564 Chapter 18 Concentration Distributions in Solids and in Laminar Flow Concentration at surface 6 Cae _ ASB (inside catalyst ~Concentration tay solid Stream with ‘concentrations enn ad oe Fig. 187-1. A sphericai catalyst that Fig. 16.7-2. Pores in the catalyst, in is porous, Fora magnified version. which diffusion and chemical reac- of the inset, see Fig. 187-2. tion occur, Aiffusion inside the tortuous void passages in the pellet. Instead, we describe the “aver aged” diffusion of the reactant in terms of an “effective diffusivity.” Specifically, we consider a spherical porous catalyst particle of radius R, as shown in Fig. 18.7-1. This particle is in a catalytic reactor, where it is submerged in a gas stream containing the reactant 4 and the product B. In the neighbothood of the surface of the particular catalyst particle under consideration, we presume that the concentration is car moles of A per unit volume. Species A diffuses through the tortuous passages in the eata- lyst and is converted to B on the catalylic surfaces, a8 sketched in Pig. 18.7-2. ‘We start by making a mass balance for species A on a spherical shell of thickness r within a single catalyst particle: Nal Am? = Naleart Smtr + ANP + Ry 4dr = 0 87-1) Here Nis the number of moles of A passing in the r direction through an imaginary spherical surface at a distance r from the center of the sphere. The source term Ry * 407’Ar is the molar rate of production of A by chemical reaction in the shell of thickness Ar. Dividing by 47 Ar and letting Ar = 0 gives CPN edlevar = FNad) mm FNadlaae = Jim, — FR, as72 or, using the definition of the first derivative, £ Ng) = PRe «i873 ar This limiting process is clearly in conflict with the fact that the porous medium is grana- lar rather than continuous. Consequently, in Eq. 18.7-3 the symbols N,, and R, cannot be interpreted as quantities having a meaningful value at a point. Rather we have to inter pret them as quantities averaged over a small neighborhood of the point in question—a neighbothood small with respect to the dimension R, but large with respect to the di mensions of the passages within the porous particle. BW. Thiele Fed. Eng. Cherm, 31, 916-620 (1999). °R, Aris, Chem, Eng. Si, 6, 265-268 (1957). 8 Wheeler, Afomices in Catalysis, Academic Press, New York (1950), Vol.3, pp. 2504326.§18.7 Diffusion and Chemical Renction Inside a Porous Catalyst 565 We now define an “effective diffusivity” for “ns Ain the porous medium by = 9, (187-4) ie ts in which c, is the concentration of the gas A contained within the pores, The effective diffusivity %,, must be measured experimentally. It depends generally on pressure and temperature and also on the catalyst pore structure. The actual mechanism for diffusion in pores is complex, since the pore dimensions may be smaller than the mean free path of the diffusing molecules. We do not belabor the question of mechanism here but as sume only that Bq. 18.74 can adequately represent the diffusion process (see §24.6). When the preceding expression is inserted into Fg. 18.73, we get, for constant diffusivity «18,7-5) We now consider the situation where species A disappears according to a first-order chemical reaction on the catalytic surfaces that form all or part of the “walls” of the winding passages. Let « be the available catalytic surface per unit volume (of solids + voids). Then Ry = —Kjac,, and Eq, 18.7-5 becomes (see Eq. C.1-6) den _ yy, Ot (4 a Kacy, (18.7-6) ‘This equation is to be solved with the boundary conditions that ¢, = cya at that cy is finite at r = 0. Equations containing the operator (1/P)(d/dnlr°Cd/sdn] can frequently be solved by using a "standard trick’—namely, a change of variable ¢4/ax = (1/r)f(r). The equation Rand for fr) is then & (8) ela 87-7) This is a standard second-order differential equation, which can be solved in terms of ex- ponentials or hyperbolic functions. When it is solved and the result divided by r we get the following solution of Eq, 18.7-6 in terms of hyperbolic functions (see §C.5): z Sos cht Pann fs (187-8) Von Application of the boundary conditions gives finally en _{R) sinhVEa/D,r ~=|-)—— = (187-9) fan (eres R “8 In studies on chemical kinetics and catalysis one is frequently interested in the molar flux Ni,q of the molar flow Wisp at the surface r= R: Was = 4aR'N gx = ~4e RD, a (18.710) When Eq. 187-9 s used in this expression, we get Wee = srtite(1 - [ie Rol oon fe x) 87-11) ‘This result gives the rate of conversion (in moles/'se0) of A to B in a single catalyst particle of radius R in terms of the parameters describing the diffusion and reaction processes.566 Chapter 18 Concentration Distributions in Solids and in Laminar Flow If the catalytically active surface were all exposed to the stream of concentration cas, then the species A would not have to diffuse through the pores to a reaction site. The molar rate of conversion would then be given by the product of the available surface and the surface reaction rate: Wary = GaRVal(—Keae) (18.7-12) Taking the ratio of the last two equations, we get War _ 3 a (d coth g - 1) 18.713) Wigs ge One in which @ = VEa/@,R is the Thiele modulus," encountered in §18.4. The quantity n, is called the effectiveness factor." It is the quantity by which Wg has to be multiplied to ac- count for the intraparticle diffusional resistance to the overall conversion process. For nonspherical catalyst particles, the foregoing results may be applied approxi- mately by reinterpreting K. We note that for a sphere of radius R the ratio of volume to external surface is R/3. For nonspherical particles, we redefine R in Eq, 18.7-13 as Rang = (2) (187-14) where Vpand Spare the volume and external sutface of a single catalyst particle. The ab- solute value of the conversion rate is then given approximately by [Wael © Voakievens (187-15) where m= 2, GA coth 3A ~ 1) (187-16) 3a in which the quantity A = VE%@7ai,(V9/S)) isa generalized modulus?