Capillary,. 4 Capillary 2 mm.Hg /N 60 Optimum rate. 50 40 30 20 10 300 280 260 240 220 200 _1 Bath temperature
v/hero k w ill "be constant fo r a given c a p illa ry . Other l e t t e r s ho.ve th e ir usual sig n ific a n c e ,
2 2 2
Then a p lo t of (P^ -P^ ) =AP a g a in st the floiv ra te as
c a lc u la te ! above w ill serve as a c a p illa ry c a lib r a tio n . The r e s u lts of th ese esgperiments are shovm in fig u re
When more th an one c a p illa ry i s used th e t o t a l flow ra te i s th e sum of th e flow r a te s f o r each c a p illa ry .
6 , C irc u la tio n Pump Plow R ate,
The mercury d iffu s io n pump f o r c irc u la tio n of th e gas through the flow system was heated by a m olten m etal b a th a t 200 - 300°G. The e ffic ie n c y of such a pump v a rie s w ith b oth th e b ath tem perature and the p ressu re in th e system (if). A ty p ic a l v a ria tio n i s shown in fig u re 10, By p lo ttin g such graphs a t d if fe re n t p ressu res of
argon i t i s p o ssib le to c o n stru c t fig u re 11 which shows th e v a ria tio n in m etal b a th tem perature allow able before th e ra te of the flow f a l l s by I t was found th a t changing the flow c a p illa r ie s had l i t t l e
or no e ffe c t on th e maximum flow ra te tem p eratu re. Under
experim ental conditions th e maximum flow ra te was always used as only under th ese co n d itio n s i s th e flow ra te s ta b le .
00 r4 M H rH M
a
o
wI
to CO r4 (H46
7* P ressure g ra d ie n t in th e flow system .
The flov/ ra te v/as doterminod as described above and, since i t v/as intended to use f o r subsequent c a lc u la tio n s th e high pressure reading of the flow c a p illa r ie s as a measure of th e p ressu re w ith in th e fu rn a ce, a determ in atio n of th e p ressu re g rad ie n t w ith in th e flov/
system was d e s ira b le , McLeod gauge attachm ents were th e re fo re made a t p o in ts 0 and D ( f ig . 21) and a t p o in ts R, A and B ( f ig . l ) .
Readings on th e furnace and bypass lin e s were a lso o b tain ed .
The d ata f o r about 1 imu (and 2 mm) t o t a l p ressu re i s shown in g rap h ic al form in fig u re s 12a and 12b. To avoid confusion on the diagram only th e d ata f o r flow c a p illa r ie s 2 and 4 are shown.
The McLeod reading p o in ts are marked along tho a b s c iss a .
The r e s u lts in d ic a te d th a t a co nstant percentage in c re ase f o r each c a p illa ry ap p lied to th e high p ressu re sid e reading of th e flow c a p illa ry would give th e p ressu re w ith in the fu rn a c e . The percentage in c re ase v/as independent of tho t o t a l p ressu re w ith in tho lim its
used. The average values used in c a lc u la tio n s were as fo llo v /s:-
Plow C a p illary % In crease Required
No. 1 14.0
No. 2 25.0
No. 4 3 .0
Ion
gun.
Magnet current power and control. Magnet. Collector. Focus box. Filament supply. Recorder. D.C* Amplifier. E.H.T. power supply. FIG. 13.
47
The r e s u lts a lso showed th a t th e p ressure g rad ie n t w ith in th e furnace was sm allj th is i s a d e sira b le fe a tu re f o r accuracy of
c a lc u la tio n s .
Experiments perform ed w ith th e furnace hot showed the same p ressu re d is trib u tio n , id e n tic a l p lo ts being obtainod f o r tho furnace cold and a t a ty p ic a l v/orking tem perature.
