The Velocity dependence of the absolute total ionization cross section for collisions of argon atoms with singlet and triplet metastable helium atoms

Material Information

The Velocity dependence of the absolute total ionization cross section for collisions of argon atoms with singlet and triplet metastable helium atoms
Woodard, Michael Read, 1943- ( Dissertant )
Muschlitz, E. E. ( Thesis advisor )
Bailey, Thomas L. ( Reviewer )
Colgate, Samuel O. ( Reviewer )
Luehr, Charles F. ( Reviewer )
Person, Willis B. ( Reviewer )
Place of Publication:
Gainesville, Fla.
University of Florida
Publication Date:
Copyright Date:
Physical Description:
viii, 110 leaves : ill. ; 28 cm.


Subjects / Keywords:
Atoms ( jstor )
Electrons ( jstor )
Energy ( jstor )
Helium ( jstor )
Ion currents ( jstor )
Ionization ( jstor )
Ionization cross sections ( jstor )
Lamps ( jstor )
Metastable atoms ( jstor )
Velocity ( jstor )
Atoms ( lcsh )
Chemistry thesis Ph. D
Collisions (Nuclear physics) ( lcsh )
Dissertations, Academic -- Chemistry -- UF
Ionization ( lcsh )
bibliography ( marcgt )
non-fiction ( marcgt )


Measurements of the velocity dependence of the absolute total ionization cross section of argon atoms upon impact with selected metastable states of helium atoms are reported. A low voltage D. C. discharge was used as the source of the excited atoms, and a rotating slotted disk selector was used for velocity selection of the excited atoms. Selection of th electronic spin state of the excited atoms was accomplishe by irradiation of the excited atoms with radiation from a helium d i s cha r g e 1 amp . Ionization of the argon target atoms by metastable. heliui atoms was studied by the gas cell technique in which all ionization products were collected. The ionization measurements were of sufficient precision to allow simultaneous determination of the cross section and the secondary electron ejection efficiency for each metastable state of helium. The secondary electron ejection efficiency of triplet me tas table helium atoms on an electroplated gold surface was determined to be 0.440 + 0.018 in the presence of argon gas. The secondary electron ejection efficiency of singlet: metastable atoms was determined to be 0.582 + 0.024 under similar conditions. The local ionization cross section for the He(2 S)-Ar system was found to increase almost linearly from S.8 A to 2.1.95 £ with an increase of relative velocity from 1162 from 9.8 ?? to 26.2 K 2 with an increase in relative m/sec to 2787 in/ sec. After an initial increase, of the cross section velocity from 989 m/sec to 2058 m/sec, the velocity dependence of the cross: section of the He (2 S)-Ar system entered a saturation region in which the cress section changed very little with relative velocity. The collision energy dependence of the He(2"S)-Ar system was also used to determine the values of adjustable parameters present in a current theory based on the potential curve model for Penning and associative Pennine ionization.
Thesis--University of Florida.
Bibliography: leaves 105-109.
General Note:
General Note:
Statement of Responsibility:
by Michael Read Woodard.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
000063267 ( alephbibnum )
04205548 ( oclc )
AAG8465 ( notis )


This item has the following downloads:

Full Text








Of the an'yY people who have contributed their time and

efFort to thiis pr~ojectl, the author particularly wii1ides to

ackn!ow~ledgeE the fal ow~ing pei~rsons. Forem~ost-, Professorr E. E.

Mu-chlitz, Jr., Chairman of his Supervisory Comm;;ittee-,wrhose

determinat-ion,, optimiismn, and enthusiasm were crucial to thle

completion of th~is research. The author particuilarlyy wishes

to extend a special note of thanks to his follow participants

inl this research effort, Mlr. Mike Secely and Mr. Ricke Sharp..

Their spirited participating and natural interact werle e.

constanti source of inspiration to thle project.

The author Al~so wishes to expre~ss hiss appgreciiatin of

the aid so capably given by Messrs. R:. J. Dugan and J. G.

Chiamblee of the electronics shop, R. Strasburger and R. Stroh-

schein of the glass shop, and E. C. Whkitehead- of the machinc-

shop. The author also wishes to thanked Ms. Ann Kennedy for

typing this dissertation.

Also, appreciation should be registeredl for the financial

assistance of the Nuational Science Foundation that made part

of this work possible.

The author is also truly grateful to his wife, M~artha,

for her support, understanding, and encouragement: throughout

thee years of his studies in Chemlistry,





A. Criteria for Chemiionizatio nI 1
B. Ionization MIodels 2
C. Quantitative Theory Basedi on the 9
Potential Cur~ve Model
D. Verification of the Classical Theory 17
E. Object 24

A. Mietestable States .26
B. Spin Selection of 2 S or 2 S Heclium 27
Atocms from a M~ixecd Beamz~
C. ecrection of M~etastable Atoms 27
D. Velocity Distribution of a ~axwejll- 28
Boltzmann Source Gas
E. Total Lonization Cross Section and 30
Secondary El~ectron Ejection Coefficients
for Mletastab le Atoms

A. General Description 35
B. Vacuum Systemi 38
C. Gas Source 40
D. Discharge Chamber 43
E. Velocity Selector 44
F. Quench Lamp 51
G. Collision Chamber and Detecting Systemns 55

Al. Calibration of Target Gas Pressure Gauge 61
B;. Conditions for M~etastable Production 62
C. Distribution of Metastable Velocities 62
D. Charge Collection P~otentials 68



IV Cont~inued
E.OperaLion of Quen~ichl L;.unp~ 68
F. P~ro-cedure of Ex:pe-irimen 73
G:. Recduction of Row~ Datl-a 76

A. Secoindary Elec~tro Ejeedon~ Coeffiic ient G2.
B. Relative Velocity D~epen~dence of the 83
Total lonization? Cross Section
C. Calculation of~ Relativea Vlocity DependencEE 94
of tihe TIotal lonientionr Cross Section





Fi~ur:e Page

1.. Elctcron Ex:channg: iodel. fo~r Penaing. l 3

2. Twoc St-ate Fatetiatjl Cu;Irve M~odecl fror 6
Penning and AIssociativec Ionization

3. Dependence of Branchine Ratio of 11
Associative to Total Ionrization on
Relative Velocity for the System

4. Dependence of the A~ssociative and 13
Penning Ionization Cross Section on
Relative Velocity for the System

5. Dependen~ce of T'otal Ionizationi Cross 20
Section on Relative Velocity for the
Sys tem Nie(2F'O ,2)..APr
6. Cross Section of Apparatus 36

7. Beam Gas Del~ivery System 42

8. Velocity Selector Rotor Assemnbly 48

9. Power~ Supply for Velocity Selector 50

10. Rotor Frequency Heasurement 52

11. Helium Quench Lamp 54

12. Collision Chamber and Detecting System 56

13. Calibration Curve for RCA 1946 Thermo- 64
couple Gauge for Argon

14. Velocity Distribution of' Detected M~eta- 67
stable Beam Intensity



1.5. Charge C7ollectionn Versus: D):.cout 70

16. Saturation of Quecnch i~Lamp Effricincrly 71

1.7. P~_i)Vru o C2S-r79
at Differ-en-i Relat-ivce Velocities

18. P(i /i4t) Versuls P for Hie(2 S)-At 81
ar Different Relatlivee Velocities

1.Dependence- of: Ablsolutle Total lonizaztion 8
Cross Section on Relative Velocity for
the Systemo He(23S)-Ar

20. Dependence of Absoluto Total Ionization 87
Cross Section on nRl~ative Velocity for
the System He(21S)-Ar

21. Dependence of the Ratic of Absolute Total 90
Ionization Cross Section of Ho(21S)-Ar to
Hec(23S)-Ar on Relative Velocity
22. Calculat-ed and Expe-rimnental Denendence 98
of Total. Ionizationi Cross Section on,
Rieltive Velocity for He(23S)-Ar

23. Comparison of Calculated Dependence of 100
Total Ionization Cross Section on
Collision Energy for He(2 S)-Ar

Abstract: of Dissertation Prese~nted to the
Graduate Cou;,cil. Lf the Unsivers~ity of Florida
in Par; tial. FulfTi l~lmont' of: thle Neou!i~lij-reent~s for the
D~eGree of Doctor of Ph~ilosophy



Milchaeul Roard W:ooded

June 1977

Ch-airman: E,. E. Muschli~tz, Jr.
M~aj or Depart-ment: Chiemistry

Measurem~ents of the velocity dependence of the absolute

total ionizati~jou cross secLL~io of argonl atoms upon impact

with selected metastable states of helium atoms are reported.

A low voltage D. C. discharge was used as the source of thle

excited atoms, and a rotating slotted diskc selector was used

for velocity selection of the excited a::oms. Selection of the

electronic spin state of the excited atoms w~as accomplished

by irradiation of the excited atoms with radiation from a

helium discharge lampI.

Ionization of the argon target atoms by metastable helium

atoms w~as studied by the gas cell technique in which all ioni-

zation products w~ere collected. The ionization measurements

were of sufficient precision to allow simultaneous determnina--

tion of the cross section and the secondary electron ejection

efficiency for each metastable state of helium. The secondary

E3cetron7 ejecftion rrfficincy of tri-jloJ t meta~stab~le helium~

at~oms on an electropl~ated goldi surface waas deterineiid to be

0.440 + 0.018C in thle precsenlce of airgon gas. Tihe secondaryy

electlron ejrection e~fficincy of singlet mneta~stabl.e at~oms was

determcinedi to bie 0.5832 + 0.024 under similar conditions.

Thec total ionirsation crose section for the Hoc(233)-A\r

ysten w~as foulnd to increase almost Iliinarli fr~om 8.8 #2

to 21.95 92 with an increase of relative velocity fromn 1162

m/sc to 2~787 m/sec. ~fte~r an initial. increase of thie cross

sescion fromi 9.8 ,2 to262 2 ith an increase in relative

velocity from 989 m/sec to 2058 m/rsec, the velocity dependepnce

of the cro~s section of tihe He(21S)-Ar system entered a satura-

tion region inl which the crcca section: changed very little

w~ith relative velocity. The col~lis:ion energy dependene of

the He(23S)-Ar system was also used to determine thec values

of adjustable pairamete~rs present in a current theory based

on the potential curve model for Penning and associativre

Penning ioni zation.



A. Crijteria for Chemriionisation

Collisions between metastable ~t~oms and target atoms:

can result inl thle dcexcitation of the electronically ex-

citted meta-stable atoml and the folrmation of an ion~ic species.

These processes are gener-ally r~eferred to as chem~iioniliation

if collision~s between t-hese neutral reactants result in the

fontlatio~n of. one or more charged produlcts.1

This ionizat-ion process is understood to be the colli-

sican1 auccionisation of a Lwo-pairticl sysile: A*"-E, inl which!

the energy of act~ivat~ion E(A*) stored in the metiastable atom

' ex;czeds the ionization potential IP(B) of thei targetr

atomr B. The twro exit: channels open to the reaction, in the

case where Ei(A*) IP(B) > 0 are as follows:

A* + B A + Bf + e- (1)


A* + B +~ AB +e(2)

w!here~ exit channel (1.) is commionly called Penning ionization

and exit channel (2) is called either associative Penning

ioniizat~ion or simply associative ionization. For collisions

occurring at -omne thermal tnergy Eik(m), associative ioniza-

tion will1 predominate if

Eel > E(A*:) IP(C) + Ek(m)

wJhere 1501 i the kinetic. energy of- the electron released by
thE ionlizatiionl pLoceSS. A~ thi~s case, the electron ca~rres

aw~ay excess energy since the~ bound s~tat AB~ has a npeative

efn~rg~y. Penning ionization occurs in? the alterna-te case where

and A and B;t are also available to carry off reaction energy.