* ‘The particular utility of the quantity A may be seen in Fig. 18.7-3. It is clear that when the exact theoretical expressions for 7, are plotted as functions of A, the curves 10 08 Flat particles os Lig einai partie Spherical particles oa a 03] 02| Fig. 18.7.3, Effectiveness factors for porous solid on catalysts of various shapes b¢ Os 08 10 2 8 $ 6 8 10 IR,Atis, Chem, Brg. Sci, 4 262-266 (19579). “0. A. Hougen and KM. Watson, Chemiza! Process Principles, Wiley, New York (1947), Part I, Chapter XIX. See also CPP Charts, by O, A, Hougen, K.M, Watson, nul R.A. Ragatz, Wiley, New York (1960), Fig. E.§188 Diffusion in a Three-Component Gas System 567 have common asymptotes for large and small A and do not differ from one another very much for intermediate values of A. Thus Fig. 18.7-3 provides a justification for the use of Eq. 18.7-16 to estimate 7, for nonspherical particles, §18.8 DIFFUSION IN A THREE-COMPONENT GAS SYSTEM Up to this point the systems we have discussed have been binary systems, or ones that could be approximated as two-component systems. To illustrate the setting up of multi- component diffusion problems for gases, we rework the initial evaporation problem of §18.2 when liquid water (species 1) is evaporating into air, regarded as a binary mixture of nitrogen (2) and oxygen (3) at 1 atm and 352K. We take the air-water interface to be at 2 =O and the top end of the diffusion tube to be at 2 = L. We consider the vapor pressure of water (o be known, so that x, is known at z = 0 (that is, x, = 341/760 = 0.449), and the mole fractions of all three gases are known at 2 = L: x, = 0.10, tx = 0.75, 2% = 0.15. The diffusion tube hasa length L = 11.2 cm. The conservation of mass leads, as in §18.2, to the following expressions: Nae az 0 =1,2,3 881) From this it may be concluded that the molar fluxes of the three species are all constants at steady state. Since species 2 and 3 are not moving, we conclude that Np, and Ns, are both zero, Next we need the expressions for the molar fluxes from Eq. 17.9-1. Since x + x) + x) = 1, we need only two of the three available equations, and we select the equations for species 2 and 3. Since N;, = 0 and Ns, = 0, these equations simplify considerably: a de hy dz By (18.82, 3) Note that the diffusivity %,, does not appear here, because there is no relative motion of species 2 and 3, These equations can be integrated from an arbitrary height to the top of the tube at L, to give for constant jg Ny f° day Nie a f de; f os zl a (188-4, 5) x Ny AL -2)\ coy NyAL — 2) = =o(- e ': mea } (18.8-6,7) and the mole fraction profile of water vapor in the diffusion column will be Nib = yal xy ex(- oD 188-8) When we apply the boundary condition at z = 0, we get xy =1-Xy oo( 18.89) 12 isa transcendental equation for N,.568 Chapter 18 Concentration Distributions in Solids and in Laminar Flow 10. nL, 2 13, 4, PROBLEMS 184.1 According to Reid, Prausnitz, and Poling,! y» = 0.364 em? /s and Dy. = 0.357 cm’/s at 352K and 1 atm, AL these conditions ¢ = 3.46 X 10° g-moles/cm’ To get a quick solu- tion to Eq. 18.8-9, we take both diffusivities to be equal” to 0.36 cm?/s. Then we get MLD ~ 09 188-10) ve(- 462 x 10 aa sax) O8e10 from which we find that N,, = 5.523 x 10-7 grmoles/em?«s. This can be used as a first guess in solving Eq. 18.8-9 more exactly, if desired. Then the entire profiles can be calcu- Tated from Eqs. 18.8-6 to 8. oas9 QUESTIONS FOR DISCUSSION What arguments are nsed in this chapter for eliminating Ny from Eq, 18.0-1? Suggest ways in which the diffusivity 2.5 could be measured by means of the examples in this chapter. Summarize possible sources of error. Jn what limit do the concentration carves in Fig, 18.2-1 become straight lines? Distinguish between homogeneous and heterogeneous reactions. Which ones are described by boundary conditions and which ones manifest themselves in the differential equations? Discuss the term “diffusion-controlled reaction.” What kind of “device” would you suggest in the first sentence of §18.2 for maintaining the level of the interface constant? Why is the left-hand term in Bq, 18,2-15 called the “evaporation rate”? Explain carefully how Eq. 182-19 is set up. Criticize Example 18,2-3. To what extent is it“just a schoolbook problem”? What do you learn from the problem? In what sense can the quantity N,, in Eq, 18.3.9 be interpreted as a local rate of chemical reac- tion? How does the size of a bubble change as it moves upward in a liquid? In what connection have you encountered Eq. 185-11 before? What happens if you try to solve Eq. 18.7-8 by using exponentials instead of hyperbolic func tions? How can we make the simpler choice ahead of time? Compare and contrast the systems discussed in $818.5 and 6 as regards the physical prob- Jems, tite mathematical methods used to solve them, and the final expressions for the molar fluxes, Evaporation rate. For the system shown in Fig. 18.2-1, what is the evaporation rate in g/hr of CCNO; (chloropicrin) into air at 25°C? Make the customary assumption that air is a “pure substance.” Total pressure 270 mm Hg Diffusivity (CCLNO,-air) 0088 ent?