In l a t e r experim ents where a lin e d furnace was used f o r h etero g en eity t e s t s th e same p ressu re d is tr ib u tio n was
experim entally observed,
8. The Mass Spectrom eter.
The in stru m en t, a conventional 60^ N ier (5 ) design, was co n stru cted in th is departm ent. F u ll d e ta ils are given by
Davidson (6 ), The e le c tro n ic power supplies however had aged and become u n stab le and were re c o n stru c te d , care being taken in th e
redesign to ensure maximum s t a b i l i t y and r e l i a b i l i t y . A box diagram of th e e le c tr ic a l requirem ents of a mass spectrom eter i s shown in fig u re 13#
Tho E.H.T. power supply to the ion gun and th e magnet cu rren t power supply wore b a s ic a lly of th e same design as p re v io u sly .
The emphasis in re c o n stru c tio n was on th e under-running of
component p a rts and th e in c o rp o ra tio n of a la rg o number of re a d ily ac ce ssib le t o s t p o in ts on th e side of tho u n it, A fter
c o n stru c tio n , when th e u n it was fu n ctio n in g c o rre c tly , notes wore tak en of th e p o te n tia ls of a l l im portant referen ce p o in ts on th e c ir c u it, p a r tic u la r ly co n stan t v o ltag e lin e s and grid-oathode
p o te n tia ls on th e v a lv e s. Where p o ssib le th e c ir c u it param eters had been arranged such th a t th e valves were operating near to th e middle
of th e s tr a ig h t lin e p a rt of th e valve c h a r a c te r is tic .
The focus box was a v o ltag e d iv id e r to supply v ario u s p o te n tia ls to th e io n gun p la te s . The magnet cu rren t co n tro l was e ith e r manual
or autom atic. The previous autom atic c o n tro l had been by an e le c tr ic motor d riv in g a 10 tu rn potentiom oter a t varying speeds and in
varying ran g es. Tho new c ir c u it was an e n tir e ly e le c tro n ic T e lle r Sweep c ir c u it w ith v ario u s scanning speeds as re q u ire d . The magnet cu rren t was v a ria b le from about 6 mA up to about 150 mi, th e l a t t e r corresponding to m ass/charge r a tio s of about 250 fo r the norm ally used a c c e le ra tin g p o te n tia l of 184*0 v o lts ,
A new power supply was co n stru cted to feed th e c o n tro l valve h e a te r in the lÆ iller Svæep c ir c u it, th re e valve h e a te rs in th e D.C. a m p lifie r and tho DBM 8A e lectro m e te r valve h e a te r. These fiv e valve h e a te rs were fed in s e rie s from t h i s h ig h ly s ta b iliz e d power supply which a lso fe d one or two referen ce p o te n tia ls to th e a m p lifie r and magnet food c i r c u its . The D.G. a m p lifie r (?) was
a
100^ negativefeedback type vd.th a n p lifio a tio n sta g es in c o rp o ratin g th e electro m eter and two
12 SO 7
v a lv e s. The c i r c u it, when c o rre c tly ad ju sted49
(whioh was a d e lic a te o p e ra tio n ), gave an output v/hioh was lin e a r up to about 15 v o lts and i t could be used on one of th re e high value in p u t r e s is to r s of 2 x 1 0 ^ \3 x 10^^ or 5 x 10^ ohms. The a m p lifie r fe d th e c o lle c to r sig n a l to a 1 second Honeywell-Brown H lectro n ik reco rd er v ia an autom atic range ci’iange device which was trip p e d by m icrosw itches f i t t e d to th e re c o rd e r.
l'o r th e major p a rt of the re se a rc h however th e B.C. a m p lifie r was rep laced by an Ekco v ib ra tin g reed E lectrom eter type N616B-4 This u n it w ith i t s a sso c ia te d decade vo ltag e u n it N659,4 was id e a lly
su ite d to observing a backed o ff output sig n a l from th e mass
spectrom eter ( th is was a l a t e r requirem ent of th e re se a rc h and i s d escrib ed under run p ro ced u re). The electro m eter c o n s ists of a head u n it which i s sealed and d esic c a te d and contains th re e input r e s is to r s of 10^, 10^^ and 10^^ ohms se le c ta b le by push b u tto n . This u n it also contains a v ib ra tin g reed type dynamic c a p a c ito r operating
a t a frequency of about 45O c .p .s . The re s u lta n t 4.C , sig n a l i s am p lified and r e c t if i e d and d isplayed on a m eter on. th e in d ic a to r u n it and fed to the re c o rd e r.
