Convective Mass Transfer and Mass Transfer Coefficient While preparing a cup of black coffee you take a spoonful of sugar and a little instant coffee in cup, add hot water and stir it with the spoon for about fifteen seconds. The coffee is ready for you. The crystals of sugar and the particles of coffee have dissolved quickly and the content is homogencous. But if the water is left unstirred, it takes more time for the sugar to dissolve and a much longer time for the liquid to be uniform in concentration. What message does this convey to us? Dissolution of a solid occurs slowly in a stagnant medium—molecular diffusion hhaving a major or sometimes the governing role in the process. But the rate of dissolution or, in other words, the rate of mass transfer increases dramatically if there is motion in the medium. This occurs because of convection. Mass transfer occurring under the influence of motion in a fluid medium is called ‘convective mass transfer’. We may try to visualize a simple mechanism of dissolution of the sugar crystals in a stirred cup of water. The water just in contact with a crystal (or, in other words, the liquid at the ‘solid-liquid interface") gets saturated with sugar almost instantly. Here ‘interface’ means the contact surface between the two phases. The dissolved sugar diffuses from the interface to the bulk liquid through a thin layer or ‘film’ of the solution adhering to the crystal. More sugar dissolves in the liquid simultancously at the interface. ‘The ‘thickness’ ofthe film that we visualize decreases if the stirring rate is more rapid. The more brisk the motion of the liquid is, the thinner is the film and the quicker is the rate of dissolution or the rate of mass transfer. The picture we visualize is schematically represented in Figure 3.1 In the previous chapter, we discussed the methods of calculation of the rate of mass transfer due to molecular diffusion. Now we are faced with another question: how to calculate the mass transfer rate when there is bulk motion, particularly turbulent motion, in the medium? And, how can we calculate the rate of “convective mass transfer"? In this chapter, we shall address these questions. Before we go further into this issue, let us discuss a little the modes of convective mass transfer. Similar to the case of convective heat transfer, convective mass transfer may be of two types—forced convection mass transfer and free convection (ot natural convection) mass transfer. In the former case, the motion or flow in the medium is caused by an external agency (for example, a pump, a blower, an agitator, etc.). In free convection mass transfer, on the other hhand, the motion in the medium is caused by a difference in the density. Let us elaborate this point ”Convective Mass Transfer and Mass Transfer Coefficient 75 sea_Mtic ofthe etal in contact ASS an eau Hee | Sucaceconcenraton,C, EEF stration cocctation inthe iid Ae soiule [ogee | Balkin guia solid (A) * ie gets Hypothetical Se] liquidfilm Figure3.1_ Visualization ofthe cissolution process ofa soli 4 little further. Consider a sugar cube suspended in a glass of water (Figure 3.2(a)]. Sugar dissolves, the concentration of the liquid near the surface increases and its density increases as a result, But the density of the liquid below the sugar cube is less. Thus, a density difference is created. This causes a motion in the medium in the form of free convection currents and thereby enhances the rate of mass transfer. However, if the sugar cube rests at the bottom, the density of the liquid at the bottom will be more than that at the top. No buoyancy force comes into play in this case and there is no free convection mass transfer [Figure 3.2(b)]. Free convection mass. transfer will not be dealt with in this text. The interested readers are referred to Skelland (1974). Suspending string Solution of higher density (because of ~ . higher concentration) Atal of Tower density Convection Solution of higher Aensity (because of ‘higher concentration) Solution of lower density (a) The sugar cube is kept suspended (b) A small sugar cube rests atthe bottom of ‘in water; natural convection occurs. a beaker, natural convection does not occur. Figure 3.2 (a) Natural convecton occurs due to adverse density gradient; o) higher densiy atthe bottom, no natural convection. ‘The mechanism is similar to that in the case of free convection heat transfer in which the density difference is caused by a temperature difference. In free convection mass transfer, it is the concentration difference that creates the density difference. Convective mass transfer is strongly influenced by the flow field, If the flow field is ‘well- defined’, it is possible to calculate the mass transfer rate by solving the differential equations obtained from mass and momentum balance. Typical examples are (i) the dissolution of a solid.76 Chapter 3 Convective Mass Transfer and Mass Transfer Coefficient coated on a flat plate in a fluid flowing over it, (i) the absorption of a solute gas in a laminar liquid film falling down a wall, etc. where the flow field is sufficiently well-defined. However, t0 a beginner of a course in mass transfer, the physical visualization and the mathematical exercise ‘may not appear simple and easy. For more complex situations such as the dissolution of a solid. in a mechanically stirred vessel, the absorption of a gas in a liquid flowing down the packing in a tower or ina liquid in an agitated tank, the theoretical calculation of the mass transfer rate becomes very difficult and may even be impossible since the flow field is complex and not properly understood. The concept of mass transfer coefficient has been introduced with a view to developing a simple and practically useful approach to the solution of such problems. The idea is similar to defining ‘heat transfer coefficient’ for expressing the rate of convective heat transfer between a surface and a fluid in motion. 3.1. THE MASS TRANSFER COEFFICIENT ‘The mass transfer coefficient is defined on the following “phenomenological basis’. Rate of mass transfer Concentration driving force (ic. the difference in concentration) Rate of mass transfer e Area of contact between the phases If W, is the rate of mass transfer (kmoV/s) of the solute A, AC, is the concentration driving force between two points, and a is the area of mass transfer, Wye aC, = Was ka dC, G1) where k, the proportionality constant, is a phenomenological coefficient* called the “mass transfer coefficient’. If Ny is the molar flux (expressed as kmol/m-s, say), we may write Wy = aNq= aC, = Na = KACy Na molar flux. AC, ~ concentration driving foree > Mass transfer coefficient, &. For the purpose of comparison, we may recall the definition of heat transfer coeffi heat lux Heat transfer coefficient, h = ————__heat flux —__ Temperature driving force, AT ‘The inverse of mass transfer coefficient is a measure of the ‘mass transfer resistance’. If the driving force is expressed as the difference in concentration (kmol/m’, say), the unit of mass transfer coefficient is m/s (or cm/s, fU/s, etc. which is the same as the unit of velocity). If the mass transfer coefficient is expressed as the ratio of the local flux and the local driving force, it is called the ‘local mass transfer coefficient’. When it is expressed as the ratio of the average flux (over a surface) and the average driving force, it is known as the‘average mass transfer coefficient’. + The word ‘phenomenological’ implies thatthe definition of mass transfer coefficient (A) given by Eq. (3.1) ‘or (3.2) is based on the ‘phenomenon’, that k, is proportional to the driving force AC, and area a,Types of Mass Transfer Coefficients 77 3.2. TYPES OF MASS TRANSFER COEFFICIENTS Convective mass transfer ean occur in a gas or a liquid medium. A few choices of the driving force (difference in concentration, partial pressure, mole fraction, etc.) are available unlike in the case of heat transfer where the temperature difference is the only driving force. Different types of mass transfer coefficients have been defined depending upon: (i) whether mass transfer cceurs inthe gas phase or inthe liquid phase, (ii) the choice of the driving force, and (il) whether itis a case of diffusion of A through non-diffusing B or a case of counterdiffusion. Convective heat transfer is often visualized to occur through a “stagnant film” adhering to the surface. The transport of heat through the film is assumed to occur purely by conduction In the study of mass transfer too, this concept is frequently used. This will be discussed in more )(bar) /(kmol)(K)] (300 K) (1.1556 bar) = 24x 10 m= [024mm 3.2.3 Typical Magnitudes of Mass Transfer Coefficients and Film Thickness It will be relevant here to mention the ‘order of magnitude’* of gas- and liquid-phase mass transfer coefficients occurring in a typical separation equipment, Gas-phase mass transfer coefficient, k. ~ 10? m/s; film thickness, 8 1 mm Liquid-phase mass transfer coefficient, ky, ~ 10°° mvs; film thickness, 6 ~ 0.1. mm ‘Once these typical values are noted, itis rather easy to determine the orders of magnitude of other types of mass transfer coefficients. For example, at T'= 298 K and ata ‘low concentration’, 107 mis = (0.08317 [(m? bar)kmot) (K)] 298K) kg ~ 4 x 10-4 kmol/em?y(s)(Ap, bat) "The concept of ‘order of magnitude’ is sometimes used in the analysis of physical systems to identify quantities which are small enough © be neglected compared to other quantities. The boundary layer _pproximation of equations of motion i a common example of order of magnitude analysis. Here we use the term to imply an approximate range of value of a quantity. Thus if we say “cris of order B (sometimes writen 4s c¢~ B), we mean that cand B differ by less than a factor of ten,82__Chapter 3 Convective Mass Transfer and Mass Transfer Coefficient For liquid-phase mass transfer in a ‘dilute’ aqueous solution, 1k, = K,(Q/M) yy ~ (10 m/sy(1000/18 kmom?) => ky ~ 5% 10° kmotlim?\()(Ax) Since the liquid-phase diffusivity of common solutes, D ~ 10° m/s, liquid film thickness, 6~ Dik, => 6~ 10m or 0.1 mm ‘Typical values of gas-liquid mass transfer coefficients in industrial equipment are given in Chapter 5 (Table 5.1). 