Scribd Logo
Importer

Bienvenue sur Scribd !

  • Diffusion Coefficients

    Transféré par

    Bharat Vaaj
    10 vues117 pages
    diffusion, constants

    Copyright

    © © All Rights Reserved

    Formats disponibles

    PDF ou lisez en ligne sur Scribd

    Avez-vous trouvé ce document utile ?

    Ce contenu est-il inapproprié ?

    Signaler ce document
    diffusion, constants

    Droits d'auteur :

    © All Rights Reserved
    Téléchargez comme PDF ou lisez en ligne sur Scribd
    Signaler comme contenu inapproprié
    10 vues117 pages

    Diffusion Coefficients

    Transféré par

    Bharat Vaaj
    diffusion, constants

    Droits d'auteur :

    © All Rights Reserved
    Téléchargez comme PDF ou lisez en ligne sur Scribd
    Signaler comme contenu inapproprié
    sur 117
    Gaseous Diffusion Coefficients 1. R. Marrero” ~ EA. Meson Brown University, Providence, Rhode Islond 02912 Diflason couiciat of binary mitures of dite esses ave comprehensively compiled, cxtically evapiel Ta soeines i's Tnlereial seone Tere re cee uy pene {eoommended wih canbence, Deviation plts re given for most ofthese systema. Allont every svn fo ccc sich vor examen Scere an eared 1970 canbe Shed tone mse sy rae cada So errno ah emer lth eae ls terme rsdn ov neyo of er on ee fo ictle Sse” cate sett Mh tempers, tn Lodo peron ctons fro tempera clk Sitsion cof rtevcal mars te ned ‘et tpn so moet nd teal inca, Cron hee won ao oe At Sts ag orate ir composton dependence andor oat let New somes! cus lol anes cece cay Key words: Sins git misures; enbcaly evaluated ast; dase; aimuson coma; Eset; letdepoct proper Contents isto abe 4 santas Aco Lito Riewes 1, Mature thermal conduct. 1. ated «, Thormal difuion factor Referenoes for action References fer ection? 2. Thenretcal background. 2 lone dion fib sn di 5 Experimental methods fr difcioncoeicions ‘31 Outline of experimental methods. sion coin. ‘ 132 Major experimental metho 1 Two-camponent itor 6 ‘Cloned ube By Atucomponent mares 1 1 Two bulb appara. 22. Moleciar theory of ifsion. 1 Poin otter 's. General background, 7 4. Gar chromato. 1b Theoretical methods 8 Evaporation tebe ©, There rest » 3 Minor experimental meio 23, Temperate dependence ofiiwion orient 11 ‘Open tae. . 1, Louprangeinterston enone 2 1 Back dita. bk Shoreangr interactions 2 Caparo © Inertedioerngelnteractons. a 4. Utendy ener. 24, Compoutin dependence o diforioncocfiione, 12 «© Difusion bridge. 25, Quantum eet ndfosin efficients B Diesoiated gases 26, Determination of ifarien coin nter- 34 Mccall eptintal etids roles fee nen “ ‘Droplet evaporation ‘4 Dilfesion colin inca for lng rage inter 1. Dalour elec. scion. “ «Thera separation rue: Dain ass itera shor rng ner 4. Kitkendal eect sins ” Sound aberptin «, Cambination les 22. Determination of diferioncoticiat rom at rapt Property measatemen. ie sing, mat Cg, Rep Dr is 1s Caaphoresis 1 Resonance method References fo section 4. Treatment of dat 41 Relinbiiy estinates 442 Correction for composition depandence 145 Corelaton for temperate dependence 4. Phys. Chem Rel Dat, Wal, No. 1, 1972 4 T.R, MARRERO AND E. A. MASON. Relrence for ssn 4 33TH Seed birtemnertre put for caren. see ne : 3 pewter = 1 Uncrainty ‘ 33, Mokeiarbeam pocatias, gle) = Ki for wp 35 Condon ermeict ncn 38 u “ 46 £3 Deviation oe 3% 18, Vel ef cy doa nesuement comeued wih 53 Date ema § ‘Siem tue econ, sad mera + Group | vine sgn 5 030) 8 ‘onc. 9 1 Gout vt su 26), Dvn econ eis aren. {Grup deistin rte fen ff Bien cron ‘8 SMES ete esi mat Ha i rs Reference for Sete 5. tio Us » 6 beknowegmetc CII yo Diaan coding fr care angen fee n 1. Gaesas difvon coins fr binary minus 2 Henan mtn 0) = for wo B70 w 197). on tm " 0 1, Motclar bean messteeni 74 Dio ais sd eid re ie a 11728 Moklarbenm pret. 9) = Rito 2 He Siete acs micelle 4 ” 5 Otter refrnces SSEmnl ne 1, London doors cots. his TV, Rete viscnaty menses hs : 1. Pinar aes ee t of Figures Pee Tr Quaiaive temperatise dependence of fain Senden ne n 2. Pcp cape meted for iin cou foe 19 List of Tables 4. Canonion dagen of emt mac oe (a) Ersnion 2-10) he iat fons oe hayman ast thay: (21) i the te opie expan ofthis eprtcvesnners 32 ‘rae 4 Somwet weeny neo a fu "Quant flat on diflson seins and on Nepanere x“ collnon ital ae AZ ref th deur ‘te folowing Ses cette eatin of experimental