Absorption for HAP and VOC Control 11.1 INTRODUCTION Absorption is a diffusional mass-transfer operation by which a soluble gaseous component is removed from a gas stream by dissolution in a solvent liquid. 1 The driving force for mass transfer is the concentration difference of the solute between the gaseous and liquid phases. In the case of absorption this driving force can be interpreted as the difference between the partial pressure of the soluble gas in the gas mixture and the vapor pressure of the solute gas in the liquid film in contact with the gas. If the driving force is not positive, no absorption will occur. If it is negative, desorption or stripping will occur and the concentration of the pollutant in the gas being treated will increase. Absorption systems can be divided into those that use water as the primary absorbing liquid and those that use a low volatility organic liquid. The system can be simple absorption in which the absorbing liquid is used in a single pass and then disposed of while containing the absorbed pollutant. Alternatively, the pollutant can be separated from the absorbing liquid and recovered in a pure, concentrated form by distillation or stripping (desorption). The absorbing liquid is then used in a closed circuit and is continuously regenerated and recycled. Examples of regeneration alternatives to distillation or stripping are pollutant removal through precipitation and settling; chemical destruction through neutralization, oxidation, or reduction; hydrolysis; solvent extraction; and pollutant liquid adsorption. 11.2 AQUEOUS SYSTEMS Absorption is one of the most frequently used methods for removal of water-soluble gases. Acidic gases such as HCI, HF, and SiF 4 can be absorbed in water efficiently and readily, especially if the last contact is made with water that has been made alkaline. Less soluble acidic gases such as SO 2 , C1 2 , and H 2 S can be absorbed more readily in a dilute caustic solution. The scrubbing liquid may be made alkaline with dissolved soda ash or sodium bicarbonate, or with NaOH, usually no higher a concentration in the scrubbing liquid than 5 to 10%. Lime is a cheaper and more plentiful alkali, but its use directly in the absorber may lead to plugging or coating problems if the calcium salts produced have only limited solubility. A technique often used is the two-step flue gas desulfurization process, where the absorbing solution containing NaOH is used inside the absorption tower, and then the tower 11 9588ch11 frame Page 117 Tuesday, September 11, 2001 10:16 AM © 2002 by CRC Press LLC effluent is treated with lime externally, precipitating the absorbed component as a slightly soluble calcium salt. The precipitate may be removed by thickening, and the regenerated sodium alkali solution is recycled to the absorber. Scrubbing with an ammonium salt solution can also be employed. In such cases, the gas is often first contacted with the more alkaline solution and then with a neutral or slightly acid contact to prevent stripping losses of NH 3 to the atmosphere. When flue gases containing CO 2 are being scrubbed with an alkaline solution to remove other acidic components, the caustic consumption can be inordinately high if CO 2 is absorbed. However, if the pH of the scrubbing liquid entering the absorber is kept below 9.0, the amount of CO 2 absorbed can be kept low. Conversely, alkaline gases, such as NH 3 , can be removed from the main gas stream with acidic water solutions such as dilute H 2 SO 4 , H 3 PO 4 , or HNO 3 . Single-pass scrubbing solutions so used can often be disposed of as fertilizer ingredients. Alternatives are to remove the absorbed component by concentration and crystallization. The absorb- ing gas must have adequate solubility in the scrubbing liquid at the resulting tem- perature of the gas–liquid system. For pollutant gases with limited water solubility, such as SO 2 or benzene vapors, the large quantities of water that would be required are generally impractical on single-pass basis, but may be used in unusual circumstances. An early example from the U.K. is the removal of SO 2 from flue gas at the Battersea and Bankside electric power stations, described by Rees. 2 Here the normally alkaline water from the Thames tidal estuary is used in large quantity on a one-pass basis. 