3.3 DIMENSIONLESS GROUPS IN MASS TRANSFER In many cases it is convenient to express the transport coefficients and other important parameters (such as the fluid properties, velocity, etc.) in terms of meaningful dimensionless groups. In the study of heat transfer, for example, the heat transfer coefficient h is often expressed in terms of the Nusselt number (Nu). The other important parameters and properties influencing the heat transfer coefficient are taken care of through two other dimensionless. ‘groups—namely the Reynolds number (Re) and the Prandtl number (Pr). Experimental forced. convection heat transfer data are frequently correlated as Nu = (Re, Pr) and the resulting. correlation may be used to estimate the heat transfer coefficient for any other set of process conditions and system parameters. The most important of such correlations is the Dittus-Boelter equation, A similar approach is followed in the study of mass transfer too. Here we have the two most. important dimensionless groups—the Sherwood number, Sh (which is the mass transfer analogue of the Nusselt number) and the Schmidt number, Se (which is the mass transfer analogue of the Prandtl number). In fact, the origin of Sh and Se can be traced to their analogy with Nu and Pr respectively. Let us discuss this analogy here. In heat transfer, the Nusselt number is convective heat flux Nu jjeat flux for conduction through a stagnant medium of thickness [for the same AT hat _ hl = Gipar = Fe = thermal conductivity] G.15) Similarly, in mass transfer, the Sherwood number is convective mass (molar) flux ‘mass (molar) ux for molecular diffusion through a stagnant medium oF thickness / under the driving force Ap, sh= If we consider the gas-phase mass transfer of A through a binary mixture of A and B (B is non- diffusing), Convective mass flux [see Eq. (3.3)] = ke Ape DysP. Mass flux due to molecular diffusion of A through non-iffusing B = rte PaDimensionless Groups in Mass Transfer__ 83 = Koby = oPaw RTE kl Paw Then. SO DygPIRT PP DigP Das P ere If we consider transport of A in a solution at a rather low concentration (gy = 1), Convective mass flux, Ny = k, AC, Diffusive flux of A through a stagnant liquid layer of thickness ! AC, 1 [from Eq. (2.60)] Then, the Sherwood number, Sh = —MtACa_ Ait = DygMDAC, ~ Dap en Here | is a ‘characteristic length’. The commonly used characteristic lengths are: for a sphere—diameter, d; for a cylinder—diameter, d; for a flat plate—distance from the leading. edge, x (say). Now let us turn to the Schmidt number which is the mass transfer analogue of the Prandtl number. We define the Prandtl number as momentum diffusivity __jt/p _ cpt Pr= “thermal diffusivity ~ kpc, ~ Analogously, we define the Schmidt number as |= Momentum diffusivity Wp | Mv * “molecular diffusivity ~ Dyy ~ PDyy Day G.18) ‘There are a few more dimensionless groups commonly used in the study of mass transfer. A list of the important groups along with their physical significances are given in Table 3.2. The heat transfer analogues ofthe groups are also given alongside for comparison. It will be useful to have look into the orders of magnitude of the above two dimensionless groups, Sh and Sc. Take the case of gas-phase mass transfer for flow past a sphere, | cm in diameter, at low partial pressure of the solute (i.e. Pau/P ~ 1). The Sherwood number and the Schmidt number may be found tobe Kod Pow _ (10 m/s)(10™ m) Das P 105 mis _ vv _ 10S m/s Das 10% m/s For liquid-phase mass transport in a similar geometry, = Fed 0 m/s) (10 = Sh~10 Se = Se~1 sno 8 | CO myo m) Sh~ 100 Das 10? m/s vy _ 10% mis sez = Wm = se ~ 1000 Das 10 m/s * The location of a point in the boundary layer is often indicated by the coordinates (x, y). These notations should not be confused with mole fractions,Table 32 Important dimensionless grupsinmass tans a ori eels eae Repeater fe MP = «eee Tresame Seiiaantr se = HS = X «Semen sy Oe a aa Stoned ner Sh A A a a Ey aE ami San ue: Su tes =H = HSE gemma “Tar * Rahsaan Pein oH = = Ts = a Fo oN ebun ate Suse» dex P= Tea Crt rw r= MARE Teme +The Sherwood number for gas-phase mass vanseris definedin Eq (816). The suffx Mor D refers to mass wansfr, Htc heat vanster. The Grashol numbers the analogue ‘of Reynolds number in ee convection mass tanser. Pr= Prandtl number, Nu= Nusselt number, (= characteris length: v= wp = momentum dius ‘thermal tusvty, p= dference in density because oa diferenceinconcentraton or temperature 8 TinjioaD TalsuOI|, SOR iD TafsUOL, SDH aap —¢ IRNCorrelations for the Convective Mass Transfer Coefficient _ 85 3.4 CORRELATIONS FOR THE CONVECTIVE MASS TRANSFER COEFFICIENT As stated before, we need to know the mass transfer coefficient for calculating the rate of convective mass transfer. In practice the mass transfer coefficient in a typical situation may be obtained from a suitable correlation if an experimental value is not available. A large number of ‘mass transfer correlations, covering a wide variety of practical situations and geometries, exist in the literature. These correlations have been developed by researchers on the basis of large volumes of experimental data. A few important correlations for the convective mass transfer coefficient in terms of dimensionless groups are listed in Table 3.3. A comprehensive compilation of mass transfer correlations covering a wide range of systems is available in Perry’s Handbook (7th ed., 1997, Ch. 5). Table33Corelation forthe mas transfer coeficientina few simple situations Description ange ofappleaton Coren Laninarfow trough acircularube Re <2,100 Sh = kdD= 162{ ReySoKan! “Turbuletfow trough atube 4000 < Re < 80000; 08 ean Putting these results in Eq. (3.21), y= Kp? W? (Dyg)' a kid _ op any ytep, yr gta x (a2N) (_w_)” = Day 7 KPH Ons) a= «(2) (6) = Sh = Ky(Re)“Sey” where a” = a+ 7 3.25) So the experimental mass transfer coefficient data can be correlated in the form of Eq. (3.25), which is identical to Eq. (3.20), in terms of dimensionless groups. If enough data on Sh, Re and. Se for various systems are available or collected, the values of the constants cr and a” can be determined by ‘nonlinear regression’. The resulting equation or correlation can be used for calculating the mass transfer coefficient in any other similar system. EXAMPLE 3.3 (Mass transfer in flow past a sphere) A naphthalene ball of 1 cm diameter is suspended in a stream of air flowing at a velocity of $ m/s at 45°C and 1 atm total pressure. Calculate the time required for its diameter (© be halved. The following correlation for the Sherwood number may be used. Sh = 2 + 0.6(Re)"*(Se)"* Sublimation pressure of naphthalene at 45°C is 0.8654 mm Hg. The other relevant physicochemical properties are: Dyy = 6.92 x 10° m/s; the density of naphthalene at 45°C 1140 kg/mm; r= 1-1 Kg/m?s ge 1.92 x 10° kgim-s, Assume that the ball retains the spherical shape all through. (Is this assumption correct?) Solution Naphthalene vapour(4) is transported through non-diffusing air(B). The mass transfer coefficient can be obtained from the given correlation. The Reynolds number and, therefore, the Sherwood number and the mass transfer coefficient depend upon the radius of the ball which decreases with. time. If r is the radius of the naphthalene ball at any time 4, the Reynolds number is 2rpv _ , 2M) ue 192 x10 Re = 5.73 x 105r (the maximum value of Re is 2864) Lu L923 108 PDajg — (11)(6.92 x 10°) ‘Schmidt number, Se = 522 > Sh = 2 + 0.6(Re) (Se)! = 2 + (0.65.73 x 105)52.522)"9 = 2 + 618(7)"*88 _Chapter 3 Convective Mass Transfer and Mass Transfer Coefficient Given: Temperature, T = 45°C = 318 K; since the vapour pressure of naphthalene is small, Pla = 1 Pow RT 2r __ ke (008317)(318)(27) PDs «92X10 Then, Sh 644 x 10%rke _ 2616x107 | 8085x107 r vr Now we have to relate the rate of change of the size of the ball with the mass transfer rate, At time 1, the rate of sublimation can be expressed as = 1644 x 10 frig = 2+ 6180)? => kg d(4. 50 p dr _(M ~ GFP) = Oe Mhele Pw) he ~F= (“)kcre @ where 'p = density of solid naphthalene at 45°C = 1140 kg/m? M = molecular weight = 128 as = Vapour pressure of naphthalene at 45°C = 0.8654 mm Hg = 0.001153 bar a= paral pressure of napialene inthe bulk ai =O Puting the values of various quantis in Bq. (), we have 4 (1) [2altat0” 85310 agsy “de ~ 140. r vr 3.3867 x10"' 1.0467 10% r vr In order to calculate the required time, the above equation has to be integrated from the initial radius, r; = 0.5 em (= 0.005 m) to the final radius, r/2 = 0.25 em (= 0.0025 m). 0.025 a 0.0925 33867 x 10 1.0467 x10* ois + r vr ‘The integral is of the form I tor put frex = drs 2adr => Is fre, make another substitution £= a+ Bx => x= (E- afor dx = de/p Is 2fieaniee ase) i eo as 2 384 saga? ang] Put the limits and calculate the integral, / = 13,820 s = [384h If there is no convection, i.e. if sublimation occurs in a stagnant medium by purely molecular diffusion, it can be shown (see Example 2.13) that it takes 76.9 h for the radius to be reduced, to half of the initial value [note that we get the same result by discarding the second term on the of Eq. (ii) above and integrating the function}. This again shows that convective mass transfer is a much faster process than transport by molecular diffusion alone.Turbulent or Bday Difusion 89 EXAMPLE 3.4 (Solid dissolution in an agitated vessel) A solution of KCrsO> is to be prepared by dissolving crystals of the solid in an agitated vessel. Twenty kilograms of the solid is charged into a vessel containing 500 kg water, The vessel is provided with a flat-bladed turbine that rotates at 120 rpm. Following data are given: density of the solid = 2690 kg/m?; average particle size (0.6 mm; solubility of the solid in water at the given temperature = 8 wt%e; average viscosity of the solution over the concentration range involved = 0.98 cP; density of the liquid = 1000 kg/m? (assumed constant); diffusivity of potassium dichromate in water = 1.5 x 10-° em*/s. Calculate the time required for complete dissolution of the solid. The following correlation for Sherwood number can be used to calculate the solid-liquid ‘mass transfer coefficient. hud She aon 2 + 0.44(Re,)"* (Se)995 where Nd3p Re, = stirrer Reynolds number = 42; se = particle diameter, 2 H pp“ P Hints: Let at any time t, the mass of the undissolved solid in the vessel be m kg. Mass of solid already dissolved = 20 ~ m. Mass of solution = 500 + (20 ~ m), volume = (520 ~ m)/1000 m®. Instantaneous concentration of the solution, C = (20 ~ m)/[(520 — m)/1000] kg/m’. Solubility, C, = 80 kg/m? Now we express the quantities k, and a (the total area of the particles atthe time, ¢) in terms, of m If do= 0.6 mm (the initial diameter of a particle), the number of particles = 6.574 x 10”. Mass of a single particle at the time ¢ = m/(6.574 X 10" instantaneous diameter, dy = 2211 x 104m)! area of a particle = md,” = (7){2.211 x 10-4(m)'9P;; total area of the particles, a = (10.1), Schmidt number, Se = b/pD = 653. Instantaneous Reynolds number, Re, = Nae = (0.09976)(m)"* Sh = 2 + (0.44)(0.09976 x m?)°5(653)"5 = 2 + 1.67m°* ky = (6.784 x 10-2 + 1.67m3%\ my "9 The rate of dissolution of the crystals can be written as dm Ge = Malls O = (6.784 x 10°92 + 1.67m?ym-"](10.1m) (s0 Numerical integration of the above equation (from m = 20 to m= 0) gives the time of dissolution. Ans. [75 min 3.5 TURBULENT OR EDDY DIFFUSION Turbulent motion of a fluid is characterized by rapid and highly irregular fluctuation of the velocity at any point in the fluid. Experimental evidences indicate that in a turbulent medium, tiny90 Chapter 3 Convective Mass Transfer and Mass Transfer Coefficient fluid clements move about randomly and are responsible for a high rate of transport of ‘momentum, heat or mass. These fluid elements, not necessarily of the same size, are called eddies. Itis visualized that eddies are continually formed and they break up by interaction among. themselves or even may disappear in the process. Eddies are, therefore, short-lived. Mass or heat transfer in a turbulent medium is a result of the mixing process caused by the ‘movement of the eddies. This is illustrated in Figure 3.4 for transport of a soluble substance from. 4 pipe wall to a flowing fluid. For the sake of simplicity it may be visualized that the effect of turbulence does not reach the wall. Transport of momentum, mass or heat occurs by diffusion through a thin layer, called the laminar sublayer, at the wall. Beyond this layer, transport occurs. predominantly by eddies. However, the motion of eddies and the phenomenon of eddy transport are not well-understood despite the extensive theoretical and experimental research done on these phenomena’ To simplify the analysis of the phenomenon, the physical laws of transport (of heat, ‘mass and momentum) are very often extended to the case of eddy transport by replacing the ‘molecular transport coefficient by an ‘eddy transport coefficient’ (also see Section 3.10). For example, in the case of mass transfer from the tube wall we can write Ep Ge 3.26) ee . Radial momentum oF 245% patil mass mate Movdment of eddies Figure 4 Momentum and mass ranspotin turbulent fw by eddy dtfusion, where Ep is the ‘eddy diffusivity’ of mass. This equation is identical to Eq. (2.9) except that Dag is replaced by Ep. The ‘total flux’ is, therefore, obtained by adding the fluxes due to molecular diffusion and eddy diffusion (Sherwood et al., 1975). acy Jasnat = ~(Dya + Eo) Fe 27) ‘The eddy diffusivity Ep is not as simple a quantity as the molecular diffusivity Dy. It depends upon the intensity of the local turbulence. The intensity of turbulence decreases as one approaches the wall. Therefore, the contribution of eddy diffusion towards the rate of mass transfer is substantially less than that by molecular diffusion very near the wall. The situation is just the opposite in the bulk of the turbulent medium. The phenomenon of eddy diffusion and a few models have been described by Deen (1998) * Advanced softwares for computational fluid dynamics (CFD) are now in routine use for the numerieal solution ‘of transport problems in many complicated geometries and flow situations.The Wetted-Wall Column 91 3.6 THE WETTED-WALL COLUMN [oo Liquid in Mass transfer from a flowing gas to a falling liquid-film (or vice f versa) is encountered in a number of practical situations (for example, gas absorption or stripping in a packed tower). The wetted-wall tower is a classical model experimental set-up for —_ ‘measuring the mass transfer coefficient. The mass transfer data Liquid obtained in such a column also throw light on the mechanism of A mass transfer in a system of similar geometry and flow conditions. The wetted-wall column is a vertical tube or pipe provided with an arrangement for liquid feeding and withdrawal (Figure 3.5). The gas flows up through the column and the liquid i o Ht olution Tube ‘wall Liquid film flows down as a film (which may have a calm or rippling surface depending upon the “film Reynolds number’; see Rosenhow et out ,, 1985) over a section of the column. Most ofthe experiments Gas inf reported on the wetted-wall column involve evaporation of pure liquids in flowing streams of air and other gases. In an experiment on evaporation of water in air, for example, the average mass flux can be calculated from the known area of the wetted section of the wall, the total rate of mass transfer (calculated from the flow rate of air and the moisture content atthe inlet and atthe exit of the column), and the average driving force. The driving force at any section is the difference between the vapour pressure of water atthe prevailing temperature and the partial pressure of ‘moisture inthe bulk ai. Since the area of gas-Liquid contact is known and the liquid-film is visible (it can be photographed), this apparatus has also been used for study of mans transfer in distillation and in chemically reactive gas-Hiquid systems. A large number of mass transfer correlations for the wetted-wall column is available. One such commonly used correlation is cited in Table 3.3. A simple calculation of mass transfer in a wetted-wall column is illustrated in Example 3.9. “The wetted-wall column has been used widely for determination of diffusivity ofa dissolved 228 in a liquid. Ifthe flow of the liquid is laminar and ripple-frec, the mass transfer coefficient can be obtained using the penetration theory (see Section 3.7.2). The diffusivity can be calculated from the experimentally measured rate of absorption and the area of gas-liquid contact (Davies et al. 1967) Figure 85 Schematic of a wetted: wall ecluma, 3.7 THEORIES OF MASS TRANSFER At the beginning of this chapter, we defined the mass transfer coefficient just as a phenomeno- logical quantity to be determined experimentally. No attempt was made to have an insight of the physical mechanism of convective mass transfer at a phase boundary. However, there are a ‘number of theories of mass transfer which aim at visualizing the mechanism and developing the expressions for the mass transfer coefficient theoretically. In fact, any such theory is based on. a conceptual model* for mass transfer. We shall now discuss the more important and simple TA model is an idealized picture of a physical phenomenon or of an object thal makes the phenomenon or the ‘object amenable to theoretical analysis using mathematical equations and tool.92 Chapter 3 Convective Mass Transfer and Mass Transfer Coefficient theories of mass transfer at a phase boundary. These theories are old and rather simple, Nev- ertheless, these are considered very important in the study of mass transfer, particularly when the mass transfer is accompanied by a chemical reaction, 3.7.1 The Theory (also called ‘Film Model’) Let us describe this theory (Whitman, 1923) through an illustration. We consider mass transfer from a solid surface to a flowing liquid. Even though the bulk liquid is in turbulent motion, the flow near the wall may be considered to be laminar. The concentration of the dissolved solid(A) will decrease from C,, at the solid-liquid interface" to Cy, at the bulk of the liquid. In reality the concentration profile will be very steep near the solid surface where the effect of turbulence is practically absent (Figure 3.6). Molecular diffusion is responsible for mass transfer near the wall while convection dominates a litle away from it. The film theory, however, visualizes a simpler picture. It is based on the following assumptions. (2) Mass transfer occurs by purely molecular diffusion through a stagnant fluid layer atthe phase boundary (here the wall isthe phase boundary). Beyond this film, the fluid is well- mixed having a concentration which is the same as that of the bulk fluid (i.e. Cao). (b) Mass transfer through the film occurs at steady stat. (©). The bulk flow term fie. (Ny + Np)C4IC, see Eq. (2-15)] in the expression for the Fick's law is small. So the flux can be writen as Ny = —Dyy(dCyldz). As stated before, thi is valid when: (i) the flux is low and (ii) the mass transfer occurs at low concentrations [or the mass transfer occurs by equimolar counterdiffusion]. For many practical situations this assumption is satisfactory. Filmy thickness, Wallor phase boundary of interfice _-,_ [-Fulmibickness, Stagnan | +3 —+! Well-mixed True concentration profile A> foam (ly / too leas SMES cu je ed. Fe | Consett fie nthe poetic sagan ln (0) Me stagnant i rr C7 eon ath itertce nerfs Cee. he bal (©) Concentration profiles ofthe solute near the phase boundary (Bindicates the edge of the film) Figure 3.6 The stagnant fim and th concetraton roles. The underlying concept of the film theory of mass transfer is analogous to that of heat transfer (the heat transfer coefficient is frequently called the “film coefficient’). Figure 3.6 shows. "Tcis generally assumed that equilibrium prevails a the interface or the phase boundary. So, Ca, is, i fac, the Solubility of the solid inthe liquid.Theories of Mass Transfer 93 the stagnant film. We consider an elementary volume of thickness Az and of unit area normal to the z-direction (jie. the direction of mass transfer). We make a steady state mass balance over this element located at position z. Rate of input” of the solute at z = Nal, Rate of output of the solute at z+ Az = Nalesac Rate of accumulation = 0 (at steady state) = Nale~ Naleese = 0 Dividing by Az throughout and taking the limit Ac —> 0, we get = aNglde = 0 =0 3.28) dz Integrating Eq, (3.28) and using the following boundary conditions () and i), (i) z=0 (ie. the wall or the phase boundary or the interface), Cy Gi) z = 8 (ice. the other end of the film of thickness, 5), Cy = Cane we get Cai Oy = Cn (Cui~ CWE 8.29) ‘The above equation indicates that the theoretical concentration profile, according to the film theory, is linear as shown in Figure 3.6(b) (where the “true” concentration profile is also shown), ‘The mass transfer flux through the film is constant at steady state and is given as Pap - a yo Bu Cw) 630) Comparing with Eg. (3.3), the mass transfer coefficient is Paw k= G31) Thus the mass transfer coefficient can be calculated from the film theory if the diffusivity and the film thickness are known. Whereas the former can possibly be obtained from the literature oor may be estimated by using a suitable correlation, the latter (i.e. the film thickness 6) is unknown. So this theory does not help us in reality to predict the mass transfer coefficient. However, the film theory, like two other theories described below, has been extremely useful in the analysis of mass transfer accompanied by a chemical reaction (as we shall see in Chapter 16). The film thickness, however, can be attributed a physical significance. Its the thickness of the stagnant layer of fluid that offers a mass transfer resistance equal to the actual resistance to ‘mass transfer offered by the fluid in motion. The film theory predicts linear dependence of ky, The notation |, means evaluated at (ee Section 27.2)94 Chapter 3 Convective Mass Transfer and Mass Transfer Coefficient upon the diffusivity Day. Unfortunately, experimental data for diverse systems show that the coefficient of mass transfer to a turbulent fluid varies as (Dag)! where n may have any value from zero t0 about 0.8. The film theory is simple indeed and that is why it has got wide acceptance. Nevertheless, it does not visualize a realistic physical picture regarding the mechanism of mass transfer at the phase boundary. Two other classical but more realistic models are described below. 3.7.2. The Penetrat yn Theory Liquid elements stiding down Before going into the details of this theory, let us hhave a close look at the phenomenon of mass transfer from a rising gas bubble—for example, absorption of oxygen from an air bubble in a fermenter. As the bubble rises, the liquid elements. from the bulk reach the top of the bubble, move along its spherical surface, reach its bottom and. then get detached from it. The detached liquid elements eventually get mixed up with the bulk liquid (Figure 3.7). Absorption of oxygen in a small liquid clement occurs as long as the clement remains at the gas-liquid interface (ie. in contact, with the gas). Similar phenomena occur in a large ‘number of other situations involving mass transfer at a phase boundary. Thus, we may generalize the picture by arguing that in the case of mass transfer at a phase boundary, an clement of liquid reaches the interface (by any mechanism whatsoever) and stays there for a short while when it receives some solute from the other phase. At the end of its stay at the interface, the liquid clement moves back into the bulk liquid carrying with it the solute it picked up during its brief slay at the interface. In the process, the liquid clement makes room for another liquid element, fresh from the bulk, on the surface of the bubble. The above visualization (or ‘model’) of the process of mass transfer at a phase boundary forms the basis of the ‘Penetration Theory” proposed by Higbie (1935). The following are the basic assumptions of the penetration theory. An clement about t9 leave the surface of ‘he bubble Figure 3.7 The penetration modelitustiated (@) Unsteady state mass transfer occurs to liquid element so long asi sin contact with the bubble (or the other phase). (b) Equilibrium exists at the gas-liquid interface. (6). Each ofthe liquid elements stays in contact with the gas (or the other phase) forthe same petiod of time. It we confine our attention to mass transfer to a gas bubble of dlameter d, rising at a velocity a the contact time of a liquid clement with the as is f= dp, Unsteady state mass transfer to a Tiqud element during this period of time can be deseribed bya paral differential equation given below ({his equation can be derived by using an approach very similar to that used in the case cof unsteady heat conduction to a semi-infinite medium, as discussed