B. Ion-iza.i~on Models

Hiot~op and Nichaus have proposed twJo possible mechanisms
which are as follows:2,

A*()- B2 (1) + B e(2) (a)

A*(1) + B(2) + A(2) + Bt + e .(1). (b)

The umbrs n paencasi ri zif elctrns.The direct'

model. (u1) is considered to be unlikely since such a process

is forbidden by spin select-ion rules.

The second model (b) shown 'pictorially in F~igure 1,

follows thec Hagstrum model. for Auger deex~citation of meta-

stables at metal surfaces. In this model, a lcrnfo

target B tunnels through a barrier of effective w~idt-h a
into t~he vacant: inner orbital of the metastable A* with the

subsequent ejection of the high energy metastable electron.

The approximate thermal energy of the ejected electron is

about: E(A*) IP(B). As shown in the figure, IP(B) is the

ionization potential of the tunneling electron: IP(A) is the

ionization potential energy level. of the vacent orbital, and

Ro is thie internuclear separation.

o~ n


A* (1) Ro 8(2)

Fig~ure 1. Electron Exchange M~odel for- Penning Ionization

The ionization- cross section o is; comnputed' by the im-

pactI parameter' mc'thodit by

a = 2m JP(b,v)bdb (3)

and th~e ionization probabili ty is wri-tten as

P(b,v) = 1- expl-2 [Wi(R)/v(R)1 dR] (4)
an-d b is the impact parameter, Relj the cla~ssical turning

poin-t, W~(R) thie transition frequency, and v(R) the~ radial


Thei tranlsitionn frequency i; then calculated from the

p-robability of tunneling: through the potential. barrier pro-

posed by thec model. This method o~f calculations is complicated

by a steric factor introdu;ced by t-he nonuifciormity of thl-e

poiential barrier surrounding thei target anid by difficulties

ini detlermnin;ing~ the effective wi~dth of: the potentia-l barr-ier.~.

A! potential curve model. for Penn1ing and associative
ionization has been proposed by Herma~n and Cermnak~, Nichiaus,

and quantitatively developed by Mliler.6- The ionization

process is thought to proceed through the autoionization of-
a reaction intermediate [A* + B] into the continuumi of A +t

B~~t e-e. The process is represented in Figure 2. The

collision partners A* and C approach each other along the

potential curve V*'(R) with relative k-inetic energy Ek m-).
The curve V*(R) is emnbedded in a continuum of potential

curves V (R) + Eel of the exit channelsfrthrecin

V (R) is taken to represent the lowest available state of

thle continuum~ of states V (R) + Eel'
Transition occurs at some internuclear distance RL by a







+ II

30 N

F-iranlck-Condron tlralsjition. In the Bolrn-Oppenheime.r alpproxi-

maetion, the vertical transition electronic enfery and the

kinec;ic energy of heavy particle miotion are separately con-

served.' Th-us, the initial trnnsi~tion kinetic energy

Ek(R ) of th~e collision partners is eqjual to the kinetic
energy q',(R ) following the. tiransition and is written as

Ek(R ) E (Rt) (5)
and: the totral energy of c-he cntranie channel. relaced to thie

transition kinetic energy Ek(R ) by

Ek(R ~l) == Ele(m: -- Ve(m) -- V*(R~t)
The energy of thle exit channel i:: similnrl~y related to

the transitionn kinctic energy by

Ek~(R ) E Ik(m) (o -i +m V (RC)
thle conne:vation of total energy du-ring the Franrck-Condon
transition i.s written as

Ek(R ) + V*L(R l) = Ek(R ) + V (R ) + Eel(R ) (8)
where Eel(R ) is the energy of the ejected electron. At
infinilte separation, the kinetic enrery of the ionization

products, E'(-,Rt) is given by

Ek(mR~t) = Ek( .) +t E, Eel(R ) (9)

Eo r vt(m) V (m) (10)

Eal(R ) = V*(R ) V (Rt) (11)
The vertical transition V*i to V' is simlilarr to the

autoionization of a mnoleculet except for the fact that the

upper state is continuous wvith respect t~o nuclear motion./

T'he rfore-, tr~ans~itions cann occurl at al~l separati~ons R. The

t~ranisition frecquency Wj(H) is~ connlected to thec widthr I(R) of

the potential; curve V*'(R) by the expressionl

W(R) = T(R)/h~. (12)

The widthh r(R) is shown in Figure 2 as an envelop~e of

values for the curve V*(R) at small7 nuclear separations. The

transition frequency has a particular value for each separ-

ationl value and does not depend upon the state of particle

motion. Furthercimore, if the electron exchange mechanism

is correct, dipole selection r-ulces are not vzli~d for the

population of the different ion electronic stat-es and FR

should decrease exponientiall~y with Ri at large R.' Thiis

conjecture has been wel~l established by numerical caic~ula-

Thle im~portance of: exrelrssi~on (9) m~ay nowc be soon whenr

aided by reference to Figure 2. The relative energy of the

reaction products is shown by E c(Om,R ) plotted as a function

of R. The ejected electron energy distribution is shounr as

a projection on an energy scale at the left hand side of the

figure. Thle length of' the arrows drawnm from energy lefvel

Ek(m) + EU to EltC(mRt) represent possible energies of the
ejected electron. Associative ionization and therefore bound

states are possible only when E1[(m,R ) < 0 and Rt < R .

Penning ionization is the only result for transitions

occurring at distances Rt > Ro since E'(m,Rt) > 0. As the

relative k~inietic energy Ek(m) increases, the range Rt < Ro

and E'C(=,Rt) < 0 will decrease and associative ionization

Such behavior- has boocn observecd for several systems.

The- fralction of associa~tivc io~nizatjon was found by Hlotop

to decrease as a function of relative collisi:on energy (20meV

to 300mecV) in the systems Ec1(21S,23S)-Ar and Ne(i PO,2)-Kr~.11
This was determined by energy analysis of thie Penning ioniiza--

tion electron (PTIES) andl marss spectrometer ion analysis..

A similart trend hacs boon repojrted for the systems Ne*-Ar, Kr,

andl Xe in which the ratio of assiociat~ive To total ioni~zation

cr-oss sections decreased from about 0).6 to 0.0) over~ the

relative collisioni energy orange 20 to 350mneV.

Pesnelle, et al. have recently reported the velocity

dependence of" the branuch~in ratio for Hef(21S,23S)-Ar usingo

a crossed beam technique.l Their determination of' thle

velocity dependence of the ratio ofl associative to total

ionizationl cross section of thle HP(2 S)-Ar~ system is shiowin

in Figure 3. The individual. velocity dependen~ces of thie

associative and Penning ioni action of He(23S)-Ar are dis-

played separately on figure 4t. A fit to the data by the

authors, using a theory developed by Nakamura, is also

displayed. 13

C. Quantitative Theory Base~d on the Potential. Curve Miodel

Miller has developed a classical theory based on the

potential curve model discussed previously The velocity

dependence of the total ionization cross can be compruted
from thi~s theory w~hen information on the potential curve

V*i(R) and the transition frequency N~(R) are available.Th

theory is classical in its development since the ionization

Figre Dpendenice of Branching Ratio of Associativ~re to
Total. lonizat~ion on Relative Volocity for~ the Sys-
tem He(23S)-Arr (After Ref. 12)

VEL~OCITY: (lO5cm,/sec)3







(6i r--


HM i
ti) 9-

ud' LLS rd
t7 1 -d

01. 6d X

,C : IC


L f i -L-------l-
(s?~i;n ~lr.l~Tcllc) r;OIW~T; SSO~D




process is describd! jin plrobabilitiesc: rathcrr than probability

The transition is ag~ainl viewed as anl autoionization

process since, for each i~ntcrnlucleai r distance R, thle disccretee

electronic state V*(R) hans an associated width r(R\), in

ene-rgy units~, for decay into its deigen7erate continuum elec--

tronic state V (R) +I Eal. Thus, the ionization procesr s is

a leakage of the! di~scretei state into the continuum state

degenerate it. Each value of angular momecntum; or partial

wave 2 may b~e treat-ed separaitely since the potentials are

spherical~ly symmletric.. Thei probability density P (R:) is

defined such- that P (R)dRK is the probability of leacage~ into
tboe continuum for ani internuclelea distance between R and

R +t dR.

The. probability of leakage, whiile particles approachc

each other, is denoted as P/n(R), and the. probability of

leakage occurring while the particles separate is denoted by

P ou(R). For the case of approaching particles, the prob-

ability for leakage at a distance between R and R +- dR is

written as

P~(~R=[-i(R')dR' ][ ]dR. (13)
I R hVR(R)

The first quantity in brackets, the survival factor, is the

probability that leakage has not occurred in the inter^val

N R.14The survival pr-obability is multiplied by the rate

of leak-age Ir(R)/t] andt by the timec EdR~/v (R) ] spent

in the interval (R, R + dR]. The term vg(R) is th-e radial

velocity at R.

Thle au:thlor solves Equai~ti.on (13) byl Converlting~ it inlto
adifferential-'- equa~ti~on to obtain

P. (R)-ep [ J----- OF](14,)

where the exponential expresse~s the survival fact~or. The

pr-obabi~i~ty of leakage into ai cont~ninuan~ state at any point

on the inward trlajector-y Rc. ER m i~s written as

Fi (R,)dR, = 1 exp [ J---- dRj (15)
Relb R19(R)

where Rel is the classical turning point or the largest value
of R, for which the radial velocity is zero.