The electro m e te r has, in i t s most s e n sitiv e co n d itio n , a f u l l sc ale ^15
reading f o r a cu rre n t of 3 .0 x 10 amp, Ih en p ro p erly mounted (and considerable care i s n ecessary ) th e s t a b i l i t y observed on th e instrum ent was 0,05 mV. v a ria tio n on th e 10^^ ohm input r e s is to r w ith the in p u t sw itch a t "ion chamber f a s t" o p eratio n p o s itio n .
I
CO E g CO d;
CO p:5 Oo
w —MAA^-e “A/WW lO CO • r4 Ü O o k M •P Üo Ü >o >o
r-« gwo
CO CQ -AWA- - m a > O — 'VWVA- • Q .nduf "peTZTloaa poq^oomg4h|l
gm oM k50
The head u n it in th e p rese n t set-u p was fix e d r ig id ly to th e bottom of the mass spectrom eter tube w ith no p o s s ib ility of fle x in g between the c o lle c tio n system and th e u n it, G reater s t a b i l i t y \yas a v a ila b le on the "voltage" p o s itio n bub a t th e expense of response tim e. The
10
instrum ent was found to bo most u se fu l on the 10 ohm r e s is to r fo r th is experim ental arrangem ent.
The filam en t supply was a conventional c ir c u it o p eratin g th e filam en t on A ,0, a t about 6 v o lts and 5 nmps and was s ta b iliz e d on th e t o t a l e le c tro n em ission cu rren t from th e fila m e n t. . Tha c ir c u it . was flo a tin g a t about 2000 v o lts fed from the E.H.T, supply v ia th e box.
The e a rly experim ents described in the follow ing se ctio n s were performed w ith th e c ir c u its as described above but w ithout th e Kkoo elec tro m e te r, ?o r th e l a t e r and major p a rt of th e re se a rc h however whore i t was necessary to back o ff a peak and reco rd sm all changes in th is peak h e ig h t, co nsiderable improvements wore f i r s t of a l l n ecessary.
Backed o ff peaks a t f i r s t showed marked i n s t a b i l i t y and quito pronounced d r i f t w ith tim e. This was suspected to be e le c tr ic a l
tro u b le cud one e a rly o b serv atio n was a c o rre la tio n of th e flu c tù ^ tib n s w ith tra p cu rren t (the io n iz in g e le c tro n beam c u rre n t) i n s t a b i l i t i e s .
A new c ir c u it, based on tr a p cu rren t s ta b iliz a tio n , was co n stru cted to elim in ate th e problem. The e s s e n tia l d e ta ils are shown in th e c ir c u it diagram in fig u re 14, The whole c ir c u it again f lo a ts a t about 2000 v o lts , fe d in from th e box lin e which i s ra is e d
2000 V. •H 0 V H.T. fine To focus box H.T.Coarse
A /W V
I Ganged I I Ganged GangedI To focus box A A A AR_,R 100,000 ohm 20 step potentiometers. Rg,R^ 50,000 ohm smooth potentiometers.
51
to th is value hy th e E*H*T, power supply* V oltages marked on th e
diagram are r e la tiv e to th e zero voltage referen ce lin e shown* E a o ilitio s were a v a ila b le a lso f o r reading tra p cu rren t (on two ranges of l{-00 and 200 mioroamps 3?*S.B*) and t o t a l em ission cu rren t (4 OO mioroamps and 2mA F.S.D *). Tho energy of th e io n iz in g e le c tro n beam was a d ju sta b le from about 8 to 70 eV by varying th e box to filam en t v o lta g e . A tra p cu rren t of 20 mioroamps was a v a ila b le fo r appearance p o te n tia l work i f required*
The re s u ltin g peak s t a b i l i t y was markedly improved over th e o ld er