/s Vapor pressure of CCNO, 23.81 mm Hg, Distance from liquid level to top of tube Vide Density of CC1NO, 1.65 g/cm? Surface area of liquid exposed for evaporation 2.29 cm? Answer: 0.0189 g/he "RC. Reid, JM. Prauonitz, and B.B. Pong, The Properties of Gases and Liguis, Ath edition, MoGraw-Eil, New York (1987), p. 5 2 The scluign t temary difusion problems in which two ofthe binary diffusivities are equal was Aiscussed by HL Toor, AICHE Joural 3, 198-207 OSSD.Problems 569 + pReiden Fig. 18A.4. Schematic drawing of a wetted-wall Water film runs coluunn. —_ down the wall Film thickness 8 Surface concentration assumed equal to the saturation concentration [Chiorine-bearing Bas 18A.2. Sublimation of small iodine spheres in still air. A sphere of iodine, 1 cm in diameter, is 18A3. AS. 18A5. placed in still air at 40°C and 747 mm Hg pressure. At this temperature the vapor pressure of iodine is about 1.03 mm Hg, We want to determine the diffusivity of the iodine-ait system by measuring the sublimation rate. To help determine reasonable experimental conditions, (a) Estimate the diffusivity for the iodine-air system at the temperature and pressure given above, using the intermolecular force parameters in Table E.1 (b) Estimate the rate of sublimation, basing your calculations on Eq. 182-27. (Hint: Assume ry tobe very large.) ‘This method has been used for measuring the diffusivity, but it is open to question be- cause of the possible importance of free convection, Answer: (a) Bh, ae = 0.0888 cm?/s; (b) Wi, = 1.06 X 10-* g-mole/he Estimating the error in caleulating the absorption rate. What is the maxitmum possible error in computing the absorption rate from Eq, 18.5-18, ifthe solubility of A in B is known within 25% and the diffusivity of A in B is known within 15%? Assume that the geometric quanti- ties and the velocity are known very accurately. Chlorine absorption in a falling film (Fig, 184.4). Chlorine is being absorbed from a gas ina smail experimental wetted-wall tower as shown in the figure. The absorbing fluid is water, which is moving with an average velocity af 17.7 om/s. What is the absorption rate in g- moles /hr, if the liquid-phase diffusivity of the chlorine-water system is 1.26 x 10"* cm’/s, and if the saturation concentration of chlorine in water is 0.823 g chlorine per 100 g water (these ate the experimental values at 16°C). The dimensions of the column are given in the fig- lure. (Hint: Ignore the chemical reaction between chlorine and water.) Annsioer: 0.273 gemoles /hr Measurement of diffusivity by the point-source method (Fig. 18C.1).1 We wish to design a flow system to utilize the results of Problem 18C.1 for the measure of %,., The approaching "This is the most precise method yet developed for measurements of diffusivity at high temperatures, For a detailed description ofthe method, sce KE, Walker and A. A. Westenbery Che. Phys, 29, 139-1146, 1147-1153 (1958). Fora summary of measared values and comparisons with the ‘Chapman-Enskog theory, see R.M, Fristrom and A. A. Westenberg, Flame Structure, MeCiraw-Hill, New York 1965), Chapter XID.570 Chapter 18 Concentration Distributions in Solids and in Laminar Flow 18A6. 1A, 18B.1. stream of pure B will be directed vertically upward, and the gas composition will be mea- suted at several points along the z-axis. (@) Calculate the gas-injection rate W, in g-moles/s required to produce a mole fraction x< * 0.01 at a point 1 cm downstream of the source, in an idea! gaseous system at I atm and 800°C, if = 50.em/sand Fay ~ 5.cm'/s. {b) What is the maximum permissible error in the radial position of the gas-sampling probe, if the measured composition x, is to be within 1% of the centerline value? Determination of diffusivity for ether-air system, The following data on the evaporation of ethyl ether, with liquid density of 0.712 g/em’, have been tabulated by Jost. The data are for 2 tube of 6.16 mm diameter, a total pressure of 747 mm Hg, and a temperature of 22°C. Decrease of the ether level Time, in seconds, required (measured from the open for the indicated end of the tube), in mm decrease of level from 9 to 11 590 from 14 to 16 895 from 19 to 21 185 from 24 to 26 1480 from 34 to 36 2085 from 44 to 46 2655 ‘The molecular weight of ethyl ether is 74.12, and its vapor pressure at 22°C is 480 nm Hg. It may be assumed that the ether concentration at the open end of the tube is zero. Jost has given a value of 3i,, for the ether-air system of 0.0786 cm? /s at (°C and 761) mm Fig (@) Use the evaporation data to find %,. at 747 mm Hg and 22°C, assuming that the arith- metic average ges-column lengths may be used for 2 ~ 2; in Fig, 182-1. Assume further that the ether-air mixture is ideal and that the diffusion can be regarded as binary. (b) Convert the result to &4g at 760 mm Hig and 0°C using Eq, 17-2-1 Mass flux from a circulating bubble. (a) Use Eq, 18.5-20 to estimate the rate of absorption of CO, (component A) from a carbon dioxide bubble 0.5 cm in diameter rising through pure water (component B) at 18°C and at a pressure of 1 atm, The following data’ may be used: Si4g = 146 X10 em?