S Full Paper Gr Absorption of Sulphur Dioxide with Sodium Hydroxide Solution in Packed Columns By M, Schalies * Introduction ‘The absorption of sulphur dioxide from effluent gases by using aqueous sodium hydroxide solution as a solvent has ‘beoome established practice in industry where a high degree of purty of the effluent gases must be ensured in a reliable ‘manner. Although it has long been known that it is basically possible to clean effluent gases containing sulphur dioxide effectively by using sodium hydroxide solution, there are ‘nly few publications to date which discuss the reaction-ki- netic processes in the wash sotution.° 9147 An analysis of ‘these publications clearly reveals that a number of experi- ‘mental studies have been conducted under framework con- ditions far removed from those found in the practice of efflu- ent gas purification. Thus, Onda et al. and Vazquez et al conducted absorption tests with a gas phase of pure SO>, al- though SO; concentr industrial effluent gas purifica- tion are lower. Hikita etal. obtained from theoretical deriva- tions with predetermined boundary conditions mathematical solutions for the reaction-kinetic processes in the alkaline solution which are only met in limiting cases. In the publica- tions to date it was always assumed that the reactions in the ‘washing agent took place very rapidly, but this too holds only under certain conditions. 2 Reaction Sequences in Countercurrent Absorption. When $02 is absorbed in sodium hydroxide solution the SO; molecules first diffuse from the cote current of the gas to the gas-liquid interface and dissolve in the washing agent according to the equality ofthe chemical potentials in the phases. In view of the low SO» concentrations predominant ineffluent gases in practic, this means that Henry's law de- scribes the relationship between the liquid concentration and the equilibrium partial pressure in the gas phase. Iti ‘only after this phase transition ofthe SO> molecules has tak- en place according to Eq, (1) that one observes a diffusion cauent toward the core of the phase, a dissociation accord- ing to Bq, (2), and the reaction with OH fons according to Eq. (3), whereby hydrogen sulfite reacts in turn with adai- tional OFT ions to sulfite as shown in Eq, (4). The reaction- (*] Privatdoant Delay. M.Schales, Raschie AG, Mundentineret, 100, Laden OVC atretaging, 29.90, Ap 1, Lass Chem. Eng Technol 21 (1988) 2 kinetic processes are also influenced by the dissociation of the water according to Eq, (5), of the sodium hydroxide sol- ution according to Eq. (6) and of the sodium sulfite/hydro- gen sulfite portions according to Eqs (7) and (8), which are formed products. phys. solution SOs+air 4 S0;+H:0 (1) disso ciation SQ:4H,0 4 H'+HSO3 (2) reacton S02+0H" + HSO5 @) consecutive reaction HSO}+OH" ++ HL0+SO¥ (4) dissociation #0 = HO 6) 1 to be recorded with only ‘very small deviations from the numerical solution. }) fortla21 (17) If the enhancement factor E is known, the height of a ran- dom packed column necessary for a transfer task can be de- termined according to the HTU-NTU model, as Eq. (18) shows. The height H is calculated from the product of the ‘umber of mass transfer units in the gas phase NTUoy and the height of a mass transfer unit in the gas phase HTUoy. ‘The latter is composed of the gas-phase and liquid-phase ‘mass transfer unit HTU, and HU, as well as the stripping factor 2. Here, 4 describes the quotient of the SO phase equilibrium He/p and the molar flow ratio L/V with He as the Henry constant p as the system pressure. The mass trans- fer units HTUy and HTU; are defined by the gas and liquid velocities uy, ur, the volumetric mass transfer coefficients ‘vere apy, and the enhancement factor E. Bete (1-onf H=HTUoy NTUov _ (Hep 1 = (Bat Ens) More recent studies have shown thatthe volumetric mass transfer coefficients can best be calculated by Eqs. (19) to (@3). They are dependent on the column Toads wy, uyy the viscosities nes ny, the densities pt, py, the diffusion coetti- cients for SO; Dy, Dy the column hold-up hy, the hydraulic packing diameter dy, the specific surface area of the pack- ing a, and the void fraction € as well as the interface (HTU, +4 HTU,) NTUov as) mse ue vin pram =ce(HEt) (Be) Pa” a as) Brars= odin (8) "(Bx)" om (21) (22)¢ er Fatt Poper 2a ros)" (BY 5 Calculating the Diffusion Coefficients ‘The diffusion behavior of molecules or ions in a liquid or ‘gaseous phase is described with the help of the diffusion coefficient which, together with the concentration gradient in the phase-boundary layer defines the strength of the diffu- sion current according to Fick’ first law. The diffusion coef- ficient of molecules in air or water can be found in the rele- vant collections of tables and converted to other system conditions as tabulated with the help of the Nernst correla- tion or the Stokes-Einstein relation"! Furthermore, semi- empirical relations, e. g, those of Fuller, Schettler, and Gid- dings of Wilke and Chang also enable a theroetical precaleu- lation of the diffusion coefficient." In the case of the diffusion behavior of ions it must be bbome in mind that their flow is always influenced by the dif- fusion of other charge carriers. In the ease of SO2 absorption {n the liquid phase, one must bear in mind therefore that, at every location in the liquid film, the sum of all postive and negative charges must be the same owing to the electroneu- trality requirement, see Eq. (24), and the flow of ion diffu- sion may effect a different charge potential inthe core of the phase than at the interface. Charge potential isto be under- stood here as the absolute value of the right hand or of the left hand side of Eq. (24). For the description of the mass transfer of the charge carriers Eq. (25) follows in analogy to Eq, (15) from the total mass balance of all ons. xn +644 = Caso; + 20308 + Cone en Foy Fey Dat Gat Pes oe ae, Doge te @) Dow Sa Boome ocr noc noe a tC Ct If we now assume a linear potential gradient in the phase- boundary layer, according to the film theory and under sta- tionary conditions Eq, (26) follows, while taking into account Eqs. (24) and (25), for the flow Nj of the individual ion. Ac- cording to ths, the ionic current is comprised of a diffusivity according to Fick’s first law which is enhanced or reduced by the second additive term, which contains the potential gradi- ent d ¢ /dis carried out withthe help of Eq. (28). Here F do- scribes the Faraday constant, R the general gas constant, T the temperature, and 2 the valency of the ions. 204 (© WILEV.VCH Verlag GmbH, 1.65469 Wendin, 1998 It we apply the penetration hypothesis of Higbie, Eq. (29) follows from analogous observations for the effective diffu- sion coefficient with the potential gradient according to Eq. (0) ae Fao 288422, 82. Ni=Dy dz Dy TRT dz se oa) ae. ey (26) WORT de de) dz _F ao az 1% er ogparas) en Fao co) RT az In case of infinite dilution, the diffusion coefficient of an jon Dj’ can be determined according to 2 method of Vino- ‘grad and McBrain; see Eq, (31), in which i describes the conductivity of the ion.®! The effective diffusion coefficients calculated according to Eqs. (29) to (31) are, therefore, to be entered in Eq. (16) for the determination of the enhance- ‘ment factor E. Das=De{leae Hits) @ a 7 i A(e2 WBF) a RTdz yz YD, o RTM pea RT en 6 Calculating the Reaction and Phase Equilibrium ‘The concentrations of the ions in the core of the phase and at the interface are linked via the reaction equilibrium of Eqs. (32) to (37) with the SO, partions, and are heavily dependent on the pH of the solution. The same holds for the Na’ portion in the washing water which is calculated accord- ing to Eq, (38) from the total alkaline concentration intro- duced at the top of the column cy = Bey, reduced by the nondissociated portion of sodium hydroxide solution ¢xsor, (Chom. Brg. Tel 21 (1998)2sodium sulphite exasoy and sodium hydrogen sulphite ‘ya1ts0, The dissociations of the sodium compounds are cal- culated from the equilibrium relations of Eqs. (35) to (37). Correlations for the determination of the equilibrium con- stants as a function of temperature are shown in Tab. 2 Suso; Sur Vso; Tae K, = 21801 Sar Vaso Yor = co Yon [molt] 2) Ce rree: K, = 8 te [mov] (3) uso; 05 {moll}? (34) {molt] (35) [mou]? (36) [mol] (67) [moun] G8) ‘be 2 Teaperatre dependence of he equim constant He, Kw and Para) In Ky = (AOS007, 178007, 2727 +085 TTR aK = (40407-36507, 2807, + 13879 TM aK = 4008381) InKe= 8053387) aK, =2.RT), lop Ke = 906-20745267 405869 og (1+ 048796 - 0900717 tn (HP) sty «277,155, «DOL TOTIIR R= 831448. 