The Process of Electric Conduction in Thin Lead Films Citation Foster, Mark Gardner (1939) The Process of Electric Conduction in Thin Lead Films. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/cyzf-m737. https://resolver.caltech.edu/CaltechTHESIS:01302025-225249706 Abstract No abstract. Item Type: Thesis (Dissertation (Ph.D.)) Subject Keywords: (Physics) Degree Grantor: California Institute of Technology Division: Physics, Mathematics and Astronomy Major Option: Physics Thesis Availability: Public (worldwide access) Research Advisor(s): Goetz, Alexander Thesis Committee: Unknown, Unknown Defense Date: 1 January 1939 Record Number: CaltechTHESIS:01302025-225249706 Persistent URL: https://resolver.caltech.edu/CaltechTHESIS:01302025-225249706 DOI: 10.7907/cyzf-m737 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 16974 Collection: CaltechTHESIS Deposited By: Benjamin Perez Deposited On: 04 Feb 2025 22:29 Last Modified: 04 Feb 2025 23:08 Thesis Files PDF - Final Version See Usage Policy. 21MB Repository Staff Only: item control page

TEE PROCESS OF ELECTRIC CONDUCTION IN THIN LEAD FID1S Thesis by Mark Gardner Foster In Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy California Institute of Technology Pasadena, California 1939 .... ovn.INE,_= ·......, ,2E O Q l ~ A. In.trodu..etion I . 'rhil.t films in general II . Detailed eonsideration of eleet-rieal conductivity 2. .( l) Thi,c lmess de:pendenee 2.• ( 2) Oritieal thickness 3. ( s ) Instabill ty c.t films 3. (4) eonnttiu ·Gf the supporting su:rtaee 5,. III. Theories proposed in the past 5. {l} Noneondueting layer theory 5. (2) Altered me elee.tron...path theory 5. (3-} Granular struetnre t.heory 6. IV., E:;s::perimental met-hods ot other workers 8. (1) Production o'f films 8., ( 2} iteasurem&nt of tbi.cknesa 9. (5) Electrical and themal contact 10. ·( 4) Metheds of measuring resistance 10. V. Oon"C.lnsiona tram previous worlt ll. :a. The erperiments reported in this thesis I . Pu.:rpose <,f the investtgati ons 12. II . Experimental methods employ-ed (Genel.'al.) 12,• .(l) Produetion o:t fil.ma { 2) Measurement .o f thiekness 12. (I) Themal. eontaet 13. ·(4} Ele.c triea.l eonta~t (5) Methods of measuring res1stan.ee 13. (B , G.011·t 'd ) III. Detailed deser:f.p,t ion. of the apparatus and tts use (1) 'fhe vacuum emamber Page 14. (2) Evaporation 16. (3) 'l,'he film 18. (4) Electrical measurements 19. (5) Calib-ration o:E condensers 25. c. Results and their significance I. -G eneral 2.5. II. Preliminary observations 26. III. Films eondensed at room t~pe.r ature 28. 1V., Films condensed on liquUl- air-eooled oo.ekground 30. v. Time changes in conducti'tity 32. VI., 0Qnclu1ons frcm the abQve observations 33. VII,- Q.ualitattv.e mechanism of thin- film ~on.duct1vity 34. ]).~ 40. E. Aclalowledgme:ate 41. Graphs ltlO; 2 to 9 Plates l Tables 1 to 3 Bibliography tG S ~ ••rk A. l,ntred:u.ftt1mt. J. ~ film$ •in. egral:: of other in.ve-s Migf! ta~: ~l"iments have been made, for ma.nr, ~ars 0 on the prope.rt!:es of thin metallic ~ s . Hamburger•) ( 1931) , Schi2l.tz~ 3 ) ( 1933) • Anclerson i,) {19133) and .A.mlrade· a> (19:56) give e.xtenal'l'e SUl!lmaries of' su'b8$4:11&nt 110r•~o., hltY& been eharaeter~ e.d bJ', pr9nouneed lack sf qit.a11tltatiw a,eeaent, altho'.l1gh c:erta1n phen&'iDiel'!a al'e :fouacl qu.ali\a.ti'V!&'.b°, by, most observe:r~. Xt is nea.rl.T, alwqp f'oimd,. dtsregardblg minor deta ils, that the condmt!ntN', phenomena :of me.tals a.re vol'Um$ phenomena, •inc& when 1n a th1n l.~r they, an ¢ .t e different from those of the sam metal v.ben in massive pieces. !he Rall effect (Mackeoa s ) ci-oes :aot find the 4evi.atian.J obtained b7, Riede eJ and others), m1d the thermoelect ric eoe::f'fieknts ( Holme• io>) have be.en ~b.serftd to show ,mexpe.c ted variati<>u for :thin ~2!'9 of metal_, ktensive work ha.9 been .(lone •Oil the: optical a.h01llal,ies exhibited by, thin f11Jns (l3a-uer 11> , l ukuroi 13 • 13 ) , llass 1.,) , Kirchner ea) , Orastei u 1;5 ).. Bouard 7;&) , and. Schul tse a) give copious reference,), and of late the q"Uestion o:f c17;s tal strnetnr:e under thin ~ r comtitlns has received diree,t trea:tme:nt a.t the ha.ml.a <>f l!laXq, inv:est1p,tors ~ing optical and ,e l-ectron-aptieal m&thotls. (References 2, 4, 14, 49, 17:..-30.) Ca.utio,a is requiJred in the 1nte~retat1'Gn, however; Wa$ :60), :f Gr example. shows tba't the distur'ba.nee to the film ,c aused ~ . exp&:sing it t o electron r q p cannot be :neglaeted, siac& changes in eonflgurat:ion are• somet,bles ·stt;l.J.'ted. prematmrely' by, supplying this mt!J.dl ,e,ner«;v,. 'lhe 100'-S't~measureii. ;property', o.f thin. films. however, has been t he electrical conductirlty~ . Sb1ee ' • this t:hesis deals with i :n.ve1:11t1gat.i u ef electrical eenducti:vi~, the lttera.tnre on this detaJ:l will be treated mom speeifieally;: lI. Det;tµl&d eonsideratien o-f el,!ctrieal ee,nduct1:v1tz: !here- are several qualitative .features common to t he l"esnlts ( l;l. The ( 1Jp&e.t fic) co.miuetivity ,o 'bHrved for t he metal 1n thin films depends $tl'ongly. 011 the thtcku.ess, the val,a for t .be. thlll11GJst films btd:n g much leas than tbat for the massive metal. Andersi>,n 1 ) finds that the, measured resist'i:vi ty of his -' napo;rated &11ver ~ l : ' & inerea.se.s by, a £actor .c ,f about 100 in the· range. from wero to 300 atomic ·~ i-s, then falls r ap141;;, toward the bulk va:t•. s•me 1oi;a bel:ov., as the tblekuss is t~rea,sed. Braunbek •3 ·1) . and Vellme1"· and li1ttermann 3a) , a-sin.g eva.pora ted ·mertmry, films; Fuk:m-oi 13 ' p) , using crulmiu and ztiw a:s we:11; Kamb':o,r.ge.,r •' 23 ' a..l. usi':ng s11ver. a.nti~n;r, nickel, platinum, and tung.sten; ltrautkraemer 3,s) .. using silver and gold; Lovell 33,, q ,a.s,se> ,. u sing nbidium, potaScsium. and caeatum; Mltehell, 37) • us:~ s.il'v.e:'J!'i lUedmi.l ler -ae> ~ using nickel; and Was 39 ) . usi.ltg gold. a ll find that thin evaporated metal. films. have speeU'te resistance• muDh h.igber than the b.u.1k metal.. !he 'VU'lous investigat.ors diffe:r as to the- she ot ~his effect, Lovell observing ( under very, clean co.ndlt1ons) .o nly, a factor of' 10 be't.w$8n a 40 .Angstrom film and • bulk -,tal,. - . the ,o thers larger fa.ctol'·s , although the p.re.lb.dna:ry, increase, no,ted b71 And$rson d0e:s not se&m to be recorded ~ r e ,e lse. Ma.Dy, inve-s-tiga:t-ors used films obtained b7, cathG.de -sput·t eri~ J.ol1-ot- .;._s ) " kam~nr ••,••) • mid Per.u.cea '"'.;••&) ) lill.Ud obtain similar remts, complicated 'hY,/ t.he ef'f-ect o'f foreiga atmas neees:saril;r, present -ti'u$ to the poor vacuwt requi~~ fG'f' ~ sputtering pr.o,ee-s s·. ·' !he ,ebemlcal. me·taod G-f depositing metal fi.:lms from a ~lut.10:n c&ntaint:ng its icna (see. for emmpl-e Wait e,>. S-tone so)) is 9 nl)w BG longer used. :for eondtlCt-S.vtt,-, experiment.s, 4• to. the e<0ntamfnati-011s whlch tM-s method pl"Q-vlde-s in a'b1mdane&. (~z· 1lhe:re ~xtsts a cr1U,eal tld.clme,ss·, in the neighborhood 0£ which tbe change in cetnd,wt_iv1 ty, with thickne.s s .is very rapid. lkperlment (See Anderson _,) tor SU1!11.11e.?7. to 1933) shows that rut a :til.m b fiirst bei'ng -d:epoalt&d th,e obse,r ved conductivity, is 1ow, seme- timee c:omple.tely, undetectable. (Although. Love-1 1 ••) observes some ,c onductbd t7, ••n with fawer atoms preeent than would be requi:red te to-~ a .complete mono~z. ) appears quite awlds.,. !hen., as th& thickn:es-s her•a.ses, eonduotie ·TM critleal thlckn.e·ss for th.ts to 000\U' ia alway, ema.11, different obser-rers finilng values ran~ from about 50 &t.gstre• (ICraut~aemer a,e)) to seTeral thousand (Braunbek 3 1.)) • although mo,,s-t wgrkeTs. agree with Lovell 3-) that films 104 ,o r mor,e Angstroms thi.ck approach the normal m&tal v-eey_cl'Qael.J', 1n &pecif1c prepe.r t:les. (3) • A thin film when first ldd down i:s usually, unstable t and irre.versibl.e ehaage.s eeeur with changi:ng tsmpe-r a.ture, with time, ¥'1th p a ~ 0£ heavy current• am with the addS.ti.O'n and removal (})f f>Qre:tga atoms. the sharpest te~ratu:re c.banges are reported by,; Kramer ,..) •· who .finds that 'films, sputtered a:t low temperatures e:rhib1t a -~ n -i rreversible deer.ease in resis·tan.-ce as they pass thro,agh a .cer-W ncha.mcter:1s,t1:e. tempera-tun as they, are ftr-med up. Other workers (ll'ukt1r,0i ~;a), Mitchell a,--,), and Suhrmatm and lla;rth sa-) ) fiml stmU.ar ol:lange-s which appear somewhat Jn&re grad:oal.ly4 allowing numero.'Q mea~nts t .o be ma.de as the re-s tstanee fall; . lia.'nlburcer aa:l. ·Su.b.:s:mann and :Barth o) and o·t her,s find t-h at the, re,sistanee- o.f' a g1ven .tum behava~ in a Tev-ers:il>le, fashteip. belov., a. madrtm:Ul temperat'tlir.e to which the film bas pn-v.:lou•l.y', b~en. r atsea,. but it the te;mpera.tnr.e is inenased b:ey~nd t-hi,s maximwn" irrevers:ibil11;y, o-e:cur-s once- ino.re-, with -ch.a.nges now rev&l"sible unde-r t~ a~ B."l'.:imm!t.- Levell as) -repo1"ts, <,n the -o-ther hand, tha:t. his r-ahidium films eouen'B-ed at "l<}0 .Absomt.a are stable and E!Di.bit no h;v'iteresis all the vq, up to 98° Absolute, prov iding au:tfteioen:t -can is t • n w.1th the ~teaning and o:u.t~ning ot if"Jua -t"ilm bat-Jd:ng. Most wo-rkerIJ. who have examined the pout agree that t-he r.esiatane& o,f a. film ~..ges with time, 'lmt ~ si.gl!t -&t ~ ehange is ,., d.s,ta &ho.wing that the red1'tivttr, et a fl.lm- increases with t.1:me, e.pp~chi:ng -stab-le. val.'Ue:S greate-r by. fa.et«)rs- a:t large aia lo3 about 40 ~ t e·s afte~ 4QpGsi:t , altheugh he obtained a de_er:ea•e by rela.:d.ng lids precautions t ,o exel:lme foraign, atoms. !!?hat ~ri:a.e.nte:rs ~1ng good vacv. (Wrl,ght ·s a), and. liGve-11 ~5 ) f,o r e.mmple) obtain an increase, and workars1p.g. llaztle'tt 5~) asi:Jtg tlle spu.t,t;er pr-1,teesa, with ! .ts eompare.tivel;y; low va-c-tmm. obtain the. decree,se- le-nils the weight of more e'riihuica to kruler:s ea•s explana.tlon. Lo-vell aad. Wright both fbtd,. ho-we.v e,r , tha.t careful ev.tgaH~ 0:f the bacldng reduce& tW.s aging '8f!eet~ in so• case-cs el1minati:ng it en.tlre-ly. :Br-mmsfurth • 1 >, 5. and Pe:rueoa 4e) have f0und eomplex anoma.lf.re,s due to a@:mitting 'VU'ifUS: gases- to tl<.1.eir sputtered films. Vand. sG) ~ however,, fuw.s a ~crease wtth time fttr hl$ evaporated films• of sil.ve:r and ~ld.,. togErtb&r wt.th :fllWtuatioins in Nsi.s tanee on a&nittilig ail- to the apparatus. -Mitehell • .,) al!:l'tl finds a decrease with time • .i!l. The observed. proparti:e,s depend strongly on the e0nditi;0n ot tlm -~ rting SU..l"faee e:n <1hinh the film is ls.id. • Rere ~ be recorded the ad.ditiot1al results of Ditebburn s:5 ) and of Was 3 -9 ) . Di:t~h:bu'n finds that a backgre:and Gf 'Ceppe,r ,t an be made 1mif.o;r.m only by Mating s-trengly !11 vacuo immediately befere use. from about 50 Angstroms t-o a.bout 100 Angstroms o,n beating the background to drive •of£ ga.se·s adso?"be,d. Amie:r:S(),n and ltta.s have eempa.:red films ee:ndensed on. various su.'bstanees a.s baeking, and find me-asurable dif:f'e·r ences be·t¥..een the prape:rties of 0thewise identical fi.lms-. Ill. 