/s, cq, = O41 g- mole/fiter, », = 22 m/s. (b) Recalculate the rate of absorption, using the experimental results of Hammerton and Gar- ner! who obtained a surface-averaged k. of 117 em/ht (see Eq, 18.1-2). Answers: (a) 1.17 X 10™* g-mol/em? 5; (b) 1.33 X 10" g-mol/em? s, Diffusion through a stagnant filim—altemate derivation. In $18.2 an expression for the evaporation rate was obtained in Eq, 182-14 by differentiating the concentration profile found a few lines before, Show that the same results may be derived without finding the con- centration profile, Note that at steady state, NV is a constant according to Eq. 18.2-3. Then Eq, 182-1 can be integrated directly to get Bg, 18.2-14. = W. Jest, Difusion, Academic Pres, New York (1952), pp. 411-413. 1G. Tammann and V. Jessen, Z.auorg.allzem. Chen, 179, 125-144 C929); FH. Garner ava, D. Fasumeston, Chem. Eng. Sci, 3, 11 (1958) 'D, Hamunerton and FH. Garner, Teams, [tt. Chem. Engrs. London, 32, S18-$24 1959)188.2. 18B3. 18B.4, 18B.5. Problems 571 Error in neglecting the convection term in evaporation. (a), Rework the problem in the text in §18.2 by neglecting the term x,(Nj, + No) in Bg. 18.0-1 Show that this leads to Nae EE beg = ad (1882-1) This is a useful approximation if A is present only in very low concentrations. (b) Obtain the result in (a) from iq, 182-14 by making the appropriate approximation. (0) What efror is made in the determination of gin Example 18.2-2 if the result in (a) is used? Answer: 0.78% Effect of mass transfer rate on the concentration profiles. {a) Combine the result in Eg. 18.2-11 with that in Eq. 18.2-14 to get eo 5”) as. (b) Obtain the same result by integrating Eq. 18.2-1 directly, using the fact that N,, is constant, (© Note what happens when the mass transfer rate becomes small. Expand Eq, 18B.3-1 in a Taylor series and keep two terms only, as is appropriate for small Nj. What happens to the slightly curved lines in Fig, 18.2-1 when Nj, is very small? Absorption with chemical reaction. (a) Rework the problem discussed in the text in §18.4, but take z ~ 0 to be the bottom of the beaker and 2 ~ Lat the gas-liquid interface. (b) In solving Fg. 18.4-7, we took the solution to be of the sum of two hyperbolic functions. Try solving the problem by using the equally valid solution I’ = C, exp(6e) + Cy expt—dd). (© Inwhat way do the results in Eqs. 184-10 and 12 simplify for very large L? For very small 2 Interpret the results physically. Absorption of chlorine by cyclohexene, Chlorine can be absorbed from Cly-air mixtures by olefins dissolved in CCI,. It was found? that the reaction of Cl, with cyclohexene (C,H) is second arder with respect to Cl; and zero order with respect to CH). Hence the rate of disap- pearance of Cl, per unit volume is kc} (where A designates CI). Rework the problem of §184 where B is a C.Hh-CCl, mixture, assuming that the diffu sion can be treated as pscudobinary. Assume that the air is essentially insoluble in the CHCl mixture. Let the liquid phase be sufficiently deep that L can be taken to be infinite, (a) Show that the concentration profile is given by oo fy, [Re cA oe, 4b) Obtain an expression for the rate af absorption of Clz by the liquid, (©) Suppose that a substance 4 dissolves in and reacts with substance B so that the rate of dis- appearance of A per unit volume is some arbitrary function of the concentration, ftc,). Show that the rate of absorption of A is given by eaten 1sB5-2) | (18B.5-1) Use this result to check the result of (b). § GH. Roper, Chem. Eng, Sci, 2, 18-31, 247-259 (1953).572 Chapter 18 Concentration Distributions in Solids and in Laminar Flow 188.6, 188.7. Fig. 18B.6. Sketch of a two- bulb apparatus for measuring, ‘0s diffusivities. The stirrers in ‘the two bulbs maintain uni form concentration in the bulbs. \ Volume ¥ Volume ¥ Mole action of Ain, Entire gaseous Mole ration of Ain Eftbubissy1-s1 “systemisat right bulbis x10 constant pandT ‘Two-bulb experiment for measuring gas diffusivity —quasi-steady-state analysis® (Fig. 188.6. (One way of measuring, gas diffusivities is by means of a two-bulb experiment. The left bulb and the tube from z = —[ to z ~ O ate filled with gas A. The right bulb and the tube from z = 0 to “+L ate filled with gas B. At time f = 0 the stopcock is opened, and diffusion begins; then the concentrations of 4 in the two well-stimed bulbs change. One measures x as a function of tine, and from this deduces 9... We wish to derive the equations describing the diffusion, Since the bulbs are large compared with the tube, xi and xj change very slowly with time, Hence the diffusion in the tube can be treated as a quasi-steady-state problem, with the boundary conditions that x= xj and z= ~L,and that x, = xj atz = +L. (a) Write a molar balance on A ovet a segment Az of the tube (of cross-sectional area $), and. show that Ni. = C;,a constant. (b) Show that Eq, 18.0-1 simplifies, for this problem, to Nao= ~eO yg (1886-1) ae (© Integrate this equation, using (a). Call the constant of integration C (4) Evaluate the constant by requiring that x, =x) atz = +L. fe) Next set.x, =a, (orl ~ xj) atz = ~L, and solve for N,, to get finally _, Bae ~ ae (18862) {0 Make a mass balance on substance A over the right bulb to obtain (18B.6-3) 18B6-4) (h) Suggest a method of plotting the experimental data to evaluate ag, Diffusion from a suspended droplet (Fig. 18.2-4), A droplet of liquid A, of radius r, is sus pended in a stream af gas B. We postulate that there is a spherical stagnant gas film of radius : surrounding the droplet. The concentration of A in the gas phase is x4, at r = r, and x,y at the outer edge of the film, (@) By a shell balance, show that for steady-state diffusion 7°N, is a constant within the gas film, and set the constant equal to rN, the value at the droplet surface. (by Show that Bq. 18.0-1 and the result in (a) lead to the following equation for x, Bye adty ae (1887-4) TN = °5.P.S. Andrew, Chem. Eng. Sci, 4, 269-272 (1855).Problems 573 Fig. 188.8, Diffusion of helium through pyrex tubing, iatural gas xen ‘The length of the tubing is L containing ~ helium’ (©) Integrate this equation between the limits 7, and r, to get 2 (hn aan) ‘What is the limit of this expression when r—* 2°? 