ee es alas of Sip om the rfrnce aon frp pan the moan So sameness 2. Catenion ef ejecta ai aR 3. Dermiaion ef Su by theclwed tie etd.” 20 4 Determine of bythe woul etd 21 6 tions of Sy by the gaschromatography Group I ee 2 1 Bocrition dB mosis Tage Ne Te Neke «Bact oy by mir inl chy 3 TE Nek a 1s Anke 9. Dacmiti o3uby iene ee iy te ie 1 Groping yc” Scone we 1a Heke ——— “ M HeXe 41 19. BeNe 2 Thoin ioc owe ibis RS Boe. 12, Canlaon parameters oot a Se 3 18, Contain rameters eq th22 fer 8 1, Chascallowtompernue asymptote vac Si aes) on 0 Doses eos aca Group I encoo a sceringte at), ” 16. Experimental determine af Sin acorn Pete te gue fe {eps partemperatre nate, TE teone "55 SoS: Mable ae ee ae) Bi Nese 3 1. Notes td nate compen. 4 ott a Bec, 3 ¢, Drsnchted gues 34 2 Hen, St B. Heo 54 i Oita comm et Se tie” St 4. Phys Cham Re. Dato Vl. 1, No.1, 1972 GASEOUS DIFFUSION COEFFICIENTS 5 Figure Poe Figure Pie BL HyNe 55 B. Hair 59 Hear 55 MHC 39 Hear 56 a. Heo, 6 Sh OHeKr 56 2 Nar 60 SD 57 B NCO 6 36 HD, 37 4 NCO SD, 58 NC ao Heco 38 6 NCO, Group Hl Pose rer ae #1. Ascii, 64 n ew nco 2 Bao 6 cs Arco, 6 n SL ASR 65 * @ xe ” Sk HO, 68 % tsk 68 8 % Cliele mM. Orc0; 16 St then, i Gere 8 CHeO, 70 2 Coa 7 B CHS. 10 CONO 77 @ NN 1 Th Cosh 18 Boner! Miscellaneous Systems Fee Pa Face Pa Se wont Te woce, “a 7 oo, 8 . Corns 8S 78 Oar 84 & CO-ca 85 1, Introduetion ‘The purpose of this report is to evaluate pub lished data on gaseous difision eoefents ‘and tatablch recommended “values, when posible Slot ‘catlogs of measured difadon, coins previously appeared. inthe” Inernational cl Tables {0 the Land mtn Tables | and the Theriophsical Propertics Research otaure Retheval Culde By Addon ated reviews are ale avaable (C8) In the preseat Tarver giscous difwein gooicnls ars sombre hensively compiled and_ critically evaluated, Hadi anaats for constr} with theory sd Totter measured molecular properties, Such Uetated evaluation Ts possible beeause ofthe exist ence of a highly de ped theory for dilute eases. "Ti ecope of this survey la Uneefore luted to the dine gts regen. In is repln the donalty fo fl that he rate of transport of tas, omentum, Ba onaraphrnr eure cored ea cee molecular collisions. In practice this means gases with densities appeosimetely coresponding. to standard conditions (ic.. of the order of 10° — 10% molecules per cm?). The emphasis here is on two- component (binary) mixtures. Multicomponent Aiffusion eau be wecarately desesibesd in terns of the binary diffusion coefficients for all possible pairs of gases in the mixture 9}, Allthe atoms or molecules considered are neutral species, The mass transport processes outside the scope of this survey are as follows: (1) diffusion of ionized particles (jon mobility), @) mixture separations in a temperature gradient (thermal diffusion), (3) mixing due to con- ection or turbulence, and’ (4) the permestion gases through liquids or solids. The sole interest the mixing of gases caused by composition gradients ‘The gaseous diffusion ewefficiems revonuuended here are succinctly reported by means of semi- empirical funetions; temperature limits renge up to 10.000 K and to a lower temperature of the order of 100 K (tables 12 and 13). There isa small compos tion dependence in the diffusion coefficients which aay be estimated {rom parameters listed in table 15. Figures 5 to 81 are deviation plots, for sixty-two gas pairs, of experimental measurements from the sein empirical equations; these graphs illustrate dis- crepaneies in the data, ls sepoHe ts a follows. AU al possible’ tempera spl diffusion coefficient values were critically evaluated on an individual gas ir basis. The data assessment was measurements From the rigorous kinetic of gases an approximation was developed to make corrections. for small composition effects. Coefi- lems could then be normalized toa specific mixture ‘concentration for comparison and subsequent corre: lation. Diffusion coefficients derived trom other transport property measurements, particalarly ture viscosibes, were uselul for the extension of ‘values to intermediate temperatures and for consistency checks. In the absence of direct meas- urements, intermolecular forces from theory and from. beam experiments served to determine diffusion coefficients at very low and elevated