11.3 NONAQUEOUS SYSTEMS Although water is the most common liquid used for absorbing acidic gases, amines (monoethanol-, diethanol-, and triethanolamine; methyldietnolamine; and dimethy- lanaline) have been used for absorbing SO 2 and H 2 S from hydrocarbon gas streams. Such absorbents are generally limited to solid-particulate-free systems, because solids can produce difficult to handle sludges as well as use up valuable organic absorbents. Furthermore, because of absorbent cost, absorbent regeneration must be practiced in almost all cases. At first glance, an organic liquid appears to be the preferred solvent for absorbing hydrocarbon and organic vapors from a gas stream because of improved solubility and miscibility. The lower heat of vaporization of organic liquids is an energy conservation plus when solvent regeneration must occur by stripping. Many heavy oils, No. 2 fuel oil or heavier, and other solvents with low vapor pressure can do extremely well in reducing organic vapor concentrations to low levels. Care must be exercised in picking a solvent that will have sufficiently low vapor pressure that the solvent itself will not become a source of VOC pollution. Obviously, the treated gas will be saturated with the absorbing solvent. An absorber–stripper system for recovery of benzene vapors has been described by Crocker. 3 Other aspects of organic solvent absorption requiring consideration are stability of the solvent in the gas–sol- vent system, for example its resistance to oxidation, and possible fire and explosion hazards. 9588ch11 frame Page 118 Tuesday, September 11, 2001 10:16 AM © 2002 by CRC Press LLC 11.4 TYPES AND ARRANGEMENTS OF ABSORPTION EQUIPMENT Absorption requires intimate contact between a gas and a liquid. 4 Usually means are provided to break the liquid up into small droplets or thin films, which are constantly renewed through turbulence to provide high liquid surface area for mass transfer as well as a fresh, unsaturated surface film for high driving force. The most commonly used devices are packed and plate columns, open spray chambers and towers, cyclonic spray chambers, and combinations of sprayed and packed chambers. Some of these devices are illustrated in Figures 10.1, 10.2, and 10.3. Additional illustrations of absorption equipment are shown in Figures 11.1 to 11.5. Packed towers give excellent gas–liquid contact and efficient mass transfer. For this reason, they can generally be smaller in size than open spray towers. A countercurrent packed tower maximizes driving force, because it brings the least concentrated outlet gas into contact with fresh absorbing liquid. These features make this type of tower design the best choice when the inlet gas is essentially free of solid particulates. However, packed towers plug rapidly when appreciable insoluble particulates are present. The cross-flow packed scrubber is more plug resistant when properly designed. 5 In plate columns, contact between gas and liquid is obtained by forcing the gas to pass upward through small orifices, bubbling through a liquid layer flowing across a plate. The bubble cap tower is a classical contacting device. A variation is the valve tray, which permits greater variations in gas flow rate without dumping the liquid through the gas passages. Sieve plates are simple flat plates perforated with small holes. The advantages are low cost and high plate efficiency, but they have narrow gas flow operating ranges. Spray chambers and towers are considerably more resistant to plugging