Chapter 15). The equation is ac, PC, wae 7 Page 3.32)Theories of Mass Transfer 95 ‘The appropriate initial and boundary conditions are: Initia condition =0, 220; G=Cy Boundary condition 1: 1> 0, 2=0; C= Cy G33) Boundary condition 2: 1>0, z==; Cy=Cw ‘The initial condition implies that in a fresh liquid clement coming from the bulk, the concen- tration is uniform and is equal to the bulk concentration. The boundary condition 1 assumes that “interfacial equilibrium’ exists at all time. The last condition means that if the contact time of an ‘element with the gas is small, the “depth of penetration’ of the solute in the element should also bbe small and effectively the element can be considered {0 be of “infinite thickness’ in a relative sense. Equation (3.32), subject (0 the initial and boundary conditions (3.33), can be solved for the transient concentration distribution of the solute in the element by introducing a “similarity variable’ m, as in the case of unsteady state heat conduction (see Dutta, 2001). The solution is 1 -erfn: wh 2 3.34) ert: ere Daast (3.34) ‘The mass flux to the element at any time ¢ can be derived from the equation = —p, [24] = [Pte — NAO = -ral | = Cu Cw) 38) ‘The flux decreases with time because of a gradual build-up of the solute concentration within the element and the resulting decrease in the driving force. At a large time, the element becomes. nearly saturated and the flux becomes vanishingly small ‘The average mass flux over the contact time f, is given by 2) Poe c, a G36) Comparing Eqs. (3.35) and (3.36) with Eq. (3.3), Daw Instantaneous mass transfer coefficient, ky = | AE 37) Daw Average mass transfer coefficient, Ky. = 2] 8.38) The above equations show that the mass transfer coefficient is proportional to the square root of diffusivity. Although this is again not in conformity with experimental observations in general, this is definitely an improvement over the film theory for a more realistic visualization. Here the contact time 4, is the model parameter like the film thickness 6 in the film theory. 3.7.3. The Surface Renewal Theory One of the major drawbacks of the penetration theory is the assumption that the contact time or the ‘age’ of the liquid elements is the same for all. In a turbulent medium it is much more probable that some of the liquid elements are swept away, while still young, from the interface96 Chapter 3 Convective Mass Transfer and Mass Transfer Coefficient by eddies while some others, unaffected by the eddies for the time being, may continue to be in contact with the gas for longer times. AS a result, there will be a “distribution of age” of the liquid elements present at the interface at any moment. This is how Danckwert (1951) visualized, the phenomenon, He assumed that (j) the liquid elements at the interface are being randomly replaced by fresh elements from the bulk, (ii) at any moment each of the liquid elements at the surface has the same probability of being replaced by a fresh element, (iii) unsteady state mass transfer occurs to an element during its stay at the interface. The Danckwert’s theory is thus called. the surface renewal theory. The model parameter is the fractional rate of surface renewal (5, the fraction of the surface area renewed in unit time). The equation for the mass transfer coefficient according to the Danckwert’s theory is as follows (the derivation is shown in Chapter 16). ky = (Daas 39) It may be noted that the surface renewal theory also predicts the square root dependence of the mass transfer coefficient k, on the diffusivity. It also has similar dependence on the fractional rate of surface renewal, s. Increasing turbulence in the medium causes more brisk surface renewal (ie. larger 5) thereby increasing the mass transfer coefficient. The expressions for the ‘mass transfer coefficient obtained from the above three theories and the boundary layer theory described below are listed in Table 3.4. Table3.4 Mass transfer coeficens from the dferen hears of mass tanser Theoy ‘Steadyunsteasy Expression or ‘Dependenceon| ed stale the coefiient ‘itty paraeter(uni) Fim theon ‘Steady tate k= 08 kao 3 (m) Pendatontheoy Unsteady state (Oxq"® k= oF ott! k= £) Stace renewal heory Unsteady stale Ky, = [al!® k= oF sie) Boundary layertheoy Steady state Eqs. (3-40), (3.41), k= inet—insanlaneous,av-— average 3.7.4 The Boundary Layer Theory ‘The foregoing theories of mass transfer have been built up on the basis of idealized pictures of the way mass transfer occurs at a phase boundary. As such, these theories, though fit well in some particular cases, do not reflect the real picture in general. The major reason is that none of these theories explicitly takes into account the hydrodynamics or flow field that characterizes 4 system, although the flow field greatly influences the rate of convective mass transfer, What the foregoing theories do is to use a model parameter (the film thickness, the contact time or the surface renewal rate) to take into account the effect of fluid motion implicitly. For example, more intense fluid velocity or a high degree of turbulence in the medium is associated with a decrease in the film thickness or the contact time or an increase in the fractional rate of surface renewal. ‘The boundary layer theory (sce, for example, Bird et al., 2002), on the other hand, gives a far ‘more realistic picture of the way mass transfer takes place at a phase boundary. Before dealing with mass transfer in boundary layer flow, let us recapitulate the formation of boundary of layer on an immersed surface in a viscous fluid in motion. For simplicity, we takeTheories of Mass Transfer__ 97 the case of flow past a flat plate (Figure 3.8) oriented along the direction of flow (zero angle of incidence). The fluid velocity is zero at the surface of the plate (the no-slip condition) but increases with distance from the plate. At a rather short distance away from the surface, the fluid velocity almost reaches the velocity of the bulk fluid along the direction of orientation of the plate (called the ‘free stream velocity’, V..). The region above the plate within which the fluid velocity changes from zero atthe surface to the free stream velocity is called the “velocity boundary thickness’, 6. To be more precise, this region is called the ‘velocity boundary layer’ (or the ‘momentum boundary layer’ or the “hydrodynamic boundary layer’). The boundary layer thickness increases gradually along the plate, It is expressed as a function of the distance from the leading edge, i. 5 = 5(4). Theoretically, the free stream velocity is reached at a large distance from the plate. However, in reality the velocity reaches nearly the free stream value at a rather small distance away from the plate. Typically, the boundary layer thickness at any point on the plate is defined. arbitrarily as the transverse distance from the plate where 99% of the free stream velocity is attained. » V—. ~~ TRigeof the veloety or momentum Bi [Edge of the concentration BL SEA C=C 09HC GCs) Te [Boundary Tayer] C4 Ca) Be) Bylo) Velocity =0 7 Flat Plate witha soluble coating Figure 3.8 Momentum and concentration boundary layers ona iat plate for Se > 1. V_= tee stream velocity C concentration; C= surface concentration othe solu ofthe costed soldinthelgid. Since Sc> 1, he momentum dusty islargerthan the molecular ius (v.> Ogg), and the momentum boundary layers thicker than th concentration boundary layer asa result If the flow occurs at zero angle of incidence and the plate is ‘wide’, the motion in the boundary is two-dimensional, = u(x,)). The general equation of motion ofa viscous fluid, called the Navier-Stokes equation, can be simplified and approximated to write down the equations of motion for two-dimensional boundary layer flow, which can be solved to get the velocity field in the laminar boundary layer (Dutta, 2001). The characteristics of the boundary layer depends upon the Reynolds number defined as Re, = p¥s/d, where Re, is the local Reynolds number 48 itis based on the distance x from the leading edge. Boundary layer flow remains laminar if the Reynolds number isles than about 3 10, above which the flow turns turbulent. Ths value of the Reynolds number is called the “critical Reynolds number’. Mass transfer in boundary layer flow occurs in a way similar to that of heat transfer. Ifthe plate is coated with soluble substance and the liquid (or ga) flows over it, two boundary layers are formed—the “velocity boundary layer" and the “concentration” or “mass boundary layer" (Figure 3.8) Itmay be recalled that in boundary layer flow over a heated plate, a thermal boundary98 Chapter 3 Convective Mase Transfer and Mass Transfer Coefficient layer is formed along with the velocity or momentum boundary layer (thickness = 6). The concentration distribution in the boundary layer is a function of position, Cy = C(t, ys nd its thickness is a function of the distance from the leading edge, 5 = d(x). The thickness of the concentration boundary layer 5, is defined as the distance over which the solute concentration ‘drops by 99% of the concentration difference between the wall and the bulk liquid. The relative thicknesses of the velocity and the concentration boundary layers depend upon the value of the Schmidt number, Se (which isthe ratio of the momentum diffusivity to the molecular diffusivity). If the Schmidt number is greater than unity, the thickness of the momentum boundary layer at any location on the plate is more than the concentration boundary layer. It can be shown that 6/8y = Sc" (Deen, 1998; Bird et al, 2002). “Theoretical analysis of mass transfer in a laminar boundary can be done following the same approach as adopted in the case of heat transfer, and the following equation for the local Sherwood number, Sh,, can be developed. sn, = kee® 12(6)" Sh = 0.332(Re)!"(Se) G40) Dap Here xis the distance of a point from the leading edge of the plate, ky, isthe local mass transfer coefficient, and Re, is the ‘local Reynolds number’. If lis the length of the plate, the ‘average Sherwood number", Shy,, ean be obtained from the above equation as follows: Kravl 1250)! Shay = ™— = 0.664(Re/)!"(Se)" G41) In the above equation, ka isthe mass transfer coefficient averaged over the length of the plate, and Re, = pV.I/i is the ‘plate Reynolds number’ based on the length of the plate / Itis to be noted that the boundary layer theory predicts that the mass transfer coefficient sz, varies as (Dap)*® which reasonably matches the experimental findings in many cases. 