temperatures, respectively. Semi-empirical fune- lions were constructed to correlate the data over three decades of temperature within the experi mental uncertainty. ‘This report is divided into five major sections, It begins with a section Theoretical Background — which includes the diffusion coefficient definition and its theoretical expression according to the rigorous, kinetic. theory of gases. The. ki theory foundations are necessary for the under. standing of temperature and composition depend: ences, and quantum effects. This section closes with. equations for the determination of diffasion eae ets from intermolecular forces and {rom other transport properties. The principal experimental techniques are described next. Methods of measure- classified hy the geometry of the apparats and their reliability is estimated. Procedures used to critically evaluate the entire body of experimental data for accuracy, composition and temperature dependencies are outlined under Treatment of Dat ‘The analyses and results related to the small compo- sition dependence of the diffusion coefficient are entirely in this section. The semi-empirical correla: tom equation sas chowon on the basis of knowledge 4 Phys. Chom, Ref Data, Vo. 1, Ne. 1, 1972 6 TR, MARRERO AND E. A. MASON of how intermolecular forces affect the temperature ‘dependence of the diffusion coefficients. Previously tuneslculared values of lye weimporatae soysmptutes of diffusion eoeffcients are tabulated. ie fourth major section—Results—can be subdivided into four areas. | First, diffusion: ‘cooficient uncertainty limits are classifled accordi to temperature and gas pait. Second, the tabulation ‘of correlation parameters. for the ‘reeommended is ‘Then a series of graphs shows the ations. hetween graphs will readily indicate that the unqualified Selection ofa diffusion coefficient from the literature may be uncertain by at least several percent. Ihe Tast part of the Results section contains detailed remarks about data appraiaals for specific systems. ‘The final masjor part of this report is the Bibliog raphy: two annotated bibliographies are given; fone contains all the experimental sources, complete through 1968, according to author (gas. pair and ‘method are noted). and the second is a supple: mentary listing of eitations according to was pair. Practically all diffusion coefficients ever measured cean be traced through these bibliographies; how- ver for many svstems. the results, are too frag nentary ot oo uncertain for the illusion coefhcients to be accepted as reliable, Additional references froin 1969, 1970, and a few irom 1971 are included, Lut the correlation is complate nly through 1966. References for Section I (2) Boynton, WP, and Brattain, W.H, jn nternatins Ciel “Tales of Nemevial Data, Physic, Chepistry tnd Tec tology MeCraw fi ok Ch New York, 1B) Vo fa] Roth, W. As Scheel, K., Bator, Lando Binsin, Phys TisiachChemische, Tabellen. 5 Aufage Je Spence, Berlin 923, 1997, 1981. (3) Teuoalian YS Geet, 3K and Moore N, ¥ Eos aide Plena Press, Now York 1967. (g] Westberg, & A. Combustion and Fae 1,346 950, [5] Westenberg AA Ady. Heat Trancor 3, 2551900. [6] Parry, Re HL, Chiton, €. I, and Kirkpatrick. D. Ealtoy, ‘EReancal pices Hpdtook. tk Eaton, core Hill "Book, Co New York, 1965), Chap. 14, 9p. 19-23 {0} Vargaty NB" Manual of Thermopinyateal Properties of ‘Gans, ond Ligue Go Ruslan), (PM, Howcom. 108), PP sh tay aol i, end Rummeticn, ML ane. eg, nem, 10 (1), €6 (i068 SB (a), 32 UB%0y, BT Cz) 94 AES], 56(12.61 1908, [b] Wrehfelder, JO Cuptie, C.F, and Bid RB. Meeslar Tiers of Gaset and Liouids Gohn Whey and Sons, New Yor obs) host 2. Theoretical Background In ths ection the presentation ofthe thera hackernmnd ig preceded hy the ohenomenologe ‘ichniton of ilusion coeficiems. Phen sn outline form expressions for diffusion coeficiens are de Hed by the Chapman Enskog. procedure for inion of the Boltzmann equation, Most mathe Thatical details of the derivation are omitted, and ussion accentuates the application limits of these rigorous. kinetetheory formolas, More Sonipete information is available in three’ mono. 2nd rocent developments, appear in [4-13], The emphasis here is on molecular physics as a prerequisite to the under- seca Pidebptod thevretiool varie whieh re ‘necessary in ordet