'l'ba,p:r1::e-s l!'l'Ol}?Ose!ii in the p~et te ·a~eount fo:~ t~¼le o,}?Gerva.UQ'ns.: ill N0n-e0ndueting la,~:r theom P•e.rucc:a ·¥) a.dvane:es the i.d ea th.at a metal surface is naturally of z:ero eondue,tivity line te interac·tien o'f e-onduetie:n eleet'1"o.tt:S wi th the potential b.arri.e·r at t1'..e surf.ace. 'fh:ls ex:planstio:n does n(;}t $'eem tco b.e a;eee:pt'ed by other l'r~mr~. and eg;d.:n~t it mt1St be set the ;experimental results o:f Lovell ss> and ot Hamblll"-g&.l" ~ 3 ) which eend\l'Ct eleetrici ty,. tzl Altered free electron-path the:ory} an \:he a$':sumpti.on ef a etmtinu.ous fi 1m of u.nii'orm thickness, r. o. l. J. jhom&Oin and. mo-re ~ntl.y. Fuebs ,svJ ha.V'8' ea.'ltmla.ted the. effect on o'b'servad eon«u<rtivitt •& f the shortening Gf tbe •en free pa.th f{)T $ndu~t1,o n eleet:rons a-, to :C 'Ollisio.ns with tlae· bonn41ng surface of a Dirac statistics, and. th& proporticm. between electrons r&fleeted in ·s peeula.r and diffu.se :fa.shicm fr&m the surf.ace. ~ . prQp.er thickness de:txn!denee e,:t thls lsst proportion. be is able t:o s,eeu:re agreement with exP&rime:ntal !iata. ~ically. this proee.s,s means tak111€ proper accoimt 0£ the .degree- of r~s-$ e,f the &urfaee, that b . departing s,l ightly from vhe primary .a.s.:s umption et a plane parall-el film. :remarked l:,mnJ ( It ~ be that F,oe-rst:e:r-1.1~ .5.·e) dse1de$, £rom r:esul t$ o-f optical exp-e:rime11,t,s. t!w. t the assumption ,e f a homo-geneou;s, is:o,t ropie ~ltletrically ideal f'tlm 1s eo.n trary to the- :f'ae:t·s.) ~ . refleetion eo.r rect1oa JmlSt be 1'!Jali:e, empirically in heh:B • th:&o:ry, ,and this limits i :t:s significan.e.e .'.but it. appear:s that there, is 'ff&.t im t:n the- f11ndamental idea. that oendu.ction is low because the eonduc-ti0n el'ect:rons ha.v-& a thickn&:ss- lim!ted ()J?portmdty to move fre&ly. This the-ory pr<>vio.ea :e.i0 simple machani.$m £er the ehange ,o f' :r.-e-si.sta:nee with t :i me. (3! Gr a.nid~ stnc' tttre theory: A &llghtl;y di'f".f.-ere:n t a.pp~aeh to the thiekne as-depe-nde:nt inte:ri'1u.-ene-e with eo.ndnoti.011 e]::e4tr.ons bas; beeD :prQpOnd by Swann 59) and adop~d by, ~ - other· investigatorS' ss1nce. 'fhts theory-. ·s ta:rt:s w:tth 1th$ faet that atoms wha,n first depoaite-d. on a solid surf.aee f"~m a. tl.•l:y divi'-ded etate are distributed in e-o.mpleteq -random f a,sZii-<.m. A& motion ( see Fl"emtel ,e.e,) ) beeomes one of sliding over the wrfatre. ( Andrade ~) , Essi.g ~cl) , Fineh :aa) , and Was 48) give :f'a irly direct e~erime:n~l :evidence that such a.n ,e ffect actually, Goeurt(. ) S.1:nee the temperature is 'below th& Eeltil'lg point of the li>ulk metal., the attraeti,ve forces between the a toms are importtmt, and atoms which eo'me -close to e-ach 0;ther tend to, eling together. !'his process resalts in the format1on of n:tanY, nuclei. or centers of ooaguJ;a.tion, so that the entire film ooeomes e.hanged f:11:om the random dist,rib:u.ti~n: to the lnO:re stable ,c onf'igm-ation ts:£ mat1;1; small patche:s ( t he dee.rea:s:e 11'1 potential energy furnishing the entropy l'e-quire-d by thermodynamies). !hi;s resul.t give,s a film wh!ch ts atI;,y thing but unif',oT'm: i1:1 t-hiekness. and alloi,s the observed time variations to be described in t ,e.rms of su.rfa.ce mbillty of the eondenaed atoma. Anderson :i) has carried out this extens ion: along the following line-a: Sine.e these patehEH¼ are distinct from each ,o ther unless there be more than a certain er tieal amunt 0:£ metal :pr.e:Sen.t, el.ecvons w'ill llave a. di:ffi-cul t time erossln.g the Spaces between, and: hi gh ob.served. rw:H3i'Stance will oc.eu.r. As time stri~ng a patch and staying there,. the resist.a.nee· ot the film w.ill in:er-ea.:se still farther t and for the thi?llle,s t films will sc>'on beec• ec1nnpara'ble with the resistance td the backgroud. and th$:re:fore unde-trEJetable. On the othe'r hand, once the patehes be00me }?ig eno"Ogh and reiristanee as obnrved exh:ibl.t a rapid drop. 'lh&rmal effects .c an be obtained by a. littl-e more ~xtendt>n of these idee.s: -0:f th&· three theowies here :presented., this granu.la.r structure. theory seems the !DO'St promising, and has received the most S14)p~'1tt :fr.o:m experiments:, lV, E!'P!l"'i11'!M\ta.l methods of J?revieus and eo:ntemp~raiz. inve$tiggtors; The prob1-em that mast be ulved in making conductivity experiments with thin metallic films is a three.fold one-. A film im~t l)e proit"IJMd unde-r e-ondit.io:ns a,s simple as pos:sible and so tba.t the av~rage thickness is known., goo'd electrical and thermal eeuta.et mo.st be established b&we:en t.ne. film aitd the measuring in:strumen.ts, and the· re:sistanee must be measu:rad rapidly ,Y methGds- which ar-e adaptable- to a ride range of val1Jes. The methods which ha.vs been ,emplo7~d in the pa.at to do- these tbings are discussed.: - ,( l) Qf the thr.e& met.hods of prod.uaing thin films outlined unde-r A, II . ( l) that of eonden.sation f'rom metal vapor in high vacuum seems to gin least elw!lee fo.r e-ontam_ j na t'le.n, al though it pa,ssesses the disadvantage that the metal at;oms are :movi11g rapldly when they, strike the backing, wbere best eoniit!.ons would be expected if they_had the smallest exces·s. energy, possible. .A:rldrade seems to• be the latest worker to 'USe sputte-red films for extensive inves:tigaticns, and claims his films vhe-n :first la.id down &xbibi t no tnhomogenieties visible under ultra.miero,s eope axamination, though a patch structure appeared later on heating. l t se$8 best, however, to have the :fewest :foreign a.tome po.s,siole. i.e. •e-ven the less volat:il-e me:tais ma,-, be evaporated from tungsten filaments ( Bass ~•} t We,$ 39 •·• ) ) , plat.iJmm ( Eembi.nska ~•) , Anie.rson ,J ) , tanta.l'Um (Bauer 1-')) , or directly, from wires ef the metal conce.med (ltiedmtii.-r ~s)) ad in the ease of more easil:y, vaporised. me,tal:s it ia p:essible to us~ a. JQOre eas1~ controlled resie:ta.nce, furnace (Prmgsheim and Pohl -e:1;) ,. Ge:etz so), Lovell n>) .; Ca.Nful eleuing -o f th& backing and surroimding apparatus is. nee&s.saJt7; if foNign atom-s a.r.:e to 'be ef'feetive,l .y, elWnat:ed. :Bauer u.) , Riede ,- .) , :&e:eeh1ng tlf), and Ande.r son 1 > -empl~d variou-s: ve-r sions (Jf the serul>-&:ff !:n. ni trie ac.id me-theid recomme:n ded by the Bureau of Sta.niiardis ea-) ana otheTs. J:.o.vell 34) and Was 891 outgassed their backings .t oJ' ~ , hours at high tempe::rature tn vacuo. No one seems "t to have used the electronic -clean-up method o-f Strong ,s a) in oonductiv1ty, invelltipttons. Ande,r son 1 > d.epos.ited some Gf his films oa, a freshl.yeleaved .c alcite surface. ill Vui-oas methods have been p'rO'pcsed for measuring the th1ckness of deJG'slt&d ~rs.. Actually .. all that can be. measved i,s: th& average thickne.&s ,o ver a eonslder~ble area. Optical • thoda of • a sm1.ng thicknes-s (:Brueck is) , Row.ml 1~)) are doubt.M .since thin films exbl,'.bit, as has been noted a.bo.ve,. impe,rfcectly, understood optiea.l anomalies. Ms,y~r ss> a.nd lio'Vell. 3.-1) measure the den:s1ty, of atoms iu the eeam to the backing plate, usin.g the io:n1za.tion from a ho:t tu.ngs:ten win plaoed in the beam a-s an in-di.e s.tor. Ander'son ,) , Fl,r-:,, 4 a}, Was s.9 ) . and Riedmiller as) measured the depo.S'it :rate b7, we.ighing a thick fibt deposited at bon se-t .t ing of' the appara.tu. then .aS-'SUU\ing the :rate was the .same throillghou.t. Methodu which giv.e 10. the nwnber 0£ a to111s per unit area ,c an be used to obtak th-e thtckr nes'S" only, i.£ it be a.~sumed that the d.en.sity, o.£ thi:n l a;r~rs is re1a:t-ed in a. lm.~wn fash.ien to that of the normal metal. Nearly all vor.kers point out in asa'Ulll'ing normal density_for thin films that it. ls really, an: assumption., which future experiments may or· ma.7, not ·d.i .apnve-. ~3) Care i,s required to insure that the film makes good e·lootri·e al and the'r.J!ial ·c ontact with. ·t he· surreundings whe·r e desired~ Ear-ly_ :methQ.d.,s of el.amping elect.rodes on the film {:Stone ,o)} are no-t sat.lsf.ac.1.0.17,, and it lla.s- be:come al.mo-st universal praettce to lay, down conducting areas on the backgroimd to which eonta.ct can be made, and d~po:sit the thin film on t .op s:o that it bridges an etherwise bare gap. '!bus, ililtimate mohcula.r contact, oeeu;rs 0ver a e-e:nsiderable area.. b ma:t&rials used tG c.'bta.in these pre-depo:&i ted eonta.ets have• been eitber· a p~vio-us d&po·s it of the same metal as investigated (Anda•rso.n i) ·• Ri.edmiller aa) . King 46)) $ Gr o:r platinum (Fukuroi ,;a). Houllevigu& t19). Subl?maa and Barth , .a )) ar o:f graph! ta ( Lovell aa>} , gold ( FerJ ..,a) ) • or s-ilve,r (Mitchell av)). h q11estion of proper the-r mal C()ntact betwe,en the t.ilm and the, material 1'11!ea to cool i t and measure 1t.s temperature backing; but does ne:t differentiate e1e.arly_'between this temperature and several thermocouples,. that his temperature gradients are, all smll between. f1lm and eo0ling fluid. t -4 ) S'ev.eml -methods o-f measuring t .he- ns-i.sta.nee of films ha.ve been 11. have used ·t he Wheatstone bridge,, though Mitchell use-s a g,a,lva.nome-ter d$:rectly, for the hi,gber resistances, .Ander.son il use.d an ele.e:t-rometer ;eirfflli t , of variable, oo.-pa.Qi tv. and mes.s ured the eba.rging rate . ».&ttlme--r 611) m&a's:t.tred his resistance$ with altenmti:n.g current ,, using mi ,ampli.tie-r and a vlm-a.tion galvanometer. !he more rec.ent eleetron tub$ D.C', a.m_p1if'1er circuits do not seem t o have bee.a m,;um. emp:loy-~d T. Gonelu-si-o:n:s •f,:.o,m ;p~vieus wor!5; wi. time dependence of re--sista.nee of a. thin metal f1-lm._ to_g ether vith • the !Jmal.l.