188. Method for separating helium from natural gas (Fig. 188.8). Pyrex glass is almost imperme= able to all gases but helium. For example, the diffusivity of He through pyrex is about 25 times the diffusivity of H, through pyrex, hydrogen being the closest “competitor” in the dif- fusion process. This fact suggests that a method for separating helium from natural gas could be based on the relative diffusion rates through pyrex.” Suppose a natural gas mixture is contained in a pyrex tube with dimensions shown in the figure. Obtain an expression for the rate at which helium will “leak” out of the tube, in terms the diffusivity of helium through pyrex, the interfacial concentrations of the helium in the pyrex, and the dimensions of the tube. ‘Brie oyeer(Criet Crea) in R/RD 18B.9._ Rate of leachng (Fig, 188.9). In studying the rate of leaching of a substance A from solid par- ticles by a solvent B, we may postulate that the rate-controlling step is the diffusion of A from the particle surface through a stagnant liquid film thickness 3 out into the main stream. The ‘molar solubility of A in B isa, and the concentration in the main stream is (a) Obtaina differential equation for ¢, as a function of z by making a mass balance on A over a thin slab of thickness Az, Assume that $4 is constant and that A is only slightly soluble in B. Neglect the curvature of the particle. Answer: Wa, = 2a Solid particle containing, ig. 18B.9, Leaching of A by diffusion into a stagnant liquid film of B. Scientific American, 199, 52 (1958) describes briefly the method developed by KB. McAfee of Bell “Telephone Laboratories.574 Chapter 18 Concentration Distributions in Solids and in Laminar Flow 18B.10 188.11. (b) Show that, in the absence of chemical reaction in the liquid phase, the concentration pro file is linear. (© Show that the rate of leaching is given by Nae = BaalCan — 650)/8 (188.9) Constant-evaporating mixtures. Toluene (1) and ethanol (2) are evaporating at z = 0 ina vertical tube, ftom a binary liquid mixture of uniform composition x, through stagnant nitro- gen (3), with pure nitrogen at the top. The unequal diffusivities of toluene and ethanol through nitrogen shift the relative evaporation rates in favor of ethanol. Analyze this effect for an isothermal system at 60 F and 760 mm Hg total pressure, if the predicted’ diffusivities at 60F Fare citys = 1.53 X 10 %,c5ty, = 2.98 10°, ey = 4.68 X 106 g-moles/cm «5, {a) Use the Maxwell Stefan equations to obtain the steady-state vapor-phase mole fraction pro- files y,(2) in terms of the molar fluxes N,, i this ternary system. The molar fluxes are known to bbe constants from the equations of continuity for the thtee species. Since nitrogen has a negligible solubility in the liquid at the conditions given, Ns. = 0. As boundary conditions, set, = y= Oat £= Land lety; = yjpand y= wyat z = O; the laiter values remain to be determined. Show that An, wee (Cy DY), wn 4S wa) y= 7? oe (E+ sPa)e +€ asp.t0a Me NN. Me Nu Nw ante Ge Sta Degli (sB102 (©) A constant evaporating liquid mixtures one whose composition isthe same as that ofthe evaporated material, that is, for which Nj,/(N,: + No) = x. Use the results of part (a) along ‘with the equilibrium data in the table below to calculate the constant-evaporating liquid com- position ata total pressure of 760 mm Hg. In the table, row I gives liquid-phase compositions. Row II gives vapor-phase compositions in two-component experiments; these are expressed as nitrogen-free values v,/(y, + y,) for the ternary system. Row Ill gives the sum of the partial pressures of toltiene and ethanol — + koox 0.096 0.155 0233 oars 0373 Te w/in tw oaa7 0.198 0202 0.256 0277 TM: py + ppm Hg) | 388 397 397 395 390 A suggested strategy for the calculation is as follows: () guess a liquid composition 2); (i) cal culate Yip. Yop, and yc using lines 2 and 3 of the table; tii) calculate A from Eq, 18B.10-1, with 2 = 0; (iv) use the result of ti to calculate LN, LB, LC, and LD, and finally wy (0) for assumed values of LN},;(v) interpolate the results of iv to y, (0) = yy to obtain the correct IN... and Li for the guessed x), Repeat steps i-v with improved guesses for x; until Ni-/(Ni, + Nad cor verges to x,. The final x, is the constant evaporating composition Diffusion with fast second-order reaction (Figs. 18.2-2 and 18B.11), A solid A is dissolving ina flowing liquid stream $ in a steady-state, isothermal flow system. Assume in accordance with the film modet that the surface of A is covered with a stagnant liquid film of thickness 5 and that the liquid outside the film is well mixed (see Fig, 18.2-2) (2) Develop an expression for the rate of dissolution of A into the liquid if the concentration of Ain the main liquid stream is negligible. () Develop a corresponding expression for the dissolution rate if the liquid contains a sub- stance B, which, at the plane x = 6, reacts instantaneously and irreversibly with A: 4 + B—> P.