when solid particulates are present in the inlet gas. However, difficulties with plugging in spray towers and erosion can be troublesome when the spray liquid is recycled. Particle settling followed by fine strainers or even coarse filters is beneficial. Another tower contacting device for absorption is the baffle tower, which has been employed occasionally when plugging and scaling problems are expected to be severe. Gases passing up the tower must pass through sheets downwardly cas- cading liquid, providing some degree of contact and liquid atomization. Baffle tower design may use alternating segmental baffles (Figure 11.4) or disk and doughnut plates, in which the gas alternately flows upward through central orifices and annuli, traversing through liquid curtains with each change in direction. Mass transfer is generally poor, and information on design parameters is extremely scarce. 11.5 DESIGN TECHNIQUES FOR COUNTERCURRENT ABSORPTION COLUMNS Although the design of co-current and cross-flow towers are important, countercur- rent towers are the most frequently used. Moreover, the design of cross-flow towers involves complicated calculations which do not enhance the theory of tower design. Therefore, the principles of design of the countercurrent tower will be explained in more detail. 9588ch11 frame Page 119 Tuesday, September 11, 2001 10:16 AM © 2002 by CRC Press LLC The two main factors to be determined in countercurrent absorption column design are the height of packing influenced by the mass-transfer conditions, and the tower diameter, influenced by the flow rate of the gas to be treated. Size and type of packing affects both the mass-transfer conditions and the pressure loss through the column. Height of packing is determined by the rate of mass transfer from one phase to another. The amount of mass transferred is equal to the rate of mass transfer times FIGURE 11.1 Packed column with all equipment. 9588ch11 frame Page 120 Tuesday, September 11, 2001 10:16 AM © 2002 by CRC Press LLC the time of contact. The time of contact is dependent upon column size and flow rates. For a very high tower, one might expect that most of the pollutant in the gas stream can be removed. Tower diameter is determined by the quantity of gas passing up the tower. The upper limit of flow occurs when the tower begins to flood. Most efficient mass transfer occurs at flow rates just short of flooding. Higher flow rates result in higher pressure losses. Thus an economic optimum column size could be established, if cost data and information regarding mass transfer and pressure loss as a function of flow rate are known. Because we are dealing with dilute solutions in the case of pollution removal from an effluent, the heat of solution will most usually be negligible. Therefore, we will assume that air contaminant absorption columns operate isothermally, and the design equations presented in this chapter will be for the isothermal, dilute gas case. FIGURE 11.2 Spray chamber. 9588ch11 frame Page 121 Tuesday, September 11, 2001 10:16 AM © 2002 by CRC Press LLC 11.5.1 E QUILIBRIUM R ELATIONSHIPS For a very detailed introductory discussion of phase equilibria see the text by Smith, van Ness, and Abbott. 6 Vapor–liquid equilibria for miscible systems is determined from the equality of fugacity in both phases at the same temperature and pressure: (11.1) For component i in the liquid phase: (11.2) FIGURE 11.3 Cyclonic spray chamber. ˆˆ ff i v i L = ˆ fxf i L iii o =γ 9588ch11 frame Page 122 Tuesday, September 11, 2001 10:16 AM © 2002 by CRC Press LLC FIGURE 11.4 Plate column. 