-puticle s.tru.e·ture ob.served with the mi-ero.se:0p.e(kidrade: a ) * Kramer. ,...) ) Wld.er -c ertain ,e - o n d : i itons. an:d the m i c r o. . . -cry1tal s.~tl.lre gi,ven ( F1neh ~:,) ., Kirchner s-a> , Waa • 0 )) by_electr.o.n ,d :i;,fira~tion ex_pe,..imentl! o:n $Ueh f.11m.s t makes it very pro1mb.1e that the pa.teh.;.st.rtte'ture picture: of the cond:ueti01n pr ocess in thin: lay~rs out- lined .t n A. Ill, ( 3) eonta.ins mw:h t nth. Disage-ement ill resul. ts be-tween ,ez;pel"imn:t:s of apparently eo-mparable va.lid.i ty:, partieularly as tot~ sign c:>if' the time variation o:f' resistance, makes it 1ntperative that more l.1.~t be S¥d on this phenomenon and on. the o-ther special features ef thtn...f1lm ,e0cnduetiv1ty before detailed theories ean be eon•tructed vhi.ch take- qtmnti tati:v.e acco:unt of the lll'. i f fer&nCe"s be:tween different me tal.s . :s.. 'f ile ..ea;,e:rime-ts !fj)prted i n ·.this •tbes~·s : I. P!!P9se of •the- inve·s t.i ptien: 'The results deseribed here were o,bta.ined, using apparatus adapte.d from that of hderson.. in an a.tte~t to ~ somewhat elea.rer the' pi.c ture ef the conduetioa ·meehanism 1n thin meta.111.c fl'lms, with partieuJ.ar empha.:sis on tbe aging p.r o.c ess and its temperature-aepend.ent aspect.s . ll. hperi,-nta1 llietbo:d~ -e !Pl~:ret: ·t o o1ftain trustwo.r thy results, the apparata. was arranged to fi.u.fill the fellowing Condi tians: (lletail-s will be gi•en be'1ov_, 'but in this paragraph is in.el'uaed under each req_-uireme~t a brie_f sta tement of th& means empl&~d t-o. meet it.) ill !'he :f'tlm 11mst be prod.wed with the smallest poS:sible 0:0-ntaminat:le., and small -e-xC$SS ene-r gy. The t.h in layer.s of lead were deposi t eil. frem vapor in high. vacuum on caref'ttlJ.y cl-aaned glass s'f.lll'faees held at various te-mpeTatttres:. Lead ·as eh-Osen be-ea.use of 1 ts appreciable vapor pressure at ue·t-too-high te-mperatur-es, and depo-s:it rates were kept ~lov so a.s to avoid extra kinetic motion in the film. as :far a:s _ po:ss-i 'ble. O.are,:f'u.l ele~»g of the entire appe;rat.ua pre~ud :ea.eh l"Ul'J:, a.11..d all metal parts. we.re thoroughly, 0u..t9,$sed wi th an indue-t ioA fu:rnace:. (2) ~ thiekness of the film wst be known as accurately, as possib~. Care. wa-s tak&n. that the 1.eaA e:va.porate at 1.udt:orm rat:e threughout ea.eh experiment. and t he rate w1:u.? merumred 13. as a. ~tion of •~ te.1"&-tm-e 01 the evapt>AtiJJg ~ by, U&ie;hi»g seve.m l. extra-heavy films w:ltb a Ja$,cml>ale~.. 4 ~tl.eallY ~tad shutter f!ftVO pJ.!&C1'Se .-ntl"O'l ovtn• eva:p()atlo• time. ill ~ b~n ~ •ntact mutJt be ob~d nth the cooling £luid.-. .I t wlU be sean from l?la.•• i am\ 4 that thA!I coolbg n-utd ts eon.tat.A in a q'Uart~-bottomd chamber. al'Ul the. :rum 'baeld.ng (.a micrbs.Qope altde cover gla-n) tc plat:u..- 1~ed qn e.a.n•fu117. t1cldered tato place. i!)_ ielSa'ble electZflcal •ri.taet mutt be lllitd.atained v.1\h th$ film to Pl.attnt.a eo11tact are~~ were provided {Plate 4) so tbe.t the. ttlm 'bw1~ t • map between t,~m, ad the ~lecutc i~a.d& • ~ solder$! to th&• pl.&M,.mse.d pl.a(NHJ,.. Plat1n1~1ng was dom by- pa.tntittg ph.t!mnn c.hi.p po.int where d$S1red (5ee Pl.ate 5A) and baldttg for, t!tt.>m& hMtts,. ~5) !the resistance of the. i"'llm $!1USt 'be mea$m'Gd _rapidly_~- en as 5-t vutes O.'!Ml" lawge ~ c. bl', apply$.ng a potential, and o'l)gerving the rat-a :at which the n -sul.t~ ClDT'ent aeMss a film eharged eou•nse.rs ot v,a;.r:10,1U) a1Bes. Part 0£ the meaaure•nt.s rtal.\'s done using And&Fl*11' s ,J met~d (?law S:L ~ ;part were do• mJil\g .an adaption .c f ,owneend' s '10) method (nata 7) . 1.fh.e ad.vantage~ ot \he le.ttet- are that it mainWns cb.nttant po-te-nti.al ao:r oss t~ fllm, and 1t. looalize& leaks end ea:paei~,.nees eo that more ~lete c-•neeti.ons fair ttl"q, ser-s we-r e -calibrated. aga.iast a sta:tlda.rd 'C3'linth"tce.1 •O ~ r of' ka(i)Wll 6:imensins. lt was, - ~ advisable- t:o &tta.ch a themeeou.ple at the po.l;-n_t 1mlleat.ed in fl.ate: 3 to ohtaia a ro~ idea ot the' temperature of the tila baeldng at va.:rto,\1$ tim&s. tU. l)e,ta.f.l&,d _ie$aC2".iptton (Jf the &f>W!\W<$ and its: ,:U@': ,,, ~ 'l'h ... ~ = ' ' __ ; e ~ -,. · · _ .. , ,_ , .. ·'.fhe assembled .a pparatus is sh.mn:1 in Fla:te 1, and is adapted f~am that o-f ~ n 'I) u.s.in:g so• ei' hi:s gl.a.s-s parts. De vacuum Qyptem lla.-s been ftJ!f$What imprcved,. however• by u:1,ng two me_~ dJ:££wd.e pumps hate.ad. et· om,. a.n.d tli~ U.q-uid air traps i n sar:te • (only, one sht,wn he-~) -. An ionizati,()n !l/J,Uf!/1, ( a,n ordin,a;ry', 201-A •l~e.t :ron tub&; grid potent'ial + 157 1/2 vol.tslt plate p0'tetftlal ... 22 1/2 volts) wa.-s insta.11.ed as shown, connecte.d by a. long tube to the spot where nncter the eva:porati&n f'u:rnace~ 'Tb.is ,-p,Uf!JJ was 110,t very pre·oi.sel.7, -eal,ib- r.atea.. hut a ela.bt of at worst. 10-6 mi:ll;ime.tt1:r:s of m-$l"GU17, !$ believ.a ;(lonse.rvati-v-e for the pressure,s obta.ining d'tll"ing all the $.Xperi.ments. J'nrthe·r co:nt·t rmation of tbe high vacua maintai:n.ed is given by_Plate eA. whieh is a pho.tami.erograph ( dark field illmninatiaa) of the edge o-t a typi:cal film deposited und.~1· nor.mal ~ondit.i'{)as. !be wlclth o:£ the pe:awnbra. observed is that caloulateli .fl'ma the ge,¼)l)letley, 0£ tbs apparatus, assuming atra.tgb.t--line paths ,o f lead. a.to-mS. from f'u;rnaee to depes.tt plate. Se.at.terlng is ,see:n to be negligible., hence oollisi<>ns between lead atoms and :r-emaWng air mole-eul.e.s mnst 'b$ wa-1mpo.r tant. 15.. All glass a.1ld. metal part$ were thorou.gi:lfy, cleaned be.£0~- eaeh run~ us~ the following procedure: ·fbe. glass parts were senbb•d with di1ute n1 t.' i'e a.eU, residua1. lead. Any, &tra.J'_bits of vacu.wn wax were re.move:d by, a thorough ,s:eN'bbing with carbon tetra;ehlorid-e. Metal parts re.cet:tved the same treatment, and. tn ad.di tion we.r e syp-temat1ea.1l;r outgas$ed. a.t r~d h&at vi th an ind:uction furnace . Tbs <::ove:r glass that was ~o :receive t-he. depos:ite-d lea.i was subjected to a parti:cula.r!T thorough eleanlng. It had. previous~. heea plati:n1zed as, required. for eo;ntactts, and wa.s son:b'bed br1skfy, with a cott.on Rab in concentrat.e d n.itrie aci--d (ac recommended by th6 lhweau o,f .S tand.ards B'S)) until the disti.11.-ed r1nBe waier did. n~t run off in.to drops, but :clung -everu.:,-, as a thin .f ilm all O-V"e'l!' the svfaee . !lum th& glass was rinsed aeva:r.al more · times in distilled water, dried in a&., and. promptly inse-rted in the apparatus. As add.ad protection \ ·o the t'ilm baek:ing:, _a n auxiliary shutter (Plate 3} wae placed. over the· ape:rture to the final s.,t op • and re.mo,ved oly, -a fte-r ,o ut gassing the rest o:f the apparatus, the iron portio-n ?'e1!tp0-nd:ing to a m&gJllet outside. NQ part of ·the glasi; or meta.]. that would ultimately be inside the high vacuma. wai touched di:reetly with the irumds in reassembling. Contamination fro.m sealing compound (.P1C$in o:t A.pie.son W,}. was avoided bw. keeping the- wax only on the outer portio-n Gf otherwise clean grO'l.l'G.d joints. and is·olated f 'rom the ma.in vacuum. e ~ r by 1&48 e,a .pill.acy tubes at the ·w:t3'e seal.-in:s (Plate- 513). 16. since this metal has a. low vapo.r pressure at higp. temperature'S,, and doas E;ot. lone its e-las·t icity on. hes.ting below 1200°0'. Outgas-sing was conducted in ap-mo-ts of a.bout one minute heating :f'ollowed by until the ioni zation gauge showed pra.ctica.ll7 ne increase 1:n: pre·:s-sure due to heating metal parts b-r.ight re-d. for a minut&. '.~2}, Eyapg_ratl~: The -f:u:rnace used ts the one described by Goeti& • 0 >., the was ne-c.e ssary., ·s ince melted lead i'll th.e ba:r.e .Mg() (!rucible gripped 'the· side-$ S:Q t i ~ that the erucible was. l.le~l;r dest-roy:ed wlMlln the m&tal $hra.nk on ·Cooling. It will "be observed from Pate, 2 that t .h& thermo- ·e'011;P'l.e Jun-ctio,n us&d. to mea.1mre tlm- ~ratur.e. -0,f the lead i's wry, Q.3.os.e t o the center of the fnrna.ce,. wi:th htit •tal 1m nea-rly all d.ireetiens :from i t . 'Renee, the- temperature a s read im:tst be cyp:1-oal Q.t f~om condensing on the· outer walls .O'f the va,euum ebatllbeir, to be la.t er re-evaporated i>nt.o the fi.lm backing. for the few hours -e xpoaures thermocouple wires did not ca.use ·s .r:tou short cireui t, but it was ne:ee-s.sary to, ol&a.n o'ff the .s tray. l~ad bef:ore .aach nm. to a.vo.U. tr~le 0:f this sO"rt. Above the furnace {Plate 2) ar& placed a tops .S t, a.s t0 -Collimate the beam. of l~a.d atoms ( that i ,s . tQ' t,r,ap a:ny whi:oh an no·t going in t.he right direetion) . ~ shutt~ ahown is tvne:d -on its p-ivot 'by a mgn.et i't-om o'll.tside t:be vaeuwn.~ ope.n ing and el-osi.ng the I7. r out9 to the depo,sit a-rea a s de·s ired. fbe. entire .s top and shield. :S1"pt:em is ~thed,, and so is one te,:-mina.l for the heater eurrent to the furn.ace. This reduces electric ;field changes inside the va.c'Ulmt chamber to the pout where conductivity measurements can be made wb.i1,e the f'ilm is be.b g deposited, without the 60 e~ie alternating eurrent used to heat the furnace d.i sturb.ing the sensitive electrometer of the measuring circuit. Yarlou:s .e vaporation rates for the :f'urnace were determined by_ di.r ect experi11&nt., weighl.ng the total amount af· lead depo:sited on an unplatli.niz-d eever glass. in a known length of ttme (ab'Gut ten hours), th& turnace· temperature being kept eonstant throughout that time~ !he result.lug eurve of mass de:po,s i ted pe:r unit area per m'lit time ~tns1i reading of the milllveltmet&r attached t·o the eoppe:r...eons-tantan therme·eCt3ple a.ttaehed to the £urna.ce is shown in Graph 1. As will be seen from the sample data Gf '!'able 1. the aeevae:,. obta1ne:d in weighing the actual t''il,m is not high.1 s.i nce &tehing r:tff the lead a:.n.d weighing the backing plate, e.lo-ne gives ano-ther weight alm&-s-t as big a.s that of plate and film ~gether. hence the accure.-ey- of the iU'ferene& tufters. Also. it appeared :f'rom comparing the.r.mocouple read:bigs with veJ.