1(n01 K) Te 29815K pete -asiacntg) . ur Pla Tek ~21a (HT) ‘The activity coefficients take into account the deviations fof the thermodynamic equilibrium of real mixtures from those ofan ideally diluted solution and can be calculated for ions according to Debye and Hickel in accordance with Eq. (G8) and for $0; according to Harned and Owen in accor- dance with Eq. (40) asa function of the ionic strength I of the solution "1 The parameters Ao, By, and C; forthe de- termination’ of the activity coefficients are compiled in Tab.3. (Chem. Eng Technol 21 (1998)2 (© WILEY-VCH Verlag Gmbt, D669 Wein, 1998 Full Paper Avi =A sql 39) wane Te 9) log Yo, = 0.0761 (40) ‘ible 3. Constants fo calculating the activity cote WOW HS0; «SOF Nt | mete tee sete tt) alu as op ‘Ao = 49,7063 + TRAOUTNT 4 10306889 a (1) 4313501 (17100) + susees cmn00} ‘aon = 450, Has0,~ 1 ‘The phase equilibrium for SOs over the sodium hydroxide solution is described by Henry's law according to Eq, (41) which states that the SO; concentration in the gas phase Yso2 is directly proportional to the physically dissolved SO; portions cso2. The sulfur dioxide molecules absorbed from the gas into the liquid is present no only inthe form of physi- cally dissolved SO, in the solution but also in ionized form as HSO3 and SOs, as well as in the sodium compounds as NaHSO3 and Na;SO3. The total quantity of sulphur bound in the solution is calculated, therefore, from Eq. (42). For the phase equilibrium Eq, (43) can be derived from Eq. (41), ‘aking into account the reaction equilibrium and the electro- neutrality requirement, Yoo, = He Ma (a) mo (2) @) “ ‘The differing solubility of SO, in an electrolyte solution in comparison with its solubility in pure water can be taken into account by the desalting parameter h and the ionic strength I cording to the empirical equation (44) of Van Krevelen and Hofijzer 9 In the SO,-sir/sodium hydroxide solution sys- tem this leads to the Henry’s law coefficient increasing asthe jonic strength increases, The Henry's law constant with infi- nite dilution He® can be calculated as a function ofthe tem- perature following a correlation according to Tab. .er Fatt Peper 7 Discussion of the Experimental Investigations ‘The ealeulation model presented isto be verted with the help of the experimental investigations. As Fig. 1 shows, ‘characteristic height segments, in which certain reactions de- cisvely determine the mass transfer rate, are formed during the absorption of SO; in sodium hydroxide solution. The portions ofthese ranges inthe total height ofa column de- ‘pends heavily on the sodium hydroxide solution concenta- tion with which the SO» absorption i earred out 1, for instance, the experimental system will operate with 8 lage surplus of sodium hydroxide solution in comparison withthe total molar quantity of SO, absorbed, the column is run exclusively with very high pH values and Ea, (9) des- cirbes the mass transfer rate. If, however, o sodium hydror- ide solution i added tothe washing wate, it willbe exien- tially the $03 dissociation according to Eq. (2) and the physical solubility ofthe SO, molecules according to Eq, (1) which determine the mass transfer rate. Any intermediate copertng state can be achieved by adding a sodium hydron- {ide solution of differen concentrations. ‘The experiments were, therefore, carried out with difer ‘ent sodium hydroxide solution concentrations between 0 ‘moV/ and 0.9 mol. Larger mole fractions of sodium hydrox- {de solution were not examined since in the case ofthe 09 ‘molar solution the very rapid mass transfer rate according to Eg. (9) was already so dominant that the resistance for the ‘mass transfer rate along the experimental column ly exci- sively inthe gas phase. Furthermore, the influence of the temperature on the absorption was examined by conducting the experiments both at 20 °C and at 0 °C. On the basis of the experiments with the 09 molar sodium hydroxide sol- ution and investigations withthe COs-airvater system fst the mass transfer eoeicients Pytoy ion inthe purely gas and liquid phases are determined. The enhancement factor Ewa then determined from all additional experiments The results of an absorption test with O01 molar sodium