The formation and nature of thin films is considered from the point of view of molecular kinetics of film and substrate o General briteria for control of some.properties of films are discussedo this document is issued for internal distribution and use only by and for Lincoln Laboratory personnel. It may not be reproduced in whole or in part without permission in writing from Lincoln Laboratory.
Descrição original:
Título original
6M-4809 Parameters of the Thin Film Deposition Process Dec56
The formation and nature of thin films is considered from the point of view of molecular kinetics of film and substrate o General briteria for control of some.properties of films are discussedo this document is issued for internal distribution and use only by and for Lincoln Laboratory personnel. It may not be reproduced in whole or in part without permission in writing from Lincoln Laboratory.
Direitos autorais:
Attribution Non-Commercial (BY-NC)
Formatos disponíveis
Baixe no formato PDF, TXT ou leia online no Scribd
The formation and nature of thin films is considered from the point of view of molecular kinetics of film and substrate o General briteria for control of some.properties of films are discussedo this document is issued for internal distribution and use only by and for Lincoln Laboratory personnel. It may not be reproduced in whole or in part without permission in writing from Lincoln Laboratory.
Direitos autorais:
Attribution Non-Commercial (BY-NC)
Formatos disponíveis
Baixe no formato PDF, TXT ou leia online no Scribd
Massachusetts Institute of Technology Lexington 73, Massachusetts SUBJECT: PARAMErERS OF THE THIN FnM DEPOSITION PROCESS To: Professor EwanWo Fletcher From: Peter Fowler
Page 1 of 37 The formation and nature of thin films is considered from the point of view of molecular kinetics of film and 5ubstrate o General briteria for control of some.properties of films are discussedo Distribution List& To receive complete memorandumg J 0 Goodenough Lo Harri:s" Room Do Smith p. Fowler remainder of Group 63 Staff to J e Raffel Haq,Room 6=031 recei va abstract <?n1y Guditz Ko Shoulders Eo Fletcher 10 Loeb Wo d tR\;ghl'el!@n This docUD1ent is issued for internal distribution and use only by and for Lin- coln Laboratory personnel. It should not be given or shown to any other in- dividualsor groups without express authorization. It may not be reproduced in whole or in part without permission in writing from Lincoln Laboratory. The research reported in this document was supported jointly by the Department of the Army. the Depart- mentor the Navy. and the Department of the Air Force under Air Force Contract No. AF 19(IZZ)-458. 6M=4809 Page 2 of37 The memorandum has been divided into the following sectiono TOPIC
Section A.o Section Section C.o Section Do Section E.o Section Section G.o Section Section 1.0 Section Jo Forces of Adsorption Molecular Adsorption Kinetics Lateral Interaction and Molecular, Mobility Molecular Velocity Rate of Evaporation Film Thickness Lattice Fit Electron Diffraction Cleaning of Substrate Using Particle Beams Condensation Coefficients .c It!hemically Bonded Films MatJi:.Bmblh:tog}."aphY i . c, .. Supplementar.y Bibliogra phy:I
25" 27 333 " f' 34 37 6M-4809 Page :3 ot 37 INTRODUCTION A problem exists as to what materials, structure, geometr.y, are reqtti..red for the optimum thin film componento It is known that various solids with particular crystal orientations will have to be firmly de= posited on different substrateso Reliable means for control ot any comb- ination of desired properties must be developedo EXperimentation to the present bas shown that the difficulties attendant in just getting a vapor to deposit and remain on a surface are numerous 0 The first and foremost problem, then, in connection with attraction and the macroscopic phenomenon of adsorptiono Since Langmuir published his Nobel Prize winning theory13J ot monomeleeular theoretical coverage of the phenomenon of multimolecular Assuming that the deposit has been accomodated by the substrate, the question then arises as to what the structure of the deposit i5 9 and how we control ito A unified theor.y describing the prooess of cr.ystal growth proceeding from multimolecular adsorption has not been published as yet,9 although et alia, have given some rate equations volving few unknown parameters,o This paper discusses observations made upon deposition of vapors of normal solids in relation to the forces involved and the molecular kinetics of the oCri teria for the choioe of parameters which effect various properties are implied to some extent o Section Ao FORCES OF "ADSORPTION The forces entering into physioa1 adsorption are 1 ,2,14,24risted belowo The dependence of energy of interaction on separation, of two particles is given in each case as r:no In general when one molecule is considered as interactingwitliasnrfaee composed of molecules of the other type in question the dependence o interaction energy varies as r 10 (London)3$)129 fluctuating multipoles .... Always attractive and orients molecules with greatest numbers of nearest neighborso* * This force is the reason for different heats of adsorption on different crystal faces (varying surface density of molecules)o Germer 31 endorses this viewlJ but i,n 1,9 PPo Rhodin states objections 0 6M.,,4809 Page 4 of 37 a o Dipole - dipole" proportional to r=6(by far the most important term at film equilibrium distances except for spme polar substrateso) . . . -8 Dipole - proportional to ro c!' Quadrupole - quadrupole J' p;roporti onal to r -10 0 , 20 Repulsive (Exchange or Valenee)9 proportional to.e=r/r o , or r=n, with n taken equal to 9 to 100 for various caseso 29 3.0 Vibration of solidlatt-:i.6eo Vibra tion of adsorba te 0 . 50 Internuclear molecular vibrationo In addition to these, polar substrates have forces acting which cause molecules to orient in a especiall:r ;tn the case of p03f'at- ornie molecules 0 These are predominant in some cases 4) . 60 Orientation mutual attraction between permanent di- poles;; proportional. to r =60 CD Attractive or repulsiye depending on orient=- ation whep adsorbedo . 128 5 70 Ipduction (Debye)' <=> Always attractiveo a o M.utual attraction permanent dipole and induced d;i.pole$' proportional tor-6o b o Mutual between permanent quadrupole and dJ1ced tor -80 Co Mutual between permanent quadrupole and in= . -10 duced quadrupole* proportional to r 0 In the case of ionic s'ubstrates, the surface energy for an bate has been calculated. as. a;function of distance from the adsorbent, in= eluding the effects of lJ 7 above, using a summation process up to a distance r = f =1(3, where f is the number of lattice points per um t volume and; contlnuopscenergy per un;i.t volumeo Other cases have been' calculated also, especially by deBoero 29 The minimum'll the curve obtained gives the energy required to take the adsorbate from this equilibrium. position to ipfini ty with tbe,ej- ception of the effects of Vbration (3.94,9 and 5 above)o The physical sur- face is thus described by a 3 dimensional potential surface 70 ,913 0 which gives us the heat of adsorption at every point for a molecule at 6M=48 09 Page of 37 its equilibrium position, being a periodic function for a perfect qr.r- stalo This is ,.the type of model used in the study of api 0 tie ef"" fects of surface heterogeni ties (cracks, dislocations!J etc 0.) has bEfen investigate9 by deBoer and Custers 24 and polyani 68 0 The find thal the variation of energy of adsorption for topographies varying from ne,dles 28 \ to needle-ltke carl ties is of one order of magnitude and thus c0'tld ..:.. prove to be a great effect in adsorptiono * Section B<t . MOLECULAR ADSORPTION KINETICS We have then that when a molecule strikes the a solid 9 (the height of the surface dependent on the energy of the ecule)!J it Jlay transfer some of its energy to the solid. lattice, a thus be left wi th than enough energy' to escape the A would thus remain 1n this site if not energizedo This is the $,ituation we hgve at low temperatureso The well that the molecule may be in thQugh pos= sibly not the deepest one that the surface offers and th, bi:nd= ing of the molecule is not as strong:as might be is we get ecule to 1ll.lgrate over the surfac80 ** The process that:Qceurs to allOw such lateral motion of the molecules is the quantized en,rgy transfer "", the solid lattice vibrationso . These vibrations are on tae temperature and thus we see how temperature enters as a inlhe sticking of'filmso We may also see that this energy in op};)osition to the Van well and thus the heat of adsorptiqnwil1 be the depth of the well as we go to higher temperatures but according * The effect of electrostatic attraction and repulsiQ:n ions and ionic substrates, tending to localize action over ion of latt1pe, has been neglected in the foregoing 0 This is of some importance in a ;f'ew . , ' ** A cluster of molecules that may by chance aggregate at some spot' will undoubtedly become unstableand:form a crystal structure in an at';' tempt to get in a lower energy stateo The absence of migration however 9 will not allow any sort of eCntinnity of the film (ioeo, any long distance ordering) but rather may have a layer of random crystals 0 The acti ve sites thus presented will be of great adsorptive ability but we must balance this gain against our requirement, in some cases, for unifornnty of structureo 6M=4809 Page 6 of 37 to some lawo Lateral motion due to incomplete transference of the kin- etic energy of the arriving molecule is of importance a1s0 69 0 Asswning we know the potential variation of the surface due to multipole effects and repulsive forces, we are thus led to consider the process or surface diffusion which is required to si tua te molecules in preferred siteso The motivating forces is as noted above = solid lat= tice vibrationso Volmer and others B ,9,71,9l3l showed conclusively that under certain circumstances, surface diffusion does take placeo A famous experiment due to Volmer, is tne formation of crystals of Hg from vaporo The crystals grew out of the substrate at a rate more than 1000 times that calculated to be possible the basis of direct addition of molecules of the vapor to the crystalso Since we may assume that during a suf=> ficiently long period of time during _. the experiment, the atomic beam deposits atoms uniformly over the surface then there is little question as to the necessity of surface migration for the occurrence of this efQ fecto Yang'O that When some area of nuclei is presented 9 direct impingement of atoms becomes an part of the nucleation processo The quantum mechanics of the migrational process were investigated for the case of gases by Lennard=Jones and assuming the Langmuirian model of adsorptiono They treat the vibration of the solid lattice as the perturbation (derived-from expansion of a Morse potential) causing a surge of energy to and from the solido The surges cause trans= itions of the atoms to various energy stateso Calculations made place the time that a molecule will remain in such a on the order of 10= 12 seoo Let us assume that an adsorbed atom (Becker coined nadatomQ21922) held by some potential well is given a quantum of solid vibrational The atom may then be excited to a higher bound level in the well or into the continuum where it is either captured by another well or leaves the surface completeiYo (Lennard=Jones 56 points out that a mobile atom may also lose energy by collision with another mobile atom or bound atomo If the energy of activation for migration is small compared with k'r9 then collisions will be rare and it the concentration is not too large so that migration is not prevented by other adatoms then theory that the . 1 mean life of activation will be nearly independent of terperatureo) 6M=4809 Page r of 37 We may assume that the momentum transfered to the adsorbate i$ dependent on the type of substrate and cr.ystal face presento since the fact that whether migration or evaporation takes place is dependent on the direction and magnitude of then we should have certain substrates causing more migration, or evaporation, than others at the same temperatureo The macroscopic parameters here are specific heat and total energy 0 In addition, there w:tllbe a temperature dependence (of exci tat"", ion) of both the number of quanta transferred per second and the energy of these quanta 0 There is a :ti-ni te probability of 2 or 3 quanta being C transferred simultaneously, but it is found to be negligible in our ca$eo A possible mechanism for evapdrationis the second order process of an ad= ditional quanta being transferred to" the molecule while it is an excited stateo It is onthis latter basis that Lennard-Jones calculated the mean =12 life of the adatom in an axci ted state to be approximately 10 sec o 'He further calculated the time that a gas atom is expected to be in an bed state and found it to be on the order of 10=5 seconds o This result is in agreement will the mean life1::"of an ada tom found experimentally on the basis of Langmuirus theoryo This carries the implication that the turbation used in Lennard=Jones theor,y is by far the most important effect in the range of temperature arid pressure used for confirming experiments o The possibility that substrates rich in conduction electrons c could cause excitations by means of the electron motion was investigated by Lennard=Jones o In metals we have approximately 10 15 co10/seco of sur""" face atoms wi th electrons so there is ample reservoir of energy for tran= sitionBo The calculation of mean time of transition is again approximately 10=12 and the electrons are 1% efficient in this transfer 0 This _ then is not as large an effect as the former effects discussed but it is interesting that it will be independent of temperature and thus may be a dominant factor for transitions at low temperature o 6M=4809 Page 8 of 31 Section Co LATERAL INTERACTION AND MOLECULAli MOBILITY , The foregoing theor,y was developed for the explanation of gaseous adsorption 0 It is obvious that we must make some changes to apply these results to the adsorption of molecules of substanc,s that are normally solido (eogo 9 metals)o A.. gas is clBracterized by "light attraction at large separation and by large repulsion at small S3paration9 whereas normal solids have quite strong attraction at small s eparationo We cone> cern ourselves with lateral interactions 0 (Lennard"", Jones has given a good qualita-tive description of molecular cohesion 23 )o gaseous adsorption9 then,' lateral interaction is negligible and$ in fact, theories of monolayer coverage by adsorbed gases postulate that it is However, the process of aggregation of from vapors of normal solids will make a continually decreasing portion of the surface available for true Langmuirian adsorptiono In addition to the variation of the number of substrate adsorption sites we see that there will be attraction (or possible repulsion) due to the presence of the adsorbate itselfo This situation then is not as simple to treat as gaseous adsorption9 for instead of th-amolecples playing a more or less part in the adsorption9 now each molecule is a source of energy, 10e 09 tre molecule being adsorbed takes over the role of the subs.}rat&J but with (in general) different adsorptive abill tyo ":1 This adsorbate=adsorbate interaction will lead to the formation of crystals while the crystal growth will be modified by interaction (termed lattice fit)- and enhanced (or not) by activated migrationo The lateral interaction 9 if negatives just com= pensate for the positive vibrational energy thus 'giving the depth of the I I multipole well (page 5) its physical meaning ot. heat of That lateral interaction is important is apparent from the marked rise of heat of adsorption and peak at monolayer coverage lateral interaction is ex= pected to be pronounced when each molecule hitting the surface must find itself confronted wi th several nearest neighbors26 s 27 o The general process of migration and evaporation of adatoms due to Lennard<=Jones as described in Section B is undoubtedly that occuring with the adsorption of vaporized solids 0 We may a conclusion from the first 6Mc::a4809 . Page 9 of 37 second order processes of energy transfer noted in that section, namely 8 there should be ? discreteness in the phenomena of adsorbed films in the following mannero Assume an atom'to be captured by a well of depth Wand further imagine the transfer to the, atom of a quanta of energy E corres.- ponding to the existing temperature e If E > W the goes to the contin- uumo If E <. w the atom goes to an excited bound stateo In this case, the second order process is required to excite to t he continuum and this will occur if 2E >W& Now, if 2E ,W but 3E / W (it is unknown whether the third order process process is required for excitation to the continuumo Now on the assumption that migration is required for orientation of an adsorbed film,9 we see that as W DE n will have critical points for orientation phen- omena 0 Since E f(T) there should be critical points for these phenomena dependent on temperatureo There are experimental effects that may be explained on this basis 9 for example, RhodinRsxoorientation of aluminum films whereby the atoms quire sufficient mobility (energy E above) to aggregate into cr,ystals o 10 That is, as Rhodin states,9 many deposited films are unstable in the sense that a cluster of adatoms will aggregate in to cr.ystal(s) provided that the atoms have sufficient mobilitYo'Brunauer (refo 2,9 Po472) states that kT must be raised So that it is as great as the energy difference between the minimum and maximum potential Further 9 Rhodinu s investigation of Aluminum deposition showed that at lower temperatures, the (110) Aloplane was preferentially deposited parallel to the mica backingo However, as the temperature of the substrate is there comes a point where the (111) plane deposited parallel to the backingo It appears that fit orientation must be modified somewhat 0 On the lattice fit basis 9 the (110) plane should always be found parallel to the substratso We see than that a "critical u temperature ex= iats above which the effects of the lattice fit are overcome by some other force{s}0 The above effect brings to mind the theory of Dixit 30 0 He gives a classical theory for such orientatrhan based on the Van der Waallls equa= tion of state for a 2 dimensional gas (n1 :: RT) and the Ramsey"",Shields'J equation (which gives surface tension as a function of temperature)o 6M=4809 Page lOot 37 This argument shows that a given temperature corresponds to a area .;:per adsorbed atom and thus. to crystal planes of varying. densi tyo Assuming an amorphous substrate J Dint was able to predict "6ri tical re- orientation for the deposition of varioue".p,!anes parallel to the 'substrateo The surface-.can be represented in our potential