•s of hea:t1JJ€ em-rent that the thermocouple might not be absolutely reliable in its response. though neithel' error S.ss so. large as to invalidate th& curve. !he chief souree o.f thermocoupl:e error is proba'blY, slight sh.lits i:n it:s positlon from run to run. al.th~ the fvnace 1.s censtructed to make small eha.nge;S in thi.s :pos1t'1on imimportant. 18. Plate ~ shows the support for the film 'bae:king, with the two supporting ho,aks insula.te,d from each other so they act a~ current leads to determine the conductivity as well as supporting the cover gl.atts while it 1:s being s0ldered. Co..ol!ng fluid can 'be poured in at t:he top, and ctrelllates 1n -close juxtapositioa to the depo,s itei film lt,s tilf. !eo cont.r ol the melt:ing 0£ sold.er used to fasten the backing down. a thermoeeU¼')le was attached a.t the point 1ndicate<i.. Proper eleetrieal and thermal contact 'betwee.n the d-ep,o site:d ,f ilm and. its S1:1?Touna.ings was obtained by using so,l d-er. Be~0re insert... ,tng the film s~ort into the vacuwn chamber 1, earefu.11.y el-erumd strips of so-l der were plaee:d under the, contact ho$-ks, and ~ the cover glass 1tse,l f. A prev:lou$l.y, tinned brass pl.ate inse,r ted as shown helped the . sol<t&r to spread e;venly ever the pla.t.lnin4 back ot the .e ever glass when - heated (Plate 4). !hen, while eutgas.si~ tile apparatus, the induction tu:rnaee was m&Te:d so that the te.m peratur& 0f the cover glass and 'brass baeld.ng :plate remained a few degrees above the melting point of solder tor several minutes. In this way, all creric&s were filled, .and dependable eon:tact esta'b1isbed both with the cooled q~tz ehambe.r :a.ml with the curreut leads. Someti•s there were a. few ·s mall bubbles mtde'r the eove'l" gle.ss, but these selt:tom cov-ered m<>l"8 than a small fraction. o'f the total area co·el&d. The film bacldng was n0t heated except as n.oted here. ln using the apparatus, after the cleaning, loading, and outga.s,si,ng proeedur.e •de.seribed above, the :f'ln-:n.aee would be brought to the desired temperature• ( using two rheo,s ta.t,s 1n parallel t.o ,coutrol the 19. eurrent preei-sely) and held there f'-0r at least an hOll?' with the shutter e-1:,osed.. Mter th&t, the shutter would be opened f<Jr a k:nGwn length ,o f time, thus depo,s-i ting a lt::ne-wn amount o•f lead o.n the backing. C'-0:nduetivit;r, measurements were made f'rom time to time while t he f'ilm wa·& growing., aniJ. afterwal'ds a..s well it growth was interrupted to st'l1Ccy" aging eff:ects. ,i ll Eleetr!eal measurem,&nt•: The electrometer was ea.libra.ted by coime:eting in pal'a.llel w.i th a voltmet-er and applying various potentials. lil1,eetromet$l'-indiea.tea p:e tentials required in the ealibra.t.io-n et the aud-1:ia.ey eonden~s ( see be,lGw) iw&re obta.inefl by direct ~a.surement fer th& point d:estred. instead. ot taking the r.e-s.u it from the .cs.ltbration chart. Some o'f the e-e.rlie-r mea.sure.ment.s were made using etre1:1.i.t Ne. l {.Plate 6). fh:1.s is esse.ntiallJ', the cirettit used~ Anders-on:~ and the ~·xpe,rlmenta.1 rrange.ment repre9.ent-s for the most p.a rt mino:r re-arrange,.;. a eur:rent &nta the elec,tromet;e r, a.ml 'by. measwbg the ti.me batveen indicated potentials the resi&t.a.nee of' the :film• can be ca.lcu.la-ted using the ordinary ,e,-xpone:ntial law of ehe.rging a c01\d:en-ser through a. resistance. Serious 'Charg,v leakage was avoided by_insulating the ele·e trometer sy~tem 1-1i t-h amber, ba.~d rubber, or lncite at ])0:1nts o:f Sl:q)port. 0a.libratien of the various autliary_conde.n ser,s 't1ill be taken up with e-ircuit No. 2, bn:t for this ea.tie we need the. eapa:c.itanee, o.f the eleet.r oseope sy~\em to the. eleetrometer a.lene was in the circuit with that when C1 was also i:n~ usin.g a typical f ilm when at the :ri~t th1ok;ne,s-s fo:r tbs P'lll'P'H•· !?he accuracy, of this ealibratiQ,n is ne-t high, bu;t the -re.snlts obtained. :from th.1-a basis are consistent, and attempts to improve th& aoeu.;raey, by, other means brought d.oubt:s eGnceming stray_ca.pacit-an.ee:s to ,grotmd &r· the electro-s ta:t!.o shielding (a.!§. that o:f 'the contact hooks). Onq, the first re-a dhgs w.ere ta.ken with cir~"llit No l~ however. This circuit asaumes t .ha.t Ohm's law holds tor the film, si.n ee the potential a.cross the film varies several volts a'S the condenser eharges 1.tp, al1d ta -e xponential variatiu in vol.ta.ge with time (a.ssmaed tn re.du.eing the data) ts a. re-s ult of the ,strict p:ropo·r ttona:11ty 'between current .and - validity_ot Ohm's law for thin films,. a circuit was soupt which would allow eurr.ent to be, mpas.ured at <eon~tant potential drop. '!hu&, if measmf.?tmetl:i;$ be eon.ducted rapidly, e.:nough, .uo that the current would. not lw.ve tim& to e ~ apprecta.-bly, due to o,ther eau'$8>& ( time. or iucr~sbg thickness of the film) •· each measurement will be valid, even if ,c urrent i:s not s-trictly, proporti-onal to po-tent.ial drop. l.f the change :t'ro.m Qn& po,tential drop to another he made conw.nient,, the experi.men-ta.l cheek of thj.o, pl'Qp~·tton b~~.mes simple and preeiso. Ma.~·..o.f the room temperature data and all the liquid a.tr te:mperature data were taken using cirem:t lro. ·2 , due fundsmentally_to Townsend 78 ). lt vill be se&n from Plate 7 that, as the potential of the film is ra1.sed. DY, allowing -charge to leak on thr.o-ugh the film, it ean b& leve-re,i by_reducing the potential o-f t.he othe-r sid& of tbe mere and m&.r e charge ~son. the. potential of th& e,leetro.meter. can 'l. be maint-a ined at zero by_manual arl.Justment ef tlle 15.-vo,l t petentiomete-r-. Hei,.ce~ leakage o•f charge fr0m th& eleetrometer systtem becemes ne,gligib.le except as it •Y. occur -a cr~ss t-he eon.denser bank:. Furthermore• ~ capacitance of the elect _rometer sy;stem disappears from tbe -e quatiou used to measure film re·s istanee, affecting only, the se-asit1vi.t7, o-f -~ circuit. fhus tbere. is no longer need te W<Jn7, about stray_eapacitance11. ?!he theory, ~:£ thia c1rcu1t is s,imple, but will be given. here to- 1llutra.t& the cerreetion fo?t 1 ~ acre:•• th~ auxiliary, condense!'. (It bas been :aot-ed above that this is th& ·&nl.y, place where the &lec-tn,ome te-r S.y~tem is 1m'luenced by, a -eonducto,r of rdi-£1'ere11t potential, except acl'O'Ss the film. and the leakage across the ftlm .i s being measured. ) Let the c.apaeitanee -o f the entire electrometer syptem be Ce w g.r01llld • Let. the capaei tance of the el.e.ctr.omete:r sy-;stem to cemlucior ( l) l (Fla.te ?) 'be Clb . (,niia can be ,rar•i ed by, connecting in more auxilie.ry_c,ondens:ers.. ) laet the am:ili.a.ry, condense,r hav.e leakage re'.s istanc:e Jl, and let this resisbanee <tbeT, Ohm•s law. Let the time of' mea:suresnt be short enoug)l so that current t .c be me.asured. will depend on'.13, on the p:otent1al drop .across the film. mm .1.f C:, be maintained at zero potent1al, Q = c;h V, The rat-e of increase of this ehar:ge is tbe film current less the c'Un'e:n.t 1-ea.ld.ng awq, aero.ss the auxiliary, eonde.n,er-. lF , lfhat is: praetiea.11$, constant. 11$11'1 we ( '!his is the only approxima.tion so £a:s-. ) If call -RC,b = T == the tiM e,o nata.nt of the auil1at7, condenser• then :1.:f T ia la:rge e!U)• in ~1son with o-v sbort time &f observation, l F - we 'J1J8,'i', write: v, O/ -..t: nce?--f t :n all but a :£ew ea.s1u1., -., ) - C,b J,/, t (1-fF) 1. ,,- will be so small that even the corre·c ti0u term can be negle:eted. ~ . only, changes n&-OOScB.al'Y, te adapt U. Anderson apparatus to this eircui t were, to insulate the can.denser bank from g?'C1md on both aides, set 11p the· potan:t:iOmat~r .r.he.e stats (which need ao,t be ea.lihra-t). and eonwtruot ·aneth$,,r amd.li:ar.Y, e.onden.e&r having ver:,, ls:rge redstance and eapaa,1tan<;e or the e,rder of magnitude of the eleetrometer sy1Jt&m (.Plate SC). !be electrometer b now ueed enly, as a null indicator. In u.s'tng th.is- eircult, one eoimeets the film to the electromear, releases the gro,lmd1ng ~,,, llowe the f"ilm to leak charges oate tlle insulated eleetromete:r sy~tem for a known length of time, movug the petentio.mete'r contact awq, from itere just .fas:t n.ollgh t .e keep tbe &lect:vomet&r reading ·z er0, cuts Qff the film at a meaaured instant, adjusts the eleetromater accura.t,el;r, tQ se"• and read,s the resulting l/, . Tile ma:aipul.atien is but sli~tly, mere .complieated than that cf' oi'?'euit. Jlloo 1., and th& errer mu.ch l,ess. leakage eorrections: ea.n be applied. :more easil.Jr, stray capacita:nees do not matter, and. .c. cons\a.nt potential drop 1 maintained acr&Scs the film. By, the use .O'l the 45-volt potentiometer (Plate 7) this potential ean b:e varied.• and 23. resistance measured tor various potent ial dTops, thus checking i§l CalibJ;"ation ~ cond&ns.er-s: 'lbs elee-trome.t er sy-Jtem is so 'Shielded (Plate 7.) that it,s ·twe ~ s aet like separate conden-se;i•s connected. !hat is• the capa.c1.tmuz.e of ( a) to ( 1) in the diagntm can be negl.ee.ted, d'1e to the gt'0\131ded shi--eltl· between. Thus, lf we keep ti- to-t al. c.ha:rge on the electr.ometer sy-~tem zero, we will alw~ ha:.ve. pa.rt of el'G'Ctrometer-· &yJ'tem = - Gia= - Q6 = charge on (b) Co ~ W'he-re ~ is the pot.&ut.ial. irullltica.t• ·oy, tha electnme,tar, and Ca is the ea.pae1tance t:o grQlmd ( i.e. to the :shiel<Uag) of the ( a) pa.rt of the sy,stem. C3 = eapaeitanee of ihe standard eond.e.1\.ser Let us eomiect h parallel ri.th Ce, and atlju:st the- po:te11tiometer te, V, the aa:me· potential 1 ao tha.-t I-' on the elec-t·romete.r obtains. Renee w& :have: V,'-V, 'this metho-d calibrates one auxiliary_condenser against the standard, but is net -convenient if ,d .iffs·r s so• little from /c,bj> > IC.rl ) since then V, I 1/, that JieaWl·ement is difficult and inaccurate. The other cond.ense,r:s we-r e. ealibrated against the kn-own anxil.1ary_in cascade fa:shien as follews-: Let the, total charge on the .electrometer sypten be Q , grounding the ele-ctr0meter,. with -o nly, the k"Qwn c;~acro-ss the· f!IJ..P between ('b) aud ( 1). Wow,. with tbe ,electrometer insulated, we 24. :connect the origbl:all7, uehal."ged ½h !n parallel with the C,1, , sad e.djus.t tbe. applied pete:nttal to V, 1 to bring the ele:c-trcmeter back to s-ero. In d&i.J:tg these things, we have thus we ealibra.te the -co-ndense_r:s in or-ieJ"', eaeh again:st th,e, 1a.s,. !he -results are shmm in Ya.ble 2. each value giv&n being tlle result et a.