(An example of such a system is the dissolution of benzoic acid in an aqueous NaOH solu- tion.) The main liquid stream consists primarily of B and S, with B at a mole fraction of ao. *L. Monchick and EA. Mason, Chem. Phys, 35, 1676-1697 (1961), with 6 ead as &y, in Table {V3 E.A. Mason and L Monchick, J. Chem. Phys, 6, 2746-2787 (1862) L,S. Tee, Gotoh, and W. E, Stewart, Ind. Eg. Chem. Fonda, 8, 356-362 (1986).18B.12. Problems 575 Fig, 188.11. Concentration profiles for dif- reaction (outer ege plane) ofstaprant fusion with rapid second-order reaction. liquid film) ‘The concentration of product P neglected. (Hint: Tt is necessary to recognize that species A and B both diffuse toward a thin reaction Zone as shown in Fig. 188.11.) (222) see) 14 6 saSDas A sectioned-cell experiment? for measuring gas-phase diffusivity (Fig, 188.12). Liquid A is allowed to evaporate through a stagnant ges B at 741 mm Hg total pressure and 25°C. At that temperature, the vapor pressure of A is known to be 600 mm Hg. After steady state has been Ansivers: (a) Na Constant. Temperature bath sete | Gas manifold with stream of pre gos armani Seven. « seen [-O- Diffusion cll | vf diameter? em. | tiandetorvotating Dison ce Four rotatae— Anat cell Rototed = | ectons sections sechons Thoce stationary Seationary eetione sections as eight of biguid suriace Taguid 4 # ! added here Somaintoin constant Aiquid level Liquid seservoir guid reserve © © Fig. 188-12. A sectioned-cell experiment for measuring gas diffusivities. (a) Cell configura- tion during the approach to steady-state. (b) Cell configuration for gas sampling at the end of the experiment. 9]. Crosby, Experiments in Transport Phenomena, Wiley, New York (1961), Experiment 10.3576 Chapter 18 Concentration Distributions in Solids and in Laminar Flow 188.13, attained, the cylindrical cohimn of gas is divided into sections as shown. Fora 4-section appa- ratus with total height 4.22 cm, the analysis of the gas samples thus obtained gives the follow- ing results @—2)inem Bottom Top of. Mole Section of section section _ fraction of A I 010 110 0.757 0 1.10 240 0.641 mL 2.10 30 0.469 Vv 3.10 410 025 ‘The measured evaporation rate of A at steady state is 0.0274 g-moles/hi. [deal ges behavior may be assumed. (a) Verify the following expression for the concentration profile at steady state: _ Nutt ne 488.1211 (©) Plot the mole fraction x, in each cell versus the value of z at the midplane of the cell on semilogarithmic graph paper. Is a straight line obtained? What are the intercepts at 2, and 2? Interpret these results. (0) Use the concentration profile of Eq. 18B.12-1 to find analytical expressions for the average concentrations in each section of the tbe. (d) Find the best value of 2g from this experiment, Answer: (A) 0.155 ett /s ‘Tarnishing of metal surfaces. In the oxidation of most metals (excluding the alkali and afka- line-earth metals) the volume of oxide produced is greater than that af the metal consumed. ‘This oxide thus tends to form a compact film, effectively insulating the oxygen and metal from each other. For the derivations that follow, it may be assumed that, (@) For oxidation to proceed, oxygen must diffuse through the oxide film and that this difft- sion follows Fick’s law. (by The free surface of the oxide film is saturated with oxygen from the surrounding air (© Once the film of oxide has become reasonably thick, the oxidation becomes diffusion con- ‘rolled; that is, the dissolved oxygen concentration is essentially zero at the oxide-metal surface, (d) The rate of change of dissolved oxygen content of the film is small compared to the rate of reaction, That is, quasi-steady state conditions may be assumed. (@) The reaction involved is xO; + M— MO, We wish to develop an expression for rate of tarnishing in terms of oxygen diffusivity ‘through the oxide film, the densities of the metal and its oxide, and the stoichiometry of the reaction. Let co be the solubility of oxygen in the film, cy the molar density of the film, and 2, the thickness ofthe film. Show that the film thickness is [Eioonot a Hot fo (188.13-1) ‘This result, the socalled “quadratic law,” gives a satisfactory empirical coreation for @ num ber of oxidation and other tarnishing reactions. Most such reactions aze, however, much moze complex than the mechanism given above." “G, Tammann, Z.anorg. allem. Chemie, 124, 25-35 (1922) © W Jost, Diffusion, Acrdemie Press, New York (1952), Chapter IX. Fara discussion ofthe oxidation of silicon, see R. Ghez, A Primer of Diffusion Probleme, Wiley, New York (1988), 823,18.14, 18B.15. Problems 577 Fig. 198.14, Side view of a disk-shaped «catalyst particie, Effectiveness factors for thin disks (Fig. 188.19). Consider porous catalyst particles in the shape of thin disks, such that the surface area of the edge of the disk is small in comparison ‘with that of the two circular faces. Apply the method of §187 to show that the steady-state concentration profile is cy _ cosh VRa/Gh,z to- oaPaee 88.141) 1 coshVRa/Sg where z and b are described in the figure. Show that the total mass transfer rate at the surfaces 2 bis QnRicy Zsa tanh Ab (18B.14-2) in which A = Vkja/9q. Show that, ifthe disk is sliced parallel to the xy-plane into 1 slices, the total mass transfer rate becomes [WH | = 2aR%e ho tanbiAb/my (188.143) ‘Obtain the expression for the effectiveness factor by taking the limit mm (18.144) Express this result in terms of the parameter A defined in $18.6. Diffasion