The outward trajectory is treated similarly where the

probabilityy of le~akage occurring over- [R, R + dRJ is
wriitten as

m .R
Pout(R) dR( = [1] Pi (R )R '.-f' P >nut('d '

x~V (16)

The quantity enclosed in brackets is the probability of

survival over the entire approach trajectory and the part

of the separation trajectory Rel to R. Solution of this

equation gives

r(R) F(R') Rr(R')dR'
Pou(R) exp [- j------ dR' J ---] (17)
hvQ(R) R (') R hv(R')
c31 cl

and the probabijlity of lea~kage at any point oin the outward

part of the trajectory is written as

Pout (R R=ep(- J ----- d~R:
Rl RcP 1(R)
vm T(R)
x [1. excp [ ----- R] (18)

The p~obabil~ity density t~hat a transitioni occur~ at a

particular R is given by

P (R) = Pin ( Pou(R). (19)

Tlhe probability of a transitiion during the entire collision

is given 'Dy

P, = rP (R)dR
or "c1

P JPn(~R Pu(R)dR!. (20)
Rel Rc 1

Upon substitution of Equation. (15)and Equation (18), the

probability of ionizatijon for partial. wave R is written as
P (R)
P~ = 1 exp [-2 / -- dR:] (21)
where the radial velocity is given by

V*(R) (L+1) 1/2
v2(R) = v [1 - - - (22)
o Ek(m) 2L1R2Ek(m)

and ii is the reduced mass for the A*-B system.
The total cross section for these inelastic transitions is

written by the partial w~ave method by

o (vo) 2 (20 i + 1) (vo) (23)


Ko kIo) *()]-h 1/2 (2 )

The full quantum mechanisml dev~elopmenft gives the ex-

pected result of el~iminating6 the singularity in the argumecnt
hv(i:) at thei classical urningg p~oint. The quantity is

still peakecd near thef classical. turning point but is not

infiite.The most probable internuclear distance for ion--

izatrion transitions i~s still near the classical turning


D. Veri ficationo f the Classi];is~-ical her~

01son used the classical. formulas developed by Mil.ler

to fi~t the velocity dependence of the total ionization cross

sections for- the system Ne*(3rgPO2)-A measured by Tang,
Mari:us,arnd Moischlitz.8,51 The author used a potential

curve of the Lefnnard-Jones foirm, and an exponential form- of

t~he couplinG width given by

r(R,) = exp [-1.53 R}.61

A very good fit to the experimental data was obtained.

The theory successfully predicted the unexpected minimum

found in the velocity dependence of the cross section;

whereas, the theoretical development advanced by Mlicha, Tang,

and Muschlitz predicted a monotonic decrease of the cross

section with increasing relative vel~ocity.1 This increase

in the ionization cross section for thermal cnergies above

50 mneV was at-trijbutedt to thec exponentlial~ly increasing nature

of the coupling~ wilt~h as the distance of closest approach

decrefased wjith increasing energy. The calculations pre-

dicted thait tbo cross section would rise to a maximnum around

10 eV and thenz- dec-rease slowly with energy. Figure 5 shows~

the fit obtained by Olson to the low\ thermal data of Tang

et al. andi the high collision energy of M~oseley et al.161

01son also used the potential curve classicall theory

as a basis for calculationn of the veloc~ity dependec~:e of the

total ionization cross section for the system Hef(21S,23S- 1

Thne long orange van der W~aal intleraction potentials were ob--

tained from the wjork~ of Bell, Da~lgarno, and Kingst~on.2Te

rep'ulsive interaction potential!s were d7-ran fromi the results
of differenti::al scattering wo~-ctcrk dn b-y Sm~ith et al. in whiichl

screening constants were in1curporaeed to account for pr~o-

nounced shell structure effects.2 The potential wsas

writteni in thle form

V(R = 2/R (8 cxp (-R~/C ) +- 8 exp (-RZ/CL)

+i 2 exp (-R!/CK) ] (Cab/R6) 1 1 exp (-x)

(1~ + x 1/2;2 + 1/Gx3 + 1/24x + 1/120x5)] (25)

where CK, CL, and CM are the screening lengths of the K,

L, Mi shells respectively, Cab is the long-range coeff-icient,

and x -: R/CMI. The coupling frequency wras written. as

\j(R) = exp[ -R/B
where thec adjustable parameter for Hec(23S)-Ar was B = 0.667 ao

and for He(21S)-Ar wzas B = 0.629a .

The potential submnitted by 01son has been used successfully







C~) r-l ri
on to

u doo

r:q m co
O 000

ri, ao m G~

d 9-! (d
o G) > H r-
H: CO ed

(6 J rd 0 e

V C) a, a

O exr X$

ChP I O 3












0 o[


by several groups attelmptinlg to ob~ta-in theoretical.

fits- to t~he~ir excperlimnctal results on thn. He(23S;)-Air system;.19

Nichaus and Illenborg~er have recently mea~sured th-e velocity

dependence ce of the total ionizationn cross secti~on for the

systems; Hoc(21s 23S)--Ar, K~r, Xe, N2, andJ Hg.2 Usinge the

same V*:(R) and thie values B -- 0.357ao and A = 7400au for
the adjustable paramotocrs in tboe transition fr:equenicy for

the He(23S)--Ar system,, ILlenberger and N~iehaus ob~taine~d good

agicrcoment: between their experimental results an-d theory.

TLhe experimental. results cannot be regarded as absoulute crosR

sections. The absolute values w~ere determined by no~rmal~iza--

tion of their relative cross sections values to absolute

destruction rate conistants obtained from the flowing after-

g9low reCSultsR of Schmielteh~opf and FebsenfeLld.23
Pesnelle, watel, and Ma~nus have also obtained data on

the relative velocity dependences of the total ionization

cross section of the systems He(21S,23S)-Ar using a crossed-

beam apparatus.12 The results were reported as absolute cross

sections obtaHined by normalizing the relative cross sections

to the results of calculations again based on 01son's poten-

tial curve and Miller's classical theory. The values A =

4000au and G = 0.360ao were used for the adjustable para-

meters in the transitional frequency f~or thef He(23S)-Ar system

with good results. Pesnelle, et al. also found good re-

lative agreement between their mneasurecments and th-e absol~utei

destruction rate constants of Lindinger, Schmeltekopf and

Fehsonfeld using the approximate expression

R~ate: Constanlt (1.empT. T)!
.- - --(26)
01?) 2.2( )/

where~c thec value 2.2 is -ian arbi~tra-ryr normalnizeti~on facto_.122

Velocity-averaged absolute total ionization cross see-

tions of thei systems He*_(2 S,2 S)-L.r hav-e been obtained by

Rich act cl. and Rundel et al.; however, comparison with
their resul-is is difficult since the velocity distribution
of thleir metastable beam is not k~nowJn precisely.22 Assum~-

ingi a disitrib~t~ion of thef ty~pe -yexp(-v2/a2) anid a beam

temperature T = 3000Ki for thec velocity avieraged absolute
cross sections, Illenberger and N~:iahaus normailized? their

de tra to the~ veloc ity-averaged a1bs lute cros s se action s .22
thre m~easured cross sections werle so niormalized, they estimnat-e

that: their- reportedc cross sections woculd' be increased by a
factor of 2 for the Hie(23S)-Ar~ case and would b~e increased;

by a factor of 1.5 for the Hef(21S)-Ar case.
The strong increase of the total ionization cross see-

tio~n with relative energy pre~dicted by Miller's theory and

01son's potential for the systems Hef(21S,23S)-Ar has been

dem~onstratedi very wecll for He(23S)-A~r, but the H~e(21S)-Ar~

results are conflicting. Th;e Niehaus and Illenberger re-

sul~ts for Her(21S)-Ar shows veryI little relatives velocity

dependence; whereans, the results of P~esnolle et al. have a
mnuch stronger energy dependence.
The strong energy dependence of the total ionization
cross section is also demonstrated by results obtained from

flowiing-afterglow measuremeint s of Lindinger, Schmreltekropf,

and Fchsenfeld.24 The temperature dependence? of the total

des truction rate constalt~s of H(3)byAanotrgss

increased stirongely over: thel themel~l energTy rainge 30300 to


The approach of th-e flowing afE~irg30w technique differs

from the molecular beamn method in that the irate of metastable

destruction is measured rather than cho appearance of reaction

products. The afterglow~ from a discharge source is allowJed

to flow downl a tube at a volocit~y of about 104 cm/sec at a

pressure of about; I torr. The aft-erglow, which persists for

about 1 x 102 see, wrill extend downri the tube for about 1

meter. The reactant is added down-stream~ from thle discharge

source, and the destruction of metastable atoms by the re-

actant gas is determined at various distances along the flow
path by selective absorption of radiation at 3889 A (2 S to

33p optical transition in helium) andi 5016 R (21S to 31-I

optical~ transition in helium) by the~ helium metastable trip-

let and singlet states respectively. Flowsing afterglow

results for Hef(21S) arte generally not as reliable as He(23S)

results since collisions of the type

He(2 S) +e He(23S) +- e- + 0.79 eV

provide an additional channel for He(21S) metastable destruc-

The relative total ionization cross sections obtained

by beam experiments can be related to the destruction rate

constants to obtain absolute cross sections. Ill~enberger

and Niehaus normalized their relative cross sections to the

destruction rate constant at the kinownm temperature of 3000K.22

The proportioi~niti~y factoL K- correcting: ol~l(v) to absolute

cror:s sections "kbr:(v) was deterwiD;ned from the expression

jK.ogcl(v) f(v) vdv = Rat:- Constant (3000K) (27)

GAbs~v = Rel

and f(v) is the normlalized Maxwell~ian relative velocity dlis-

tribution at 3000K.

Relative cross sections may also be normalizedd to the

absolute veloci~ty-aver-aged total ionization cross sections

OAbs obtained by beam techniques by the expression

jK Rel(V) f'(v)dv =- ?Abs (28)

where f' (v) is the velocity distrjibution of the metastable

The object of the present w~ork is to report the velocity

dependence of the absolute total ionization cross section for

the systems He(2.1S,23S)_Ar. These results differ from those

previously discussed in that the cross sections reported are

absolute rather than normalized values, A gas cell technique,
in which the metastable beam is allowed to enter a chamber

of static target gas, has been used rather than a crossed-

beam technique. The gas cell technique is somewhat limited

since ionization product branching ratios cannot be deter-

mined; however, the method should give reliable absolute

cross sections at accurately known beam velocities. The

experimeintal procedure is basically similar to thatr of Tang,
although the arrang~ements for rignal collect on anid the

data analysis technique have been changed.28


A. Metastab-le States

Low voltage discharge sources produce many electronically

excitedi states by collision of ground state atoms with

01ectrons. While a large numbe, r of diiffrentr excited states

are possible, the lifetime of mnost states is very short
where electric divole radiation is allowed. Excited states,

ho:.ever, ex;ist~ For which the, electric dipole m~atrix rransi-

tioni cenleentl for transition to the ground state is zero.5

These stable excit-ed states, with lifetimes greater than a

microsecond, are reifered to as metastable ex~cited states.

The selection rules govrerning electric dipole radiation,

where Rcussell-Saunders coulpling is assumed, are as followus:29

(a) 6 S = 0,

(b) n L, = 0, +t 1,

and (c) n J = O, + 1 excepting J = 0 f ,J = 0,

whero S = total spin angular nmomntum?,

L = total orbital angular comen~tum,i

and J total angular momentum.

Considering the case of heliumn in particular, two inetastable

states exist. The 1s 2s (23S) level in helium is the lowest

possible tripleti level since the only lower level is the

siniglet 1s? ground st~ate. Selection rule (a) forbids triplet

to ::inglet t.ransi.tions anid trhe res~ultant lifertime of the

state is about 6 x 105 seconds.3

The second helium- mecastable state is the sing~let state

den~oted bsy ls 2s(21S) in which the transition to the singloct

ground state is forbidden by selection rule (c). The life-

timec of the singl~et state is of thes order 2 x 10-2 seconds.30

B. nSel~ection of2S r23 eliu Meata7 Atom
froml a Mix~ed Beain

A beami of triplet helium metastable ztomns can be ob-

tained fromn a mixed beam of singlet and triplet helium neta-

stable atoms by irradiation of the mixed beam with radiation

fromt a hel~iumi discharge lamp. The 21S to 21P, 2,0)6 p radia-

tion emitted by the helium discharge promotes the 21S meta-

stable atomn to the 21P' state.f The 217 state atomis then

decay preferentially to the ground state, and the sin~gl~et

state is said to be "quenched." The triplet 23S state atoms

also absorb radiation from the helium discharge lamp; how-

ever, the triplet can only radiate back~ to the 23S spin state

since no lowYer level exists for the triplet systeml.