surface picture as a planeo Why notJ then assume that our crystalline' :. structure approaches a fiat potential surface with increasing temperature? The model that must be assumed contains no new forces but rather only the effect that vibrational pushes the potential wells up until we have effectively a plane surfaceo EXperiments like. those of Dixit.on several substrates could prove the correctness of this 0 Rhodin states that the '. maximum orientation and rate at wpich orientation increases with temper= ature is typical of the substrateo This is what we would expect, since the potential surfaces of various substrates at the same temperature are cer= taiIlly different and the effect of heating of these different substances should not give the. same rate of increase of vibrational energy (flatten- ing of the surface)o Dixit rr S. shows that as temperatpre increases, each atom requires more areai ioeo,9 less dense of the cry.stal are deposited para- llel to :the backingo If we assume,Sl as is logical, that at the temperatures involved, formation of a will proceed (migration is talti.ng place and&crystal is a lower energy state than an amorphous cluster)s then the mechanism that seems to determine is, that the surface is pushing as many atoms away from it as possibleo That is, would appear to be repulsing more adatoms as the temperature is increased and repulSing in such a manner as :to allow only a low density of atoms to re= main parallel to t he substrate 0 When enough atoms to form a unit cell aggregate these atoms will form into their crystal state with a plane par- allel to the substr?te corresponding to the density of atoms by the above mechanismo Carry;i.ng this to its logical conclusion, we should have all molecules repulsedo The evidence for this is in the phen= omenon of evaporation at sufficiently high temperatureso has that for highe;r substrate temperatures more oriented films are producedo We therefore, that since the of temperatJ[re is to increase mobility, then with more mobility the atoms 6M-4809 Page 11 of 37 are able to reach more stable positionso This is in with the earlie;r conclusion that for more mobile adatomic motiol'l on a crystal subCl:t strate,9 the temperature must be such as. to give the adatoxns mo- bility to get out of the potential well& The smaller.orle:h.ting forces (less stable orientations). should require lOW'F for such orientations to take place" This is a minimum 1fr is required for a particular orientation, K9fdg-h6 showed. that germari:ium films appa .. , continuous when deposited a substrate held at room tempera:ture, but granular when t:qe suljstrate is heated during dep"" ositiono Sennett and. Scott47 57 found from ,lectron microscopy that silver deposited at 70 to 300 0 C atomic aggregation than when deposited at room temperatureo !feating of (2 Sec"'" onds) films produced the same Bannon and CurneyS3 found that CaF 2 , crystals grew ,from 170 i to 2.50 i as the was increased from 100 0 to 30QO C, necessitating atomic mobilityo They noted different orientations appearing as the temperature. was. increased just as did Kupo and Miyake.54 in their investigation of Ca, and Ni on PbS" Levinstein 32 investigated the deposition of 3.5 metals(. Results show .... ed decreased orientation for substances of increasingly higher melting point 0 This is in agreement with Rhodin for if' the substrate (amorphous) temperature kept constant and substances of varying melting points are the metals of lower melting point will have more mobility than those o:f the h;tgher melting point,9 (the closer to, the melting point'..;. the more mobility)o His results predipt that a greater observed orientation should occur for. more mobile atomBo If we were to raise the temperature of the we expect the metals of higher melting point to become moreorientedo Experimentation for proof of this effect was carried out by LeVinsteino When we speak of mo'DilitY.9 it must be remembered that the motion will not of a perfect fluid, but rather more like that of a viscous fluid j ioeo, mobility is characterized by short. distance motions <> That mobility is limited is pointed. out by Schultz,2o He shows that for an atomic beam obliquely incident, crystals grow inclined somewhat toward the axis of the beam thus demonstrating that the dts.tance of mobile Page 12 of 37 motion is of the order of magnitude of the size of crystallites grown by him 0 Schultz also argues that aince all crystallites at aQY time during deposition ane of the same approximate height (even those from projections of the substrate) there/ore mobility is else one might expect some. crys.talli tes to be towering over others 0 The la t ter argument, however, is not sufficient to validate limited mobility since if one allowed completely free motion of all molecules one would expect something like a distribution of crystal heights which may be the distribution Cockcroft 37 showed experimentally the idea of sbme limited mobility by placing a wire near the substrate and noting that a aggregates of the condensed vapor appear in the region of the wirens shadowo In addition to the former effect, we would. also expect the metals; of higher melting point, if given enough mobility,to form more.strofiglY oriented films since a high melting point implies a stronger bond= ing or cohesive force" Uo Semenoff 33 has showthe relationship of the energy of cohesion, U" (eftectively the latent heat and the energy of adhesiong fI to the problems of deposi tioD. ot filmso He :Sl1ows the critical pressure to be a fll11ction ottemperatureg p := ! exp( =A/I!.f) where .A =. U .. 6 and ! is a constant which depends on the geometry of the system,9 . tpe . Size of condensing and a linear temperature correctiono The quantity A is said by Rhodin to be the Van der Waal 8 s energy between adatom.and substrate. and experi- mentally found by him for aluminum films o We see then, of the attractive forces acting on an a datom increase, then the of impinging atoms need not be as, great for condensation to occuro Rhodin also foundg as we could expect9 tpat orientation is greater for pairs of materials with greater A() Thus, in general, we might expect a ) more oriented crystalline film to be fonned from of sub ea stances of higher melting point, dependent, of course, on the mobility imparted by substrate vibrationo 6M",,4809 Page 13 of 37 Section De MOLIDULAR VELOCI'lI The effect of yarying velocity of atoms being deposited was inves- tigated by Levinstein p Using a velocity selector apparatus, he was able to find no difference in the size or structure of aggregates of the de= posited meta1s e . In the case of Antimony, a variation of size of aggre= gates was found,9 but further investigation showed that this was due to the presence of some polyatomic molecules of Antimony which naturally would a different crystal size or structure than the monatomic molecules 0 Bateson and BaChemeyer 51 attempted to on the fact that slow atoms were used However, it would seem that the velocity-would be an indirect effect the true factor being the mobility of the atoms was so low that at concentrations (pressure) \lsed by them, well oriented aggregates could not be The randomly or= iented crystal atructure resulting "WO\ll.d naturally provide for or ... i . ientation in. later layers than would a well crystallized formationo Beeck$ and Smith?4 investigated in gaseous (0005 = 10 Memo Hgo ). atmosphereo* The collisions of beam atoms in tran- sit form clouds and clusters of slow atoins before reaching the substratee The randomly oriented wh;ich formed in the:tr experiments when higher gas pressures (2=10 memo I:ig& l were used c01l1d be explained by the low mobility argument of the preeeeding paragraphlO ?ddi:liion, it is notised that the presence of a gas of pressure 05-2 moM" Hgo induces the deposition of planes of low density parallel to the s,ubstrate implying that this is the. lower energy configuration rather than the planes de- posited in vacuoo that. there is no apparent reason why this should take placeo Beeck s et alia" suggest that the presence of adsorbed gas would the vapor atoms into corresponding to planes of least densityo 'Theygo onto say that the low density plane. phenomena occurs under conditions for which a complete surface layer is highly improbable and thus the effect of gases on orientation tnust be . . . \ * See also references 55,86,.88-90,162-4 for other experiments on de- position in gaseous atmosphereo 6M-4809 Page 14 of 31 purelY'" kinetic fl. ei through dissipating the condensation potential energy ot the surface, by effecting the distribution of the metal atoms. they reach the substrateo The answer to this problem can be partly answered when we know the .,meohanism that orient.,., ation of vapor deposits in vacuoo A good start would be constructive or destructive proof of Dixitns theor,yo gave a theory which a dependence of sublimation and condensation on the, velocity of atoms;" However, there appears to be an incorrectassumptiono He postulated a potential barrier" for sublim- ation ( transition to continuUlll) much as was shown at 'the beginning of thi s paper but then goes on to postulate ? barrier for condensation without giving the physics of such a Since the forces an atom is sub- jected to on approach to the surface are attractive and not until adsorb= ed doe,S it meet the repulsive effects of vibratio:m.,9 and other forces, this assumption would seem fals80 The atom which is adsorb- ed (for aperiod,as given by Lennard-Jones theory) comes to equilibrlum with the, surfa.ce, and in this state its velocity of approach is completely obscured 0 Th,e deciding factor of the effect of velocity is evidence and Levenst'ein IT s conclusion is that veloei ty is of no effecto Section Eo RATE OF EVAPORATION The fact that experimentali,sts may have concluded that velocity is a factor, is probably because veloeityand rate of evaporation are independent 0 The effect of one could be for that of the othero Levinstein!s investigation and qualitative theor,r shown the effect of rate of evaporationo That a "critical" pressure exists,below which condensation does not take place has been shown by severa1 35 ""'38 o It is the pressure at which point the rate of evaporation of adatoms is greater than their rate of arrival thus being due to the that substrate-adsorbate force is lowo It is found in many cases that i t presence of adsorbed impurities partially satisfy and thus lower the substrates attract- ive potential o Frazer 16 .5 says that if a surface is completely the va .... pors of most normal solids should experience no -critical point" even 6M=4809 Page 15 of 37 with the weakest of substrate attractive forcese> Many theories have been given covering "critical pressure" with referenpe to its, tem- perature dependence including those of Frepkel 35 , FOJorler 41 ,9Semenoft;-126 Peier1s 42 fi Devons hi re 43 :j and reference variatipn witb atu.re is quite amenable to experiment (eogo., references 8,936-38,9104-106)0 All such theories depend in essence on the presence and magnitude* of the forces of physical 'adsprption discussed earlier although various methods of treatment are usedo A high rate of evaporation however should be expected to have the same effect as a dense atomic beam of slow atoms, for both processes have the effect of putting atoms on the surface in . great enough quantities for them to atgregate during their meart life Levinstein 9 s work shows that slow evaporation produces (for Antimony)'large patches which have diffuse diffraction patterns whereas rapid gives ISmaller patches with sharp diftr..acti:en patterns (crystalline form- ation)0 He interprets. the diffuse- pattern as showing the Antimony to be in the amorphous phase" Levinsteinfurther<hblds that we caa.not expect \1 this to occur in general since the of many sub$tances is not found probably because crystallization takes place sofasto found that slow evaporation gives diffuse electron. t diffract1Dn rings whereas fast evaporation gives sharp patternso ion in gaseous atmosphere produced the same resultso The rate ofevapor= ation then effects the structure of the deposit, by effecting the manner in which the molecular aggregates form and grow o Sennett and Scot%7 9 57 showed that rapidly evaporated (2 seconds) films tend to remain thin and cover the substrate more continuously as compared to slow evaporated (2Q minutes) films which grow thick in aggregates before They concluded from their experimepts that normal migration time is intep. upted by the arrival of more The ensuing collisions are the means for nuclei formationo In Lennard-Jones 56 predicted a shorter mi= gration time for gaseous adsorption than is given by his earlier quantum mechanicairesults o His conclusion, was based on the fact that mobile at= oms, when present in great numbers, will collide otten leading to trans"'" i tions to excited states or the continuum (whereas in the case of l;, '* For example:; the statistical mechanical theory shows"'that as ;a greater lateral interaction (reaching to far:bhest neighbors) is assumed, the more definite is the critical temperatureo 6M=4809 Page 16 of 31 adatoms of normal solids, nucleation may take place instead)o Their eon= clusion in terms of experimental parameters is that migration is a func= tion of beam intensitY9 in our earlier conclusion that i' is a function of substrate temperatureo Levinstein explains the preceeding phenomena as follows: The nftm- ber of a toms in motion on the is proportional to the rate of arrival of atoms and these adatoms will move qver the surface until they collide with other atoms and form_ patches (depelldent on subetrat!3 tbrees especially the presence of cracks, implirities, dislocations, Assuming evaporation has been in progress for a while with some patches the atoms arriving suqaequently co11tbde with each to fom new patches or with old patches to inOrease their sizeo With a high rate of the former event place since there is er probability of hitting other adatoms than the relatively few patcheso With a slow rate of arri,al, the latter evetftt$kes place since greater probabilij}y that the mobile atoms'WiIT, strlki":-;e. patch before thEfT leave the surface and thus make fewer but larger patches 0 Carrying this to logical end, there will be a rate, of a rrip'al whereby the mean lli.fe !J \ 'L of an a tom in the adsorbed state,? is shorter than the reciprocal of this ratetand thus condensation practically stop except for some direct additions to the nucleio .Levinstein's experiment was con= fined to Antimony and show.d we use a less -mobile 8nbstahh$, the same ef= - \. - . fect might not-be obse;rvedo By increasing-the temperature, the mobility .could be increased the noted for An= - timony-might occur for these other substances also o Sennett and Scott varied the evaporation rate-of their deposition by a factor pf 2000 and yet ,the thickness of the film only changed by a factor of 100 Evaporation rate then appear to be a valuable para- meter for varying the properties of filmsb 6M-4809 Page 17 of 37 Section FG FiLM THICKNl!SS ' was 48 concluded that thin metal films are amorphous up to a cer- tain thickness somewhere in the range 50 i to 200 i and at this crystallization takes place if temperature conditions are favorableo. His argument is based ong (1) the increase in conductivity of at this point which he says follows from cryst,allization 9 (2) the fact that adhesive forces are than cohesive for- ces in very thin and thus the substrate forces prevent formation of In reference to point (1) W'e may conSider Sennett. and Scott R sin:", vestigation' of the growth of metal films and in, the region of 50=200 io Electron micrographs shpw th?t the phenomenon that is taking place is . that small aggregates (0001=<>01 microns) are joining togethero We con= elude that this is the phenomena that was called crystallizationo Sen= nett and Scott found that adsorption of light at this point-where the aggregates begin to joino The voids that exist between aggregates when the film is thin would be detrimental to and thus as more aggregates join together conduction would increaseo Repeating Was 9 ex- periment and using electron microscopy could clear up the pointo Concerning point (2), agree that the first few atom layer;s of film are strongly influenced by the substrate put contend that after several atoms are deposited in close proximity, the lateral interaction will lead to a ggregation, (granting that there must be sufficient mobility allowed) 0 Yang j et 801i8949950in developing their thermodynamic""'statist= ica1 theory