and heterogeneous reaction in a slender cylindrical tube with a closed end (Fig, 18B.15). A slender cylindrical pore of length I, cross-sectional area $, and perimeter P, is in contact at its open end with a large body of wellemixed uid, consisting of species A and B. Species 4, a minor constituent of this fluid, disappears into the pore, diffuses in the z direc- tion and reacts on its walls. The rate of this reaction may be expressed a5 (11 1.) accae = fle); that is, at the wall the mass flux normal to the surface is some function of the mass fraction, Ome OF A in the fluid adjacent to the solid surface. The mass fraction ws. depends on z, the dis- tance from the inlet, Because A is present in low concentration, the fluid temperature and density may be considered constant, and the diffuston flux is adequately described by j= —/2,s¥irg, ag Side View End View Well-mixed ‘aid Fig. 18B.15. (0) Diffu- siom and heterogeneous reaction ina long, non- circular cylinder. (6) Re- gion of thickness Az over which the mass balance is made.578 Chapter 18 Concentration Distributions in Solids and in Laminar Flow 198.16. 188.17. 188.18, 188.19. “where the diffusivity may be regarded as a constant. Because the pore is long compared to its lateral dimension, concentration gradients in the lateral directions may be neglected. Note the similarity with the problem discussed in $10.7. (2) Show by means of a shell balance that, at steady state, dita P Ge = Erte (188.1541) (b) Show that the steady-state mass average velocity v, is zero for this system, (c) Substitute the appropriate form of Fick’s law into Eq. 18.15-1, and integrate the resulting differential equation for the special case that f(eac) = K'e,q, To obtain a boundary condition atz = L neglect the rate of reaction on the closed end of the cylinder; why is this a reasonable approximation? (d} Develop an expression for the total rate w., of disappearance of A in the cylinder. (@) Compare the results of parts (©) and (@) with those of §10.7 both from the standpoint of Ansivers: (05 Effect of temperature and pressure on evaporation rate. (a) In $18.2 what is the effect of a change of temperature and pressure on the quantity x4? (b) IF the pressure is doubled, how is the evaporation rate in Eq. 18.2-14 affected? (©) How does the evaporation rate change when the system temperature is raised from Tto 7"? Reaction rates in large and small particles. (2) Oblain the following limits for Eq, 18.7-11; LR WK Reng gB.17-1) (mR)? cng (188.172) Roo: Roo: War Interpret these results physically. {b) Obtain the corresponding asymptotes for the system discussed in Problem 185.14. Com pare them with the results in (a). Evaporation rate for small mole fraction of the volatile liquid. In Eq, 18.2-15, expand : y Gr. ts txa)l9 1-x in a Taylor series appropriate for small mole fractions of A. First rewrite the logarithm of the quotient as the difference of the logarithms. Then expand In(1 ~ xq)) and In(1 ~ x.) in Taylor series about x4 = 1 and x7 = 1, respectively. Verily that Eq, 18.2-16 is correct. (286.18-1) Oxygen uptake by a bacterial aggregate. Under suitable circumstances the rate of oxygen metabolism by bacterial cells is very nearly zero order with respect to oxygen concentration We examine such a case here and focus our attention on a spherical aggregate of cells, which has a radius R. We wish to determine the total rate of oxygen uptake by the aggregate as a function of aggregate size, oxygen mass concentration pat the aggregate surface, the meta- bolic activity of the cells, and the diffusional behavior of the oxygen. For simplicity we con- sider the aggregate to be homogeneous. We then approximate the metabolic rate by an effective volumetric reaction rate ro, = —ky and the diffusional behavior by Fick's law, with an effective pseudobinary diffusivity 20,4. Because the solubility of oxygen is very low in this, system, both convective oxygen transport ard transient effects may be neglected. © J, A. Mueller, W.C, Boyle, and. N. Lightfoot, Biotechnol. aud Bicengr., 10, 331-258 (1968).8c. Problems 579 (a) Show by means of a shell mass balance that the quasi-steady-state oxygen concentration profile is described by the differential equation BaelZ) where x = po,/2» & = r/R, and N= KR?/piom (b) There may be an oxygen-free core in the aggregate, if Nis sufficiently lange, such that x = 0 for € < , Write sufficient boundary conditions to integrate Eq. 188.19-1 for this situation. To do this, it must be recognized that both x and dy/dé are zero at € = é. What is the physical significance of this last statement? (© Perform the integration of Eq, 188.19-1 and show how é may be determined. (4) Sketch the total oxygen uptake rate and f4 as functions of N, and discuss the possibility that no oxygen-free core exists, Nae Np(l a or Xe(l 9 3e¢,(1_3)_ a-e+ @ Roe Diffusion from a point source in a moving stream (Fig. 18C.1). A stream of fluid B in lami- nar motion has a uniform velocity v, At some point in the stream (taken to be the origin of coordinates) species A is injected at a small rate W, g-moles/s. This rate is assumed to be suf ficiently small that the mass average velocity will not deviate appreciably from ve. Species A is swept downstream (in the z direction), and at the same time it diffuses both axially and radially. (a) Show that a steady-state mass balance on species A over the indicated ring-shaped ele- ‘ment leads to the following partial differential equation if Zi,.is assumed to be constant: (18B.19-1) Ansiver (0) y= 1 tion of N from 1) for & = & = 0, where é; is determined as.a func- fea) | Bey sofia] mem (&) Show that Eq, 18C.1-1 can also be written as 2%, Ms) og [raf eds) Pea, a2 Pea ofe+ 8) af 12 (6 me woe casc.a-2) inwhich # = 99 +22 velocity “origin of coordinates placed at point of injection; Wi, s1oles ‘of Aare injected per second, Fig. 18C.1. Diffusion of A froma point source intoa stream of B that moves with ‘a uniform velocity.580 Chapter 18 Concentration Distributions in Solids and in Laminar Flow 18C2. 