to achieve the fallowing: (Q) suggest the mathematical form for the cor relation of diffusion coefficients a3 function oftemperstorey (2) Gorrect diffusion coeficient measurements for composition dependence, {@) estimate quantum effects for low-temperature ‘fusion cote, (4) calculate diffusion coefficients directly from Knowledge of intermolecular forces, and (6) caleulate. diffusion coefficients from other {ransport properties, 2.1, Phetiomenological Definition of the ‘Gaseous Diffusion Coefficient In a nonuniform mixture the diffusion coefficient isa proporonalty constant between the molecular flux and the composition gradient of & species. Diftusion coefficients are defined by plienomene- logical equations for two.component and mult ‘component mixtures. 1 Ne, 1972 a, Two-Component Mixtures In two-component mivtes, in the sheance of temperature and pressure gradients, external forces, and chemical reactions, the flux equations are ue Seen FyVan. Anan, eu (2.2) Each opecios (or component) ie labeled by cubcoripte Lor 2. The flux densities are Jy and Js (molecules) com?-s), the total number density is " (molecules) fom), and the composition gradients are in terms of file fvactone sy ad sa. Theme exyeeions hold ly in the ease of zero net flux, Jy-+J,=0. If the net flux is not zer0, eqs (2.1-1) and (21-2) ean be considered to hold in a coordinate system moving ‘with the met flux, Uiat is, atu velocity equal ww Gh Ja)in. "The diffusion coefficients Dip and Dyy are posi tive constants with units of em*/s, Its easy to show from eqs (2-171) and (2-12) that Dy 2s, be joause Jy + Ja=0 and x;+19= for a binary mixture. ‘Thus diffusion in a binary mixture is described by a single diffusion coefficient, Molecular aiitusion, stnetly, speaking, vane ‘eur under conditions fn which both the net fox and the pressure gradient are simultaneously ero. If the pressure is uniform, then in general fluxes fare ditferent for different spectes, aud lw wet fax isnot zero. HF the not fax i zero, small presse igradient must exist in order to counter the tendency for the different species fluxes to be different [VE-IG). For instance, im a closed system the ai ference in the species faxes causes the number Aencty and hence the pressure to increase on one Side of the system and decrease on the other side side Oe Sting pressure erueut Torces tie nay GASEOUS DIFFUSION COEFFICIENTS 1 flux to be zero. Ifthe fluxes were to remain unequal in a closed system, then the pressure would con: tinue to, increase on one side and decrease on the other side ‘The pressure gradients in diffusing gas mixtures turn out to be very small in magnitude, however, infact they are altoatimmeastrably mall except in capillary tubes: where they ave been mest ured [17236]. Because of thie its unnecessary to include in the flux equations any term directly pro: Pertional to 2 gradient Tha whol affect fof any presaure gradient is simply to modify the net, flux, and this is the only term that needs to be directly included. "The gencralisation of ogo (2-1-1) and (21-2) for nonzero net flux is therefore simply Sa n9uTerband, b= na Int wd, (2-8) iy swhete tha net lus ie T= a+ Te. Equations (2.1-3) and @.1"4) define the diflusion coefiient in @ sta: tionary coordinate system. [can be easily shown, as before. that 9y2= 2 ‘Phe reouering squatiane sehich define inary dif: fasion covffcients, are applicable to any fuid, and hold regardless of any dependence of the diffusion coeficient on composition, pressure, or tempera dependences ly simple-the ai ticpendences. are especially simple~the difusion coefficient is inversely proportional to pressure and is only slightly dependent on mixture composition ‘The tomperstire dependence is more complicated [Ali these factors are treated in more detal a subse- ‘quent sections the precenre anal enmpaeition ‘The flux of a species in a multicomponent mixture ot conveniently expressed in terms of composi- ‘gradients. a0) in the foregoing equations for binary mixtures. The reason is that the multicompo- nent diffusion coefficients thereby defined have an ‘xessively complicate composition dependence ‘makes the calculation of a flux a formidable Task. A simpler set of equations for multicomponent diffusion is obtained by a different arrangement of terms—the composition gradient