of nucleation from the -vapor pp.ase concluded that a unit cell aggregation is required for crys tal growtho The unit eell is < 10 atoms for many substances and thus we see that only very .. ll aggregates are required before we could expec-.t; the transition of the film to go from amorphous to crystal struetureo Sennett c:tnd Scott did not define the state of the aggregates that were present before the joining process at .50=200.i but we would conclude that they were crystallineo Was u con cO elusion that the films changed from amorphous to crystalline then is thought to be in error although the joining of aggregates may change the 6M=4809 Page 18 of 37 type of crystal structure perhaps leading to long, chains of oriented crystals 0 As a confirmation of the state of very thin films we have the re- sults of Germer and White 161 who show transmission patterns of films. in random crystal struct- ure=strong apparent even at 30 i thickness o Schultz52, investigating kali Halides,'-'was unable to get reflection patterns for films thinner than 100 i9 but transmission patterns showed random crystals with (100) planes parallel to the sUbstrateo (The explanation for (100) orientation is giv- en as the result of cube faces lying on a smooth surfaee o ) Thicker films showed (Ill) orientationo Harris, et investigating thermal sta"", bilization of gold smoke deposits (prepared in inert gas atmosphere) sug- gest three possible structures for the filmsg 10 Each aggregate consists of a number of crystallites surrounded by disordered regionso 20 .Each aggregate consists of a large number of smaller nuclei o 30 Each aggregate consists of a single crystallite having a non- equilibrium latticeo From energy considerations, an excess of vacant sites is the most probable type of lattice o They point out (1), which is the probable structure at very low tem= peratures (/'IV 20 o K) is untenable from the point of view of rate of approach to equilibrium and that diffraction rules out (2) in favor of (3)0 Other structures that one might have to consider might be 10 random crystal structure with non-equilibriym latticeo 20 amorphous throughouto An example of 2 in the literature is Levinsteings work noted on page 15 of this papero Antimoqy was fOUlld to be one of the substances for which the amorphous phase can be formed during slow evaporationo We might take this substance as representative of one end of the scale of thin filmso Next on this scale would come the randomly oriented films I and thence to the well-oriented films and this whole scale verying with film thickness as notedo An important here would be an arg- ument, pointing toward the relation between this scale and a scale of molecular orbital shapes and energies or some other microscopic quality of the deposito 6M=4809 Page 1<9 of 37 Rhodinus 10 investigation of thin aluminum films sh9WS a variation of observed percentage of orientation with film thickness o He graphed thickness against and found maximum orientation to occur for a small thickness, the films becoming less oriented as more atoms are depositedo We 1IlUst keep in mind that temperature is a big factor in the orientation and 9 in theory," proper temperature can produce maximum . orientation even to a single crystal throughout the thickness ofa It is reasonable to believe that the maximum. conductivity point9 maximum adsorption maximum orientation point, and joining of the small aggregates in the. experiments of Sennet and Scott" and Rhodin are all closely related" To compare Was' and Sennett and ScottUs we might vary the evaporation rate and measure conductivity at the same " thickness for fast and slow evaporated films o Sennett and Scott srlft' that the aggregates donatjoin for slower evaporation rates until a greater thickness of deposit is obtained and thus at a given thickness the 90nductivity could be expected to be lower for the slowly evaporated filmo An account of the probable kinetics of very thip deposits during the period in which nuclei form is given has wri ttenin the kinetics of the formation of 11'll01e1o McLauchlin,9 Scott' and Sennett l 5'8 have. watched the formation of nuclei and "aggregates in an electron microscopeo It is possible that as separate patches of cf __ growing" o:ut over the substrate will have boundary when they jo1n.9' un"'" less we provide the atoms with sufficient mobili tyo . tt has been 1" '. that long distance ordering of structure can be promPted very thin . " . (discontinuous) films by means of deposition in an atmosphere of gaso For light vacuum deposition the. thicker films begin to show long dis- tame order At liquid ai;r temperatures:"hN.e probab= ly occur but as is more crystalline structure of very long distance order should appearo A phenomena probably allied to specific heat has been noted by some in that certain substances can not be, deposited beyond a certain thickness em let us say in the niicr9n range, yet other substances can be deposited. to millimeter thicknesso If we imagine a molecule strlking 6H=4809 .Page ot 37 a deposit of some insulator whereby the energy transferred is not \ pated by conduction j then we expect an elastic collision resulting in the repulsion of the moleculeo That a, sink: for the energy is' requirep. ,has been rtumerous tim.es in the MoI-o To film laboratory, iSrotesso;r*o Harris o A good electron conductor is used for a backing for the film and !ood deposits are formed hut if the back- ing is not a good then it is in many cases impossible the vapor to sticko Section Go LATTICE- FIT, The effects mentioned on page g of the of the ited material, with the $upstrate cOIl:cernesi many. (eogo, references -1$1 79 37, 49, 50, 58=67.9 l07 i 116.9 1$5, 156, 159; 169)&, Basically we re quire a t to have a complementary to that of the subs trate 0 Naturally the spacing required "aecomodation is not that of the substrate1 but t. that is of greatest effect in adsorption of the particular adsorbateo As examplefj _strong dispersion force will the molecule with the tnO$*,. neigbborsj an ion will be attracted to a lattice ion (if the strate is ionic)o The best fit of an adsorbate lattice then could be with a spaCing different from the of the substrateo There- fore:; ,in considering lattice fit, we may not say (wi inyestigation ot'the forces acting,in'Jl particular an adsorbate, deposited on, a substrate with identidallattiee' will be easily accomodatede Some have concluded that the maximum difference of lattice spacings that still of a deposited film is This again is dependent on the strength of forces of eogo 9 ",the adhesion of an ion to a metal substrate is very and .y accomtnoq,tlte a pair of materials with much greater than 1,% differenceo An example of the determining f?ctors of lattice fit is seen in the adsor- ption of aluminum on 'Aluminum.,(llO) plane will fit within a. few percent in"the plane of the atoms residing in ions of maximum Van der': ' Wail H s attrac:tion6 It seems likely that if we were to ionize alumin1.lm and proceed with deposition there wo uld be much 6M=4809 Page 21 of 37 less accomodation due to electrostatic attraction of Aluminum to the opposing the position With most number of nearest neighbors, etco Due to the fact that the alvminun lattice is so distorted in such a ion, little if any crystal would be expected at normal (The situation for higher temperatures is discussed ear= lier with reference to theory)o Each case of lattice fit orient= . . ;' ation must be judged on its own basis using as cr;tteda the relative nitude of the various of adsotption, such as have been calculated by Sect:J..on Ho ELECTRON DIFFRACTION There is some dispgreement as to the effect of the electron diffraction process on films o Definite evidence has been given that the crystal (or amorphous) is changed by the electron and other equally evidence shows no change in the films so 6 64 '. treated 7 ,91 $ The variable that must Ibe taken into account when com= paring the results of variQ:ps' experiments are} 10 Pensity of 20 Energy of electronso 30 Adhesive force of deposit to substrateo 40 Cohesive forees of deposit() With regard to (3) we must realize thlt eleotron bombardment of a physically adsorbed film may require very little energy to break th e bonding and allow mobility of the atoms, whereas the energy required tor disrupting a chemically bonded film may be considerable o et alia 162 .9 provide ,an example of (4) in their 'Work on the sintering of gold blacks o Gold being heated stabilize and change optical and electrical characteristicso The energy required diffusion of a lattice atom to the surface of the smoke is found to be about 1/3 of the 2705 kcalo/gmo atom for diffusion in massive goldo We expect that the transfer of electrons should produce the same effects as heating did in this caseo We must make a careful comparison of the parameters noted above when conSidering the