hydroxide solution at 20 °C are described by way ofan ex- ample Fig.2 explains the influence of the Hatta nember Ha on the enhancement factor E. The acceleration factor ina reac- tion E of afinite velocity serves a the parameter, If there is a pseudo first order rapid reaction, i.e. 3 Sr 2 oa if i : g \ ap tHsog} * oe }- - ~— | ea ay 4 8 12 16 20 p-value and concentration Cont: Coy Crssa,s Cscf in molims ig 3. Calelated conceatation profes slog the hight of «mass tater ‘column for selected measurement pont with 25 rum pase Hilow Rings, ‘SOzair/ 0,01 molar sodium hydroxide solution T= 203K: p= 1 bar 7 =| Nit 02 = 3 3 ow : : i aalteaaees 3 I Sean g 08 i—t ; wou [ne 0,8] L 1 I 7 eth ea mas arta MTU ave guie mass vans esttance AU i Fe oe aaa en ‘he mas transfer lon the eight ofa as reer column for selected ‘easuremsen point wit 25mm plat Hiiow Rings. gram includes the resistance in the gas phase for the mass transfer HTUy/HTUoy: It can clearly by seen that in the "upper segment of the column there is a resistance almost ex- clusively in the gas phase for the mass transfer, and the (Che, Eng, Techno 21198) 2 (© WILEY-VCH Velag Gt, D.49469 Weinheim, 1998 < Full Poper Ker HTUoy values are very low, The resistance in the gas phase drops already in a height range in which very high pH values atl predominate, i.e, the sulphite formation and the break- down ofthe sodium hydroxide solution has not been com- pleted, The HTUoy values increase along the column to about triple and the mass transfer resistance inthe gas phase drops to about 32%. The greatest changes take place in a pH range between 11.5 and 4, in which, for instance, the Wellmann-Lord process is used for desulfurization of flue gases. In this process, sodium sulphite is used forthe absorp- tion of SO; and is converted into sodium hydrogen suphite. The experimental investigations were conducted in an ex- perimental column with a diameter of 0.16 m and a mass- transfereffcient dumping height of 1.12 m and was filled with 25 mm Hiflow rings made of PVDF. The ambient ait was frst saturated with water vapor and charged with maxi- mum 4% sulphur dioxide. Thereupon, the gas mixture flo- ved into the bottom of the column and upwards through the dumped packings. ‘The fresh sodium hydroxide solution prepared with deion- ined water was introduced atthe top of the column via a liq- uid distributor with 2000 drainage points per m? cross-sec- tional area of the column and trickled down internally toward the gas stream, ‘The tests were conducted at 20 °C and 50 °C with various liquid and gas loads as wel as six sodium hydroxide solution concentrations so that over 100 measurements were avail- able for the verification of the model. The Hatta numbers ‘were determined from the enhancement factors determined in the experiments with the help of Eq. (17) and the con- stants of the reaction velocities ky and kz according to Eq. (14). The evaluations showed that the constants ky and k, can be best described by Eqs. (45) and (46) as a function of the temperature T. On the basis ofthe measurements the kz values published by Wang and Himmelblau were con- firmed 4 For the velocity constant k; in [Us] Eigen"! gives the value 34x 108, Roberts and Friedtinder'" 20, Phipps 2.8 and Wang and Himmelblaw*!! 2.2 x 107 at 20 °C, i. e., they differ by a total of 8 powers often. The value of Eq, (4S) lies inthe mid-range ofthe values given inthe literature with ky = 101 Ue 280 exp(=22") (4s) k, =2738 exp| 2) 6) On the basis of the calculation process described, the height of the dumped packings was precalculated theoreti- cally according to Eq. (18) for the set experimental operat- ing parameters, and compared with the mass-ransfereffi- cient dumping height installed in the experiment. The mean deviation between the calculated and the experimental heights was only 10%, whereby 80% of all the values com- 207Er rut Paper pared display a deviation of less than 15%. This good agree- ‘ment confirms the algorithm of calculation presented. In many desulfurization processes under industrial condi- tions, higher sodium sulphite concentrations are being used than described above. Under these process conditions itis shown that, even in the case of medium-range pH values, the influence of the mass transfer resistance in