C. Detection of Metastabl~e A*tomo

Mletastable atoms can be detected indirectly by counting

the electrons ejected from metal surfaces by incident meta-

stable atoms. This process: of dfeeci~tation is thought to

occur by Augor deexcitation of the metastable atom in which

the metastable atom is ionized by the tunneling of its ex-

cited electron into a vacant energy level in the metal. The

resultant ion is; then~ neutr~ializel upon impi;act with the metal.

surface andi a free electrron subsequecntlyy ejected.' Tihe

ratio of free electrons ejected froma thie metal surface for

each incident metastable atomn is called thle secondary elec-

tron coefficient andi is usually denoted by the symboL y.

Dunning and others have measured this ratio for several

different mietal, surfaces and have studied the effect on the.

ejection coefficient by surf ace contamination and by the

angle of incidence of thc mnetastabl~e with the target sur-

face.31-35 Te authors conclude that the secondary electron

ejection coef-ficient varies with the condition of t~he metal

surface and thiat the coefficient is only reliable when deter-

mined from surfaces under actual experimietnal conditions.

Using a gas cel.l technique, the authors measured t:he value

of the secondary electron ejection coefficientr for singlet

and triplet helium metastabl~es to be 0.46 and 0.63: respec-

tivel~y. The metal surface wlas gold plated and the target

gas was argon.

The values for the secon-dary electron ejection coeffi-

cients for singlet and triplet helium metastables in this

wJork were determined from the same experimental data used

to determine the ionization cross section. The expressions

used in this analysis are givein in the following section on

total ionization cross section.

D. Velocity Distribution of a Maxwell-Boltzmann Source Cas

The metastable states of helium are produced by a low

voltage DC discharge w~ith source gas pressure of about

0.1 torr. A Mlawlc l-B~oltzmann veloci ty di stributiio n my be

expected at these gas pressures since the. very shlort mean

free path of the atoms leads to the-rm~al equilibrium~. The

resultant velocity dependence of the intensity (particle

fluxc) of the mietastabl~e atomic beamn is given by

I~~d =4 0 2exp ( 4-)dv (29)

where I is t~he total beami inten-sity, ~22 2kT), and mn i~s the

mass of the atom. The velocity distribution, however, is

altered by thle chamber exit slit wjidt~h which is much smaller

than the mean free path of the discharge gas. The gas will

escape the chamber by effusive flow w~ith the escape of

higherl velocity atomas being favored.36.37Tbm velocity

dependence of the modified M~axwell-Goltzmann gas is given

21 2
I(v)dv = --v3 exp (-j )dy. (30)

The velocity selector further alters t~he velocity

distribution of the transmnitted beami. In cases (such as the

present one) in which the selector transm~ission band is

narrow in comparison to the width of the function Itv), the

distribution function of the transmitted beam intensity


Ttv) vI(v) (31)
and the velocity distribution detected at the collision

chamber is given by

Ttv) v^ exp ( %). (32)

Exjper~i~mentlc~ 1 verifi~cati~on of t~his ve lo0city dependence e is;

given in Chapter IV.

~E Total lonizacionn Cr-os Se~ction andu SecondaryElectron
Eicention Coeft~iclents fo~r Meftactabl~e Atom~s

The Ltotl ionication cross section and the secondary

electron eject~ion coefficient can be evaluated from qluantities

measured by exrperiment. The. ion- and elect-ron current, re-

sulting from the reactive colisioni between the met-astabl.e

beamn and the static target: gas, are measur-edi separately anld

supply` the basic data needed to dectermine the ionization

cross section~. The reAct~ionl path lengthl is thre distance

from the entrance slit of thl~e gas cell to the" rear w~all of

the collision cell. The length, 1E, of the reaction path, is

2.25 cm. The number density, nI, of the target gas is assumed

un~iform!: throughout th:e entire collision ch~amiber. Th~e temper-

ature of thle target gas wras assilmed to be 240C the ambient

temperature of the apparatus.

The loss, due to chemiionization, of metastable beamr

incensity, dl, along a small interval, dx, of the reaction

path is written as

-1 = -0 ndx. (33)

Upon integration over the total reaction path, thle intensity,

IQ of- the. remaining metastable atomrs including those

elastically scattered is given by

I = Iecent" (34)
R o

where 01C. is tzhe total ion,?iza~i~on cross sectioni, n is tihe

target gas number density, 10 is thie intensity of mietasrable
atom;s entering the collision cell, and t is the total reac-

tion pathi length.. The ion current, i+ collected from this

process is obtained f-romi the~ differences between the mecta-
stable besam intensity entering the collision chamrber anrd the

intensity at distance L or

i+ I IR I (1 e- t) (35)

Electrons come from two sources. Ionization will con-

tr-ibut-e a~n electron cur-rent equal to i ~. Electrons are also

generated by unreacted metastables striking the rear of the
collision cell. The fraction of mectastables detected by

Augerr deexcitation is given by the secondary electron ejec-
tion coefficient, 71, and; thec electron: current generated

from the deexcitation process is Y(o - i ). The combined
electron~ cur-rent, i. is therefore given by

i_ = i+ + Y(Lo i ). (36)

The target gas number density is corrected to the den-

sity at 3000K~ and is related to pressure in mtorr units by
13 ,_3 3000K P(mtorr) (7
n = (3.2193 x 10 #/cm1 )(290 tr

where 3.2193 x 1013 #/cm3 is the number density at 3000K; and

1 mtorr pressure. For convenience, the argument ofr the ex-

ponential. termn in expression (35) is consolidated and re-
written a-s

i+ = I [1 exp (-cp) i (38)

c = 7.317 x 10 10t (A2) ] (39)

and the pressure variable .s in mtorr units.

The total ionization cross section could, in principle,

-be determined from expression (38) if the pressure decpendencee

of T0 wer~e known.-l It is mrrile expeldient, howJeve~r, to develop
expreessions inl which I_ is eliminated and involve only the

experime~ntall~y measured values of i+ and i_ at k~now~n tar-get

gas pressures. This is accomplished upon division of (36)

by (38) with the result

exp (-cp)

1 exp (-cp)

sim~plif'ying to

= 1 + V [------- ] (Ic08)
i+ exp (cp) 1
FollowingS Tlang, the expansion

1 3 4
1 1 x- x x
exp~~ (x 12 720 720

can be used to obtain

i 3 4
(1 Y) + Y 1- e ~-P ye p3 ye p4... (40b)
i 2 c P 12 720 720

and the linear approximation

S(1 Y) + 1(40c)
1 2 c p'

where ep
developed by nultiplication of (40a) by pressure and is-

writteni as

1_ P
( )F P +- y [------- --- (41la)
1 ~exp (cp)) -1

where the same expansion gives

i 3
P ) 1+ 1 )P+ e2 ePh (41b)
i c 2 12720

and finally the linear approximation

P (-- ) + (1 ) (41c)>
i c 2

The linear approx;imation (41c) gives very little error

for target pressures up to 5.0 mtorr. Thli s was determined

by comparison of calculations using (41a) and (41c) and

appropriate values for 0 and y. Both (40c) and (41c) appear

to be suicablef for data analysis, but (41c) should give a

more reliable intercept. It way be showni that

i Y
Lim ( --
P'-o i+ c

for both (41a), using L'Hospital's rule, and (41c), whereas

for (40a)

Lim P (--) = 1.
P.,m i,

Thus, if Equation (40c) is to remain reliable, care must be

exercised to insure that the condition cp << 1 is met. Ex-

pre-ssion (41c) was therefore used to determine both y and a

fromi experimental meiasuremen-ts of i+ and i_~ for a series of

common target gs3 pressures.

Othecr expressi~ons can be used w-ith varying success to

evaluated c; and y Ulsing (36) anid (38), thie expression

1 1 ex (cp) 1 ](42a)
1_ 1+ 7

canl be written, w~hichl upon expansion of the exponen~tiail term


1~ 2I p2 c3 P3 (42b)
i_ iS Y 27 6Y

Ther values of a and y can be dtc--~enrmnd as adjustable palram~-

eters from a least squares fit of experimental dat-a to the

forme of expression (42a). Relative cross section canl be

determined from the linear (first approximation) part of

ex~precssion (4-2b). Th~is final- procedure is re-l~iable only for

data taken at vYery small target gas pressures.


A General Decigy

Th-e molecular beam apparatus shown in Figure 6 is a

modi.Fication of the appa-r~aus previous-ly described by

Tang.28 The aipparatus has four separate chambers: the dis-

chargo chamber, the fore chamber, the post chamber, and
the collision chamber.r

Admission of the beam gas into the discharge chamber is

controlled by an automatic pressure controller. Production

of mietastable atom~s in thie discharge chambler is achiev-ed

using a lowi-voltage DC( discharge. The geas exits the dis-

charge chamber th-rough slit S-1l and enters the fore chamber.
The beam~ gas then passes through a second slit., S-2, into the

post chambers:.

Charged particle formed in the dischar-ge are removed
from the beam when it passes between two parallel deflection

plates. A 450 VD:C potential is applied across the plates. A
solenoid driven beam stop, BS, is located directly behind the

deflection plates so that background readings mlayy also be
taken during the experiment.
The beam is further col~limated by slits S-3 and S-4

which are positioned At the ends of the velocity selector.
In addition to mechanically selecting the velocity range to

be studied, the selector prevents any photons gen~rated in

the discharge from reaching the collision chamber. Some










I nz

j. -

zz' an z~

i In
10:i, -ii n

L _..e

an a 0 ------

P~~ -r-

za aC zz a

SLates wjjith hig1 cl~ect~ronic excitation energy are also pro-

dcen~d by thle discharge; however,, their radiative lifetimnes

are short compared to che timre required to traverse the

dist-ance to the collision chamber. Their fast decay leaves

on-ly netastable and ground state atoms inl the beam. Since

thle det~ectolr sys.temi inr th-isr appara~tus is insensitive t-o

ground state helium atom~s, the beam may be considered a pure
metas~tabl~e beam.

The quench lamip, located between the final two disks of

the velocity selector, is coilerd around the beamn path to ob--

tain maima~LILn photon f~lux across ther beam path. Heliumn 21S

metastabl~e atoms are removed froma the beam:i during the opera--

tion? of the Lamp in a process where the 2s electron of the

21S state is promloted to a 2p state by 2.00 pi phioto;:s front

theo lamp. Transition fromn this state to ground is heavily

favored and the 21S state is subsequently quenched. Thus,

measurements are performed on a beamn of mixed metastable

atoms or on a pure beam of triplet metastable atoms. Infor-

mation on singlet state metastable atoms is obtained fr:omm

the difference of the two experimental measurements.