18C3, (©) Verify (lengthy! that the solution Ws _ - 3 > Faye NPL O/B — 2) aasc.sy satisfies the differential equation above. (d) Show further that the following boundary conditions are also satisfied by Eq, 18C.1-3 BCI: ats=%, c= asc.1-4) ay BC? 45830, Ansys 52> Wy 80.15) BC.3: atr=o, = 184-6) Explain the physical meaning of each of these boundary conditions. (e) Show how data on ca(r, 2) for given % and By may be plotted, when the preceding solu- tion applies, to give a straight line with slope %/22qq and intercept In 5. Diffusion and reaction in a partially impregnated catalyst. Consider a catalytic sphere like that in §187, except that the active ingredient of the catalyst is present only in the annular re- sgion between r= xR and r= R: In tegion 10 0 asc2a Such a situation may arise when the active ingredient is put on the particles after pelleting, as, is done for many commercial catalysts. (ab Integrate Bq, 18.7-6 separately for the active and inactive regions. Then apply the appro- priate boundary conditions to evaluate the integration constants, and solve for the concentra- lion profile in each region. Give qualitative sketches to illustrate the forms of the profies. (b) Evaluate Wax, the total molar rate of conversion of A in a single particle. Absorption rate in a falling film. The result in Eq. 185-18 may be obtained by an alternative procedure, (a) According to an overall mass balance on the film, the total moles of A transferred per unit time across the gas-liquid interface must be the same as the total molar rate of flow of A across the plane 2 = 1. The latter rate is calculated as follows: 1p? og en( f ‘ Explain this procedure carefully (b) Insert the solution for c, in Eg. 18.5-15 into the resutt of (a) to obtain: Wie Wintio SJ (fF pee OPP) AS) de NTE = Weantn 2 spat f (f° ple) au asc.) In the second line, the new variable v = x/V4DjaL/0q,, has been introduced. (© Change the order of integration in the double integral, to get 5 . < Wy = Whegy ae af. o-(f' du jae (18033) Explain by means of a carefully drawn sketch how the limits are chosen for the integrals The integrals may now be done analytically to get Eq. 18.5-18. ute) = Won f cabeude ascaay Jo18C4, 183. 18D. 18D.2. Problems 581 Estimation of the required length of an isothermal reactor (Fig. 18.5), Let abe the area of catalyst surface per unit volume of a packed-bed catalytic reactor and S be the cross-sectional area of the reactor. Suppose tha¢ the rate of mass flow through the reactor is (in Tby/bhr, for example) (a) Show that a steady-state mass balance on suibstance A over a length lof the reactor leads to Ee SaN My “dl w (aca) (b) Use the result of (a) and Eq. 18.3-9, with the assumptions of constant 6 and Sq, to obtain an expression for the reactor length L needed to convert an inlet stream of composition + (0) to an outlet stream of composition x). (Hint: Equation (P) of Table 17.8-1 may be useful.) weMy ) fe as Answers to) L = (2M iswer: (b) (a 1) (Mitay + My — x0)F In Steady-state evaporation. Jn a study of the evaporation of a mixture of methanol (1) and ace- tone @) through air (3), the concentration profiles of the three species in the tube were mea- sured" after attainment of steady state. In this situation, species 3 is not moving, and species Land 2 are diffusing upward, with the molar fiuxes N,, and N.z, measured in the experi- ments. The interfacial concentrations of these two species, x and 2, were also measured. In addition, the three binary diffusion coefficients were known. The interface was located at Gand the upper end of the diffusion tube was atz = L. fa) Show that the Maxwell-Stefan equation for species 3 can be solved to get xy = aye «8c 5-1) thug + Posy With vyg, ~ NyL/eSp, and ¢ = 2/L, {b) Next verify that the equation for species 2 can be solved to get in which Cay A-B age BB 8 + (et ~ ety asc5.2) where B= uz tz and C = v2 Yas (© Compare the above equations with the published results. (@) How well do Eqs. 18C5-1 and 2 fit the experimental data? Effectiveness factors for long cylinders. Derive the expression for 1), for long cylinders anal- ogous to Eq, 18.7-16, Neglect the diffusion through the ends of the cylinders. ey) modifi ions” Asien: my = Frapay Where and hare “modified Bessel functions Gas absorption in a falling film with chemical reaction. Rework the problem discussed in 5185 and described in Fig. 18.5-1, when gas A reacts with liquid B by a first-order irreversible chemical reaction in the liquid phase, with rate constant ky. Specifically, find the expression. for the total absorption zate analogous to that given in Eq. 18.5-18, Show that the result for ab- sorption with reaction properly simplifies to that for absorption without reaction. oy, Vi iu Answer: Wa = Went | in which 4 = KL/P.nu. DH. A. Wilson, Proc. Cam Phil. So., 12, 406-423 (1904). *R Carty and T. Schrodt, Ind, Eng. Chem, 14, 276-278 (1975).