-t lea.'8.t ten capaeitaaee read.inge. ! he cal:ibrat1011 •f$,s the resw.j; 0£ 60 trials,, c" against Cs 'b-ut the s ~ is such a.a to make the probable- ·e nw 2$. A .f.mtthe-r :eh.eek is ~lable nea.-r the o.ther e1ld ot tbs caseaae: ~ bast auxtli.&17, eo.ndenser· (No. 4) is marked 0.2 ptd. by, 1 ts :make'r.· {-Oarpe-nt.1 -. Paris) • while the value obtained by the ind.i reet meth<>d <1.eaeribed Jtar,e is 0 .. 201 p:ta. 1f oarrid out to three decimal plaa~s. the- agree.meat with the expe-rimeatal ,ralue 1.s encouraging, but $011le of the apparent accm-acy must be called good .f'o-rtune. Leakage r.a.tes and t-ime ·eon&rtants were determined by_keeping the -ele<:trometer sy:ate,m, charged, at zer-o poten..tia.l, mea.-suring at various times the value of \/, reguired to do th11t. lfh& results are iMluded in Table 2. Since observation times were kep:t unael"' 20 ~e,eonds for- the mst part, only, the readings ta.ken with ~ have· appreciable leakage eorreetions. C:;· -~ t • and the<t.r· sipifi:canne· ! I. (}eneml ! 'The films U.8ed for the data r.ee0:rd:ed. here were selected from those actua.11¥, made by', the method described as the films which exhibited uniform mirror-like appea.ru:ce uml:e:r es,reful visual 1nspeet-i0l'J.. lt was folmd tha:t., among other things., the slightest traee o-f pease irapor., even or s,lightl,y, ·warm Apiezon W.. inside the vacumn chambe'r i t:nlf pve rise· to films ,o f uneven thickness and non-metallic- appearance. All w:eh films were discarded tor use 1n cal.cu.lating ·remilt,s . On the o,t her hand" esse.a tiall7,, all the readings taken with :f'tlms it was decided to use are f-ound. en the &raphs; 1n Grder to ·t nel• the d:o.zen that do not appear, !hi.c kness of the- films was ob·ta1n.ed. b7, dfvid1ng the mass of lead p&.l' square eenthneter (known fro -the evaporati,o n rate of th,e furnace. G:raph l ) by_tlle densi~Y. of bulk lead (U. 3 gms. per ce. ) . A brief diseu.ssio-n o.f this as-sumptlon will be, found abo-ve, on page 10... ~ ~ &tl',fle,~ asSW!led by~ t~se lead. film& vh$n a1l·owed time to, come to e~ltb1"1um was photo.gra.plaed with ti. ul~oseape. Plate 8A has al:read;.r, been referred to as a. photograph. of th& edge ef a film tabn b3, this method. Plate 8l3 is such a photograph, taken from t:he eent:ral. reghn of a typteal lead film deposited 'b3", thb apparatus a t ttomew-ha.t higller magnifies.ti on than SA. 'fbe g;rain.v.. B-truc~ i .s Te.l'Y, o'briGue, ·m id is ty:piea.l e:f the appearance of the:se films lJ.nde: r uttramiero1!copic ex.am1.n ation. fhe colld.11Ctivity, pl,O,tted is the qu.antitJ", 0- = L R, wd R is the measured r esist-a.nee of' the film (ratio of po;tenUal drop. ~he• to eurrent). L ts the l ength 0£ the film between electro-des,•. W 1s the wi dth of the. film, and d i.$ the thicbess det ,a-rmi:med as above,. ~s q'lil.aniti't y, will ,e,e~tain1Y. ~ve the eeadnet1~i 't1', i n: illho~ per eentimete,r -cube. Gt a. thick piece of me.tal, but it ls by, no means obvio,us that the ex.pr&s:-s-ion means mu.ch mol"e than it aetus.l~. say, for a thin la~r . 'fhe values of L and: W Y-<ere mea..mired for each film., 'bat wel'e mabtta.ined --re-17, uarly', 1.0 .e entbieten eae:h b7, the con.stl'Uetio:D:. of the appe.ratas. Also,, eai-e wa.s taken: that tJle> eicb;rs of the f!lm b.e a.& near"J.T, a:s possible. pel/p.e ndicul-ar to the elttctr<Jde e.dges, but no ;c o;rrectio,n was applied .l f tbe $sul\ing :film was 0:ut by, a. 'te.w degre:e·s . since i~ is ea sily, shown that the co.rreetion factor to be· applied £.or- a pa:amllelo.gram film whose corner angle is «. i :s. somewhere between siaO(. end -osc: al.. . For angles .ne.az-· ,p,• , bo.--th these functfons have val~s ve-.r.r. nearly, 'Wd~ Catting off the potbtta.l f:rom the film .£or a time wa,s observed. to have negligible &-f;tect on subseqll&.nt obaena-tion~.; ps:s-sing ~ n t thnugh the ·t tlm seemed to increas.e .its resi.st;uac.e sU.gh.tl:;r, f'.•r a t .i llle, lmt if the film was allowed to rest a few .minutes, without c.urr.ent ( with .or wi:tho'lllt being -e leetricall;;r, ehal"g$d) • 1.t seemed tu ~cover. Il. P~limi:MJ:Z obS.!ntitt lon11: 1'he results given by, eight prel.im1n.a!Y, l'm1$ of the apparatus are shown in Graph 2. ·!fhe films were obtabied b7 exposing the water-eaoled backing ,e.0 ntb111,0u~ t-o the beam o:i lead a.t oms. meaSUll:'ing the eoll.d1wti.vrit-y, f'nm t:ime tto time. ·mie spee·d G'f !eposit wa.s kept as neaTlJ', eonstant a:a pe:Ss:ible tturing each rtm , lmt d.iffe:nnt 'Spe&ds were used for· 41:tf:erent runs; tlJe; rate of 4epe-s1t 1n ~trom• increase o:f thi;elmess pe:r mfnute ts a..ttaebed to e-ach C'\U"ve on the- ,graphs.. I t will be obwrnd that the eond.uctirl. '7, of tlle 'thilm•r film• de.erease-s 1Jli.glltly: vi.t h thiekne·s:s, a re-nl..t that is in. harmony, with the preUmi12ar.r, 1.ne.rease tn resistiv1t:,, noted by, Anderson (page 2 ) .• although it nmst ba em,pbasize-tt that tMs ~ with thicknes<& really; m&ans. that the meaSlU"ed :resiatane,e: ot th& film ts aot decrea sing as fa.st as- thethie.kness 1a grovtng. In m•t ca•s·. the fil'm imM'b:i ts det~etable et-'.!l(l.v.cti1>n: a:tte~ a. -short time. o-f depos:it.., and then the ~si s ta.nee remains almost -eo~:staat for a considerable mta'tb:er ot· Angstroms increa• tn thie~ss. Soo-n. ho•ve1" t there· eeme,s a petn-t { depe:ndin_ g lriOft., apparen t .q ,, on the con-di tions Qbtah!.hg 'for the intiridU&l film thaa ea th~ ee;,ntr0llri varia.ble-s,, e. g. speed of deposit; as will be seen. hette and later, c:orrelaU.sn between depo1tit spaed and film pr.'Opert'les is v.e,r,-, ~ll(skt ) where the eonducti?it}r., begi-8 te ris& ·r apidly, toward the ~ul.k val• abcmt 1013 higbe.r·. lt will fu.rt}¥)r be .o bsene,d tha.-t thette po-int:$ are in th& m.1,glibD:,rhood Qf 150- Angst.:rmn& tor mo.st -o f the ft.b is, there \le,h g thr:e;e, at about 100 .bgstrall!$, and tt1& at abo~t 200 Angst-rems. lt will a.lsa be seen that f-aT several. of the £ilms the ·C'\ltVS· begins to have an infl&eti.on point. as t h ~ eoncluctiviq, began to e.ome prominent en these ,~ s , but will be more so- im those. to eome later, ffl.; ·ltbu •CQ~ttntf •~1f•pl).:m }te.grature: . . . ,, .( . eone'SpQflding inl;,rovemeat in the v.acmum obtab\ed in. ti., appara.tu. heue t t ·was tielt that the :tilms obtained under these uw con.di t1Qn {!l!J,ve a.u,ffiehn.tly, a-re rel.table retrults te j'1tst.:tt:,, their being d1-s.eu.sse.d separatellt The recsults an fi'.,v;e mon .1"\mS of the appara.tu-s are plotted on Graph 3. l t wtll be. obsened hEtr$ that. the thiekness where, a film begins same £er the diffe.1"$nt spee~Mens mea:SUI'Sd than wa.11 the case f'er the fi:lms described und8'r Il , be:lng abo-11t 60 Angst.~• for f'our o:l tlw films. and only, a little .~ l-00 for the -other e.ne. hrthe.:rmare~ th& Wl-eeti<m point is mo:re p,:"eminent. in these data. showing ds.t1n1te,1J,', that tbs eonductivitw, of th& fJlms depo&f.ted "lUldel" these e&.nditton-s starts its rap·l d :rise• then s"lews down a little,, but l .s st.ill in.c.reastng very rs.pi.al¥~when 1t ,s val• beeomes too, large to 'be measured u the- apparatus in its present fo:rm. lt i'& partleul.e.rlY', tn~ti?lg to not"iee how tar the co~tivit)-, of a f'ilm deerea.se'S with ti:me. when the beam i :s cut 0.ft, and the :film left to 1 tsell. Two of tlJe,s e· CVV'e:S ·&how la.ow the .eono.ueti.vi t7, deccNases 'by-, •re tha-n an order of magnitude under these circumstan:ees. ,(Stable lGWe'!i valne.·s we·?!$ a.pproa.e.he.d after about half an h~). When th& lead ls turned -on agai11i, it is seen tha.t the r1se tows.rd bulk value is pa.rtieularly', rapid.,. imtil. the cu:ne becomes p:rae-tieal)¥, a continJila.tion of the form.er po-rtio.n. !hese &ff'ec:t s were te!tted more thoro~. by the subsequen.t :experiments depicted 1n Graphs 4 and 5. Graphs 4a. and 5a. are the results ob,ta.ined from \wo rwis of th& a;pparatus. und:e:r the same high va.emun conditions as the othe:r films. diseussed in thi.s seeti.o-n. the dif':f&l"e:nce being that -deposit was interrupted at ngnlar intervals·. and the film allowed to come to equilibr.ium after each partial d$po-s·1 t. fhe curves ,o.f eonduetivity, against t,ime while this approach to ~qutlibrium procews was oe~ing are; ,giv.en in Graphs 4b and '5b . The 81Jllbo l :B i's attached 1 to the. cunes at the polnts where the, c.o mucti:v1~. becomes so low that th&. conduct:i.o n of the film wa.s lost in that of the baekgro.uni. l·t will b:e s.een. that the corulac'tivi ti'; of the thirmer fi.lm$ rises rapidly, to • value bl the neighbe:'!'hocd o-f 1-0.-5 lllBO C!Vi:'1 • then deeJ:!eases '$lightly, as the thie,lm&ss .b icreases. until the film is cut aff f:r-em the lead beam. then swittl:7, :falls, back to be.come ind1stingtdshable from the backgro'Uffli. On the nsxt elt]')o~, t1w ·same process i,a repeated almost as on a fresh backgouml, and so on un:tll the t:otsl average thi.ckn&n is se~ra.l bumi.nd Angstro•· f1um tbe rise1t be.come higher and the falls noit quite se lov. G:raph 5b ·e xhlbits a veey, intere,s ting :t:ea.tm-e that will be seen a.gain in the data ~o be c1U.s~s,aed below. The curves ob.tain:ed using 20 volts ac;ross the- film te measure its ·c onduetirlty, 11-e alway~ below or eoi:neidant with tbe QU!'Ve·s obta.tned by, ub:g 40 V&lta B;CToss the same film. '!he thick fihn does net show this separa:t1on within the enor of the e:q,eriment, but the thinner ones- quite deitn1t&J.v, dise'b.ey, Ohm's law-, in that as the a._p-pl:lerl. potenti:al is halved the current is reduced by, a factor· slightly, lar~r tba.n 2. 