9588ch11 frame Page 123 Tuesday, September 11, 2001 10:16 AM © 2002 by CRC Press LLC and in the vapor phase (11.3) According to Equation 11.1, therefore: (11.4) This system is structured so that only the fugacity coefficient φ i and the activity coefficient γ i depend upon the compositions at any given temperature and pressure. The standard state fugacity f o i is a property of the pure component only. The Gibbs–Duhem equation which involves partial molar solution properties is extremely useful in dealing with phase equilibria. It can be written as follows for n components, FIGURE 11.5 Baffle tray tower. ˆ fyP i v ii =φ yPxf ii iii o φγ= 9588ch11 frame Page 124 Tuesday, September 11, 2001 10:16 AM © 2002 by CRC Press LLC (11.5) Here M is a solution property — M i and is the corresponding partial molar property. By using excess Gibbs free energy changes, this equation can be written in terms of the enthalpy and volume changes on mixing ∆ H and ∆ V, respectively, and the activity coefficient: (11.6) In the case of solutions of nonpolar molecules or for solutions at low concen- tration, the enthalpy and volume changes on mixing are small and Equation 11.6 for a binary mixture becomes: (11.7) For a solution at low pressure where the vapor is an ideal gas, and presuming f o i to be equivalent of the pure component fugacity at the temperature and pressure of the solution, Equation 11.7 can be rewritten at constant temperature: (11.8) Note that (y 1 P) = P 1 and (y 2 P) = P 2 , the partial pressures, and x 2 = 1 – x 1 , therefore, (11.9) (11.10) Therefore, from partial pressure data of one component of a binary solution, the partial pressure of the second component may be calculated. Use of this technique is recommended when limited experimental data are available. 11.5.2 I DEAL S OLUTIONS — H ENRY ’ S L AW A general relationship for fugacity of a component in a liquid mixture in terms of the pure component fugacity is given by the following equation: ∂ ∂       + ∂ ∂       − () = ∑ M T dT M P dP x d M PT ii i n ln 0 −+= () ∑ ∆∆H RT dT V RT dP x d ii i n 2 ln γ xd x d i ln lnγγ 12 2 0 () + () = xd yP xd yP 11 2 2 0ln ln () + () = dP dP P PP o ln ln 21 0 2 2 1 () =− () ∫∫ ln P P x x dP P o P 2 2 1 1 1 1 0 1 2       =− − () ∫ 9588ch11 frame Page 125 Tuesday, September 11, 2001 10:16 AM © 2002 by CRC Press LLC (11.11) According to Amagat’s Law, an ideal solution is one in which (V i — – V i = 0), and the value of the integral in Equation 11.11 is also zero. This makes (11.12) for an ideal solution. This relationship is known as the Lewis and Randall Fugacity Rule. More generally to be consistent with Equation 11.2, the ideal fugacity would be defined as: (11.13) where f o i is the fugacity of component i in the standard state at the same temperature and pressure of the solution. If the standard states adopted were that of the pure substance designated by i, then f o i = f i . Equation 11.13 is plotted in Figure 11.6 in such a manner that both broken lines are valid representations of the equation. Each broken line represents one model of the ideal solution while the solid line represents the value of ˆ f L i for a real solution FIGURE 11.6 Defining Henry’s Law and the Lewis and Randall Fugacity Rule in relation to the true fugacity. ln ˆ f xf RT VVdP i ii L ii P       =− () ∫ 1 0 ˆ ffxf i L i id ii L == fxf i id ii o = 9588ch11 frame Page 126 Tuesday, September 11, 2001 10:16 AM © 2002 by CRC Press LLC [...]... Tuesday, September 11, 2001 10:16 AM ( ) * N A = K y y A − y A = k y (y A − y Ai ) = k x (x Ai − x A ) = ( kx * y − yA m Ai ) (11. 37) also ( * N A = K y y A − y Ai + y Ai − y A ) (11. 38) Then combining Equations 11. 37 and 11. 38, N mN A  NA = K y  A + kx   ky  (11. 39) Then dividing both sides of the Equation 11. 39 by NA and rearranging, 1 1 m = + Ky ky kx (11. 40) 11. 5.7 VOLUME-BASED MASS-TRANSFER COEFFICIENTS... Press LLC 9588ch11 frame Page 130 Tuesday, September 11, 2001 10:16 AM such as Equation 11. 4 For this case of a dilute solution, Equation 11. 24 will be used * * along with Henry’s Law Equation 11. 20, yA = m xA Both the equilibrium line, 11. 20, and the operating line, 11. 24, will plot as a straight line on an x–y plot 11. 