of a species is related to differences in fluxes of gus pat (2, 3,0, 21], The outstanding advantage of such a relation: ship is @ description of multicomponent diffusion in terms of diffusion coefficients for binary mixtures. These equations, credited to Stetan and Maxwell, va SC) (21-5) where i and j denote the species. For a mixture of » species there are » equations, but only v—1 fre independent. "The diffusion ‘coefficients 2 depend primarily only on the nature of the species Hand j, but are not quite the same as the correspond ing binary difuusion coefficients %y (hence the rime}. However, the difference between the 3, and the Qy lies only in their weak composition dependence, the exact value of Jj depending slightly on wmpocition of the whole Component mixture and not on just the relative amounts of i and j. The variation of the binary 2y ‘with composition is empirically a few percent at ost, and is of the same masnitude a3 the ext mental uncertainties in the few available raul component £9, or even in most measured binary By for that matter. Therefore it is reasonable on an’ empirienl basis ta. take j= y for. mn component diffusion, This is also justified theoreti cally, for in the frst Chapman-Enskog, approxi mation %j and 9y are identical and independent of eomposiuion [2, 3, 7] The apecial case of a trace species diffusing through @ uniform multicomponent mixture is of interest for two reasons. It provides a simple test of. = oy and makes possible the calculation. of diffusion coefficients of various species in air. First, denote the trace species by 1 and assume the absence of a net flux (JO), then eas (21-5) reduce to.a single equati vara 8 (21-6) If the trace diflusion coefficient 9 is defined to be the constant of proportionality between J; and Vin, then 7) where ® specifies the diffusion coefficient of the tracer in the multicomponent mixture. If 2y re- places 34 then eq @A~1 becomes wn expression of Blane’, aw (28), Detaled calculations (29) of jj and Dy for this special case show that the ‘ikitons om Blane’ law are sal for orn gases, This further justifies the application of binary diffusion coefficients to eq (21-5) for multicompo- nent diffusion. The second case of interest, the cal- culation of, diffusion coefficients of a species in fir (when direct measurements are unavailable or inadequate), is especially convenient by the appli- cation of Blanc’s law with available flu. sign coefficients of the species in nitrogen a oxygen. 2.2, Molecular Theory of Diffusion General Background ‘This presentation of the molecular theory of dif- fusion briefly outlines some major points of the Hgoivus huesle slvury. of yasem. -Kineaie teeny postulates transport due entirely to molecules in ‘motion. In diffusion the individual molecules them- selves carry mass through the gas. Since there are immense numbers of moleciles moving abou Mn a gas itis to be expected that molecular encounters (or collisions) are of cardinal importance in. con- trolling the overall rate at which transport occurs. The collisions in turn are controlled by the forces of interaction between the molecules. By the for ‘ula of kinetic theory, knowledge of these funda- ‘mental intermolecular forces can lead to gaseous diffusion coefficients. ‘The importance of molecular collisions in dif. fusion can be illustrated by some typical numerical values. At ordinary conditions of temperature and ‘pressure molecules in gases have molecular speeds 4 Phys Chom Rel ote, Vol, Mo. 1, 1972 8 T.R, MARRERO AND E. A. MASON of the order of 10* ems, which is about the speed ‘of sound, In contrast, actual diflusion velocities iin) are much less— about 1 em|s. This great do- ‘reaze in apparent molecular speed occurs because diffusion is dominated by collisions whici cause the limes the net distance Uaveled during diffusion, For gases at ordinary conditions only binary col lisions are important; temary and_ higher-order ibely Binary eallieiane, 10 particle encounters, are characteristic of gases with the ratio of mean free path to molecular diameter ‘of the order of 100. Transport phonomona diffucion, vicoosity, thor mal conduetivity, and thermal diflusion— arise by deviations, however slight, from the equilibrium molecular’ velocity distribution function known as the Maxwell