3.7.5 Other Theories Quite a few other theories were proposed at different times to describe mass transfer at a phase boundary (for example the film-penetration theory). Theories based on eddy diffusivity have also been proposed. Although the theories of mass transfer provide an insight into the mechanism of mass transfer at a phase boundary, these are not directly useful for predictive purposes (the boundary layer theory is an exception in many practical situations). However, the penetration theory and the surface renewal theory have proved to be immensely useful for mass transfer calculations when the dissolved solute undergoes a chemical reaction, Consumption of the dissolved solute reduces its concentration in the solution, increases the driving force and thereby increases the rate of absorption compared to the case where no such reaction occurs (i.e. where only the physical transport of the solute occurs), In fact, any of the above theories can be used. to calculate the degree of enhancement of the rate of mass transfer as a result of chemical reaction, This aspect will be discussed in detail in Chapter 16. The penetration theory has been the basis of the Billet model of mass transfer in a packed column (as we shall see in Chapter 6).Theories of Mass Transfer__ 99. EXAMPLE 3.5 (Gas absorption in a laminar jet) Gas absorption in a laminar liquid jet is @ ‘model experimental technique for the determination of the liquid-phase diffusivity of a soluble gas and also for studying the kinetics of gas-liquid reactions. A sketch of a laminar jet apparatus is given in Figure 3.9. A liquid jet is created by forcing the liquid through a narrow vertical tube, fitted within an enclosure. If an appropriate flow rate is maintained, the liquid comes out as @ {jet (that very much looks like aliquid rod or cylinder). The jt is collected in a tube of a diameter ‘marginally greater than that ofthe jet and having a slightly flared opening. This liquid collection tube is oriented vertically below the jet. The solute gas, diluted with a cartier if necessary, flows through the enclosure. The rate of absorption of the gas is determined either by measuring the inlet and the outlet gas flow rates or by analyzing the concentration of the dissolved gas in the cexitliquid, In an experiment forthe determination of the diffusivity of hydrogen sulphide in water, pure H,S gas is passed through the enclosure of the jet. The following data were collected in a Particular experiment at steady state: temperature = 25°C; total pressure in the jet chamber 1.08 atm; length of the jet = 5 cm; rate of flow of water = 13.2 ml/s; absorption rate of H3S = 4.42 x 10 gis. The solubility of Ha$ in water at 1 atm pressure and 25°C is 0.1136 kmol/ (0m°(atm), Calculate the diffusivity of H,S in water from the above data. Absorbent liquid in Gas out (for sampling ‘and analysis) _- Exposed length ofthe jet, + Gas Enclosure of the e laminar jet Gas in |— Collection tube Receiver of Solution out for the solution sampling and analysis Figure3.9 A simpltid sketch ofthe laminae apparatus Solution ‘A liquid jet has a rod-like flow; ithas a flat velocity profile. In other words, the velocity is uniform over a cross-section of the jet. Let us use the following notations: r = radius of the jet Q= the liquid flow rate (= mv) = 13.2 mls = 1.32% 10° mils; 1. = contact time = Uv. Molecular weight of H3S = 34. ‘The solute gas gets dissolved at the surface ofthe jet and penetrates into it by unsteady state diffusion. The situation fits in the penetration theory almost ideally. Accordingly, the average rass transfer coefficient [see Eq. (3.36)] and the gas-liquid area of contact are (uniform) velocity of the jet; length of the jet (= 0.05 m);100 Chapter 3 Convective Mase Transfer and Mass Transfer Coefficient Daw Dap v. o = Aig = 2) Pat = 2fPAET: the area of contact = 2ar @ ‘The rate of absorption of HzS = (2271 ki.,)(Cui = Cas) wo where C,, = interfacial concentration of H,S in the liquid, and C,y = H3S concentration in the bulk of the jet. The interfacial concentration is the same as the solubility of the gas at the given temperature and pressure, and can be calculated at the given pressure (1.03 atm) using the Henry's law: Cn Cure 1.03)(0.1136) = 0.117 kmov/m® (because of the small contact time, the depth of penetration of the solute in the jet will be small and the bulk liquid free from HS) The given rate of absorption of H,S = 4.42 x 10~* g/s = (4.42/34) x 10 x 10%, i.e. 13 x 10-* kmols Using Eqs. () and Gi) and the given rate of absorption [note that r = (Q/¥)'"], 13 x 10° = 201 je 2 [Pus (0417-0) = 4 Dyy 132 x 10)(0.05) (0.117) = Dag = [17x10 m/s, [Diffusivity of HyS in water at 25°C reported in the literature = 1.21 x 10 m/s} EXAMPLE 3.6 (Gas absorption from bubbles) A mixture of 50% CO, and 50% Np is bubbling through water in a laboratory column at 30°C and 1 atm. The depth of water in the column is. 30-cm. A single-nozzle gas distributor is used. The gas flow rate is 15 em? per minute and the bubbles are of I cm diameter on the average. The bubble rise velocity is 20 cm/s. Calculate the rate of absorption of carbon dioxide. The diffusivity of CO. in water is 2.19 x 10° ems Henry's law can be used to calculate the solubility of CO, in water at the given temperature, ‘p= 1860." (p = partial pressure of CO,, in atm; x”is its mole fraction in water at equilibrium). Hints: Contact time of aliquid clement with a gas bubble, f= dy, = (1 emy(20 em/s) = 0.05 s ‘Mass transfer coefficient, k, = 2(Dap/tt,)'? = 0.0236 cm/s. Residence time of a single bubble in the liquid = (Liquid depth)/(bubble rise velocity) = 0.3/0.2 = 1.5 s. Volume of a bubble = 0.5236 em’; area of a bubble = 3.1416 em?, Solubility of CO, in water (from Henry's law), C, = 1.493 x 10° gmoliem? (peo, is approximately taken as 0.5 atm). Concentration of COz in the bulk water, C, = 0. Amount of CO> absorbed from a single bubble during its residence time of 1.5 = (ky)(area of a bubble)(C, ~ C,)(1.5 8) = 1.66 x 10% gmol Average number of bubbles formed per minute = 15/0.5236 = 28.65 per min. Rate of absorption of CO, from the bubbles per minute = (1.66 10° gmol per bubble)(28.65 bubbles per min) = 4.757 x 10S gmot/minTheories of Mass Transfer 101 EXAMPLE 3.7 (Gas absorption in an agitated vessel) In an experimental agitated contactor, pure carbon dioxide is being absorbed in water at 25°C and 2 atm pressure. Water is pumped into the contactor at a rate of 1 litre per minute and the carbonated water leaves the vessel ‘continuously so that constant volume is maintained in the contactor. The outlet water contains 2.3 g CO; per litre. The specific interfacial area of gas-liquid contact is 80 m°/m? of the gas liquid dispersion; the volume of the gas-liquid dispersion is 8 litre. The liquid phase can be assumed to be well mixed. The solubility of CO; in water can be calculated using the Henry's aw. At 25°C, the Henry's law constant for CO» is 1640 atm/mol fraction and its diffusivity in water is 1.92 x 10 ms. Calculate (a) the thickness of the liquid-film if the film theory is applicable, (b) the contact time between a liquid element with the gas if the penetration theory is applicable, and (©) the fractional surface renewal rate if the surface renewal theory is applicable. ‘The density of the liquid is 997 kg/m? (Le. the same as that of water at the given temperature). 0; out B lite dispersion \ FOG GO [238 C0, petlive Gas bubbles Figure 3.10 CO, gas bubbles through a tank Solution The Solubility of CO, at the experimental temperature is given by Henry's law, p = 1640." Atp=2 atm, x° = 2/1640 = 0.00122 (the superscript” is often used to indicate the equilibrium concentration). We shall first calculate the mass transfer coefficient ky, and then use it to ‘determine the required quantities. Molecular weight of the solution = (44)(0.00122) + (18)(1 ~ 0.00122) = 18.03 = 18 Moles of solution per m? = (997 kg/m?)/(18 kg/mol) = 55.4 kmol/m? Moles of CO, per m* solution, C, = (55.4)(0.00122) = 0.0676 kmol/m* Concentration of the carbonated solution leaving the vessel, C= 23 g COpfite = 2.3 kg COs/m? = 2.3/4 = 0.0523 kmol/m? = Concentration of CO» in the liquid in the vessel (since itis well-mixed). Volume rate of input of water = volume rate of output of the solution = | litreimin = 1.667 x 10-5 mis102 Chapter 3 Convective Mass Transfer and Mass Transfer Coefficient The inlet water is COx-free. So at steady state, the rate of absorption of CO> in the vessel = (1.667 x 10 m*/s)(0.0523 kmol/m*) = 8.718 x 107 kmols 8 litre = 0.008 m*), @ = specific 80 m*/m?), and k, = mass If V= the volume of the gas-liquid dispersion in the vessel interfacial area of contact between the dispersed gas and the liquid transfer coefficient, Rate of absorption of CO» at steady state = Vazk,(C, ~ C) = 8.718 x 10-7 kmol’s => (0,008)(80)(k,)(0.0676 ~ 0.0523) = 8.718 x 107; when k, = 8.903 x 10° m/s (a) IF the film theory is applicable, film thickness, Day _ 192%10 K, ~ 8903x10 (b) If the penetration theory is applicable, the contact time [see Eq. (3.38)] is ADsp _ _(4)(192 x10) mk; (#890310)? (©) If the surface renewal theory is applicable, the fractional rate of surface renewal [see Eq. G39), b= (0.0216 mm 03085 KE _ (8903x1075)? Day 1.92107 (This means that the surface is renewed 4.13 times per second on the average.) 3.8 MOMENTUM, HEAT AND MASS TRANSFER ANALOGIES Transport of momentum, heat and mass in a medium in laminar motion are all diffusional processes and occur by similar mechanisms. The three basic laws in this connection—Newton’s law of viscosity that governs the rate of transport of momentum, Fourier’s law of heat conduction and Fick’s law of diffusion—can all be expressed in similar forms. Newton's Law: Momentum flux, r= —u ts Etpu) 42) Fourier’s Law: Heat fox, 4 = (pe) Ga) aC Fick's Law: Mass flax (ata Tow concentration), Ny = ~Dyy Ac oan) In Eq, (3.42), Tis the shear stress which isthe same as momentum flux, pu isthe “volumetric ‘concentration of momentum’ in the x-direction (i.e. momentum per unit volume of the liquid), and v= [/p, is the ‘momentum diffusivity’. In Eq. (3.43), peyT is the ‘volumetric concentration of thermal energy’ and ct (= k/pe,) is the “thermal diffusivity’. Thus, all the above three equations sate that the flux is proportional to the gradient of the quantity transported (momentum, heat energy, or mass), and the proportionality constant is the corresponding ‘diffusivity’ (that has the same unit, m/s, in all the cases). A negative sign is included in each equation to indicate that transport occurs in the direction of decreasing concentration (of momentum, heat, or mass)..‘Momentum, Heat and Mass Transfer Analogies 103 In turbulent flow, similar equations are used to express the rate of transport except that an ‘eddy diffusivity’ is included to account for the contribution of eddy transport (see Section 3.5). Thus, = -+8,)4(pu, Momentum flux, t= —(V+E,)-E-(Pu,) G45) Heat tx, g.= (0+ Ey) (pe, 6.46) aC Mass flux (at alow concentration), Na = = (Dyg + By SC oan Here Ey, Ey and Ey are eddy diffusivities of momentum, heat and mass respectively. Since turbulent transport is a much faster process than molecular transpor., all the eddy diffusivities are larger than their molecular counterparts by about two orders of magnitude (i. a few hundred times) or even more ‘The similarity of the three transport laws has been extended to relate the appropriate dimensiontess groups involving the heat transfer coefficient and the mass transfer coefficient 10 the friction factor. Such relations or ‘analogies’ (Skelland, 1974) can be used to estimate the heat transfer coefficient (fh, of the mass transfer coefficients (kz, Ky ete.) if the friction factor of the flowing medium is known, Here we mention a few simple analogies without going into the theoretical formalism (Soc Skelland, 1974; Hines and Maddox, 1985; Dutta, 2001). The oldest such analogy in the case of heat transfer was proposed by Reynolds applicable to transport in pipe flow. Extending the analogy to the case of mass transfer, we may write siy= Beef and sty = SH = EL G48) Here Sty is the Stanton number for heat transfer and Sty is that for mass transfer, Nu is the Nusselt number involving the heat transfer coefficient, and fs the fiction factor for the fluid flowing through a pipe. “The Prandtl analogy, also applicable to transport in pipe flow, is given by #12 $12 = — 2 at Sy =§ — 3.49) Dsyae-p w= Ty SJ Ge) ee ‘The Prandtl analogy reduces to the Reynolds analogy if Pr = 1 or Se Colburn related the mass transfer coefficient to the friction factor by proposing the well- known ‘Colburn analogy’. He introduced the Colburn j-factor and suggested the following analogies for transport in pipe flow. Sty Nu ju = Sty PP® = — Se = 0.023Re°? 