'lb.is resul..t is in qualitative agreement with a ;r.e sul:t g1~n by,, An.dere:on1.) ,. although the effeet ebse:rved hete b pr.evious vol ts.ge h1:st037: o.f the. filf!,. An at.teurpt will be lllade in behavior in terms .o f aeeepted 1<1.ea'S as te the. ~1eal state of :metals i,n thin £1lms.. i v. •Fi1ms •c:oadense,a ·\on ·ltgu!d-a:ir.;..co()led •b~~ound: A. ·si.gnif'ica:n.t nwab&.r 0£ e:x.perime:n~s were: made· while cooling the ba~:eund with liquid air. A t•rmoeott_ple was a:tt:aeh.ed at various plaees on tbe fllJn backing itself ~nd on ·t he bra..as underplate {Se•- Plat.es 3 and 4) .111 an at.tempt to de01.ds what relation e~isted between tbs tempera ture o:f the liquid ai.r an.4 the te:i,peratura 0£ the film :t.tselt.. No more def'i- n.:ite eone1-usion was .r eached than that temperature differences ai; high as .5 0°e. iJ0.u l.d $Xi.st be·t.ween the film and the fluid, .and this e.itf&rence- inereased when the film was exposed. to the:rmal. r.adiation from the evap,n•ating hrnae~h Therefore• the evapoi-a.tio:n: was conducted in sp,mts .0 ,:f a tew minutes eeell. with the shutter el-o,s.ed in betwee,n intervals cf depo.sit.l.ang .eno~ so that the film approached e quilibrium, 'bGth theniall,', as indicated by, ~ thermocouple and in eon1'1gura. ti.on as hdicated when the eonduct:J.vity settled down tmr.a.rd. a eonstant value . t he film was r a t her imeel"ta.in, but i t is b:el1ev.ed .from the read.ings ou+.ailll)d that t~ method.s described above maintained the fim tempe.r a.- tur. below -100-ocG. thr~ut ea;eh liqttid-atr-aoO:led. rrua. !!?he· r.asul t-s obtained in f 2:v.a rtm.S wt th the apparatt:JS with liquid air C8i>li»g are plott-¢d on flraph& s. 7., 8,, and. 9. !he·s e plots haw b'ean made a s Mar:l:r •s elf...&xplanato17, as possible.:· With each e.urv& i,s included the rate of' ,dep.o-sit. with the aging curve$ the averag& thiekne1us of each f illn meaSV&i, ana the C11.'i'V8''S taken with ·20 vol·t » potential 4r.op; ae:ros.s- t.he film an e1ea.:rly 4is·t i~sh&d :tram tltose' talten with 40 volt.a wher'ev.er sw,h tiisttnetio.n ex:lsts:. 'lhe most stt-iking dif'!ar,~rwe between these e1lr'V'es and the simllar 011$-s (Graphs 4 and 5 ) obtained £-or the higher -~mperature•s is that ta the pl'e$$t cas.e the- ph.$mnnena o,_f 5p1d ~a:se .o f conduotivitr. with th1eklti.e&s toward the bw:k valu., and o.t ~id decay, wUJi tima in:to the background eend.wtion., seem to o-ecur nmc-h ·s.ooner, and $n more i:nt-ens·e fashion. ':Che :po1-nts where cond.llctivit--.,r, begin;& to ineJ;"ea.:se rapidly, a.rs 1n ,e very case in the netghbor.ho.o.d o:f 30 Angstroms average thickne-ss-. However. when the beam $.s cut &f£-', the co3lld.uetiv1t.y 'falls a.pin to lOll valuee. ira.ph ·9 a bet;ag a pa:rtiaula.rly good ex.a.mp.le a.f the high pea.ks and. l&w valleys obt:aineu. For SOE ·o f these fihts,, howeve~, t,t Wli,U be nnted that,. whil.e the f'irst in.er.ease .o f ·e.onducti"Vity_still begins veey. ear ~., la.tar the film seems uarly, stable a.s lt is laid domi ,(.e.. g . Graph 6a er 8a}. II Depa.:rt-ures from Olm s law a:r.e s.ee.n to be le-ss s·t:rild:1ag for· the w.,s.t pa.r-t . 'The ,s lope of the aging O'tQ."V$ o;t fi];m COllld"t»tiVi:V, ( a quantity_pr.& portional to the trae·t iO¥l rate ot c.nange. ot eonduc.t,1'1it,-, with time, si'-nee these curves er& pl-otte·d en -emi.'l,og eo.o ,rdinate·s ) will be o:b·s-er-v~id t;() bee.om• le:o steep with i:aereas:tng thickness. !his. pomt will be b ~ t out :more pla:inl.y, tn -&~a.ph io ( saane• page as Graph l ) .. 1fhl4h 1:s a direet plot ot ·t he W ti.al :rate of decrease In 10:g O-- ·w1th time against the thickness fecr the va.rl.mui fibls fttJr wblch a :stgn;ifieant naber of &•h ttata exi&t. v. Rat.e • •ft£ : t!!!::::d'ec:•a;$a:'·b,t ·c~•tl.vtg: Graph 10 ·t s, as noted a'.b&w,, a pl:O't of tbe :nega.tive »nmei-ieal slope. of the initial })al"t of tbs, agi11g eurve against the th.ick:i:unis of the flbl a.:ing aged. i -t will be seen that thi$ in1 tial slope: (;propol"ti<>nal to - : ; ; ) bee:o.mes small.el" tor the eo,l d 'be.eking w!.t h il.aC'reaf!ing th1ekneS,s ef film;. and ref~rence to ~ . b-gr-a.ph. will show· that this rate o:r decrea:se ef le,g <:r de1tteases with the time-.. 1l'he- ·ma.:rlm.a. •b:sened 1,n the curves for :room,..tempeyafmre back1ng >€>?1 Graph 10 are qtt1 't$ amazin;g, th~ngh eautioa whispers that wo data. -an hardly <$not'Jdl to 'UB a~ a ba.s.i.s fo~ a new th.eo,ry_~ ~he se l!la:ldma mig):lt., o:f eo"tCse, b$, iue to ins1!ffie-ien:t agi:,ng ef previO\:tS lay:~n.•s., that i's, the: lead be applied to an uderceating act 19t Mly in. equil1brium:.How:eve·:r that this 0 1mde~ng" p-henoman<Jn s.holild ha.ppe11 pre:f'e.r e:ntial:tl', at hi&h temp&ratures wemld. be, amazing a11JO. slncM11 presumabll', the eq:a1l.i'brta state shettld ~ a.ppros.~d mo-re ra.pld- ly, at hi~e? temperatures. !fhe total time of ~lag wa9 ab.Gut ;SO minutes fG:r a.ll these t!bts.. Jar the li:quf-d air, te.mperatnr.a data, it is seen that 1 fa,r 100 Angstroms~ say" t.h~ i"ilms depos:ited more rapidly. tend to haft a higher rat:e (if Wtml decay in oonduetivity,. fhis tendeney, is so isligb:t,.howeve.J;>, t.ba.t it could n:ot 'be ~garded it it, were not the re.wit to be expe,e:ted on the basi• et the pie-tu.re to be presented in paragraph VIl .~ 'Q: ·~oll~!\i.Siffli -~m~ ·~ 'above -~l:fs,ei.rva.t1o'n$:: Appa.retttlf4 with these experiments as with :many_others tn the field. the quantitative. re:suli;a depend in such mea·s ure 011 s:ee~1-Y. minor de.tails o:f experllnenta.1 tedlniq• that. unt'il m&l!'e s t ~ b e d methods arec woiked ou.t, 1't 1;.s oi" doubtful valUity-, t0 leok much deeper inw the actual fflJllled.-eal re·s -alts 't han thelr orders o.f· magnitude.. i t seems, however, that ·the uends enibitei:l in the:se data are, qua11tati'V'$.- 17, true:.. and the r.emdts 'ther.efeT.e• not without applicatl.on to the ;pl't>blem at .hand. ln ad.di tiQn to the gene.ral ,s tatements a.hov,,t ,e lectri~al e0:nd:ae:tirtt7, undaT whic:h. the results of &tl!.ler expe,ri•nte.r s. were elasstfi~d 1n A. 11, ( 1.&. thickness depemdenn ,o f ,a0Jlduct'1vi tv, eri-t ical. thick:ne;Ss for :r apid ~ge.t .• ins.t ab.ilit.;r leading to 1n-e11ersi'ble ehang&-.s 'W ith time and. temperature• and the 11:(rp.o rtance of deposit co:nditiona) it.he .r .es:ults de-s cribed 1$.ere allow the foll0wing mo-re $pe:Cifio statements to be- made. true under the eMditio.n;s. o'bta.Wng for the experiments: (l) Dep&sited metal films ultimately exhibit a non- u.n·1f-9rm, 01' gamy s.tructun under ul.:tra.miero-seopie ,e xamination.. ( 2) 4fhe ,e rittca.l thiclates '1 for mpid .change of eonduttivtty f f' • 0 a eontimiou.l;r. cu,pO:Si:ted lee.d film is sma.llmr for lower 't&mp&l'atm'.e 'of de:pQsft . Afte,x, thi.s thi-,¢knes-s. is reaehed. the increase i.s at f1:rst v.er:,, ·r apid, then lr&se .-apid. (.3;) Films deposited at room. temperature- show a slight ~ a n in co·na.uet,lv-1:tcy, with thi,e:Jm&ss whe.n this i-s le:ss ·t han ~ critical thteknes·s o,f ( 2) . (4) 'l!he cond-«Ctivi~, o_f these filt!m deereases vi• time a:ft~r the film: t& depast ted.. the fraot.i eal rate ~, decrease: be.ing mo:n r.a.pli: f'Qr th& tb:tmier film.s., t:hough there ts indication of a ~nmm 1n thi:.s rate tor iu.t&rmea.iate thiekness at hlghe:r ttnltp&r- attaes. (5) !'he ~tal ehange o.f eonttuctivicy-, with time t:s larger ta :som~ tJases for smallu appl1e4 potent1al uncl to meas,- the eondueti'Vi i,,,. and is lar~r· :h most ¢:a.us for lowe-,r tempera.t~.. ~ beariltg of the'.Si8 ideas on. the. pneess 1<,ff e-l ee:trtcal c~tion in thi:m: metallic fill!ls ,w ill mw be exam!Qd. V~I . 93!:!l$,'t$t"i,.. ·.e.e ~ltm ':IWf Wt!,""1t;lba:':t::o ~t!.-:v:t:~ti;: On· the basis o-t tbe elltpe,r imental fact.a quo,te:a and obtained in the abo:ve sections 0f thts the..s,1:s. the fo.ll<>W'ing p.ictlll"e· ef a:t.omi-c behavi~r giving rise tG the e:b.served :eonduct!:v1:t;y, pbeiwmena can. be U::ffitl,J:1-! The .f'reshly, dep&'Site.d t1lm will Cl>nsi;at o£ a.toms eon:d.ens,ni :t:nm the vaPQr,. d;ispe:rsed at ramtm.n on a coo:ltJd ao.l id baeldng slll'face. tn view ef the large ene:rgv, changes availabltl if the· a.toms 'b&c»me, adsorbed. en the backing plate, or mo.-ve close. t0 eeeh &the~· a.ru;. fom group1J, the ra;n;d-Qm eolltf·i g,u"a,tio.n em:mot be expec·t ed t• be· e.n .eqm.U:'brlium one. It is mown., howe.v:e.:,. tb:at only, a very, few a toms o:f the met-al (eornpared to the number available in a. few minutes depo.s'lt with th~ lead 'heps used) a.rs suf!.ieie:nt to n.eutraJ.tz,& the Yan d.er Waals adsorpt-ive i'Ol'Cas. partieululy if 0:th&r at~ms (e . g. material not removed by the outge.ssing pro-ss) ha:ve bean adsorbed t'i.wst. llenee the next p<,s:a i:bility, fn:r a more ,stable: con'figllrat:ion is t.ba.t e.f f~rndng ~~ups of a.toms ., the added stability be'.i.ng .(lue to 'r &duetien of free SW!"faee and to lat-tle.$ fQ'rmaU.one ·The· eq;uJ..lib-rium state aetue.lly, observed in. tb,e:se e.x_perime'ats t:13, th.us the ene to be ~et•d ,on this pteture, since the pre:auma.blV :Uxed first. 1~w lq.-1'& are ba.eeessibl♦ te accurat& oos.e.n ation with this apparatus. and tha,refore the atoms w.Me:h ean be observed are ~se coming a liU.le later. u.eas with 1.llQ:'N greups. fhe r a te at which 'this aggregation rill take plaee will depend en the mo-l>il:i ty 0-f thee tttom$: on ~ s'ttt'face:;, the equlU.brium e-onfip;n.tion. will be appreaehed •re ?"apidly 1:f tli..e a.toms are lt!Qn mob:118. Since 1:h$.:s a.pproa:ell to e-quilibr1mn bas be-en ~in:g on all the thle dep!i>$.1t was taking :placa., the degree &£ instab.:tlit.y existing at the. inCJtant ~ 11t9p:ping the} deposit will de.1'end on. tbe nls.tive rat1!'-tJ 0 ' -0.£ d.epos-it ~nd aggregation, being pre•&UI!ie.b.ly_greater 'tor faster depas1t. 'Jlwo, maenani.:sms o.f ce,onduotion wq, be eons14-ered as p-o,s.-s1lJ.1e f o:r the th:tn lq,~r-s 'llfflder tit.~,s:e ,eo:nd1tl0\'mts: ,e,i ,\hei- eleetronie ,o,r 1om,e 0oin.itte:ti0:l1. On this p.l,e,t,ure ot gl'GtfP f <)rma ti0:11~ ei,ther mechanism will give a. -~ he rate ot aepo-91 t can be contr0lle-d with satlsfn,ng aec11raey 'bry tbe me:-tbods d.es-e:ribad above:. but .adequate e~intton of the depe-tit surta•e w give really, :repMd.uc.lble values of the :rate :o f a,ggre:~~ion was e,ppare-ztt'.ly not assured: by, the methods used here . lt •1. be that wcce-s~f 1n this d:irscttott l1e-s along the lhe, ot :rigG~l3,, wte;using ~ dep~si t. surface. a.~ a.one by, lteve-11., al tlu:mgb 1t ts 'Wle&rt1d.n h.ow nmeh neh he~ole be.at tl'eatment will. uJw.e the st.ruetur.e: of th& s:urtaee :i~lf' while d'r1rtng e/tt mtsorbe:d at-oms. -e olJ.dutivity, c'leQrease w:1 th time:· A eonru.mti.on electron 1s .able to l.