results of electron diffraction experiments o For 6M=4809 Page 22 of 37 instance, the experimenter with a high energy, high density im- I pinging on aphysically Adsorbed,thermodynamically unstable film will more than li:kely report structural changes in t he film or at Ie , a change in optical and electrical Wright and WOOds 167 ' investigated the decomposition thin films of various salts and oxides under slow electron bombardment o Decomposition was measured by the chahge in secondary emission:which after bombardment 0 It was foUnd that the energy required for did not agree with the 'heat of formation of the compound as predicted by some o The energy for decompositipn t&nded to ;gree the optical qbsorption energy 0 The mechanism o-f some qases\ras c0I19luded to be the of free and electrons which to the surface where thay may release atoms from the compound thus change the cathodic emission propertieso In this article, the work of ists in this particular fiel.d are discussed 0 Bombardment of Barium 'Oxide films is covered by lmai an' '!f1zushima 171 0 i The use of 'electron diffraction for investigating crystalstruc= ture of films is covered in references 61, 77=83,87, 1240 Section 10 CLEMJING OF StJBSTRATE USING PARTICLE BEAMS The preparation of substrate utilizing electronic or ionic beams is found to be helpful91-93.9166o The 'desired effect of these beams is to drive off chemisorbed material such as gases, whose masking of surface for= ces is well 0 Yang was to have investigated the effects of such forces in 1954 but no data has been published as yeto In the course of a symposium on vacuum evaporated films 97 at least 10 methods for " cleaning substrates were presented and all agreed that the final' step should be cleaning with a beam of charged particles 0 Possi ble mech cmisms for the process of this cleaning may beg (1) Local intense heating by the beam causes a surge of energy ejecting the or (2) Direct satisfaction ,of a chemisorption bondo Experimentation with variable density and particle en- ergy beams could clear up this pointo The use of electron beams is said to be superior to the use of beams of other particleso An ion beam could be detrimental in some cases Page 23 of 37 due to its reactive effecto Protons of sufficient velocity could form By= drogeno energy of is generally low enough so that heating drives it off quite readilyo However, if one must wait for the subc:o strate to cool before proceeding with deposition, the cleansed surface has time to become dontaminated againo Carr 93 describes the effects of treatment of Gold, \ which could be detrimental to vapor depositiono When Mercury vapor is de= posited on such a surface it is adsorbed to all but points where the tron beam had falleno Mro Ko Ro Shoulders gives the interpretation that the beam had reduced various organic substrates pre sent 9 possibly to thus presenting a low adsorptive spot compared to the gold surfaceo A recent book by Holland 166 gives mUch good discussion of the cleaning proce'sses o An important point that he discusses is the requirement for a relatively high pressure to exist in order for glow discharge clean- ing to be used 9 this paradox leads us back to the requirement for particle bombardment as a final step in the cleaning processo* Dro Ao Lo Loeb describes a method for post mortem investigation of cleanliness of the substrate surface o Visible radiation is made incid= ent on the film and the optical cons'tants are from transmission and reflection experimentso The system is then reversed so the light is incident on the backing and the optical constants recalculatedo Disagree= ment of the 2 sets of optical obtained implies the presence of impurities at the interfaceo The theory of reflection and transmission of radiation by thin films and methods for analysis sre"given in the two art= icles by Harrisj) 'Beasley and Loebo168Jl169 Section J o CONDENSATION COEFFICIENTS now bring into the discussion the coefficient describing the ability of a film to be depositedo Knudseri lOl ,102 a Thermal odation Coefficient as a = (8 3 """E:i)!(E 2 wher$ Er is the energ1 ! . * An analogous paradox has been suggested by Shoulders o One must keep I in mind when using 'getters" that their vapor pressure may be greater than the amount of vacuum 6M-4809 Page 24 of 37 the molecule possesses before striking the surface,9 E2 is the energy it I, would have if it were in thermodynamic equilibrium with the surface, and its energy after it leaves the surfaceo MUch data is available giving the accomodation coefficients for gases on solids (eog o9 see ref= erence 99)0 The simplicity of the process of gaseous adsorption compared to that of evaporated solids due to the relative absence of .attractive interaction between gas molecules makes the definition and measurement of accomodation coefficients quite For the chemisorption of is defined, it being chemisorbed (by forces of about 10 gases, a sticki:Qg ,probability (or ; the ratio. ot, thirnumber of molecules , , ' times those of physical adsorption) the number of molecules physically adso:Ubedo Since the molecule is defi':' nitely bound in such a case, rather than hopping around as does a physicr ally adsorbed particle, and since the activation energies for evaporation and chemisorption are amenable to experiment, therefore, this coefficient can be found 0 Becker 100 has given an, expression for its temperature de- pendence which agrees with experimento .Becker and Hartman 21 have qual- itatively interpreted the coefficient to })e the probability that the physically molecules will dissociate into its constituent atotnS9 This interp:retation can be seen to hold only forpolyatomicmolecules 0 For any molecule the qualitative statement should be that S is the prob- , ability that the physically adsorbed molecule form a chemisorption bond with the The monatomoc gases are then accounted foro In the case of polyatomic gases it is necessary to break the molecule into its. constituent atoms before chemisorption takes placeo Therefore, we see . .l from Becker and Hartmans statement that they assume that once the mole"" cule is broken up into atoms the probabili ty for chemisorption of the ate> onus is 10 This is reasonable since when the molecule is present in the form of its separated constituent atoms there are strong unsatisfied bonds for the substrate to satisfyo For the case of deposition of solids,a6ondensation Coefficient is postulated, defined as the ratio of the number of molecules on the substrate to the number of molecules striking the The complexity of the' events occuring and the fact that simplifying assumptions (such as homogenous or periodic surface characteristics) do not closely 6M-4809 Page 25 of 37 enough describe the surfaces investigated, does not allow as yet explicit expressions involing only known parameters 0 Devienne has made many meas .... urements of the coefficients using radioactive tracer studies o Ai short summar" with bibliography is given in reference 1030 Continued experimen- tation along this' line combined with study of crystal nucleation and will lead to the determination reproducibility of results as is now possible with Accomodation Coefficients o " Yang,? ey a1ia 49 .950}l have studied crystal nucleation and sation Coefficients 0 In addition to the rate 'equation for nucleation which , \ they derive,. they give data on varying condensation coeff;tcientso Ref .... erence 155 gives a plot of lattice fit against condensation coefficient.- which turns out to be linear and for 15 to 20% lattice fit the condensation coefficient is almost zero for' their particular experimental conditionso The increase of condensation coefficient for increasing thickness of de= posit is explained as due to the increasing availability of 1t sinks u for crystallization after nuclei have formedo As was noted earlier, another effect present after a rmono1ayer or so coverage is the fact that usually the cohesive force (which is prominent at this stage) is much greater than On page 473 of Reference 155 it is shown that the con- densation oefficient is greater for an increased crucible temperatureo This is quite in line with Levinstein.vs conclusions , given earlier,-where- by at low evaporation rates, the molecules have a greater probability of evaporating'before aggregatingo Yang discusses this effect throughout from the standpoint of supersaturation of the atomic base o Section Ko CHEMICALLY BONDED . FnMS I. ! Throug}:lout this paper the retention of the adsorbate by the sub .... strate by chemical means has been little more than hinted at& This is due to the fact that references to this effect in the literature are scaIfceG In the course of conversation with thin film experimentalists:; it has been noted that examples of probable chemical bonding are sometimes lllQre!.ap.par- ent than the literature implies o This bonding is thought not to be the chemisorptive type whereby the adsorbent and adsorbate atoms retain their identity but rather the formation of a compound at the interface that acts 6M=-4809 Page 26 of,: 37 analogously to an "impedance match" between substrate and adsorbat9o An example of this bonding is Chromium or Aluminum deposition on glass which is believed to form a metal oxide at the interfaceo The adhesion of these films to glass is far greater than that of the metals which would not be expected to form a compound at the interfaceo A simple quantitative measure of film adhesion is the "strip test" whereby an adhesive tape is plac ed on the film and being stripped off the ease of removal of film from substrate is notedo Aluminum and Chromium are effected less by the stripping than are the noble metals o A more precise qualitative measure of film binding is that used by Rhodin 5 C) The critical pressure (lowest pressure required for films to form on substrate) for a temperature T is, as stated in Section 0, p 1: ! 