3.50) ji 8 = Ne e 6.50) and Jp = StySe? = — SH. = 0,023Re-%? G1) Re Se Here the subscripts *H” and "M’ refer to heat and mass transfer respectively [the subscript ‘D* in jp means diffusion and is synonymous to the subscript M in Eq. (3.49)]. The importance of the analogies lies in the fact that if the heat transfer coefficient is known at a particular hydrodynamic condition characterized by the Reynolds number, the mass transfer coefficient in104 Chapter 3 Convective Mass Transfer and Mass Transfer Coefficient a system having similar geometry and at similar hydrodynamic condition (ic. the same Reynolds. number) can be determined just by putting jz = Jp: EXAMPLE 3.8 (Mass transfer in a packed bed) Water at 25°C flows through a bed of benzoic acid spheres of size d, = 8 mm ata superficial velocity of 0.022 m° per second per square metre of bed cross-section. The fractional void volume of the bed is 0.4 and the depth of the bed is 0.7 m. If the inlet water is free from benzoic acid, calculate the concentration of the acid in the effluent from the bed. ‘The following data are available: diffusivity of benzoic acid in water = 10-° em?/s; kinematic viscosity of the liquid = 0.95 centistoke; solubility of benzoic acid in water at the given temperature = 3.01 kgim*, For a packed bed of sphere, the specific surface area (ie. the are of the spheres per unit volume of the bed) is given by: 7 = 6(I ~ eld, Solution “The concentration of benzoic acid inthe bulk liquid, C,, avin, ems Liguit increases along the bed from the inlet value of Cy =0. Cone. _Aisributor ‘The concentration of the liquid in contact with the lr surface of a sphere is, C, = 3.01 kg/m>, As a result, the driving force, C, ~ Cy, changes along the bed nelementan Let us use the following notations: Q = volumetric pace flow rate of water; a = cross-sectioal area of the bed the bed Then, Q/a = liquid flow rate per unit area of the bed 0.022 m°/m?s = 2.2 emis (this is also the linear velocity of the liquid if the entire cross-section is available for flow and is called the “superficial liquid velocity"). If we consider a differential thickness of the bed, dit at a distance / from the top of the bed (see Figure 3.11), the concentration changes by dC, and a differential ‘mass balance at steady state over dh gives Solution out 0 dCy = (kT KAANC, ~ Cy) Figure’3.11._An elementary section ofa packed bed, |__Packed bed of spheres Cone. = ¢, In the above equation, &,7 is the ‘volumetric mass transfer coefficient’ (ie. the amount of mass transfer per unit time per unit volume of the bed. per unit concentration driving force). Integrating the above equation over the height /, aC, I ee hag san = me @ Now we need to calculate the value of &, at the given flow rate of the liquid through the bed. We use the correlation Ejp = 0.25Re°*" Given: diameter of a sphere, dy = 0.8 cm; superficial liquid velocity, v, = 2.2 cm/s; the kinematic viscosity, v= 0.0095 cm*/s; diffusivity of the solute, Dyy = 10-° cm/s; bed voidage, e= 04.Momentum, Heat and Mass Transfer Analogies 105 4, = fo¥e _ 0892.2) _ (0.0095) 185 and 3 eo 25¢18sy9%= 0.0405 n= OOBSUBSESON _ 55 = sh{ Pas) — 295(10%) - cm/s w= (2) = 2(!2°) «on et Specific interfacial arca of contact, 7 = 8U=€) . 60-04) «4.5 cation? q, 08 Potting the values ofthe different quantities in Eq. (i), the effluent concentration, ya = [1 kein? EXAMPLE 3.9 (Mass transfer in a wetted-wall column) As stated in Section 3.6, one of the experimental devices for the determination of the mass transfer coefficient is a wetted-wall column. In an experiment, benzene is absorbed from a stream of nitrogen in a wetted-wall column in which the absorbent is @ non-volatile mineral oil. Under the conditions, the liquid-phase mass transfer resistance can be safely neglected. The mass flow rate of the feed gas is 13,500 kg/h-m? at 30°C and 1.013 bar pressure. Mole fractions of benzene in the inlet and outlet gases are 0.02 and 0.0052 respectively. The interfacial concentration of benzene on the gas side is negligibly small. Calculate the gas-phase mass transfer coefficient Can you predict the mass transfer coefficient for the absorption of ammonia in a dilute solution of H$O, in the same wetted-wall column ifthe flow rate of air is 4.51 kg/m?-s at25°C ‘and 1,013 bar, and the partial pressure of ammonia in the feed gas is 10 mm Hg? Given: id. of the column = 3.5 em; thickness ofthe liquid film is small; height of the wetted section = 3 m; viscosity: for nitrogen (at 30°C) = 0.018 cP, for air (at 25°C) = 0.0183 cP; diffusivity of benzene in Nz at 30°C = 0.0973 cm7/s, that of NHg in air at 25°C = 0.231 em’/s. Since the concentration of benzene in the gas is small, the Benzene cone change of flow rate over the column also remains small. So one calculations may be done on the basis of the average gas flow — Oilin rate. It is also given that the Colburn factor jp varies as Re“, Lal Hints: Consider an clementary section of the column of thickness ea de ata height from the bottom as shown in Figure 3.12. Suppose fc mol = local concentration of benzene (the solute) inthe bulk gas . (mole fraction; y,= interfacial concentration atthe same section; yi G = molar gas flow rate; and d = diameter of the column. ‘Then the elementary area of gas-liquid contact = (7ed)(dz). If ijout __¥ the gas concentration changes by dy over the section, a differen-—, ———F tial mass balance can be written as ~Giy = (nites ~ Fie 22 0a stein na ea ‘wetted-wall colurnn,106 Chapter 3 Convective Mase Transfer and Mass Transfer Coefficient Assuming thatthe gas flow rate remains reasonably constant over the column (this assumption is valid at low concentration, as in the given case, and an average gas flow rate may be used) Integrating from z = 0, y = y; = 0.02 (bottom) to z = L = 3 m, y = y2 = 0.0052 (top), we get ky = [G In(yy/y) Vora) Average mol. wt. of the gas at the bottom = (0.02)(78) + (0.98)(28) = 29; at the top it is 28.2; average molar gas flow rate at the bottom = (13,500/29) = 465.5 kmol/m?-h, that at the top = 458.6 kmol/m?-h. Average = 462 kmol/m?-h, Area of the cross-section of the tube (0.035 m dia) = 9.62 x 10-4 m*, Actual gas flow rate through the tube, G = (9.62 x 10-*)(462) = 0.4448 kmol/h Then, = 1.786 kmol(hy{m?\Ay) Now, if ‘I’ denotes the Nz-benzene system and ‘2° denotes the air-NH system, calculate Rey (= 7140), Re2 (= 8630), Sey (= 1.5) and Sez (= 0.675). Then Joy .(Rey"* Shy (Re) (Sey Joa (Rep Sh, ~ (Re, Take pau = P in Sh, caleulate ky conditions .25 kmol(h)(m?)(Ay) for the ait-NH system at the given NOTATIONS a + molar concentration of A, kmol/m? 4, dy + diameter of a tube or sphere, or particle, m Pap + molecular diffusivity of A in B, m?is E + Eddy ‘diffusivity’ [suffix v—momentum, H—heat, M—mass diffusion], m/s. ky, oy kes ky + gas-phase mass transfer coefficients for diffusion of A through non-diffusing B, Eq. G3); a prime is used to mean equimolar counterdiffusion kg kus ky + liquid-phase mass transfer coefficients for diffusion of A through stagnant B, Fg. 3.4) 1 + characteristic length, m m mass, kg Ms + molar flux of species A p + partial pressure, bar, atm, et. Past + log mean partial pressure of the non-diffusing component P + total pressureShort and Multiple Choice Questions 107 + fractional rate of surface renewal, s-! hte + time, contact time, s uy + velocity (in the longitudinal or transverse direction), m/s Ve : free stream velocity (far away’ from an immersed surface), m/s yw + mole fraction, mole ratio of a species in the gas phase aX + mole fraction, mole ratio of a species in liquid z + distance, m Re, + local Reynolds number, xV_p/ss x is the longitudinal distance Se + Schmidt number Sh +: Sherwood number, Eqs. (3.16), (3.17) 6 + film thickness; velocity boundary layer thickness, m Sy + mass transfer boundary layer thickness, m u + viscosity of a fluid, kg/m-s v + momentum diffusivity, 4/p 1,2 + positions 1 and 2 Subscripts > + bulk of a phase i + imterface between two phases av + average value SHORT AND MULTIPLE CHOICE QUESTIONS 1. Explain the basic difference between the mass transfer coefficients ky and k;. 2. Obtain the relation between ky and kj. Write down the units of mass transfer coefficients. Discuss when and how the mass transfer coefficients depend upon concentration. 3. Give the physical significances of the dimensionless groups in mass transfer. 4, For mass transfer from a sphere to a stagnant liquid, show that Sh 5. Explain the underlying assumptions of the theories of mass transfer. 6. Derive Eqs. (3.34) and (3.39). 7. The film theory gives a linear concentration profile of the solute within the stagnant film. Is there a point within the film at which the gradient of “true concentration profile” and that of the ‘theoretical concentration” profile in the film are the same? 8. Discuss the importance and applications of mass transfer analogies. 9. A gas A diffuses through a non-diffusing B across a stagnant film of thickness 2 mm. Given: mole fractions of A on the two sides of the film are y4y = 0.3, pressure = 1 atm; molar flux, Ny = 5 x 10 kmol/m?-s. 42 = 0.05; total108 Chapter 3 Convective Mass Transfer and Mass Transfer Coefficient 10. Calculate the Colbum-Drew mass transfer coefficient. Also calculate the molar flux if transport of A occurs through a stagnant film of B, 1 mm thick, for ys, = 0.2 and Yaa = 0.03. Make the correct choice from the altematives given against the following questions. (a) Identify the quantities that have influence on the mass transfer coefficient: diffusivity, (i) solubility ofthe solute, (ii) molar flux, (iv) hydrodynamics of the system, (v) viscosity of the medium, (vi) interfacial concentration gradient, (vii) the interfacial area of contact. (b) If the mass transfer flux is expressed in terms of a mass transfer coefficient, the driving force should be (the concentration gradient at the interface (ii) the concentration difference between the interface and the bulk Gi) the interfacial concentration. (©) How does an increase in the liquid flow rate in a packed tower affect the rate of surface renewal of the liquid in a packed tower? The rate of surface renewal will (i) increase ii) decrease (iii) remain unchanged. (4) How does an increase in the gas flow rate affect the gas-phase mass transfer coefficient in a packed tower? The mass transfer coefficient will (i) increase ii) decrease (iii) remain unchanged. (©) How do you define the Stanton number for mass transfer’? (0) How do you define the Peclet number? () Pe=Re-Se (ii) Pe= Re/Se (ii) Pe = Sh/Se (g) What is the expression for the Colburn factor, jp? @stse™ Gi) sh Reso® WD Rese (h) The Sherwood number, Sh (increases with the friction factor, f Gi) increases with the Reynolds number, Re (il) decreases with Re. (i) Air containing a solute flows over a pan containing water at a velocity of 20 fus. Which of the following all in Ibmol/(f?\(h)(atm), is a probable value of the average ‘mass transfer coefficient? 