-

tion Experiment (LACE) flown to the moon by APOLLO 17 - J. Hoffman, principal

investigator) provided information about the abundant neutral molecules and atoms, Theneutrals entering the ion source of the mass spectrometer are ionised by an electron beam. The

ions are accelerated and collimated to a beau which passes into a magnetic analyser by which

the ions are separated according to their momenta and detected in one of three mass channels.

Figures 16 (a) and (b) show a set of measured mass spectra (signal current versus anal number)

for two energies tic, of the ionising electron beam (25 and 90 eV, respectively). The saluenceof the measuring conditions in each case is

(a) background gas at 9 X 10" pascal, with no CO 2 introduced via bell jur and no plasma;(b) CO2-pressure at 2 X 10' 2 pascal, but piasnw switched off;(c) CO2 -gas stream as fn (b), but with plasma switched on.

III case (c) the neutral gas pressure increases by about one order of magnitude, mainly

because CO, which is produced by dissociation in the discharge, cannot be pumpedcryogenically at LN 2 temperature. The pressure im-rease was measured independently by anionisation manometer in the tank at d by a Penning manometer within the mass spectrometer.As a result of this pressure increase, the base line of the spectrogram is enhanced and (owing tomultiple scattering fn the ion source) the p ine width of the detected peaks is increased (whichaffects the anus selectivity of the im(r unenl).

The spectra in Figure 16 (a0 and (b) show the following features:

I. When the plasma is switched on, the CO and 0 2 peaks strongly increase; CO beeonlcs thedominant molecule. This holds for both electron energies (/;' L, = 25 and 90 CV). I he

:^tt

f

I& MfASUIt1:StFN'rS

9 Background Ee=25oV

'b

-2 Plasma off Ee=25eV

•e

•9

I(^1

5

Plasma on Ee-25eV-e

..g

toli

105

16 28 29

,L 15 17 20 29

160 MASS/amu

f9'r^ty=

CBackground

Ee=9OeV

t

i

^^ J

i

ii

a

6I

-2 Plasma off Ee =9OeV

'0

9

-t0 IL7 " tt

0

6

-7 Plasma on Ee=9CO-e.

9I

_ t0

-1

12 U tG 15 76 177 10 22 20 29 90 72 74 1445 46

t6bMASS/amu

Figiu•c 16. (al Moss specd•a with 25 f' acceleration roltnge applied to the electron bean) of

the ion source. (b) Mass .spectra with 90 L' acceleration t ,ollage applied to the

eleen•on becnn of the Ion .sotnre.

n

CONCLUSIONS

19

previously measured cracking pattern of. the instrument for ('0 2 (private communicationby J. Hoffman) shows that only about 105r of the incoming CO 2 molecules aredecomposed into CO and 0 in the ion source at G C values of 70 ... 90 eV. Tile strongenhancement of CO lilts implies that dissociative excitation: CO 2 + e -i CO* + O' + efollowed by 0 + 0 -+ 0 2 , Is a process which is very efficient In the CO 2 disellarge, Sincethe neutral gas pressure in the discharge volume is in the order of I pascal, the density ofthe CO molecules coming From the discharge can exceed tine ('0 2 density in the lankbecause CO is not cryogenically pumped at LN 2 temperature.

2. For E. = 25 cV and the plasma switched off' (CO2 pressure of the order of 10 -1 pascal),no carbon 12C is detected. This holds also for CO 2 pressures of 10.2 pascal withoutplasma (separate measurement). Thus (lie appearance of 12 C when the plasma Is switchedoil that 12C is produced to a small extent in the discharge, recording to thereaction CO2 + c -)- CF + 0' + 0" + e.

Fur J,e = 90 eV the 12 C peak is strongly enhanced. This is due to the dissociation of CO2by elce'ron bombardment in the !oil of the mass spectrometer. According to theabove• mentioned cracking pattern, about 9% of the CO 2 molecules will be cracked downto produce' 2 C at these high values 01, Ec.

3. The CO, peak (anmi = 44) is followed by two isotopic lines at anti = 45 and 46. These aredue to tle molecules 12 C 1602 and 12C 160 "0, respectively. The abundance --relative to CO 2 — of these isotopic molecules is about 15c and 4%. There are cor-responding isotopic lines at amu = 29 and 30. These lire related to 1 X 1 1 0 and 12 C 1 M O,respectively.

4. The peak at amu = 18 is due to residual 1-1 2 0 vapour present in the vacuum system; themass 17 peak is produced by dissociative ionisation of 1-1 2 0 in the ion source leading toOM

S. Al re = 90 cV, double ionisation of CO 2 is possible, leading to lines at amu = 22(COz) ,22.5( 1 4 Os ) and 23(12Ci601800).

S. CONCLUSIONS

The investigations reported here show that carbon dioxide has properties that favour itas working gas for the production of a :rollisionless Maxwellian plasma with low electrontemperature (T,, in the order of ionospheric F-region values). CO 2 leads to higher plasmadensities at lower pressures as well as a larger Maxwellimi component of the electron popul-

i

IY:,t

20 ACKNOWLL' OGEMENTS

ation with lower electron temperature as compared with other gases such as Ne. A. and lie.'fhe fact thal ('02 can be pumped at high pumping speeds by a liduid•nitrogen•cuoled %;ryo-

genic wall enables a steep neutral gas pressure gradient to be achieved between tine dkdwrbc

volume and the measuring volume. however, in the difl'usiun type of p la cmn system used :u

the present work, the CO concentration relative to the CO, density is strongly incivased it thetank because CO is not pumped cryogenically at LN 2 temperature. The ubumlau^;a of uaygeain the plasma naturally imposes restrictions on the choica of materials liar diagnostic prob.,,

etc.

It may finally be mentioned that studying CO 2 plasma in the laboratory may contribute to an

understanding of some of the properties of the Martian and Venustan ionospheres.

ACKNOI1,1!rDGEMENTS

The author expresses his gratitude to D. Winningham for providing the RAA•!'ensur, to

J. Hoffmann for providing the mass spectrometer and to B. Milan for his engineering assu;wuce.

ance. Many thanks also go to W. Ileikkila for his constant interest in this project and valuablediscussions. The work described in this report was performed during the author 's tenure of an

GSRO External Fellowship at the University of'Cexas at Dallas 1974. The work represents part

of a laboratory plasma project supported by NASA Headquarters Grant NGR 44--0044 030.

6.

I