The beam is colljimated a final time by slit S-5 in the

collision chamber superstructure and then enters the

collision chamber. The target gas is admitted into the

collision chamber by a feed-through in the rear of th~e

collision chamber. Also located in the rear of the collision

chamber are the heater and thermlocouple from an R\CA 1946

thermocoupl.e gauge which are used to measure the target gas

pressure. A: moree det~aild description of the experim~ental

apparal~tus" is give~n in th~e loinscios

B. Vacuumn System

Fo~re Chamb~er

The fore chamber was constructed from a 304 stainlecss

st-eel tee 6" jini lenthl wiit~h a 0.109" walL thickness-. Thle

vacuum equipmernt- w~s a nom~inal. 4" fractlonating oril diffusion

p~ump, containiing Convalex-10 Puicp fluid, backed by a 1.5

litrs/eemecanialfore pumlp.394 An air--cooled chevron

baffle was insertied between the ch~amb~er and diffusion pu~mp.39

A chilled baffle w;as unnecessary since the pump fluid has

a very low vapor pressures at rooml temperature. Contamrina tion

of th~e dirffedon pump f:luid by~ me~rclanicral pump, olil. vapor was

prevented by an Ultek fore line trap installed between thie

two pucpsi.

Post C~hambler

The post chamber was constructed from heavy aluminum

alloy of 1.25" thickiness.42 The chamrber walls were welded

together. The top and bottomn of the chamber, made from the

samte metal stock, wiere secured to the sides by a number of

bolts. Arrangements for o-rings in the lid and bottom in-

sured a vacuum-tight seal. The chambers inner dimensions

were 10 1/2" x 12 1/2" x 25 3/4". A 5 3/4"-diameterr hole,

covered with a lucite flange, provided a viewport through

one side. wall. TIwo vacuum-tighlt feed--throughs for the gas

input and output tubes of the quench lamnp were located in the
window of a simnilar v\iewipor-t in the lid.

A nomo~inal 6i" lractironainr-g oi~l diffusion piump, usin3 Con~-

voil.-20 pump~ fluid, backerd by a 6 .iters-/sie mchanica1l ore

pum1p cos'prised7 thie vacuumi system of t-he post chamber.39,0

Thef diffusioni pumnp had a perapinjT speed of about 1,000 liters/

see at 1 x 10- torr. Pump fluid bac~kstrerningn was minimized

by a refrigierated baffle. Tihe refrigeration systemi cooled

the baffle to about -400C us'-ing Freon-12) as the coolant.

Prressu!e iHeasurement

Pressures in the fore andl post chambers were mneasur-ed

by vacuum; ionization grauges. Wi th no gas entering the two

chaimbers, the fore chamber pressure wns 2 x 107 torr and

the post chamber was 6 x 107 torr. The fore chamber an

tainted a pressure of 2 x 105 to-rr when beam gas was ~dmi~tted

into the discharge chamberr. The pressure i~n the post chnimiber

fluctuated slightly with the pressure of the tar-get gas in ihe

collision cell. The post ch;amber maintained a 3 x 106 torr

pressure when the target gas pressure was 4 mtorr.

Several methods were used to protect the diffusion pumrps

during ope-ration. Sensing tubs for a thermrocouple pressure

gauge controller w~ere installed betw:een the backing pump

and diffusion pump of each chamber.39 If the backing pressure

of the cost chamber exceeded 50 mtorr, the controller auto-

matically shut off the power to both diffusion pum~ps. The

system wias also protected from any malfunction of the servo-

mechanism controlling the gas; pressures in the discharge

chamber. A full-scale reading of 10 x 105 tor-r on the

ionlization gauge control~ler m~onitor~ing the fore chamcber ter-

m~ina~ted a sw;itchedd output to the Aultomaltic Pressu-re Con-

Proler.4,44The controller automnatical~ly closed its leak

The pressure in the wlatcEr livecs to the two diffusion

pumps" was monitrored by "F~lolt~rol" un~its. Thecse devices

automaticaally termin~lated power1 to thie diffusion p~umps if thne

water~ line pressure dropped below. a critiical level. These

units have some shortcomings in their ability to protect the

diffusion7 pumnps. The units m~easur~e tbe.water pressure drop

across a narrow~ orifice which also restricts the flowj rate

of the wate~r. The flowj of water- is stopped completely if'

the orifice becomes obstructed: by debris or mineral deposits.

Unfortunately, however-, t~he unit will continue to sense a

sa~f pres-ure drop across the orifice and fail to shut dowjn

the diffuvsion pum~p.

C. Gas Source

Auxi~liar~y Vacuum Syst~em and GasSore

A glass vacuum system with 1_0 glass 12-litor- storage

bulbs was used for gas manipulation and purification. The

system was pumped by a mercury diffusion pump backe-d by a

mechanical pump. The gas wJas admnitted into the storage

bulbs after purification by either a degassed, activated

charcoal absor-ption trap at liquid nitrogen temperature in

the case of helium or a degassed wool trap at acetone-dry

ice bath temperature in the case of argon.

Target Ga-s_ Delivery
T~he target gas, argaon, w!sn admlittecd into thle co.i~sion

chamber from its storage bul~bs through- a manually controlled

Vactronic leak valve 46,(17'Two~ or mcre~ argon st-orage bulbs were

used simulraneouslyy to enlarge the volume of~ gas backing the

leak valve. In spitie of~ tlis pr~ecaution, the target gas

pressure still decreased slowly w3ith time. This difficulty
wans circumnvented by makingg pressure readings as quickly as

possible before or after pressure related measurements.

Be~cam Ga Delivry_

The flowj rate of the helium beam gas from the storage

bulbs to the diischarge chamber was controlled by the servo--

mechian-ihs of a Granville-Phill~ips Automatic Pressure Con-

tro!ller (APC)! Th-e. Autom!aitic Pressure Cnrle n e

l~ated equipmcntr are= shlown in F~igure 7. The servomechanismT

consists of a motor-driven leak valve. A high pressure

NL-7676 ion gauge, placed between ~te leak valve and discharge

chamber, measured the linef pressure. and the gauge's ion

current served as an input signal to a picoamm~eter.484

The picoamm:eter then supplied the APC witilh a signal. propor-
tional to the line pressure. Supplied with this signal, the

automatic controller functioned as a proportional controller,

adjusting the leak rate by driving its servomechanism. A
500 ml bulb buffered short--termi pressure fluctuations in the

line following the leak valve.

S. i-... .. -

[WL- 7G7

i~~~~ ~ ToP SslT.O6Cl'l?? so`c

Figure 7. Beamn Gas Delivery System

D. D~ischailrg Chamber


The discharge chamberci was1 patterned after a low voltage

DC, discharge sour-ce designed! by Rot-he et al.5 Basic details

of the dischaRge chamnber miay be seen in Figuiro 6. The

cham~ber wras amounted to th!e removable lid of~ the f-ore chamber.

Tihe chamber con~sisted of a Pyrex tub~e C, water-cooled copper

end plates A' and Bi, and a filament F, supported by Inconel

rods R. The ends of the Pyrex tube (5.1 em OD and. 6.5 cmn

long) we~re sealerd by plates Ai and BI. The source sli~t, S,

was located on t-he anode plat~ie A. Plate BI contained the

source gas inlet: and t-he electrical feed--throughs for the

Inconel rods supporting the filament.


The ends of the tungstun ribbon filament woncr spot: welded

to nickel tabs wh~'rich~ werce in turn spot welded onto the

flattened ends of the support rods. The f~lame~nt- wras then

cataphoretically coated witih thoriaa following the procedure

described by M~uschlitz et al.51 The dimrensions

of the filamenit wlere 0.0254 mm thick, 2.5 mmr wide, and 3.8

cm long. The distance between the filament anld anode w~as

2.0 cm.


A Lambda regulated power supply, operated in the current

regulated mrode, supplied cu~rrenit to thle filament.5 To re-

duce backsputtering, the fila~ment: was floated 50 volts above

ground by a Henthkcit power supply.53 Th ano~de voltage wuas

supplied by a Houilatt Pack~ard powder supply operating in t~he

constan~~t cu:rfrent mode,5L

E' Vel oci~ty Selctoricto


Thef velocity selector was patter~ned after a sl~otted-disk

design orf Kinsey. in whlich the belical. slot path in; approxu-

imated by mazchlruing slotis tangenitial to th~e desired h~l~ical

pathii This procedure offers several advantagess over other

slottefd-disk desiigns. The advant:ages are as follows: 1.)

simpilif-ication of machi~ning pr~ocedures; 2) accommnodati~on of

higher~ helical angles w~ithoutl serious loss of transmitted

intensity; and 3) incorporationn of one dlisk of greater thick-

ness than the remaining d~iskcs Ireduces the~ total num;ber_ of

disks .

The disk~s wrere made from sheets of very hLIgh tensile

strength aluminum alloy, Alcoa 2024-T13. A tungsten carbide

saw, 0.022" thlickl with 20 teeth, was used to cut the slots.56

The disks were stackedd in the appropriate order and clamped

between thick end pieces. Thus, the respctive slots in each

disk wJere cut simultaneously. Two 1/8" diameter holes were

then dlrilled thr-ough each separate disk 1.800 apart to serve

as aligniment aids during th-e assembly of the~ velocity


Table 1 contains information pertinent: to the physical.

configurAtion and operating paramefters of the velocity

selector. Thle rotor assembly is shown in Figure 8.

Table 1. Characterist~ics of the S10tted-Disk Velocity

Parameters Set V

Total number of diskis

Diametecr of dis~s

Number of slot~s per di~sk

Length of rotor, L-id

Total angular shift Q

H~elical pitch dz Ld v

Conversion factor for angular
velocity v to transmitted velocity v ,

Length of slot (tin radial direction)

Slot- wiidth, 1

Angular width at base, 64

Tooth thi~ckness betwJeen slotis
At base of slots
At top of slots
Average value, Ez

Average opening fraction'
n = ~11 1 2)







vo(cm/sec)=737 .2v(Hz)






Thickness of disks
3rd: disk

Average radius, r

Cutting angl.e for slots,
X = tan^( T)




Table 1 continued

Parameters Set V

Position of alignilng holes
(clockwise wiith respect~ to
the same cut)
1st diskl 00
2nd disk 5/160
3rd disk~ 5/80
4th disk 1 1/40
5th disk 2 1/20
6th disk 50

Velocity spread
R-1 ma mrin Y 0.1.01
2vo 2vo 1- Y2

v n/v = 1/(1; + ) 0.909

v /v 1/(1- 7)1.111

Geometric factor 007
G = n7
1 -? Y2

T(vo) = G vg 1(g)

Figure 8. Velocity Selector Rotor Assem~bly (Af ter Ref 28)
1,2,L,5,6 ....1./32" thickness disks~
3.............1/8" thickness disk
7 12. ... .. .snacers
13........... .bearings
14,16........ .collars
15............. k-28 N. F. Allen screw
17............. 3/8-24 N. F:. nut
18.............3/8" rotor shaft

Is 8



L_ ~ i~i:13

'"--- ---- -6

3.%111~ -11


1.000" 10

--- -3--

0. 250"
.250"- e--J--1

R!otor M~ouinting and Bearings

Bardlen iiartemp, bearing:; SRGSSTD-DB:2, lubricated by Ball

Brother's vacuumn coating~ process, wrere press fitted onto the

rotor shaft575 Each bearing was then mounted isd

brass ring enclosure. Each. end of thle rotor assembly w~as

supported by 4 tensions springs attached to a vertically

moveable block and to the brass enclosure.5 Rubber tubing

w~as placed around each spring to dampen mechanical vibrations

associated w~ith the spring suspension system. This flexible

support system allowed the rotor to spin about its own axis

of rotation, elimrinating the necessity of balancing the rotor

to a very high degree of accuracy. The~~ rotor assembly could

be low~ered out of the beam pathi by a gear shaft passing

through an 0-ring seal inl the vacuum chamber wall.