-e-ap .s mall gaps,. wt when the film draws up into .e l.osely, packed ~ s the path oit a eoll.d~tioa elect.?'on be:come:s 1.n:t.a n-upted by_a ,sma;iie:r aumbe·r of larger gaps-. and the:Se iM,ibi:~ conductivit,',. Oll the o-th.e:r htuad, Mi, elec.tll'G,.,..M:e -e-endu~tiou which occurs will doubtless be motion of !ontzed ....,, point so that e;,;q, atom ·which j,o-l,:ns one G-t the gollp :i 's unlD:&;cy-, to &scape fram the s.t:roa& f'.ome• birultng i ·t there. fbat ts,, ·on:]T, tmeombined a.toms will haff app:re,eie.ble 1tt0b·i lity,. .and as t.be- numbe-r. ♦'f su.eh at-omt decrease-11 the eonduetb1:t¥, d1tcmases. !l:18 ionic mechanism 61ves, on t-his pi0tul'le., e-etDplete absence: e.f co.mitmti:o a whem. the .atoms are a ll comblned. Many_o-f the thicker filn\s 'be~ame pmetua.lly; ·~ ~ t,., and tbefl.f ore e-qu.ili'.bri'Wll eondi tions had M U supposedly, been r-eached., and all the tree atom$ bad gi~a up their f'Nedom in e l t ~ for stabiltV,• f his -residual conduction indicates the e.le.c tron pi.c-ture. wheB no· more evreat is ca1>r1.$d by the, i'ilm itself (i.e. -resistance is that of the ba.eldng plate). it means that a~egatio,n has pulled the. film into- separated pi~s 1\.0 th&l!O•l.7. t:bat the. different ;p1ec.8's h&ve lost all connectJ.on wt th ea'Ch o·th&r. fh&t room-temperature films ( See Graphs 4 and 5 ) do- not become as eond:uet.ing at the start as do those deposited at lower te.mpe,ratur.e (Graphs 6.7t8• and 9). a.ltholl@: both kinds fa.11 rapidly. back to the eondw:tion ,ef" the baekgro:uml, may; mean that the extent of bs-ta.btli:tJ', in the high--tempera:ture film ts less than that in the low- t-empe-ratu:re: while: the high-tempera.ture film was 'being depo:Sited.. !.h$. results of .Aad.-erson1) -support his asse,rtion that the rate at which atoms -d is84)pear into the eenters of a.ggr,egatiGn is N-ughl;y, propo-rtional t-o the number of atoms as " t un:eombined.. tntppQ<se.1 l t ts reasona..b1e to then, that when depos'1 t ts first s'tarted on a plate eontalning no, 'WiiCombin-ed atou ( i.e. all previG"aS l&;y~rs have been allowed to e.ome to e qtd.librta ) ., the average space oe·tween atoms will decraa-se very, ra.p1d:lv, at firs=t, ttn.ti'l e~:µgh uncombined atoms M"e present se that the rate. o.£ a ~,gat-ton: beoo~.s comparable w1th the rate of depo-sit. ! his vieW;POin.t is- ·eonstst&-nt with the rapid 1n1 tial rise in .e-ol'!d'Wetlvi t7_. no too. on Graphs 4. 5 , 8. and 9 particular~,. When the s:tate 1 s :reached growing ·tso-1.ated patches, oonductlnty, should become san-s1b1". eoncstant and the d-egree o-f d1spettion exis't ing when the su;ppl.1' ts eut oft will be measured by_the total dflereaS'8 in eond.uctivit-y_, oe:etirri.n,g a:tte-r this as the film approaches eq'ldlibrium. Measure-.ment• -of' this ,s o:rt 011 a properly, staadarohed batkground sheuld give mneh iaf'onmtion ab0:ut the rate (11f aggregatl.on o-f the a.to.mt in a newly, deposited ,,thin .f ilm\, altheugh the eha.ngtng aizi:e ot the patches -o,t me-tal eompltca.tes the pieture. Cons1.d.eratio-.n, however,, .of G:ra.phs 4:, 5 ,, arul 9, shews that under proper conditions, ,e#en f'or maD.Y, ~ r s of fully, a.ge:d lead, new depo1.t'its behave almost e-xa.ctl)'_a s on a fr..e·sh 'baekground. hom this it. f.ellowca that th& pat-eh.es eover ol\lly, a very, &mall pert1on of' the baeld.Dg at such ti-me.s. !l!ha more sudden appearance of eo.~ti0cn ( i.e. smaller er1ti-eal thickness ) for e-o ld films than for :roem,..tempe-rature :films (e. IV ) would be attributed on this p.i.oture to the smallei- degl'ee o,f rea.rran&ement oc·e -urring while too cold film is 'being deposlt:ed. ~ smaller conductivity, plotted fer 1nereasing thickness o:f the ftdnne•firt £11.ms has been discussed above as being d:ue mainly_ to the iJICrease c,f the thi.c kness { d in the formula for 0-- in paragraph e,I.)., the resistance remaining essenti:a.117, a.:t the same high value while. the thiek:nes:s i:ne:reases. When the pa:tehea of' Qt.al on the bae1dng become large eno.u gh. so tha. t e.ppreeia'ble f-o.r.ees are· exerted by_ the ed©3s o-f 02 gremp on tho,se of atbar groups in the neig}lborh.ood., these surface fQcree,s will t81'1d to close the gaps between the patches. 4'hus we should have a -sudden increase in the equilibrium cenduetiTi ty,. It seems rea.so-nable to suppocSe that the first few 0.ecurrenees c.f this sort wi l l crea te a very few paths .o f easy, 1 eondu.etion ( ttthrough hi,ghway;Stt , so to speak) across the filln,. and that $Ubsequent ine~a ses in conduet!v:i.t y _co.me from forming a more and mre dense network of such paths . When this procescs begin-s, the higher eonduetivi ty, obtained,after all the free a.tom have eompleted aggregation,: would be taken e-n the ba,ds of the .mechanism considered here as an indica.tio.n of' the, degree to which the patches are e<mneeted.. Whether l:'iiUCh a path as is considered above is a condueting path or not would be expe,e ted ·to d.ep.e:rul s~mewha.t on the &Jilerg, available -t o push th.e ,e leetro.n s acress the gaps. le·nee~ when the film is a.gtng~ ·seme path~ ~ , ver:, well beeo.n , f due ·to slowly widening gaps,. non... conducting :te:-.r 20 volt:s app~iedrpo;tential befere thts happens for 40 ,t:. volts applied.! Thi s is the sign of the deviation fo:ad hem Ohm1 ,s law. A e:onduetio-n mechanism :O'f this kind re.quires t;be, erletence of a voltage th,-esh:old.. ~ apparatus de·seri'bed d:oefit no.t,. h.Gwever, pe:rmtt the aee1U'ate determination of conductivity, in t he region &f contact pot-entials that would be neeeasary, to investigate this expected tl:11.•1Mihold. 40. D. Summar, The apparatus of Anderson has been improved and altered so as to allow experiments with lead films. Thin layers of lead were deposited under clean conditions very carefull.y-, achieved; the rate of condensation on glass surfaces was varied, and the surfaces were held at room temperature with circulating water, or below -100°0. with liquid air. The electric conductivity, of these films was measured, and the following correlations between results and experimental conditions obtained: The critical thickness at which the low-conductivity film begins to increase its conduction rapidly., approaching the conductivity of the bulk metal a factor of 1ou away, is found to be abou ':. 60 Angs t r oms for room temperature, and about 30 Angstroms for liquid-air-cooled films, if obtained by, continuous avaporation. If the deposit is interrupted ( i.e. the film isolated from external influences), the conductivity of the metal lay~r falls rapidly, toward zero, for the thinner films becoming undetectable very soon due to its being masked by the conduction of the backing plate of glass. The rate at which this decrease starts was found to be greater for thinner films as a rule, and the total range covered by the decrease was almost alway~ greater for low temperature films. These oonduetivity~time curves gave lower conductivity for 20 volts measuring potential than for 40 volts used to measure the conductivity, for all cases where the two could be distinguished. The observed effects were fitted into the following picture of the conduction process in thin metal films: !lllie time changes observed ill 41. resistance, are due to structure changes in the film, such that its tendency, to draw into small patches removes the mobile atoms from cir.culation, and opens gaps which obstrnet the conduction electrons. The effects of temperat'm"e- and time on this aggregation process have been disC'Ueed, with regard to the stability ot the group•, and the possibilities of ionic and electronic eond:action. The smaller atomic mobility expe.c ted at low temperatures has been shown to give the type of results actually, found. The deficiency_in quantitative reproducibility is ascribed to lack of sufficientl.1'. good control of the pby~1cal surface conditions on tba plate where the film is condensed. !~. Aeknovledgmei+ts: i1he writer has drawn extensively on the ideas and experience of others 1n carrying out these experiments. He has tried to note in detail his dependence o·n the results ot Dr. A.B.C.Anderson, but there has been other valuable help which ca.nnat be M definitely specified. To Dr. Goetz,, who proposed the problem, utmost thrulks are due for his many fruitful suggestions during the work and in supervising the :prepare, tion of this record; the help 0£ Mr. E.L.Armi was most welcome in severa l vital places; the glass parts were skilfully made and kept in repair by Mr. W. Clancy; Mrs. W.J .Rodekohr bas given much help in typing this report. hble l. §!mm1e WEligbj.¥ ot ealibre;tion 1'1lm. Date ot observation Time ot deposit 480 minutes Themocouple reading ll.'70 d1Visions Area of fillll. 1.53 sq. c::iw. sensitivity of' mieroba.lanee 7. 7 1110 Weight or fillll. and cover glass 0.217798 gms . -6 . f§D.S/div . o .21'7804 0.21'1805 o.21'1804 o.21'1815 0.217909 Average Weight of' cover glass ( lead etched off'} 0.21,sosgms. o.217759 ~s. 0.217756 0.217700 0.2177?1 0.217755 o.21'1755 Average 0.217759 grams -5 Weight of film 4. 7 Grams per square centimeter per minute 6.4 "'10- 8 I( 10 grmns. Table 2. Condenser capacitances and time constants Condenser ntnnber Capacitance to electrometer, ;rrd -5"' Co - 2.3 I( 10 c, - 2.16 xlO 0~ - 1. 81 .xlO 03 - 1.55 x lO 04 - Os - 2.16 -4 -3 -2. -I 2.01 X 10 Time constant, seconds 2040 3300 8600 220 4300 9540 Rheostats BLS710 lat8' 1l2 m85!61 518565 m.. 85'9 BU330l :818145 Bl.8052 ». o.Voltmete:ra '.BiBSSS m.8ll4 D. Q,.Mi lli vol tmeter Bl804.53 D. 0.M1Uim:neter BIB0lS BL8&36l.0, Eleetr<neter Bl.9345 (5) Schultze ( 4) liamburgel!' :{ 5,):'.M'.s.ekeG'l.im Phys... Rev. ;!!!, 85, {1:, 24); also!!_, ?08, (1923) {:6) Riede ~tseh . t . .Phys.• ,!!, 1'17, (l.924) (8) van Yem.pt Ztseh.. f. Anorg.. Ohem.. ·-ll4, il5. , (192t>:) . ( 9) ti:t:@btirn Proe,. Camb., Phil.. Sue. !!;t 131, (1933) (10) Ho,lJnes Phys .•R~ • ~ ' 15'1, {.1923.) (11) kll~r ~ •Ann. d. Phys. ~, 481, (19M} (12} FakurO.i. Ho:114.a Anniv. Vol. ( 13) l!"'uku.roi Tokio Sci • Pap. 769, 196, ( 1938) :p ..,80 - (J.5) omstein (16) Rouard Alm. de Phys . ? 361, {l93'1J { l 'l) .B eeching Phil.Mag. _e!, 958, (1936-) (18) Brflcl.t (19) Dembinaka ft..nn . d. Fbys. E!, 233, ( 1936) Ztseh. :r. Phys..,!i, 46~ (1929} (20) Dreiseh and Rfitten {21.) Essig Phys . Rev. !!,, 229y (1939} .(22) Finch Proe. Rey.. S-o.e~ l45, 67G, (J.934) { 23) Hamburger - Ann. d. Phys. 10, '189, (:1931) ( 24) HamitnlrgeT Ann. d. Ph.ya. Ml, ·905, (l.931) (25) Hass Nntu.~ss. §!, 232, (195'1) ( 26) Riedmiller Ztsch. t. Phys • .!Q!, 408, (1936} {27) Riedmiller Ann. d,. Phys. ~ ' 37'/, (1934:) ( 28) Rff.digel.' Ann. d. Fb,ys ., !Q., 505, (1951) ( 29) 'f'a.nmann Ann. d. Phys. !