6) IKT We make a plot of lilip versus l/KT 0 s,lope of the curve then gives (U+4), the binding energy of the film (including the effects of impurities adsorbed at-the interface)o It is questionable whether this experiment would show .p the binding energy for all "r any) chemically bonded' films however, since the atoms may be adsorb- ed at first by physical adsorption means and possible after some finite time the chemical The slope of the l/KT would not therefore indicate the chemical bondingo It would be of great assistance in discussing the possible forces acting to bind the film and in predicting film qualities such reflect- ance,9 resistance$ etco to know the 'binding energy for e,ach pair of mat- erials C) An experimenter in the optical properties of films would have an advantage if he knew that possiblY an oxide or other compound existed at the film-substrate interfaceo The distribution of electrons and space charge which exists is intimately connected with the surface boundary con- ditions as is shown by Skinner i7 00 A few binding energies are given by Wexler 112 , having been compiled from the few references that the literature . holds 0 A project which would possible be of great worth would be a more complete of such data and an attempt at a classifying anal- ysis of the aata o signed: -P.duv PF/gk Peter Fowler 6M-4809 MAIN BIBLIOGRAPHY Annals, fioYo AcadoScio Vol 58 Art 6 ppo 144-" (1954)0 2.0 pf 9'ases and Vapors, ppo180-217 PrinoUoPress(1943) 39 London, Z.o P9 C.o,B11,. po 222, (1930) 0 49 .Keesom, f 0 f.o, 2.2 P.o 129 po 643. (1921)0 5$) Debye 9 .Zo.Fo.Po 9 21,. po 178 (1920)0 6.0 Orr 9 To Fo SO.9 35, P.o 1247 ().939)o 7.0 Gomerlc Smith,Strueto. & ;prop.o of Soli9. Surfaces, Uopf ehio po)34{1953) 8 9 Volmer & Estermann, fit r.o, B 7, PPo 1, 13 (1921) 0 90 Volmer. & Adhikari, ro Po, B 35 po 170 (1925)0 10,0 Rhodino Do F 0 So, 5, po 215. (1949).0 Lennard-Jones & Strachan, Po !to .150 ppp (19j5)o. 12.0 Lennard-Jones & Devonshire,PgR,o S.o. PP.o 6-44 (1936)0 13.0 f;:t.. Po lio So, 158, ppo 242-79 (1937)0 14.0 Strachan, Po Ro So, 158, I+ (1937)0. 15.0 Lennard-Jones & Gpocl1dn j . Po R.o S,o 163., po 101. (1937)0 16 9 Jackson & Mott,P.o R.o 8.0, 793, (1932)0 . 17.0 Jackson & Howarth, PoRo .S.Oj 142, po 447 (1933)0 18p Lennard=Jones & N?:ture., 137 po 1069 19,9 Lennard:zoJo+les, & To. Fo.So,? 24, po (1928)0 20ft Jackson, ro 90 ro S.o, :P(t 136 0 21.& Becker.9 To E.o So 55, P.o. 153 (1929}.o . 22.0 Anna1s,9 No Yo. pit.o, po 23.0 P.o P.o Sf> L p .,9 43, fp 461 (1931) 24.0 deBoer. & Zo. Po Co., B25. po 225 (1934)0 25.0 Germer., p.o RpfP 56$. po (1939)0 .' 26.0 Annals 9 No Yo po 271' Becker. & J o Po 5,. po 159 28.0 Annals., No Yo Acado Scio,9 loco citj,., po 7630 29p deBoer, in, Colloid Volo3, po 1, (1950)0 30.0 Dixit,. Po Mo,. 16, (1933)0 31p Germer, Po. R4J 19330. 320 Levinstein, Jo AoPo, 20, po 306, (1949)0 6M=4809 310 Zo P,o Co P.o 411 Beeck, Wheeler, Pp Ro S. 177"po 62 (1940)0 r Frerike1, Z .. Fo ro J .26, 10 117 (1924)0 36.0 Wood, Po Mo,.32, P.o. 364 (1916)0 37!' Cockcroft,Po Ro So, Chariton . 25, po 287 (1924). 39.0 Beeck, R.o M.o p.o, 'J-10 p.o 6'J-, (1945)0 . To, Fo. 29,. po, 794. (1933) . Fowler.$). Po. 80S. 32,. po. 1#4 (1936)0. '" . 42,0 Peier1s, P.()C,O Pp 5.0" 32,. po.4n'r 43.0 Devonshire, Po 'Ro So, 163" p . 132 (1937) 0 440 Barrett, Structure, of Metals, Crysta10graphic Data,p.517 . McGraw Hill (1952) 0 . 45.0 loco cit, po 3150 46.0 fUr PP-1S a,9, 1 J. 7 (1944)0 & Scptt., J p 00 So Ao, uO, po 203 (1950)0 Was o , Physica,. 6, po, 382. 49.0 alia., Zoo PP.o 462, 470 (1954) 0 50.0 Yang, et alia, J o 00 Po, 2,2, po,.12'J-5 (1954)0 . 51,0 Batson. & Nature,,'J-5B,. po 133 (1946) 0 52,0 Schulz, Po R09 75, po 17, 1'0 'J-153 '(1949)0 53!> Bannon & po'J-36 (1948).0 .. , , 54.0 Kubo & ;Io P.G 2, Po. 15(1947). JoPoSoJo 3, po114 {1948)o 550 Beechnung, p.093.8 (1936)0 5qo Lennardo:>Joness> PaPo SoLo.9 490 po 140.(1937) 0 57 Comments on Sennett. and Scott J. O. S 42, po 686 (1952)0 , 5a.&< deBoe;r apd Veepemans" Physica 1" po 953 (1933)0 59.0 BrUck, Ao '"fo po (1936) 0 60,0 Ao do Po 30, po, 50, (1937).0. 61.0 Finch, ap.d To F 0 So, 31, po 105!" (1936)& 62.0 Barretto, loco 01 .0, po 518 p . .Sloat lind. Mellzi .9.Jo Co POi 35.9 po 200, (1935)0 64.0 Schulz-J) A,o 4" pp 483 (19'+)',0 6'.0 A,'o Co 49 p.o497' (1951) 0 . 660 Finch"nd Quarrell,Po Ro So ,141, po 398 (1933)0 . ,Finch, Po,?o $0 49 11:) (1937),0 68p To Fo, So,. 28.9. po 316 (193?)o 69.111 Blois, Science, P,o 175 0 .. 70,0 F 0, So, 28, p. 333-358 (1932). 71.0 Volmer,9 To. F,. So,. 28,. po, 359 (1932)0, Levinstein,9Po. Ro J 71, po 476, (1947)0 Was, 390 (1939)0 74.0 Kramer,9 Zo Foro )..06, po. 675 (1937) 0 .. 755' Krautkramer, Ao d.e Po 32, po. 537. 0 76.0 Sennett and Scotts o.SeA.40 9 P() 206, (1950) 0 Page 29 or 11> 770 Tbeory and Practice of Electron Diffraction, Thompson and Cochrane, . MacMillan and CO o .9.Lopdop.., {1939).0 ... 785' apd Wposte;r, Ao .. Co , 6, po 543 (1953)0 799 A& Co, 5, ppo 626,9634, 782$ ()"952).. . 80p Picard and Duffendack9 JO.AOP09 14, po 291, (1943)0 8I.1t .Matthews and Wilman, I3ri:\io.Jour.o App1oPhys,o 6, p0277 (1955) 111 82.0 Newman. and Pas}:tleY5)Pp M.o 46, PC) 927 [1955.)0 8 3.0 Bryant , Webe;r 0 J" A 0 . P. ?5, .p 0 1343 (1954) 0 84.0 Winke1mann$ Zo 7 J p,o 85,0 Ansbacher. an9 EPrenberg, Po Po .. So, Lo., 64 po 362 (1951)'. 86.0 Nelson, Jo CG Po 5, po 252 .. (1937)0 . . 87.0 Finch, Q.ua rre 1)... and. Wi).man, T G F 0 $G 31,ppo 1049-80 (1935)0 88,0 Reynolds $ Po Ro" 24" ()..924)p 89.0 SnQ.th,. 10c() 9ito, po 451.() 90!t Rhodin" DG Fo 5, po,221 (j.949)o Strong,. in. ExpO' .Phys 0 f1 po 166 .(1948) (I 925' Duslunan, VacoTe9hniques, (1949)0 93.B Carr, Ro SoIo 1, po .724. (1930) o. 94.0 Langmui:r ap.d I3109gett,P.o R.o , 40, ppo, 7S-104 (1932). 95!tRoberts,. Po .. Ro , 129 . .A 9 . po.146 (1930)0 ' 96;0 Jo Po Ro,9 l3.,. po 53 (1952)0 97 163, po (1949) 0 98 . Dunoyer, Co RoAoSo 233 pO' 9),,9 (1951). 990 Temperature dependence'of AccoCoefo on W, Ni, Pt, 1954 MIT Chem P1:iIID Jhesis, Lo Ao 1000 Gomer and Smith, loc oci t, 0 , ).<' 6M-4809 30 of 37 .Knudsenl1 Ao Po. 34, PI) 593 (1921).0 102.0 Knudsen,. Gases,. PP.o 46052 Methuen (1934)0 103,0 Devienne,. Va9" 3,. Po. 392 (1953)0 . 104.0 Becker, Po 34, po. 13?6.. (1929) o . 105.0 Esterman., Ze f!1Po 33!1 po 320 (19?5l.o 106!1 Bikerm.an,Surfo Chem for . 219 .. (1948),. 1070 Burton, and rank, Po Ro So, 243 A, po 299 (1951)0 D ) .108,0 Frank, 0 Fo. S() 59 po 48 (1949 0 .Smoluchowski 9 trpnsfprmations .. in PPo.149-183 Wiley(1951) llO.o Ind, and Eng, Chemo 44, PPo.1270-133a (1952)0 111,0 Turnbull and H911om?n,.Physo of Powo.Meto 109-143 McGraw-Hill (1951) ll2.0 Quarto Rev.o 5,p.o 315 (1951) 0". 113.0 Holloman. & TurnbUll, Progoin Meto-,PhYSo.Vol- 4,. pp.,333 ... 88 (1953) 0 1140 HollolWl, Thermodynamics iti Physical Metaurgy, ppo 161-77, !mo Soco . for Metals, Cleveland 9hip,.(19'O) 0 ". 115p Coleman. Ro65p.o 57 (1944)0 116!) Bunn, Po RQ So 141, po 567. (1933}o 117!> Frank ;md. _p. p.o Ro .. So J..98 p. 216 (1949)0 118.0 Volmer, to ;Po. 9" pe 193 (1934). PC) 0 So 5, po 48 (1949}() . 120.0 Buerger, . 32, po 593 (1947) 0 121.0 Wells, Po Mo 379 po 122!J Gomer and cito.po 50 123,0 Cunningham, iTo ro 9& 39, 69 (1935)0 124.0 Lashkarew, To F I) So 31, ppo 1081-95 (1935)0 1250 Rideal, Introduction to Surface Chamistry, po 205 Cambridge Uo . Press (1930)!t . . . . .' . 1260 -Semeno,ff j . To. Fo. 50$ 2B, po 170 (j.932)o Becker$ 5, pp 55 (1949)0 128.0 Magnus., T.o S.o 28 P.I) 386""94 (19,32) I) 129p Zoo fio. p$.63 po 245 (1930). Hill, J.o. Co. P.., 19,. p,o. 394. ,(1930) 0 I LangmUir, P .. 132._ Andrade,. To. Fo. So. 31, p&,1137 (1935). 133.0 Langmuir. J 0 Ao C. 50 p .. ltJQl. (1918)._ .. 1340 Brunauer, Emmett and Teller, Jo Ao Co S. 60, po 309 (1938); 62, P,l 1723, (1940). 6M-4809 Langmui;r.9 ;Io 1.0 0.0 So." .. p.o' 2798 .. (J.932) 0 136$) Onsager,9 Re Mo Po, 25, po 353 .. Page 31 of 3rt 137.p Wiley (1940)0 138 9 Ono . .