5.4 ORIGIN OF VOLUME-BASED MASS-TRANSFER COEFFICIENTS Interphase mass transfer takes... Equation 11. 73 Ab = B ( L G)min = B (1 − y A1 y A 0 ) m (11. 78) Substitute Equation 11. 76 into Equation 11. 75 and the following equation results, 1  * yA =  y − y A1 )  Ab  ( A (11. 79) For the dilute case, Equation 11. 79 is substituted into Equation 11. 59 Equation 11. 59 is now integrated with the following equation resulting,  1  N OG =   ln (y A 0 y A1 ) (1 − 1 Ab) + 1 Ab  1 − 1 Ab  [ ] (11. 80)... concentration can be defined, y BM = (yBO − yBL ) = (yAO − yAL ) ln(y BO y BL )  1 − y AL  ln (11. 32) 1− y   AO  Then a mass transfer coefficient can be defined as k my = D ABρm Ly BM (11. 33) and Equations 11. 32 and 11. 33 can be substituted into Equation 11. 29, N A = k my (y AO − y AL ) (11. 34) 11. 5.5 THE WHITMAN TWO-FILM THEORY Mass transfer in real absorption equipment resembles a molecular diffusion... defined in Equation 11. 36 For a dilute solution, (1 − yA ) ≈ (1 − yA )LM ≈ 1.0 (11. 58) and Equation 11. 56 reduces to yA 0 N OG = ∫ (y y A1 dy A * A − yA (11. 59) ) The operating line, Equation 11. 24, may now be rewritten as L y A =   (x A − x A1 ) + y A1  G (11. 60) Rewrite Equation 11. 60, multiplying and dividing by Henry’s Law constant, m, L  yA =  mx − mx A1 ) + y A1  mG  ( A (11. 61) – – Note... defined as Ab = L L = mG mG (11. 62) Then rewrite Equation 11. 61 using Henry’s Law and the absorption factor and * solve for yA * yA =  y A − y A1  + mx A1  Ab  (11. 63) Substitute Equation 11. 63 in Equation 11. 59 and integrate N OG = © 2002 by CRC Press LLC ln [((y A0 ) ] − mx A1 ) (y A1 − mx A1 ) (1 − 1 Ab) + (1 Ab) (1 − 1 Ab) (11. 64) 9588ch11 frame Page 138 Tuesday, September 11, 2001 10:16 AM When... Equation 11. 60 for the operating line may be rewritten, yA =  L +y A1  G (11. 74) m (y A − y A1 ) (L G) (11. 75) * yA m * and then solved for yA , * yA = Define Ab = (L G) m (11. 76) which is the same as Ab in Equation 11. 62 Then, define B as a multiple of (L/G)min to set the actual (L/G) Thus, © 2002 by CRC Press LLC 9588ch11 frame Page 148 Tuesday, September 11, 2001 10:16 AM (L G) = B (L G)min (11. 77)... similar to 11. 29 for molecular diffusion can be used to describe the mass transfer However, the concentration difference is expressed in terms of mole fractions at the interface N A = k y (y A − y Ai ) = k x (x Ai − x A ) © 2002 by CRC Press LLC (11. 35) 9588ch11 frame Page 133 Tuesday, September 11, 2001 10:16 AM FIGURE 11. 9 Illustrating new pseudo-equilibrium mole fractions 11. 5.6 OVERALL MASS-TRANSFER... mass-transfer coefficients (Kya, etc.) Table 11. 6 presents mass-transfer coefficient data from Strigle.9 Mass-transfer coefficients for structured packing are generally greater than dumped packing, indicating improved mass transfer Manufacturers of structured packing should be consulted for available mass-transfer data TABLE 11. 6 Mass-Transfer Coefficient Data KYa = KGaP KGa Values for a Liquid–Film Controlled... in most air pollution control conditions, (1 − yA ) ≈ 1.0, (1 − x A1 ) ≈ 1.0, etc and G B ≈ G ≈ G A1 L B ≈ L ≈ L A1 Equation 11. 26 reduces to Equation 11. 24 which is called the “operating line” since it represents concentrations in an operating absorption column Equation 11. 26 is a general equation which can be used along with a general equilibrium relationship © 2002 by CRC Press LLC 9588ch11 frame . Press LLC (11. 37) also (11. 38) Then combining Equations 11. 37 and 11. 38, (11. 39) Then dividing both sides of the Equation 11. 39 by N A and rearranging, (11. 40) 11. 5.7 VOLUME-BASED MASS-TRANSFER. can be defined, (11. 32) Then a mass transfer coefficient can be defined as (11. 33) and Equations 11. 32 and 11. 33 can be substituted into Equation 11. 29, (11. 34) 11. 5.5 THE WHITMAN TWO-FILM THEORY Mass. AA =− () =− () =− () =− () 11 11 11 11 y y G x x L x x L y y G A A B A A B A A B A A B 11 11 1 1 1 1 −       + −       = −       + −       110 1 10 1 − () ≈− () ≈y x etc AA .,