Moto and PwrSupl

A Globe tw~o-ph~ase, two-pole, hysteresis synchronous motor

was used to drive the rotor.0 Thle motor was secured wJithin

an aluminum housing for shieldi~ng purposes and the motor

housing mounted on four posts extending6 from the rotor

assembly housing blocks. Thius, the mlotorr and rotor could

be moved as a unit. Piano wi~re (Nio. 13, 0.031" diameter) was

used to couple the shafts of thec motor and rotor assembly.

Garden Bartemp SRASSTB5 bearings with Eall Brother's dry

vacuum coating were used in the motor to insure long operat-

ing lifetimes.575

The power to the motor was supplied by twJo CMIL variable

freqluency (50 1800 Hz.) power amplifiers.61 The input

signal to the motor was controlled by gangped variacs as seen

in Figure 9.



o qo
ol rf
"i 0



`i e
~I f vR

coo ou

.c~r~aI ~Or .
a (Li l0
s 900
1,~ C0

The rotor fre~quen~cy wasi measured by taking advantage

of the opticallyr clear path? through the~ alignment holes of

the rotor disksr. A high-outprut lecns-ond miniature lamp was

iinserted into one of the housing blocks and a photodiode in-

sorted into the opposite housing blockc.26 The lamp an:d

diode wenre positioned in the housing blocks at the samne

height as the alignment holes. Thus, the photodiode censed

chopped light pulses which w~ere proportionall to the operat~-

ing f'requency of the rotor. The diode became conductive when~

ex~posedl to the light. puls:es an~d pr-oduced a pulso in th~e

current which is shown ini Figure 1.0. These pulses were

counted and displayed b~y r. Transistor Specialties, Incorpor-

ated counter.6

F. QuenchLapp~E


The design of the quenlch lamp was somewhat restricted

since the quench coils wecre placed b~etween the last twyo disks

of the velocity selector. This placement wJas chosen to

limit the beam path, thus avoiding loss of beam intensity.

The glass coil part of the lamp assembly was constructed

from 5/16" OD Pyrox glass tubing of standard w;all thickness.

The coil consisted of 4 1/2 loops, its length 1 1/2",

and inner diameter 3/4". The glass- envelope for the eloc-

trodes was formed from 7/8" glass to metal Kovar seals. The

alumiinum electrodes were 1 1/4" in length and 1/2" in width.

tz U ~~
n N
I i -1-'
r-- I .~i't:;----------

c~ ri
O r~

s---ll:~~----------------- ----------

I 1
--- --I -i--;-------

!---I ----;
~:~i~n O
'' "
R ~-,..~.~~~-~


i ;1
- p ol


The ends of the electrodes were threaded into brass plugs

as may be seen in Figure 11. The br~ass plugs and the elec-

trode metal envelopes were soldered toGether to insure a

vacuuml-tight seal and maxiioum obimic cont-act. Small 1/8"

holes, drilled through each of the. brass plugs and the

thr~eaded ends of the electrodes, allowed a steady gas flow

through- the electrode--coil assembly. H~ard dirawn copper

tubes, 1/4" OD, were soldered onto t~he brass plugs and

passed through vacuum-tight fi~tting3 in the main vacuum

chambers wall.

The ground elct~rode of the I-mp was clamped directly

to an alumiinum plate support to insure a good electrical

ground and to provide a good beat sink!-. Th'le floating el~ec-

trode wa:s clampecd to the plste wi~th insulltjng? pleZigllss


Lamp Vacu~um Line~

Vacuum rubber hose was used throughout the quench

lamp's vacuum manifold. The entire manifold could be pumped

dowin to 0.2 torr w~hen no gas was admitted into the manifold.

A constant flow of gas was3 admitted to the mnanifold from a p

tank of commercial grade hclium~. The gas was then purified

by a liquid N2~ trap. The lamp operating pressure was reg-
ulatecd by a Vactronic leaked valve placed between the trap and

the high voltage electr-ode of the lamp. Pressure measure-

ments wrere taken with a Hlastings Vacuum Gauge placed between

the gr-ound electrodef anld th~e mechanical pump.65

The lam~p operated with a constant flow of helium gas to

Gan Gas
inlet outlet

./8" glass to metac;l

3/ Prx lscol

(536 OD.

E------ lb" ----~

Figure 11. Helium Quench Lamp

remove imrpurities resulti~lle fromt the elevated temperature

of: th~e lamp du~i~ng Th-e dlischlargee gas flowecd from

the positive, high voltage electrode toward: the ground elec-

t:rod~e. The reasons for th~is a!rra~ng~fement w!ere as follows:

a.) most of the energy produced f~romr the discharge was

carried by posl-itive ions to the groulnd! e:lecrodle whichi had

an excellent heat sink-; b) aluml-inum sput~teledd from the ground.

electrode by positive ions was carrie- d downstream, awray from

the coil by the helium carrier gans; and c:) electrons, which

carried much less therrml. energy, wjent toc the positive elec-

trode which had little heat sinklinge capabilities.

TIhe voltag~e to start and mnairltain diischarge was supplied

by a Hewlett-Packard DC power sup~ply.6 A high-power., 10K(

obsi, _500 watt~ resistor wras placred in series with t:he lamp on

the high violtage line to maintain a base resistance during

the initiation of the discharge. This wras necessary to cir-

cumvent the instrument's current: limiting protective cir-

cuitry whichi would have been act7ivated by the apparent short

applied across its output termiinal~s.

G. Collision Chamber and Dete~ctine Systems


The collision chamber wjas a modified version of chambers

used by Smith and Muschlitz and later by Tang.72 'The

basic details of tlhe collision chamber are shown~ in Figure 12.

The chamber consis~ted of 4 parts which wJere as follows: a)

gas cell, b) collision coll, c) superstructure, and d) pres-

sure-sensitive thcrmnoccuple.

o vI

!_ t Ll I 6 1
j- 1 .s--5
1 0 --




-~ CI
) "" i

o j:

i, /
t~ ~~___

The gas cell, GC, wias a bra:Ss iyl~inder wi~t-h a siinle

slit opening, S-5, through which the metastable beamn entered.

Table 2 is a listing of the dimensions of slits used to de-

fine the beam. The gas cell nli~t was mluch smaller than the

expected mnean free path of any atomic gas contain-ed within

the gas cell.. Thus, the ga~s would ex~it the cell by K~nudsen

fl~ow and a high pressure differential between thie gas cell

and the main vacuum chaimber be miaintained.6

Target gas wras introduced into the cell by a feed-

thr~ough in the rear of the collision cell from the Glass

bulbs of purifiedd gases. The target gas pressur-e wras con-

trolled by a Vactronic leak~ valve placed betwJeen the gas cell

and; the glass bulbs. The gas ce~ll wias elecctrically insulated

from all other parts of the collision chambers and could be

electrically biased with respect to ground.

The collision cell (CC), mounted inside the gas cell,

was a thrfee-part assembly. The cylindrical side (CC2) and

the cylinder bottom (CC3) were elecctrically common and could

also be biased with respect to ground. The beam gas was

admiitted into the collision chamber through a slit, S-6, in

the cylinder top (CC3). This slit was slightly larger than

the gas Cell slit to avoid obstruction of the metastable

beami path.

Signal Detection

All electrical charges resulting from collisions of: meta-

stables with target atoms or with the rear collision cell

wall were collected at the collision cell lid (CC1) when

Table 2. Dimensions of Slits

Slit S1 S2 j3 S4 SS S6
Width (mrm) 0.6G4 0.84/ 0.64 0.64 0.48 1.56

Length (run) 6.315 6.35 5.49 5.49 3.00 6.39)

approp;-iate repulsive pote~ntials er~el used on all other sur-

facesi. These chanrges wecre then meas-rured by a Cary VTibrating

Reed Electromieter usingg an input resistor of 101 oh~ms.9

Positive ionls resulting from ionizing collisions werre

collected on CC1 by applying equal, repulsive, positijve

potentials on GC, CC2, anid CC3. Tlhe potential app'li~ed to the

gas cell prevented ions focused by thle cylindi ically--sym-

metric field from exiting the collision ch-amber. Coniverse-

ly, all electrons were collected at surfaces having a

positive potential.

The electron signal was collected at CC1 after applica-

tion of negative voltages on GC, CC2, and CC3. The voltage

on CC was again primarily designed to insure total charge

collecction by preventing anyr foculsed beam~ of electrons from

exiting the collision chamber through slits S-5 and S-6. The

voltage on G;C wa~s larger than CC2 and CC3 to prevent the

creation of additional secondary electrons through collisions

of highly energetic electrons with the surface of the gas

cell near slit S-5.

The superstructure, mounted on t~he exterior fronlt of

the gas cell, w~as a precaution against inadvertent collection

of electrons formed by mnetastable collisions with the edges

of the gas cell entrance sli~t. This structure w~as insulated

from the gas cell and biased positive 90 V w;ith respect to


The output from the electromoter was monitored by a

chart recorder and integrated over a period of time to

average out noise.7 An operational amnplifier was the main

compnr""nlt of' the integ~Frator and r: logic control circuit con-

trolled thie integr-ationi tina.71 Thec integrated signal was

displayed by a Nlewport di!:ital voltmecter and recorded on a

-Digtec printer.727

Presrsure Mieasurement

The pressure of the target gas inside the collision

chamber was controlled by a. Vactronic high vacuum~ leaki val~ve.

The theater and thlermiocouple elements of an RZCA 1946 thermo-

couple gauge, located in the rear of the gas cell, measured

the target gas pressure. The two heater leads wiere connected

across a DC constant current: suppl.y. The DC current level.

was determined by measuring: the potential drop across a

precision 2_ chm resistor located in series wJithi thee thrmio-

coup'le gauge.