$! J!, Cp p.O.9' :L8.9. po, 397 _,' 139.0 OnO,9 .. Jo. Po, So, Jo,5 J po (1959); 6; p. 10 (1951). 140,. Dole., J!, e,o Pp 16". pp25 .. (1948) o. Hill, Jo 90 Po ;t6" (194.8)0. 1429 Fowler,)} Pp Op P!, Sp 3:1.:" po. 2,26 (1935). ' 143p Po. C. O P'o S.o ppo (1940). 144p Hill, Jo 00 :Pojl.l5$ po 167 (1947)0 , 145p Cassie s Tp F.o S!'9 po 450(1945) 0 146p Halsey, J 0 Ao Co 8.0 73.9 po 2693 (;t9?1) o .. ' 147'0 Huttig and Theinierjl Ko PPo 69-73, 157-60 (1950); 121,9; . pp,,50:=57 St (;t95:t.)0 . 148 0 Ha1seY9 J o Co PO,9 16,9 po 931; .(1948)0, 149.0 ;r 0 C.o Po" 57, po 946, (1953)0 150.0 Hill.$! J.& e,G 1?!, 14$) PP.o 293, 441 (1949). 151,0 Robe;rts,9 p,(} R!t. So 141 (1937) 0 152p Hill, J.o 0 .. .9 59 po. 1065 (1955)0 . 153.p Gang11l1, 71,.ppo.275;=9 (1935). 154.9 Ha1s$y, Jo Ao. Co So 74 9 . PP.o 1082-3 (1952),0 155p Yang, et alia" A,9ta, Me:t;o 2, po 475 (1954)0 156.0 Van der Merwe, D.c- F!l" S." 5Jl p.& 291,"(1949)0 . 157.0 Egli and Zerfoss, Do F6 So 6:L (:L949). . . 158.0 McLaupb].afl, $co:t;t Senpett, Cano, J.o Res o 28, po 530 (1950) 0 159.& Thirsk"P!J Fp So 'J.o 63, po sa3 ()..950) 0 . 160.0 Engles, Uo So Bu;t"o Jo 50 ppo 2#9-61 (1953) 0 161.0 Germer, and ?& pO, Ppo. 447 Harris.jl at alia., J.o C.o IS" po. (1950) {) HarriS., et alia., J.t) 0 9 Sp AOj) 3S, P6 582 (1948)0 Harris" et aliaj Jo Ao. P&.19 9 po 791 (1948)0 165.0 Frazer,. Molecular Rays, po 95 Yo (1931). '1660 Holland, Vacuum Deposition. of Thin Films, po 74, Wiley (1956). 1670 and PoPoSoLo 66 Po. 1073 (1953)0 1680 Harris.}! Beasley$) and Loeb, 41, po 604 (1951) 6H-4809 1690' Harris ana Loeb, J o OoS. Ao 45, p. -179 (195')0 170 0 Skinner, J.AoP. 26, Po498 (1955)0 Page 32 of 37 171. !mai and Mizushimt, J.F oSoJ 0 9, pp. 28-1032, 1034, (1954). 1720 Wexler, Condensation CoeffieientB of Metallic A t o m s ~ held by P.Fowler. 6M-4809 Page 33 of YI . SUPPLEMENTARY BIBLIOGRAPHY I As a guide for continued investigation into. the properties of thin films" the following list of .r.eferences is gi veno The work cov- ered by these articles is less general in nature than that of the main bibliograpw ahd should have Borne interest to those looking for thin film components 0 1.0 Resistivity of thin films, Clark, Brito Jo Appo.Physo. 6,.po.158. (1955) 20 Electric Strength of Dielectric Films,Plessner, P. Po So Lo' 60, po 243 (1948)0 -. . . . 30 Electron Bombardment and' Conduction of Dielectric Films, Ansbacher and Ehrenberg, PoPoSoLo 64,.po 362 (1951}o . 4.0 Thin Film Electronic Mean Free Path,. ,Advo in Yol, ppo 1...;42 Path J.ength, of Electron Beam in Foil, Yang, Po Ro '9'0,84 po 599 6. Elect. Condo of Metal: Deposited" on Die1.ee'tiric 'in Relation to Crystal Structure, Feldman, CoRo-.AoSo.t 0 . 70 'Relatibn between Film Res1st*o Varied by- Coverage with Diel- ectric, ... and Vodar 1 Co R.Ao$o., .. 23$, po 414,. (1952)0 8.0 High Freqo Resistivity of Thin Films, Pj).,Ro 9l+,. pp 285 (1954) . 90 Metal Film Resistance Factors, Appleyard, PoPoS.!.,. 49, Supplo po 118 . (1937)0 100 StrUctural Changes during Growth of Metal Films, Quarrell, PoPoSoL o , 49, po .279 (1937)0 .... _ 110 Conduction of- Thin Granulated Ag Films, Haas Scott, J o Po Ro , 11 po 394 (1950)0 . . 120 Some Properties Evaporated Films on Glass, Rood, Jo 00 So Ao 39, po 854, (1949)0 .. .'.' 13,0 Ther.rrmldlTl..agnetic Properties of Thin Films,. Blatt, "?o 95,. po 13,(1$4) 140 Conduction of Thin Films of Alkali Metals, Lovell, PoP oSoLo 49, po (1937) 0 ,. .:.., ' . 150 Conduction of Thin Films of Indium, Cosslett, 49, p. 121 (1937)0 . .. '. .. '. ' 160 Adhesion of Films, 00 8,0 Heavens, JoPoRo 11, p. 355 (1950). 6M-4809 Page 34 ot 37 SUPPL!m:NTARY BIBLIOGRAPHY II The following is a listof some references oftecbniques and crit- ical data that should be of to the experimenter investigating vacuum evaporation of thin films o 10 Bibliography of Solid Bureau of (1944).& . ... .- 20 Brewer, andPhys. Propo of Elemants, Rpto for Manhattan Proj- ect MODe 438-0" Also other. reports 9f 436.serieso . . .' 3'0' Brewer, Chern and of Misco Mato,Nuclear PbYSo Series, McGraw ._Hill Vol,. IV 19 00 Bo 4.0 Buo. Mines Vap:>r Pressures, etc. 50 Stull, '& Chem.o 39,9' po 517 (1947) Vapor Pressure of Organic 0, 6Jj Hayer"Jo.Co.Po, 1, 270 Lattice enfa;rgy of a1ko halides. , . ," \ ,1 ." 7 Law, R&.So Io, 19, po' 920" VFlPor- 80 Barrett" Structure of- Metals', -Crystallographic' data, McGraw-Hil1(1952) . Ta b1e pf Text, of Evap 0 'FiJ-PlB" etc Ct esp., , Po 5;1.9 .. ' 90 Wexler, Condensation Coefo,. Binding Energies, and Critical , . of Metallic Films..? held by Po. FO'JIler,o. .' " ' lOci Marshall,9Dornte, and Norton,' J o A,o C.o s., 59, po 1161 (1937) Vapor . PressUre of Fe, .... ; 110 Schuman 'and Garrett,' JoAoCoS,9 66, PG 442 (1944); 67,po2279 (1945) . Vapor press'ure of, Beo 12.0 Preparation of t;hin unbacked films,. Nature, 159 (1947) 0. 13.0 Yacuum symposium, Committee for Vac o TechniqJ}es (53303 L73 140 Thin films at low temperature, McLennan,'Rpt o of Prog. in Physo 1, po 198 (1934)0 . .... '. . 15.0 Deposition at film interface, ,T.GE.$o,. 55,. po15) (1929)0 Prep." of Metal Films., Ro Io,. 21, po. 6g3 (1949)0 17<t Prepo of Unbacked Films, Carpenter and Curcio, RoSolo' 21 po675 (1950)0 ',.. 180 Density Change of Rhondium Film after Evap, Stopped. Zehden, Vac, 1 PPo 38""9 (1951) 0 "" . Tech.o for Evapo Zns :Films" Rpod, JoOoSoAo4f, p .. 291 (1951)0 20" Evapo Films on Glass, Strong, Astrophys, Jo8j,po 401 (1936).0 6M=4809 Page 35 ,01 37 . . 210 Satisf'actory Filaments for Evaporating Metals, Caldwell, J 0 Ao Po 12 , po 719 (1941) e, 220 Evapo Filaments Usedo Olsen,Smth and Crittendon, J.A-.p.. l6i p.-425 (1945}o '. Tecp, in ,Makipg Thin Films, 2 ar:t,icles, Mar. 15, 19470 240 Rpto for Techo'of Shadow Casting, Cambridge Uo Summer School (Ref. Vaco.lo. po 30)0' 250' Techo and Plant Set-up for Vaco Evapo,. Holland, Supp.1l. (1951), Jo Scio Instro 26 Engineering. Index, Under 0 2.7,0 Vacuum. Metallurgy,: Elec,trochem..9" (1955)0 ,28,0 Contamiilatioh,by Fllni Heavens, PofoSoLo 65B, po 788 (1952).' 29$1 Use of SiO, Labo Practice a , 3, po 4.41 (1'1.!:lJ)St 30.0 Molecular 18, po 300 (1946)0 310 Radioactive ShadoW' Techti, and Dep-o of Monatomic Beams, Simnad, Yangj Pound, JoAoPo!J 25, p. 268 (1954),0 320 Soldering to- Thin FilMs, (Few. it)!) Boetticher -and Thun, Optik 11, No.1 po' 2? (1954)0 .. '--.'.. ..' . . 33.0 Prep.o Thin Filrn:!, JoioPo "?6,, po 975 (J.955)i> Proto of Mo by SiO, Fe1dmana CoRoAoSo, 706(1952)o . 350 Contamo of Film by Source SUp}:lort Material, Heavens, Nature" 168, po 694 36.0 Masso Evapo Formula, ;Holland, Vac-o--l,no-o. 1,p.o23-36 37.6 Thin Films on Liqu1ds, Nature 162 po 892, (1948)0 380 Effect of Deposi,tion Tempo on Magnetism of Ni Films, Fdwards, PoRo 27, po 252 (1926)0 390 Adsbrption of Sr and Ba on WaMoore, and Allison, JoCoP023, po 1609 (1955)0 400 Determination of' Object- Thickness in Electron Microscopy, Marton and Schiff; J o Ao Po, 12,9 po 759 (1941)0 410 Silicon Monoxide Protection for Mirrors; Haas and Scott, JoOoSoAo, 39 po 179 (1949)0 " ..... 42p Increasing Adhesion of Metal Films, Scott, J oOOS()Ao 9 hP, po, 804 (1950) 430 .Production of' small aperture by evaporation . po 115 .(1951) 0 , . '. . . . 44J:1 Multiple beam interferometry, To1ansky', Vac.o 1, P.o 75 (1951)0 . 450 Preparation and Structure' of SiO- Film,. Haas', Data 16,9 . No o 4, po 21(1951) J oAoCerlt $oco . .33, P6 35.3 (1950). ". 460 MeaSuring' Film Thickness 10 i to 3r.t, Scott, McLachlan and Sennett, JoAoPo '21, po 843, (1950)0 6M-4809 Page 36 of 37 , . . 410 Measurement and control of thin fih thiekness, Greenland, Vaco 2, ppo 30 ..... . . .' . 480 Prepo and properties of thin l'i02 films, Hass, Vac. 2, ppo 331-45 (1952)0 .. - .... . . 490 Preparation of thin'unbacked 810 films, Sawyer, R. S. Io, po 604 (1952)0 500 Adsorbed films of Os on W, Langmuir,PoRo 43, po 224 (1933)0 Page 37 of' 37 BIBLIOGRAPHICAL ABBR.EVIA nONS . Abbreviations for' periodicals listed in the preceeding bibliographies areg
A.od!,P.o
. J.oA.oC.oS o J.o.A$)P.o JoCoPo 0"
J ,
. .
P;oH;O . . PpN}>A!,S!)
. PpR.oS.o PpT,oR.oSo R,oM.oPp
i'.oE.oS.o ToFpSo V?cp . Z!'tpp.& ZoPoCo Acta C1f.i.stallographica Annalen der Physik Academie des Sciences Comptes Rendus Discussions of the Faraday Society Journal of, the Am.ericanChemicalSociety Journal of Applied Physics .. Journal of Journal of the of America Journal of Physical Chemistry - Journal de Physique et Ie Radium Journal of Physical Society of Japan Kolloid Zeitschrift Proceedings of the Cambridge Philosphical Society. Philosophical Magazine ! f I Proceedings of the of Science" , Proceedings of the PhySical Soci 'ety. of London Physical Review Proceedings of the Royal Society of London PIrllosopbical..Transactio.ns:of' Society Review of Modern Physics Review of Scientific Instruments Transactions of the Electrochemical Society Transactions of the Faraday SOCiety VacuUm Zeitschrift fur Physik Zeitschrift fur Physikalische Chemie .