distetbution. At equilibrium conditions an isolated gas mixture has no gradients in composi tion, pressure, or temperature; thus no_ fluxes. ‘Therefore to obtain transport coefficients on a the- rede basis kuowledge Of « nonoqullibsiun veloc ity distribution funetion is a nocessary requirement. bi. Theoretical Methods Diffusion coefficients can be caleulated from a flux derived from a molecular concept-—the integral of molecular velocity over the nonequilibrium ve- locity distribution fwmeton.. The Velocity distribu tion” fimction represents the probability for a molecule to have a specific velocity and location Some instant. The changes in the velocity distr- Bation due to molectlar interactions: must satisly the nonlinear Boltzmann integrodifierential equa tion. ‘The basio problem of rigorous Kinetic theory isto solve the Boltzmann ecuaton. ‘A Solution of the Boltzmann equation was inde- pendently obtained by Chapman and by Enskog [1-3]. Both used a method of successive approx mation. and even though procedures by Chapman tnd Enskog difer in detail the results are identical ‘The transport properties appear finally. in. the ChapmanEnskog theory as, solutions of infinite ‘alerbraie equations, and the transport properties can be expressed formally as {atios of infinite determinants whose elements are uations. The co plieted functions which depend on the species and the composition of the mixture, and on integrals related to binary ‘molecular’ interactions. ‘These sets of equations La by solyedy fortunately, by sapilly converging approximation schemes. ‘An outline of the Chapman-Enskog procedure is as fallows, First the velocity distribution function is expanded bs terms of « pertubacon fuedin ade to the Maxwell (equilibrium) distribution. By the assumption of a small perturbation, the expansion substituted back into the Boltzmann equation leads {0 a linearized intogrodifferential equation for the perturbation (ef. 2, see. 7.3b, 0) ‘The perturbation teria i assumed proportional to gradients, and expanded in a series the series ex: pansion coeticionts are fmetions. of molecular Yelocities (rel. 2, see. 7.34). The assumption of linearity in the gradient of composition ie precisely consistent with the brecedine ohenomendlagioal efinition of diffusion ‘coefiients; other transport 4 Phys. Chem. Ref Date, Vo. I, No.1, 1972 coefficients may be accounted for by additional appropriate gradients, The diflusion coefficient now appears as an integral of the expansion coefficient over the molecular velocities (ref 2, sec. 7.4a). The expansion coefficient satisfies a linear integro- differential equation obtained from the Beltzmann equation. This equation is solved by a second 5 For the second expansion it is convenient, but not necessary, to use orthogonal functions ‘because orthogonal properties li Cation of the ealenlations. The orthogonal functions: usually used are Sonine polynomials (tel. 2, sec. 7.3d, g). When this second expansion is substituted back nto the integral expression for the coefficient, it turns out (because of the orthogonality) that the diffusion coefficient is exactly equal to just fone of the coefficients in the second expansion (el, 2, sec. 74a). The problem now is to find the coeiticionts of the second expansion. To do this, the expansion js substituted back into the linear integro- siferenial equation, which is then solved hy a infinite set of algebraic equations in which the un- Knowns ae the coefficients ofthe second expansion, and the coeficients of these unknowns are compli: Gated multiple sptegals over molecular velocities ‘These integrals result from the moment formation; ost of the integrations can be carried out explicitly, ‘but not al, until the law of intermolecular force is specihed (ek, chap. 9: ro. 2, sec. 7). ‘The diffusion coeficient is thus equal toa single unknown in aa infinite set of algebraic equations This, get cannot be solved exactly exceDt in very special cases, and some successive approximation procedure mist he used. The set is systematically truncated in some plausible way (wo ways are oot. monly used. one due tp Chapman and Coline, nd the other to Kihara) [39 the simplest truncation fives the first approximation to the diffusion coe Cient, the next step gives the s Seda an Tn ther Bret appeoniovation