The EMFI measuirement was taken across a second pair- of

thermocnuple leads. This measurement was made by a L~eeds

and Northrup universal potentiometer and galvanometer with

an EpplcjyLaboratory standard cell.757 The w~orkcing voltage

source pow~ering the potentiomieter wrias obtained from! Leerds

and Northrup.75


A. Cal3ibration of Target Gas Pressu; re Gauge~

Thle RCA 1946 thiermocouple, miounlted wjithin the collision

chamber, was used to meas-ure thei t~argoti gas pressure. The

collision chamber and therm~ocouple pressure gauge assembly

was placed inside a vacuum bell jar and the thermocouple
calibranted at known argon pressures with a MIKS Baratron.77

The calibration procedure was as follows. A current

of 70 met was supplied to the heater leads of the RCA 1946

gaugei, andc thet EM~F (EO) wa3s meliasur~ed at a high vacuum of
1 0 or Thle bell jar w~as theni isolated from its

diiffusion pump~ by a butterfly valve, and a small amount of

argon introduced into the bell jar. Following equilibration
of the argon, the EMF (Ei) of the gauige and the Baratron

pressure reading P' (mtorr) were taken simultaneously. The
bell jar and its contents were evacuated after each pressure

reading to minimnize contamination by atmospheric gases. Con-

tamination by) atmospheric gases, honwever, was not a serious

problem since the leak rate of thec system was about 2.2 x
10-3 mtorr/min with the butterfly valve closed. Since the

absolute values of Eo and E vary w;ith ambiont temperature,

the pressure dependence of the RCA gauge was determined as

a function of (Eo E)/Eo = AE/Eo which depends on gas


pressure only. Several. pressure rea~i~ngs weocre taken be--
tween~ 0.2 rorir~o and 8.0 mtorr.

The linear relationship between the quantities A~1E/

andl P (mitorr) is demonstrated by F'igucre 13. The parameters

mn anld b of the relationship F (mtorr) = m (hTAE/E b wiere

determined by a linear least squares fit to the data with

the results being

P (mtorr) = 148.502_ (CIE/Eo) 0.2098
with a correlation coeff'icientt of 0.9999.

B. Conditions for Hectastablee Production

Tihe filament was operated +50 volts with respect to

ground at a current between 12.0 aind 14.5 A2. The beliuml

disch,-rge gas pressure wJas mnaintained at: 0.09 torr as

detelrmined by thle WL 7676 high--pressure gauge. The cor-

responding reference setting on the automatic pressure
controller was 880.

The anode current was set at 4100 mA to obtain the op-

timrum7 metastable intensity. The. anaode voltage varied be-

tween 60 and 65 volts depending on the condition of the

filamnent. Any charged particles produced by the low-voltage

discharge were deflected from the beam by a 450 volt poten-

tial across the deflection plates (D).

C. Distribut:ion of Iletastable Vel~ocities

The height of the 50, 6-disk velocity selector wJas

adjusted with respect to thle beam1 path to insure optimum

Figure 13. Calibrationn Curve for RCA 1946 Thierocouple
Gauge for Argon

rl '0:


0.9 i

o. 3 I7 0 0 30 4


beamn cranrission.r N~o signal was dletecterd in the collision

chamber wlhen the rotor cwe not turning; therefore-, no

optically clear path t-o the collision chamber w~as open to

photons or metaestablce tons produced~ by the discharged. Thie

vel~ocit~y spread transmitted by the selector waas calculated

to be

R! ma m .101.

whe-~ re v = .111 v and vi .- = .909 v

Figure 14 showJs a plot of measured and calculated
values of the relative beam intensity for the helium meita--

stable atoms8 at various nominal velocities. The points

represent~ the ex~peri~mental~lly measured transmitted beam: in-

tensjity adjusted to unity at m~aximum incc-nsityj and thle solil:

linec is the corresponding theoretical velocity distribut-ion

of the intensity predict-ed by

Ttv) myv exp [-v2 ,2

where ((2 = 2kT/mn, kc = 1.38 x 10-'16 orgi/oK, and m = 6.6488 x

10-2 g/atom for helium. The characteristic beam temperature

was calculated to be T = 4630K; where u2 = 1.92235 x 1010


The good theoretical fit strongly supports the assump-
tion of a Maxwell-BolItzmnann velocitly distribution inside the

discharge source. The disagreement~ between the theoretical

and experimental intensity at the lowecst beam velocities is

due to a cloud-like formation of helium atoms outside the

exit s~it of the source.36 Thiis necumulationl of gas will

selectively scatter the lowJ velocitry beamn atoms.





r :

0 O

O dO






O .j



o a~ u, Zr .-t o O o O *- L

Di. Chalrget C.ollectjio ra~tc;tial]s

To insure comonlete co'llcction of~ the ionization~ aro-

ducts~, the cu-rret collectecd on CCi w~ac studiecd for several

biasing arrangeme-r~nt~ on CC2, CC13, andC GC. Fr~Jom these5C se~ie: s

of intens-ity saturat~ion studies, it was found necessary to

keep VGC at a1 voltage7 equall to or greater than thie voltages

VCC and PCC3. The collection plate CC1 was essentially
sit~uatte In a p:otenti~-al well by) Lhis aranugem~ent. Furthermore,,

any charged particles focusedl by the cylindrical field in-

side the colLision call were prevented from exiting the

col.lision chambers. by the repulsive potential at: the fronti

of the gas cell.

The saturation curves shown in Figuri :1.5 indicated thiat

virtually al.l the desired ractcion products were collected

wi'jth thr voltage arraingemntns givLen in Table 3.

The beam velocity during thi; study wacs 1290 mn/sec.

The saturat~ion charal~ctieristics were studied wyith 3 mtorr

argon target gas pressure and also with no target gas.

Depending on which charged species wjas to be collected, the

voltage oni GC was mainltained at tihe voltagnEe given in Table

3 and the voltages on CC2 and CC3 were varied f-romn +120 V

for positive ion collection to -120 V for negative: charge

collection. Changing the voltage on the superstructure had

no noticeable effect since its function was essentially

duplicated by the potential VGC.

E. Operatlion of quenchc ~ Lampl

Thle qluench' lamp wias placed betw:~een selector disks 5 and


*ct tO 0



> *H -
Sr3 -
*r-I 0
r 80 00


C) I



O Ql

( OT E LIGantrsr



h co

h II


O 3
r _.

Cj 4

*H 9
4J cJ
r-; C

v, GJ

6 aLnd carefully alig.ned along clo beam!nl axis ton avoidl ob-

struction of the me;r~tantale be~am. The quenching cfficiency

of the lamp wans studied: by the saturati~on tecchnique. TLhus,

the ratio of the quen~cheid to unqluenlchedd metectable bcamn

inlt~ensri-y was mreasu-red at dliff~erent d-ischarge currents-:.

Th~is procedure owas repeated for a series of diischarge gas

pre"Ssures aLnd dlisch-arGe gas' f~low~i rates.

Although the flow rates were not measured, thle effect

on tboc quenching ef-ficie~ncy by flowJ ratec could be observed;

by partially closing a vacuum shut-off valve betwoonc the

.lamp, and the mechanical vacuurm pum~p. A conitant- di~schiarge

gas p'ressure wlas anailtinedd in the l.amp during thiis series

of studies. Thc qjuenchin;; efficiency imiproved noticeably

with- i~ncreases in the discharge gias flow; thierefore, a

maxlri~mumn pumpl)ing rate was mai~nta.iined throughout; the remnain-

ing studies. Efficient remo3val of dogassed im~puritie:; in

the l.amp, cooler operat-ingl tempe~ratures of the lamp, and

transport of sputtered Al away from the lamp coils were

other benefits; of fast flow~ rates.

The effect of quench gas pressure on th~e quenchinge
eficiency of thle lamp w~as then determined at different

discharge curr-ents. The quenching efficiency w~as studied

over a pressure range of 0.5 torr to 7.0 torr. The optimum

gias pressure fell. between 1.5 and 4.0 torr, since the onset

of quench saturation occurred at lowrii discharge currents as

the pressure wans increased. The maximum range of discharge

currents over whiich saturation occur-red was realized at 3.0

tor-r discharge gas pressure..

Figure 16; shows thle ratio of quen~chedtl to unquc;: ched

bemint~ensityy at various dischargee current:s and a difschlagrg

gas pressure' of 3.0 torr. Quenching saturat~ion is seen to
extendi fromn 30 rst t~o 50 mAP. TIhe b~eam~ velocity was :1327 mn/

sco~;nd; the Ilnaxmum: becami current: wa~s ap;.prox;imatelyr 10-1 A.

Froml tIhese studies, the optimu~m operating condiiirons of the

lamp~ wrcie as follows: 3.0 torr dlischarge gas pressure-, 40O mAi

Irlamp dj;aischarg curr^ent, and-i maxj~~imm flowl rate- of quelnch gas.

Also, th~e lamp's ef'ficliency was severely affected by impu-

rities in the dischanrge ga~s supply; theref~o-e, the h-lium

w!as purified by passage through a liquid N2 trap before
entering the Inmp.

F'. Procebeci~ of E:lgr~iment

The charge drawiout voltages we;re first set for thie

collectioni of electrons (VTGC = -135 V, VCC2 = CC3 =~ -95 V).

UsingE thle quench lam~p, signals proportional to the unquenched

(singelet +e triplet) and quench-edl (triplei-) mnetascable beam;
intensity w~e~re measured without target. gas. The difference

in these two measurements gave thle signal pr-oportional to

the singlet metastable beam intensity. The relationship be-

twoon the beam intensity (Io) and the measured signal (i )

is given by jo = YI A background therm~ocouple EMIF reading

(E ) was also taken in the absence of the target gas.
Target gas was than introduced into the target chamber
and allowed to equilibrate and a second thermocouple EMF

(E ) taken. The unquenched and quenched: electron signals
w3ere measured and the singlet electron si~nal. obtained fromn
the difference of the t~wo experimentally measured signals.

1.0 m.

'- ~"~~~ '~--- -('--------C-----G -----3-- ..---;1~

0.2 i--

10 20 30 40 50S

Figure- 1.6. Saturation of~ Qunchi Lamp Efficiencyy


The ciarge chi:uwoui- poten n11ls- wire thoui changed to( the

approprpiat ~ ar~rangemenrn t fori collorcLion of positive charges

(VGC '\CC2 = CC3 = +45I V). Th~e ion current result~ingi from;
the unquenched aIiind qenlched metastabl~le beamt were again

m~easurei d anid zn additional t~ernmocouple EMIF reading was taken.

Th~le t-arget= gas flow t~o t~he co~llision chiambe?~rwas int:erLupted~

anda a backg:-,;roun EHFI meazsurementt takenf~.

T~he second: pair ofC I:Fi thermro-couple measunRrem-l:ents was

takeni to obta:;n th(re target. pressure:i dulring thei collection of

positiLve charges. Althiouglh the setting on thle leaLk valve

regulating gas f~lowi into the chamiber wJas not al~terel, the

final! ta-r~gt Cas pressure wJas invar-iably l~ower than th~e

initial targret ga~s pressure by approx;imately 0.2 mtorr.

Th~is wase not unexpecctedl in view; o~ t~hie long pleriod oif time

betwieen thc twro paiirs of thelrmrocouple readings.

A different leak ratee of target gas wjas admirttd into

the col~lision chanmber andi another s-rie~s of EMFI and ion

charges curren-ts measuredl. After chianging th~e drawout potePn-

tial., the electjron currents and EMiF's were again measured.

The target g~as fl~ow was again- Intefrrupted and background EM~F's

and: metastable intens~ities measured. This pr~ocedure was

reneated uncil ion and electron cu~rrnts were measured for

at least 10 different target gas pressurefs in the range from

0.5 mtorr to 4.5 m~torr of argan. Using thiis procedure, the

target gas pressures and the beam flux were always knownm

accurately both during the col~lect~ion of ionric charges and

during collection of cloctrons.