045 — Gi) 0.00345 iii) $62 (The probable value of the Schmidt number for diffusion of NH in air at 0°C may be @) 40.5 i) 0.665 Git) 0.0155 The probable value of the Schmidt number for diffusion in water at 15°C may be () 5000 Gi) 0.132 Citi) 650Short and Multiple Choice Questions 109 (K)_ Dissolution of a solid is being carried out in two identical stirred tanks at the same solid loading and stirrer spm except that particles of 1 mm size are charged in tank 1 and particles of 0.5 mm size are charged in the other tank. In which tank is the dissolution of mass transfer coefficient expected to be larger? () Tank 1 i) Tank 2 (ii) The mass transfer coefficients in the tanks should be practically the same. () What is the significance of “film thickness’ in mass transfer? It is () the actual thickness of the stagnant film that adheres to the phase boundary ii) the depth of penetration of the solute in a liquid element (ii) the thickness ofa stagnant film that offers the same resistance to mass transfer as is actually being offered under the given hydrodynamic condition. (im) For gas absorption from a bubble in steady rise through a stagnant liquid, which of the following theories is most appropriate? (The film theory (ii) The penetration theory ii) The surface renewal theory (n) Consider a thin sheet of naphthalene placed in a stream of air. If we assume that ‘mass transfer of naphthalene occurs through a stagnant film, then the “thickness of the film’ on the sheet will be (i) greater than the average thickness of the boundary layer (ii) less than the average thickness of the boundary layer Gii) equal to the average thickness of the boundary layer. (0) A solid coated on the surface of a flat plate dissolves in a liquid flowing past the plate at zero angle of incidence. Convective transport of the solute occurs through the laminar mass transport boundary layer formed on the plate. In such a case the thickness of the momentum boundary layer is expected to be greater than the concentration boundary layer if the Schmidt number is () greater than 1 (ii) less than 1 (ii) equal to 1. (p) Which of the following is a unit of fractional surface renewal rate? (cm/s Gi) f/m Git) Wt (@)_ How does the “age” of aliquid element at the phase boundary affect the rate of mass transfer to it? The rate of mass transfer (i) increases with the age of the element (ii) decreases with the age (ii) does not depend upon the age. (£) A student measured the liquid-phase mass transfer coefficient for absorption of A in water. Applying the film theory he calculated the thickness of the liquid film, 8 = 0.02 mm. If the diffusivity of A in water is 1.1 x 10-* em*/s, what would have been the value ofthe surface renewal rate (per second) if he had applied the surface renewal theory to his study? (55x10 Gi)325 Git) 2.75110 Chapter 3 Convective Mass Transfer and Mass Transfer Cosficent (8) Consider a bubble of a pure gas A rising steadily through a stagnant liquid B. The following data are given: diffusivity = 3.14 x 10-* cm?/s; bubble diameter = 0.5 cm; bubble rise velocity = 20 cm/s; solubility of the gas = 0.2 gmol/lite. What is the “life’ (in seconds) of individual liquid elements that come in contact with the bubble? 10 G01 Gi) 0.025 What is the average rate of surface renewal?” (@) 40 per second Gi) 0.2 per second ii) 20 per second ‘What is the value of the average mass transfer coefficient in em/s? 02 G04 Gyo (© A.10 mm diameter bubble of pure gas rises through a quiescent liquid ata steady velocity of 25 emis and the average mass transfer coefficient is found to be 0.013 cm/s. What would be the average mass transfer coefficient in emv/s if bubbles of 5 mm diameter rise through the liquid at a velocity of 10 cm/s? (0.0104 G)0.0116 Gi 0.0162 (u) The mass transfer coefficient for absorption of CO> in water in an agitated tank is found to be 0.006 cm/s. Which of the following is a probable value of the mass transfer coefficient in cm/s for absorption of HS in water under identical hydrodynamic condition? (0.005 @i)0.05 Gin) 0.5 (¥) Iman experiment on mass transfer through the laminar boundary layer over a flat plate, I'm long, the average mass transfer coefficient was found to be 0.00705 ‘emis. What would be the average mass transfer coefficient in ems if the plate length is doubled? (0.0141 (i) 0.0035 Git) 0.005 ()_ The Stanton number for heat transfer in a certain system is 0.0048. What would be the Stanton number for mass transfer under identical geometrical and hydrodynamic conditions if the Lewis number is 1.22 (0.00576 Gi) 0.0041 Git) 0.00425 (x). The mass transfer coefficients obtained by the Reynolds and Prandtl analogies are the same if ( Se Gi Se=1 Git) () Consider mass transfer of a solute in laminar boundary flow over a flat plate. The local mass transfer coefficient at a distance 0.2 m from the leading edge is 0.012 em/s. What would be the local mass transfer coefficient in cm/s ata distance (04 m from the leading edge? (0.0085 ii) 0.006 Git) 0.017 (How does the mass transfer coefficient vary with the diffusivity according to the boundary layer theory? @D? ap GyoProblems ULL 3 32 33 34 38 PROBLEMS. (Calculation of mass transfer coefficients)’ Aramonia(A) is being absorbed in water from a mixture with nitrogen(B). The partial pressure of the solute in the bulk gas is 40 mm Hg and that at the gas-liquid interface is negligibly small. Diffusion occurs through a stagnant film of thickness 1 mm. The total pressure is 1 atm and the temperature is 25°C. Diffusivity of NH, in N, is 0.23 cm/s. Calculate the absorption flux of NHg as well as the mass transfer coefficients key ky Ke kys and Fg. Unterrelation among mass transfer coefficients)! A stream of nitrogen containing 7.5% benzene vapour is scrubbed with @ nonvolatile absorption oil in a tower at 35°C and 1.2 bar total pressure. The gas-phase mass transfer coefficient is estimated t0 be kg = 9.8 x 10° kmol/(m?\(s)(bar). The mole fraction of benzene at the gas-liquid interface is y; = 0.01. Calculate the gas-phase mass transfer coefficients fy, k, and ky. ‘Also calculate the values of the mass transfer coefficient in terms of Ibmol/ft?)(n)(Ay), IbmoV(tt)(s)(Ibmol), and Ibmol/tt?)(h) Psi). (Dissolution of a bubble)* A 0.8.cm diameter bubble of pure CO is injected into an excess well-stirred liquid at 25°C. The bubble diameter shrinks to 0.2. cm after 80 s. Given: total pressure = 1 atm; solubility of CO. in water = 1.45 x 10°? mass fraction; diffusivity of CO, in water = 1.9 x 10° em’, Calculate the average value of the mass transfer coefficients, ky and k,. How long will it take for the bubble to vanish if the mass transfer coefficients remain constant? (Is this 4 reasonable assumption”) (stass transfer coefficient in counterdiffusion)® ‘The gas-phase reaction A > B occurs on the surface of a catalyst in the form of a wire gauge. The bulk concentration of the species A is 0.002 kmol/m? and the rate of reaction of A is measured at 0.08 kmol per hour perm? area ofthe catalyst. If the reaction is instantaneous and the entire diffusional resistance to the transport of the reactive species from the bulk gas to the catalyst surface is offered by a stagnant film of the gas, calculate (a) the thickness of the film; (©) the mass transfer coefficients, kk and (C) the concentration of A halfway the gas film, Given: total pressure = 1.5 bar; temperature = 250°C; Dag = 0.16 ems at 1 bar and 100°C. (Evaporative mass transfer from a drop in free fall)’ A raindrop of initial diameter 0.8 mm starts falling from a sufficiently high altitude. The terminal free fall velocity depends upon the properties of the medium as well as on the diameter of the drop. As the drop moves through the air, its size reduces because of evaporation. The gas-phase ‘mass transfer coefficient for the evaporation of water changes as a result. I is required to calculate the time required for the size of the drop to reduce to 0.6 mm. The terminal velocity of fall for the given range of drop size can be calculated from the following equation: v= Hee $3. 10%; te dametr din mee ain ms ‘The following assumptions may be made: (i) the drop is always at its terminal velocity of fall corresponding to its instantancous diameter (this is also a kind of pscudo-steady112 Chapter 3 Convective Mass Transfer and Mass Transfer Cosficent 36 37 38 39 state assumption); Gi) the ambient air is at 25°C, 1 atm total pressure and 70% relative humidity; (ii) the shape of the drop remains spherical all the time; and (iv) the drop temperature is 22°C (which is nearly equal to the wet-bulb temperature). The mass transfer coefficient can be estimated from the Ranz. and Marshall correlation given by Sh = 2 + 0.6(Re)"*(Se)* ‘The following data are available: diffusivity of moisture in air = 0.257 em*/s at the mean film temperature; density of air = 1.18 kp/m®; viscosity of air = 1.8 x 10° kg/m-s; vapour pressure of water at 22°C is 0.02605 atm. (An application of dimensional analysis)? It has been observed experimentally that the rate of toxin removal in an artificial kidney (see Section 14.5.2) is a function of blood velocity, density, viscosity, diffusion coefficient of toxin and the diameter of the dialysis tubing. Using dimensional analysis, determine the dimensionless groups and the type of correlation that may be applicable. (Mass transfer in a wetted-wall column)' In a wetted-wall tower, HyS is being ab- sorbed from air into water at 1.6 bar total pressure and 25°C. The gas-phase mass transfer coefficient is predicted to be k; = 3.42 kmol/(hy(m?)(AC, kmol/m’). Ata given point in the column, the mole fraction of HS in the liquid at the gas-liquid interface is 1.8 x 10° and that in the bulk gas is 0.15. The solubility of HS in water at the given temperature is 0.00337 mass fraction per atmosphere pressure of the gas. Calculate the local flux. (Evaporation of water in a wetted-wall column)! Water is being evaporated into a counter-current stream of air in a wetted-wall column, 30 mm in diameter. Estimate the evaporative flux of water ata point in the column where the water temperature is 305 K and the mole fraction of water vapour in the bulk air is 0.0098 ‘The following data and information are supplied: air rate (dry basis) = 1 kg/min; water rate = 0.2 kg/min; air temperature at the location = 38°C; air viscosity (at 38°C) 1.85 10° Ibift-s; total pressure = 1 atm. Vapour pressure of water at the given temperature can be obtained from Antoine equation (see Problem 2.2). The gas-phase ‘mass transfer coefficient can be calculated using the Gilliland and Sherwood correlation given by (note that it resembles the Dittus-Boelter equation for heat transfer) Sh = 0.023(Re)%Se)? ‘The Reynolds number of air may be calculated at the bulk condition (to be more precise, the air velocity should be taken relative to free surface velocity of water). Gas absorption in a falling film)? Air mixed with a soluble gas flows countercurrent to water flowing down as a thin film covering the inner wall of an experimental column, 5 em in diameter (refer to Figure 3.5). The water-film thickness is 1 mm and it flows at an average velocity, v = 2.1 mils. The diffusion coefficient D of the solute is 1.7.x 10° em*s. If the water leaving the tower is 10% saturated, calculate the height of the tower. The mass transfer coefficient fora falling film can be calculated using the correlation (kyg/D) = 0.69(20/D)", where z is the position along the film measured from the top.Problems 113, 340 3a 342 343 ‘Assume that the gas flow rate is large so that the concentration of the solute in the gas remains essentially constant. (Mass transfer in erystal growth)* In an experiment on the growth of a single crystal of NaCl at 30°C, a supersaturated solution of the salt containing 39 g NaCl per 100 g of water flows at a velocity of 0.12 mis past a suspended erystal of 2.5 mm size. For the sake of simplicity, the shape of the crystal may be assumed to be spherical. Ifthe experimentally measured rate of growth of the crystal is drfdr= 9.5 x 10°° mvs (where ‘ris the equivalent radius of the erystal), determine whether the growth process can be considered to be ‘diffusion-controlled’. The following data are given: solubility of NaCl in water at 30°C = 36:3 g per 100 g water; density of the solution = 1300 kg/m"; density of solid NaCl = 2160 kg/m’; viscosity of the solution = 1.0 cP; diffusivity of NaCl in water under the given conditions = 14 x 107 ms. The mass transfer coefficient at the solid surface can be calculated from the Rowe's correlation, Sh = 2 + 0.79Re!Se!