the difeton Gveficient is ‘independent of composition; the second and higher approximations introduce Composition dependence, Since the approximation procedure converges rapidly, the third epproxime tion for the diffusion coeficient is almost identical ‘ith the second approximation (30) “The solution of the Boltzmann equation by the ChapmanrEnskog proceduse depends on the fut Jowing assumptions Binary Collisions. The Boltemann equation itself has @ fundamental assumptionbinary col. lisions. “This assumption—that only two-molecule interactions are important—limits the application of theoretical results to transport properties of Silute eases. Small Mean Free Path. The Chapman-Enskog solution aoa that the eae of es me ontainer are large compared to the molecular snean freo path, In gacoy at extremely low daneities Inolectles collide more frequently withthe walls of the container than with each other. When molecular olisions with a container surface are significant, the theory fats. Small. Perturbation. Tn the Chapman-Enskog thoory the assumption of @sinall pertnbation fang, tion deseries small departures from the equilibrium Wehoaey disinbuton tonction; in other "wordec at GASEOUS DIFFUSION COEFFICIENTS . onions slightly away from equibrium the trans port property fluxes are linear in the gradients. Classiral Mechanics. Historically, classical me- chanics was necessarily used by Boltzmann, Chap- man, and Enskog; however, their theory can be reformulated to account for quantummechanical ffiects. The modification needed is merely to replace fan integration over classical impact parameters for molecular interactions by an integration over de- flection angles involving the quantal differential Elastic Collisions. The original Boltzmann equa- tion and its solution by Chapman and Enskog were Umied to elastic coliaions ‘between 'molecules ‘occur between molecules with internel degrees of freedom, and kinetic energy is no longer conserved, although mass and momentum are conserved. Thus ‘diffusion and vigoosity are not etrongly affociod by the presence of internal degrees of freedom, but thermal conductivity i, The theory may be reforrm- lated to account for inelastic collisions. mn of collision integrals, and a number of complementery definitions related 0 Binary malcula collisions. The extension of the i inelastic cal mn effects i also given. “Approximation Scheme for Diffusion Coeficients ‘The higher approximations for diffusion coefficients in's dilute gus binary mixture with epesiee of type 1 and2 are written [@y]u=[9uh.f™, where [Se]: ia the first approximation, fac: Counts for the effects of higher approximations, and IM indicates the order of approximation. In the first pproaimatign, fon diffusion coefficients, f=; the effect of higher approximations is described by (2.2-28) (2.2-2b) (22-1) po where Asis the first correction term to [2] First Approximation for the Diffusion Coefficient. The expression for {ish is ile) Gaga) where jazz mma) (my+ms) isthe reduced mass of a alr of molecules, 7m 1s the molecular anaes. of & ies, Fis the Boltamann constant, and T is the lute temperature, The diffusion collision integral hhas units of azea and is dependent on the temperature and the forces of molecular interaction ‘of the gas. The collision integral for diffusion is 2-3) Soo =b ens ["eamesnaar, a2-4 2 ‘where B is the initial relative translational energy of two molecules in a binary collision, B=} pase, 0 Deluge inital relative spect! of ie suvlecales ps nd he ifn rant cs tn soya [PO niin 8 9 where 1(% E) is the differential scattering cross eto, For ceo! systema TF) sn a sthiere & is the impact parameter~ the perpendicular distance between one molecule and the initial line of relative approach of the other molecule. The clas al scattering angle for a pairof colliding malacnlos -of:#h-0) where re, the distance of closest approach, is given by 22-9) Gt In 0g @2-6) 7 te the internaalear « distance, and g(r) is the spherically symmetric intermolecular potential ‘The expression for (9, ls in practical units is c2en-ocone (et 2 aw where T isin degrees, Kelvin, pis the pressu atmospheres, Mr and Ms are the molecular weights in grams per mole, and Qj is in angstroms opereds "The frst approximation for the diffusion coe

    Vous aimerez peut-être aussi