G;. Reduction of kCnw Datal;

The data w~ere corr.ected~ for var-riations~ ini the metastable

beamn intensity b~y division of thle pressure dependent signal

by th-e signal obtanined w:ilh no target gas~ in the, collision

chamber. Thle p7es;sure dependen~ce of the reduced ion current

canl be w~ri-lcn as

1 1 -- exp (-cp) ] (4.3)
o O

and the precssure dependenlt reduced elecctro~n culrrent written as

i_ E(P)
--- -- 11 ( ) ex (- p)](44)

where E(?P) is the target gas pressure dependenice of plastic

and inelas:.ic scartterin~g of the metastable beamn in the m~ain

chamber, Yo thei secondary el~echton ejectioni coeffilcient: with

no tar-get gas, and yt:he secondary elecctron ejecti~on coeffi-

cient w:ith carget gas in the collision chamber. Equation

i402 is regained uponi division of Equati~on (/i4) by Equation?

Tbc experimental pressures and the reduced ion and

electron currents w~ere fit to a 4th degree polynom~ial in

order to obtain data at common pressures. Comparison of

the excoerimrental daita and data obtained from thie pol~ynomial

fit showed no discernible bias resulting from the polynomal

fit. Although some smoothing occurred, the polynomial was

of sufficientlyy high degree to reproduce any major variations

in thie slowl~yvaryingnf pressuree data. Folr each beanm velocity

used, the pressure depenrdent data wecre thien fit to thle


(--) F = ( (41c)
i 2

~using a linear least squalres techniques. T'he secondar-y
elect-ronl ejection cooffi-icientt 'i was obt-ained from the slope,

m, by thle exprecssion 'i = 2 - ?mn. 'The absolute total. ioniza--
tion cross section at eachl be~am velocity wasC derermilned from

the inforcept (Y/ic) by the explression

o0( 2 _c Y (15)
Y 7.317 x; 10-3

where y is the average of all values of Y(. Reipresen ta tive

samples of th-e least square lit to the data for several
relative beam velocities are shown~ for the system~ Ho(23S);Ar

in Figu~:re 17 -nd for- t~he systemi He(21S)-Arr ini F~igure 1,.

Figure 1,7 P- (i fi ) Vercsun P' Tr He(23S)-Ar atl Different
R~el~ative Velocitife

3 0 0.

1877 m/zec

2604; Im/see:

L__I __._. i._l.......... 1_ .1 L I
1.0 2.0 3.0 4.0

Figrurec 18. PI.~-(i /1 ) Versies J' for H-e(2 S)--A4r at Diiffrentl
Relative Velociies



13338 m~/'sec

,/iJ: !77 m/se,

j~,~27817 n:/:;ec



PFESSUMiL (m;torr)


Tihe second~aTry electron cjectioni coefficien;ts for

Hie (21S,23S) on the electrop:laced giotld surfaces of tihe

collision chamboer- were determine-; d experimeintally as discussed

in thie previous chapter. Thie average: value~ of the secondary

elecctronl jection~ coeffiicien of 23JS HE*c for all. beam vailo-

cit~ies was i- = 0.582 + 0.024t whfer 0.024 wJas t-he stan~rdard

devlnation of- all values. The average vallue for (2 S)He"

was; is040+008 The ratiLc ofr theC tw~o comff~_~icints

was 7 S/Ye = 0.756. Tlhe results of other exper;imental detr--
m~inations of Y on different surfaces and writh different

target gases suggests caution when comparing published values.
Dunnin~g, R~undEl, and Stebbings repocrt r = 0.53, y

0.69, andl ?is t = 0.768 on a stainless steel surface and
S= 0.5!. Yt = 0.63, and Y /Yt = 0.31 onl a Cu-Be surfa~ce.
Dunning and Smith have also published val~ues of the coeffi-

cients for an electroplated gold surface where i', = 0.63 +
0.07, Ys = 0.46 +- 0.09, and YBs t = 0.73. These last re-
sults agree w~ith thie present results wr~ithin experimental

error. This shoulld be expected since a gas cell method

w~ith argon as. the target gas was similarly employed. The

variation of Y wit~h diiffr-ent target- gases reported by

Dunnjing et ;:l. indlicated an ardsorbed g~as sur:ace effect.35i

This is somew~ati since at~omic gases such as Ar and

K~r apparently do not a-dso-rbj apprec~iably on metal suirfraces

atl rooml temnperature and low~ pressures. However, further in-

dication of possible adsorbed gans surface e-ffect mlay' be seenr

in; 'ig;u~re 17, where th-e ex:per-imeniLa data i~n somef cases deC-

pa"rt. from the linear relate nionsip wi~th p~ressu~re at lowJ target

gas pressures.""

B. R~elativei Vclocity Deaendence of the Total Toni zati~cn
Cross Section!

Present: Resultls

The relative velocity dependence of the absolute total

ionization cross section for the H:e* (23S)-Ar and Hie" (2 S)-

Ar sys~items are: givenl by F~iiure: 19J and~ Figure 20 respeal~~-ively..

The experimoenta~llyr determined d cross sections are represented

by thle points joined b~y tLhe solid line within the vertical

bars representing estima~tes of the experimental error. Thie

error for each relative velocity wvas determined from the

standard deviation of' th-e inlterne~t in the lineanr least

squares fit of

i_ Y Y

1, 2 c

The uncertainty of Y is not reflected in the crror bars.

The values i'. = 0.582 and Y = 0.440 were used in the
c s
calculation of the absolute total ionization cross sections.

The methods for calculation of the relative velocity aind of~

cor-rectioni for the random motion of the target gas are






ed a
r G

me r

0) 0

O ~

"J O

G) G












,o :



0 CI


at 0







. Fr.. l.~.. '. ..l-- L-

Giv-en in the Appe,~ndix. T'wi ratio oif the triplet to singlect

cross sectiions at: various relative volocit:ies are shown in

Fiigur-e 21

It should be noted t.hat the uncerralinty inl thle values

for- the cr~Ss sections are large in the 100r relative velocity

rag.Thi~s is due: primacrily to the vecry smell metastrabl~e

boonm intensity~ at low vcolocities. The HIe:.(21S)-Ar cross

sections ait law~ beam; vocloci~ties are~ particularly diff~icultr

to obta~in slince 'the iiata for- this syste~-m are determ;1ined f~rom

the- differences in twro ex~perimnen:al measuremrents.

The qualiati~ve beharijor of' th-e ioniization cross sections

nay~ be exp~lailled in terms~ of the potential cure anid ir-s

associated coupling frequency ton the continuum. T'o the

lowe~s:- orier of appj-sroximtion, th-e classicala" theory of

ioni'zatrion givecs the appDro:i'mat~ion

of-v) 1 Rel W(P'clrl

At very lowi co'llsion onergies (Ek(co) << ", where c is the
well depth1 of the potentially enerGy of interaction), the

collision is sensitive only to the attractive part of' the

potential curve. If w;(R) 1 exp (-R), then the coupling be-

tween the discrete state V*;(R) and the continuum of states

v -(R) + Eel is very w~eak~ and onlyv "close col.isions"" will
result inl ioniization. Therefore, ionization wiill be dominated

by tbo v-1 factor anrd the ionizationl cross section will de-

crease with increasing relative velocit-y.

Th~e region (Ek("') c) is viewejtd as a transition region
in whlichi the collision energy is approximnately equal to th~e







4J *

H (d-

O :

* IJ



O J~


p G)



/" O



we!cll dilpth ofi the in:teractioon piotentrial..Inhiron

theF cros-- section will~ pass thought: a minimum value. Hiow-

.vel, t~he collisiOTn ene"rg~y at which the cross sectioni is ait

a In!inimium doesR not c-orresplondi exactlyi to the~ minimum~ energy

of' the potet-iajl curvie. TChe~ eact: posi~tjon of th~e m~inimumi

involves the degree of couplinig Giv~en by Wi(R) as well as the

inter-action potential. .

As thec collisi~on enlergyr is L~cincrsed; b~e~yond thle mnlimumilr~

of the potential curvee (Ek,(a) << c), the int~eracti~on enters

the "hari3 d core" region. In this regnionl, the collision

one:-gy is suff:icientlyj high to probe thelf retpulsive part of

the potentialj curve. The factlor v1 is overcome by the

t-ransition n coupling Factor. Th'ie classical turning point

decreases withi in-rerasing coillirion~ enegy, and the e~xpon--

entiial nlatur~e of W~(R) w\ill bei strongly p~robjed. Thus, toe

cro;s section w~ill follow~ the coupling frequency and increase

wijth collision energy.

At very high collision energiies, the repulsive potential

becomes very steep. A~s a result, the classical. turning

point Rel. and therefore j(REl.) wijll remain relatively con-
stant and the v1 factor will again predominate. The ioniza-

tion cross section will decrease with in-creasing collision


The present measurements are the only absolute cross

section measurements so far reported. The results reported

by Pesullel et al. are relative cross sections wJhich have been

no-rm~alized to theoretical calculationss of the cross sections'

based on Miller's classical" development anld the potential.

curve suggested biy 01son, for- thel Hc(23S)-Ar systflm.12 2 Fesnelle

E~t al. emplT~oyed;; cIrossed beamii (IlchnliqCue in~ wh.!chl both beaims

wier-e generecedi''"' by effusive soril~et:. The~I results for) HoC-

(23S,21S)-Ar: rep'olted by Ille~nibrger and,~ Hiehaus are also

relatives c?oss sections w:hi~ch weire nrmrlal-ized to the destruel-
tion rat costntof (23S) by -rgon at 3000K. 2'% The

technique~ of: crossedl beams from effusi-ve sources was as!.o

elmploye~d b =Ilelenbrgeir 3nd NiLChraus.

The cross sectioll r~eported by blothr Pesnell.e et al. and

Illenberger- et al. for the system He(2 S)-Air ae relative

to thie nouna~lized cross sectionls for the He(23S)-Air system.

Th-is procedure macy be subject to systematic error- in tbo

case of P~esnello et al.., for ex~ample, since the authors'

assume- Ys t~; = 1 for the fli~rst Cu~e dynode of' theirr detefctor.

Dunn:ing, ~un~del, and Stebbingls have found! ex~periitimentally

that: Y / t = 0.81 for metastanble he.iumn atoms incident on a
Cu~e surface. In thef prese-nt w-ork, t-he absolute cross

sections for both the Hr(23S)-Ar and He(2 S)-Ar systems

alre measured ini-dependent;lyr.

ROiol, Howalrd, Rundel, and Stebb~ings have measured the

velocity-averaged absolute cross section for the Ho(23S,21S)

Ar systemn.2 Since their absolute cross sections are

vel~ocity-av;eragedl, direct comiparison wi;th their results is

difficult. For a mean cente~r-of-mass collision energy of

about 60 m~V, their volocity-averaged absolute cross section

for H~e(23S)-\r is 8 = 16.9 2i and folr Hoc(21S)-A1 r E 22.7 2
L: s
and ,-he ratio of the two values is sast, = 1.34.