®, (An application of the penetration theory)? Air bubbles of 0.005 m diameter rise through an otherwise stagnant pool of ‘oxygen-free’ water at 30°C. The steady velocity of rise is 0.1 mls. If the diffusivity of O, in water at 25°C is 2.1 x 10 m2, calculate the liquid side coefficient and the rate of mass transfer from a single bubble. The solubility of O> in water at 30°C can be calculated by using the Henry’s law coefficient as 4.75 x 10° atm/mol fraction. The viscosities of water are 0.911 cP and 0.817 cP at 25°C and 30°C respectively. (Mass transfer coefficient from the penetration theory)” For the experimental confir- mation of his penetration theory, Higbie devised a small set-up comprising a narrow vertical glass tube filled with water in which pure carbon dioxide was released at the bottom at a regular frequency. The bubbles moving through the tube were of the shape of a capsule. The length of a bubble, its residence time in the water in the tube, and the velocity of rise were measured. The contact time of a liquid element on the surface of ‘a bubble with the gas was estimated from these measurements. The CO; content in the exit water was determined by chemical analysis. In an experiment to test the validity of the Higbie’s theory, carbon dioxide gas (1 atm, 25°C) is introduced at the bottom of a narrow tube, 3 mm in diameter and 0.6 m long, through which water is flowing downwards slowly. The bubbles are of capsule shape, 0.05 m long. The thickness of the water film between the bubble and the tube wall is small. The bubble rise velocity is 0.2 m/s. Bubbles enter the tube at 'fy second interval and the rate of absorption of carbon dioxide is estimated to be 40 mg/min. The solubility of COp in water at 1 atm and 25°C is 1.5 kg/m®. The concentration of CO> in bulk water is very small. Does the Higbic’s thoery appear to work in this case? The bubbles may be assumed cylindrical for calculation of the area. (Gas absorption from rising bubbles)’ A ‘pure gas’ is being bubbled through a column of liquid. The bubbles are spherical, 6 mm average diameter (assumed constant), and rise with a constant velocity of 0.18 m/s. The volume of gas-liquid dispersion is 2.5 litre and the fractional gas hold-up (the volume fraction of gas in the dispersion) is 0.02. The liquid is well-mixed and there is no dissolved gas at the beginning. The diffusivity of the4 Chapter 3 Convective Mass Transfer and Mass Transfer Cosficent 344 348 316 347 348 ‘gas in the liquid is 1.8 x 10° m*s, The liquid-phase mass transfer coefficient may be calculated by using the penetration theory. Determine the time it takes for the liquid to attain 50% of the saturation concentration, (Boundary-layer mass transfer)? A plate, 0.5 m square, coated with a layer of benzoic acid, is placed in a stream of water flowing at a velocity of 0.25 m/s at a temperature of 25°C. Calculate the average rate of dissolution of the acid per unit area of the plate and also the equivalent thickness of a stagnant liquid film that would offer the same resistance to mass transfer. The following data are available: solubility of benzoic acid in water at 25°C 3.01 kg/m?; diffusivity of benzoic acid in water = 10°? m*/s and the viscosity of water at 25°C = 8.9 x 10 kg/m-s. (Boundary-layer mass transfer)? A rectangular pan, 40 cm long and 30 em wide, full of water is placed in a wind tunnel (in effect, it is a case of boundary-layer flow on a flat surface). Wind flows at a linear velocity of 5 nv/s through the tunnel, Calculate the rate of evaporation of water. The resistance to mass transfer is offered only by the boundary layer on the surface of the ‘The following data are available: The air temperature is 300 K and has a humidity of 60%. Vapour pressure of water at 300 K = 0.0353 bar; diffusivity of moisture in air at 300 K and 1 atm = 0.262 em*Is; Pair = 1.16 kg/m?s vie = 1.59 x 10-* m2/S, Take the water temperature as 300 K (although, in reality, it will gradually approach the wet-bulb temperature; see Chapter 10). (Boundary-layer mass transfer)® A stream of air at 101.3 kPa pressure and 300 K flows over the top surface of a sheet of naphthalene of length 0.25 m at a velocity of 15 mis. Given: diffusivity of naphthalene vapour in air under the given conditions = 6 x 10° m/s; kinematic viscosity of air = 1.6 x 10-° m/s; vapour pressure of naphthalene at 300 K = 0.235 mm Hg, Calculate (a) the molar concentration of naphthalene at the plate surface, (b) the average mass transfer coefficient, and (c) the rate of loss of naphthalene from the surface per unit width. Itis known that for boundary layer heat transfer from a flat plate, the average heat transfer coefficient can be calculated from the equation: Nu = 0.664(Re,)" (Pr) (Gas absorption in a falling film)? Pure chlorine gas is being absorbed in water in @ small experimental wetted-wall tower, 2 em in diameter and 30 cm in height. The free surface velocity of the water film is 40 cm/s. Under the given conditions, the solubility of chlorine in water is 0.823 g/100 g water and its diffusivity is 1.26 x 10° ems, ‘What is the rate of absorption of the gas in gmol/h? Assume: (i) the “depth of penetration’ of the dissolved gas is small; (i) a flat velocity Profile of the liquid near the free surface. (Oxygen absorption in water from air bubbles)® Ait at 25°C and 1 atm pressure is bubbled through water in a 100-litre aquarium for oxygenation. The tank is initially filled ‘with water having 0.001 kg/m? oxygen. At any time there is I litre of air dispersed in the water in the form of 5 mm diameter (on the average) bubbles that have a rise velocity of 0.26 mm/s. The solubility of oxygen in water at the given temperature and pressure is 0.0085 kg/m’.Problems 115 349 3.20 3.21 (@) Calculate the average mass transfer coefficient. (b) How long should it take for the oxygen concentration in the water to reach 60% of the saturation value? (c) Calculate the maximum and the minimum oxygen flux over this time (Problem 3.18 continued) Let there be a total of 50 fishes in the aquarium described in Problem 3.18. Each fish needs, on the average, 1.5 x 10°* kg oxygen per second for healthy survival. What should be the fractional air hold-up in the tank if the oxygen content in the water is to be maintained at 60% (this is reasonable for survival of fish) of the saturation value at steady state? The average bubble size, rise velocity, physical properties are as in Problem 3.18. (Mass transfer in flow through a pipe)® Ait containing 5% CO, enters a tube of 40 mm id. at the bottom at a velocity of 4 m/s. The inner wall of the tube is irrigated with a strong solution of NaOH. The tube has a3 m wetted section and 75% of the entering CO; is absorbed before the gas leaves the tube. Calculate the gas-phase mass transfer coefficient. Make differential mass balance over a thin section of the tube taking into account the variation of the gas flow rate. Also estimate the mass transfer coefficient using the correlation k.d/D = 0.026Re°*Sc"=5, Use the gas velocity relative to the free surface liquid velocity (that has a value of 0.15 m/s) in the calculation of Reynolds number. Take D = 0.165 cm/s. The pressure drop of the gas across the tube may be neglected. (Mass transfer in a laboratory stirred cell)’ ‘The “laboratory stirred cell” is a simple device for the study of mass transfer accompanied by a chemical reaction in a gas-liquid system (Danckwerts, PY., Gas Liguid Reactions, McGraw-Hill, 1970). The apparatus is a small cylindrical vessel partly filled with the absorbent liquid or solution. A stirrer enters the cell at the top and the stirrer blades just touch the flat liquid surface. The surface of the liquid gets continuously “renewed” as the stirer rotates (Figure 3.13). The ‘285, pure or diluted with an ‘inert’ carrier, flows through the space above the liquid. Gas out Stierer The stirrer blades Gas stirrer. LF) veh the gu sure Gas in Liquid sample Figure3.13._A simplted sketch ofa sired cl (gs absorption occurs atthe fat gas-iqudinterface). ‘An experiment has been conducted with such an apparatus on the absorption of pure (CO, at 1 atm pressure in water in order to measure the rate of surface renewal at a given stirrer speed. When 100 ml of water is taken in the cell and the experiment is run for116 Chapter 3 Convective Mass Transfer and Mass Transfer Cosficent 3.22 3.24 3.25 12 minutes, the concentration of COs in the liquid is found to be 0.112 mass%. The gas liquid contact area, which is virually the same as the cross-sectional area of the cell, is 31.5 cm?. The temperature ofthe system is maintained at 25°C. The solubility of CO, in water at I atm pressure can be calculated from the Henry's law: py = Hx, where H = 1.64 x 10° atm per unit mole fraction, The gas-phase mass transfer resistance is absent for absorption of a pure gas. Calculate the liquid-phase mass transfer coefficient for the given experiment and the fractional rate of surface renewal. The liquid in the cell may be assumed ‘well-mixed’ and the solution density may be assumed to be the same as that of water. The inlet CO> gas is saturated with water vapour at 25°C so that there is no evaporation of water during the experiment. The diffusion coefficient of COp in water at 25°C is 1.92 x 10° em’. (Gas-phase mass transfer coefficient in a stirred cell)’ In order to measure the gas- phase mass transfer coefficient in a stirred cell at a given stirrer speed, absorption of ammonia diluted with dry air in aqueous sulphuric acid is cartied out for a given period of time. The gas mixture flows into the cell at a constant rate and the outlet gas contains 3% of ammonia by volume. The gas in the cell is ‘well-stirred” and the cross-sectional area is 31.5 em?, Ifthe volume of acid taken is 100 ml at 1(N) concentration and at the end of 20 minutes the liquid is found to contain 0.82 gmole ammonia per litre, calculate the gas-phase mass transfer coefficient. (Mass transfer to suspended catalyst particles)? liquid-phase catalytic reaction A> B + C occurs on the surface of 50 g suspended catalyst particles (mean diameter 1 mm; density = 3800 kg/m?) in an experimental batch agitated reactor. Two litre feed solution containing 0.5 kmol of A/m? is taken in the reactor. Given that: (i) the reaction is diffusion-controlied; (i) the catalyst is non-porous (the reaction occurs on the external surface only); and (iil) the liquid-solid mass transfer coefficient, k, = 3.5 x 10° mis, calculate the conversion of A at the end of 30 minutes. (Time of dissolution of suspended particles)* In an experiment on the dissolution of salt in water, 500 g of salt of average particle size 0.7 mm (assumed spherical) is suspended in 10 litre water in an agitated vessel. The dissolution rate is measured by collecting and analyzing small volumes of liquid samples from time to time, The density of salt is 2160 kg/m? and its solubility in water at the experimental temperature is 350 kg/m?, Ifthe initial dissolution rate is 20.8 g/, calculate the dissolution mass transfer coefficient. Also, calculate the time of dissolution of the solid if the mass transfer coefficient is assumed to be independent of particle size. The density of the solution may be taken to be essentially the same as that of water (Dissolution of particles in a stirred vessel) A 5% solution of a reagent is required in the process of synthesis of a pharmaceutical intermediate. The solid reagent, available as crystals (density = 1580 kg/m?) of average particle size 0.75 mm, is dissolved in a stirred vessel to prepare 1.2 m? solution in a batch. The requisite volume of water and the calculated quantity of the solid are charged into a 1.15 m diameter tank fitted with 0.6 m diameter impeller.