@., 73, ( 1.935) ( 50} Thomson .:Nat.ure ~ ' 912, {1929} (31-) Braunbek Ztseh. f. Phys. _!!, 191 (1930) (32} Vollmer and Est.ermaim Zts,e h. -1:. Phys .. 7, 1, {1921) (33) Lovell Nature 13'1, 4-93, _ {1936.) (34) Lovell Pro.c . Roy .. Soc. 157, 511, (1936) (55) Lovell l?roc. Roy. Soc . .!!?.§., 718, (1937) {36) L&vell Proo. Roy. Soo . 1:§!, 270, (1938) (5'1) Mitchell Phys. Rev. !!_, 2ti0, (19'38) ( 38) Krautkrfmle.r Ann,. d. Phys. 52, 537, {1938) {59) Was Pnysica VI !, 382, (1939) (40) Was Physica YI!, 390, {1939) (41) Bra.unsturt.b. / Alm. d. Phys • .2,, 595, (1931) {42) F$ry Jour. de • Phys. !, 3B, ( 1928) (43) J'oliot Canptes Re.ndues 18G, 1526, {1928) ( 44) Kramer Ann. d. Phys. !!, 3¥/, ( 1934) ( 45) Kramer Naturwiss . 20 1 '192, ( 1932} (46) King Phys. Rev. 12,, 291, (1917) {47) Peruc.e a Ann., d. Phys. !, 252, (1930) (48) Perueca Ztsch. f. Phys. ,2!, ll8, (1935} (49J Goetz Ztseh. t:. Phys. ~ . 118 , {1935} (50) Sto-n.e .Phys . Rev. VI !, {1898 ) (51) Stth'.rmann and Barth Ztseh. f . Te.eh .. Phys . .!§_1 449, (19:35) (52} 8 ~ and Banh Zctaceh. :r. Phys.~, 1 33 , (19Z6} (455) Wright Nature ~ ' 107, (195 '1) (54). Bartlett Phil .. Mag ,. .~ 848, (1928) (55} Ditehburn 'trans •. Far. Soe •. ~' lla.7 , (1935) (M) "!and Zt seh. f . Phys . 104, 48, (1 956) - - {57J Fttchs (~) Jfflrsterllng Amt. d . Phys . -30,· 745, (19317) ( .5 9) SWrum: Phil.Mag• .!§_, 4611, (1914) (60) Goetz Ztsch. f . Phys. ~.t 528, {l957j <Ql) Pringsheim and Pohl V'erh . Dent . Phys. Qe.s . _!!, 506; {1912) (62) Letter Circular # 32, u.s ..Bureau of' standards :{63) Strong Astrop.. J"our . 85t 401 , (1936) ~ · . {M} Riede Alm. d . Phys. 45, 8$1, {1 914) - (65} Mayer ( 66) Houllevigue (6V°) W.a.it - Oamptes Rendues 150, 1251, (1910.) Ph;fs .. Rev. !!, 615., (1922.) '(68}. Kirelmer Z-t-seh.. ~. Pl:i.ys . !§, 5116 , ( 1932,) ( 69) Frenkel Zt seh. f. Phys. !§., 11?, (1924) flfO ) ~Q'WllSend Phil.Mag. 6,5-98, (1905) - '° I ..... - - I Gr111/'h 1 I I EvRfort:1/ion It) ' rc1re of Furnace I Gram• - - - - A r,s,s tf"o ,w,i ... ~ -- - - - · ~ t ~I -- ·- - - r - - - - - - - - - t - -- I ~ () "' 'to. 0 GMph JO S/opes~1!!'Y c-11rw:>s -,Q 0 - - - - - - wo~r ~ool~tl ~ 0 - -- - ~ ""' "' ~ ::. :..:: '-~ .~ ..._ ~ liquid ,,; ,,,. - - + -- - l l 0 ♦ tt: 0 Q) ..... - -t- -- --+--+-----,.-+- - --l ('A N · • - -- ~ --- ·· - ··- - - Q _ _ _ _ _ _......_ O 60 __,i_ _....i.,_ _. . _ _ . . . , . ,_ _ _ _ _ ___._ _.....,_ _. . __ ,ao Avera,e li /n1 fhit"kne~s ,so 0 I A,,,,7strom.s. {Nofe : circled numi;ers indict:Yfe s~e~d of' 240 300 . _ _ ___ _ 360 Graph 2 - re,,,,per11t41'r-e ~40~C. '::"' I - £/Oi---+ - --+---+--1'-H---+-- ... -+-----¼-t--,f--+--+-+--------=--+--,,4>--+----+f~~ I C) -,C: ·S ~/Ot----t---=---+---F'fr--i'---+---+-,(.__-+---d-.¥----"~----1f-------+--~-~ ~ f,ot--;;:;r-f""-s::;~±-:--:~;:?t~..::-~-r.::-.:;-/-t~~~1L-1-.--t-~--+- ---+----+k------!---I '~ 10 t-_______:_:_k~r-7 ~~~~~~~Icl~:i:•±v~~+--+--t <., -7 C ---- (l /08 _ _....______________._~...,_--'-----"--.1.--......__ __._ ___, 0 I 00 20() 300 400 500 600 Avera1e film +hickness, Anssrroms (Nofe: cit"C/ed nurnhers indtC'afe speed of eleporif) Ang-sfromsperl77imde. ) I ,-----r---.---.------r---r------.---r------,-------,---""T""------,--.., - 1 ~IOt-----+----t--~~+---+---+-----,,---+---+-F-~ +---+-------1- ~ 0 0 CJ -2 . ,oi------+-----1H.----+---+----+-if=----+----+-----,,+--~""°-cc-+---+------+-- ~ ~ " ·3 . ~IOt---+------.hr----+--~c.:.._-+-_~,...,_--,,,...-+-----+-----+----+---+-----1 ":"'\ I •of llOi---+----t-+-----+---;.t---+--r.ff'&--+--;-+f---+----+----+-----,f-----+---+ \# I 0 ~ -s ~/0 ~ _, ...... /0 ....,.~_ :::.. .... ...... " ____,,.....,._-b"'-_~- - + - -- - + - - - -+------+---+------+---1 ., ~ to 1 - - - - - - +- - 4 - - - - - - - 4 - - - + - - ~ - l - - - - - - - - - - + - - - - 1 - - - - - - - 4 -1 ~ () Ti>mpert11fu,,.#' ~ 40 oC - lJ JO ..___....__ __.__ ___.,_ _...__ 0 100 200 _ . _ _ _ ~ - - ' - - - - . J . . . - - - - . 1 . . ._ 'JOQ Avera9e ftlm fluck11ess, An1stroms 400 __,1._ _...______,1 500 (500 I ! Grtt;,h 4a I !I Teimp. ~ 40C. Speed= 2.5 I ; A "!J s frr,m ~/ rn:.., "le I\ 1A.~rr. h ~ :B ~ ; ; , ✓: , ,..(L u ir .B i .l3 \t 'B II . . / :B j c .v A' h i: ,. B 300 4()() -4 1 0 . - - - - , - - - - - , - - - - , - - - - - - , - - - - - - - , , - - ~ - - - -- ~ (Nofe: num/;;ers in rec fanr les ylve for averaJe fhiekn ess af f i/;,n fhe curve concerned) - - - --- .- -- -+------+- - - + - - - - - -- - +-- - - G,,-aph 5b Asi"j c"rv~s lo,- 5d --0-0- 40V.fl t:---~- ~-:--:::::-- - - + - - - - - - l - - - - + - - - + - -..::::::::::::,,1--X--X- 20ft'. -::....... I § , o ~ ~ ~ ~-~+---+--__:::::,_~:::,;,___~_j_,_---l• (:) ~ ~ ~-7 ., -~ JO t--- - - +-41---\------ - + -- ~ \t - ------+- - --- - -- - --- - -·--- - -- - - - - t ~ () (.j /0;:;---,..___-~;;::------'-----1-:::-------1-----1---1--_j O 5 ~ ~ 20 Ti,,,e sincfl dep11sil ,fopped, w,inufe~· (Nole: numhers in recfan_rles 11ve avPrafe ll11ekness r;!m tDr fhe carve concerned ) of I - . I I I -· - --·--- ;· // - ; I I 1= - '- - I L_ - ! .....,....--is ~ ---- ----- - ~ v~ ~ " I- - -- r- . -- - --- - - · - - - - ~ - - -- .~ .I 1- , -- I I Graph 6 a ,-- Temp< - 1o(J°c ~ S~ed= 6 AHJ.Jlrom.s/mt"nu le ,,/ ·- - I : 30Q 200 it -- - - . . - - - - - . - - - - - - , - - - - - - , - - - - - ~ - -- , - - -- - r··- . -- - - i ~ - - -- - -- -- - - - - - + -- -----+- Gretph 6 b - Afi"1 t'Ur'/'#S "' ....~ - lor 6t::t ~ 71 rn . -" -0-0- K -~ )( 4011.pd. 20 II. •o-------. ~ ·, ~/0 i ~ - -- - -- - = ---- ==--== ----=-=-t -=:~_,.,.g _ ___""i: _ ~ ~ : d - - - . a . . a·'·-'--- 4-··- ·---- - - ---· 10------------------------o .5 10 IS 20 25 30 35 r,mc 6ince d~posif sloppl'd, ,ni111,1/(}$ (Note: numbers in r-ecfan'lles 9ive averaqe /hicknes~ of Film f'or the curve Col"lsider~d.) 40 I I r _/ 1 j /..' I I ii It V. I :/ r ; ,. / V' !' a .,: . -- Graph 7~ Temp.< >--- -100°c. Speed :: 4.5 >--- Ang-sfmms/Mrin,dt! 50 200 150 100 250 300 fhic-lc11ess, A11ssfroms _, I 0 ~ - - , - - - - - - . - - - - - - - - - . - - - - - + - - - - - , - - - - - , - - - - . . . . . -- 1 ---,- Gra11h 71, Afi"J c"rves \, fo,,. 711 -0-0- 4{)v. ,,.,,...._----t -~ " [ml >< 20v.~.-- '::' I • ~m~--f~~r::::::7F~~==~ ==~d==:;=i==7 ~ 1 0 61 ~'~ ~ ! ~ -~/Ot----- -+--'"'c----1----- 1------ -+--- ---1----------<~---+--- ----1 ~ <., i ~ (.j/0 !BJ 0 0 ~ S 10 I 5' 20 25 30 ,6 n,,,,,e sinc-e deposi-1- sfoJ#JHd, tninuf-t!'s { Nofe: numher.s in recfe:tnyles yive crvera5e for fhe curve considered.) 1-/Jickness of film 40 / ,--~----.-----,----,----,-------.-----...,.._----,------r--~-~ f,0 ,---+--+- -j--:- - ----+----+---J // ~- ~-:1r-~:.· ---+---+--+---~ ; ~o , . _,, _ 4 - -- - + -3 ~-:· V ~/O ._, '0-i 1: * t -+-_.;_,,_____ I _ -_~.!----+.- !1 I I' I C / : / 0 -t: , ~ 10 1 /- ~ -6 ~I , n ~~ I • - - - + - - - + -- r . } L "-'(11 --,.a_lu_e_a_,_ it-f-er- - r -_- ____,........l_ B fu'"LJli. ~de 1 , \1 i ~ , ~ :_'._:= r. --+---+-- Graph Ba ....-----. Temp. < -100 c. Speeds: 'l /L .~ /0 't~ -] :I ·.ui_ . --<>--¢- .J -+----4- 7 A .U/m,n, "ts / 0 ll----------l---,--1...---4--- + - - - - - - + - - - ~ - - - + - - - - I - - - - - - + - - - - - -- -+---- /v1I ~ v,~'-_.....8_ ;' .. - - . - - --- - - ------ - -- }' V 0 -- A-U./min_ '---- } _.__......1_ _~ _......._ _ _ _ _....._ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __ 50 100 150 200 25'0 300 Ave,,-a1e fhic-knes-$ , A111.rfrom.s ~ Graph 8 b C Aji"'f C-llrve for 811 g <.,;. -~ ~0 1-...:,,....__--1- - - - - - - l - - - -- ~ - - + - - - - -- - + - - - - - + - -----1 ~ ....,,._. ------=~ - - - + - - - - - - t - - - - - t - - - - - - - - r -- .....~ --------------4 40 V. f'.4. - - - - t - - 7 - x - x -20v. -t--+- r"-~__,,_ ~ ~ ·t ,,l--------~• t <) '--.: (Nole: numbers in rt?cfanfle.r g-ive overa.fe fh1'ck11eu of' /'tlm for fhe curtl'e co11I1a'ered.) I I I I I i I I l I I I I ! ,A: 11 !J /i ·-- ··- I: /~l .. . ~f -·! '' _J__ ; :· ; 'I p i , 1 .. ~ :, -~ - I I. l lVah ~~ at1 Jer --t I I l ,; l/3' •r.s. a1 20c; / ' I I' . - · ···- - · -I-- Graph 9a I- n.mp. <-IOOC. Speed :::: 2 ~ ---~t . ;. : ~o I 1------.-......--~ -· - • 0 b . V - - - Anss-froms-/mint-< te i. :. j i' i 25() 300 Gmph9b Ajinj cur11e for 9a \ -o--o- 40 v. pt/. " )( 2()v.pd. - · 30 40 \ \ \ \ \ \ - - + - - - ~ -\-- B:11 \ 25 Tim~ since dt>posil STOJ'f'NI I Minuh~$ 10 15 20 (Nofe: numbers in rectanyles qive gveraye fhiel<ness for /'he curve considered.) of' r,lm Plate 1 A. Photograph of apparatus Plate l B. / Quartz Confacf hooks / P,yrex ~ I I --Film backing .I -- Stop '5hc.tfkr , -~~-1 I I I I l z Furnace and shields / Tube from gau9e ' ~ 'k,nizalion 9au9e ...,.._,Soft 9l11ss ~~.., Large bore stopcock // --- - - - - --'V \ To pumps I I Liruid air frap •I ~ I AP~A~ATLis] Plate 2. Shu tier---~, __ ~~ _ . .. - "t.,... . - - / ------z--- Shield - -- --. Furnace - -.Bands for conlacls / ' Shield / _,,.-i,_/ Thermocouple leads ,.. - -:- ------ --·--t -···--·--·· = - ~ ..,./ To ) This lead groundtd gau.je I DETAILS OF - FURNACE Plate 3. &' / Qut1rlz To elecfromeler from /J')lie11fio :•::" ..... • ....., .-.. Coolin5 fluid inside Brass I fi/rn ha~/<1119_ · (MitrOSt'O~ <"Over 9/d~$) ---=-=--~--.- Platinized areas Stop Au.xili11131 shullerc ~ --Iron ( verJ sofI) 1 I I ~ ' ./ , . r - : a - _ - - ' - -·- · - i ,' ( \ 1. / I ,: ,-- - I : I. -t --·-- ·: ' ' .5ize of hole in Stop i i ; : ·'l · . A _,' . 1 ' ~/ ~ ; ; • I I I L. ____ I I I - - - ···· I DETAILS OF FILM SUPPORT Plate 4. Bmss pl11fe- Size of ftlm as limiled by final slop CONTACTS Mefhod of ohft:tinln9 reliable thermal t:tnd electrical cont4cfs. Piafinized areas \ Nofe: solder is me/led ·in /he v11cuum /Jy &11~ful applic11lion of fl,e induction furnace. . Plate 5. ~ --· A: METHOD OF" APPLYP ING PLATINUM CONTACTS """'"'-., /'Apiezon W CoHon ✓-- 8; METHOD OF' SEALING fN WIRES lucife J, Bras, Threaded fo fif ., :1u.ppor f _,,/1../' CONSTRUCTION OF Co I I I I I Plate 6. ~ Ct,/ihrafion conde11.1er I C3~_ C4 ====:I-~ ( fnfire opparerlus shielded ) Cs Plate 7. 1111111 I/1 ~ rl11 II I 111 45 v. 15v. Calihrahon condense-r [lecfromder 45 v. JJ Plate 8 A. Edge ot film 130 diameters Ultremicroscope illumination Cardioid condenser Achromatic objective (Zeiss) Photo eyepiece B. Center of film 850 diameters Ultramicroscope illumination Cardioid condenser Glycerin immersion objective (Zeiss) Photo eyepiece