the gas phase on the SOz absorption rate prevails and the resistance in the iq- uid phase remains low. This relationship can also be taken {nto account in the calculation model if the velocity constant kis also a function of the ionic strength. Analogous rela- tionships were already recognized in the case of comparable chemical mass transfer rates. This still remains to be docu ‘mented by additional investigations. Finally, Figs. 5 and 6 illustrate once more the powerful in- ‘uence of the enhancement factor and, in particular, of tem- perature on the gas phase mass transfer unit HTUoy and on the gas phase mass transfer resistance HTUy/HTUov. The diagrams were made for a plastic Ralu-Flow No. 2, whereby the volumetric mass transfer coefficients were calculated for an operating point at the loading point of the internals with, the help of Eqs (19) to (23). ‘SO sir / sodium hyo soaion, + ar 2 e ra ~ 2 un = Raney 5 eh 3 aan we : wo $4 a 2 ° ‘ = 10 0 100 —~—~S00 100 Enhancement factor E Fig. 5. Dependence ofthe height of a ass transfer uit onthe temperature and onthe eahancement factor in the ese of chemical absorption of SO, ia ‘odkum hydroride elton fr plas Rale-Flow No.2 It can clearly be seen that the E-factors below 100 lead to very marked increase in the gas phase mass transfer units and to a serious drop in the gas phase mass transfer resis- tance, From both illustrations it can be seen that the changes increase as the absorption temperature increases. Fig. 5ilus- ‘ates furthermore that a constant HTUoy value can be achieved at increasing absorption temperatures only with in- creasing enhancement factors, i. ¢, increased concentrations of sodium hydroxide solution 208 yp 185 my 3 10 0 100 Enhancement factor E 500 1000 ig 6 Dependence ofthe distribution of sistance on the eabsceent fa tot forthe mass rae of $0. in sodium hyérovide slain inthe ase of Airing absorption ermpersures fora plate Rae Flow No.2. 8 Conclusion ‘The theoretical calculation approaches presented and the experimental results described illustrate that, during the course of the purification of effluent gases containing sul- phur dioxide with sodium hydroxide solution, the reaction velocities and reaction equilibrium in the solvent can change ‘markedly. This must be taken into account in the sizing of flue gas desulfurization plants and leads to the conclusion that the height of a gas phase mass transfer unit increases and the gas phase mass transfer resistance along the path of flow of the liquid drops. The changes are dependent on the absorption conditions and increase as temperature rises. The experiments can be reproduced well withthe help ofthe cal- culation methods described, Receive: May 5,197 [CETSI2} Symbols wsed Ao [Uimolf’® Constant 2 [axtim’] Total surface area per unit packed volume am ain’}__Eifeetive interfacial area per unit packed volume By Wimoll® Constant ¢ — [kmoVm'] Concentration in liquid cucy Constant Cy mol] Constant én ton] Hydraulic diameter D— (ms) Diffusion coefficient E Enhancement factor (Chem. Brg. Techno 21 (1998)2fe. - — Full Paper Ger F —[Omol] Faraday constant: F = 96485:31 Col sg (m4) Gravitational constant Hm} Height Ha Hatta number He [bar] Henry constant hmm’ Liquid hold-up hb [imol} Salting out parameter 1 moll} Molarity HTU [ml Height of mass transfer unit HU, [om] Overall height of mass transfer unit K__ (mol}, {molt}? Equilibrium constant for reaction ks) Velocity constant for reaction ke [m'kmols] Velocity constant for reaction L fmol] Molare low of liquid M = — [kglmol] Molecular weight N — [kmol/m’s] Molar flow of transfering component 4 [kmol/m’s} Molar flow of transfering component NTU, Overall number of transfer unit p bar] Pressure R [fol K] Natural gas constant T iki ‘Temperature ti Time uy {m'/(m’s)} Liquid toad uy [ous] ‘Superficial gas or vapor velocity V—[hmot] Molar flow of gas or vapor Y [kmot/kmol] Mote fraction in gas or vapor phase 2 [a Coordinate 4 Valency of on Fm Mass transfer coefficient ? ‘Activity constant e {m°/m] ‘Void fraction n— [kpims) Dynamic viscosity i Stripping factor # Conductivity of ion y [mts Kinematic viscosity p— keim') Density o Tkgis*) ‘Surface tension 6 fic) Electrical potential Chem. Eng, Teck 21 (1998)2 Subscripts B Base eff Effective 1 Ion i —Instantancous L Liquid Ph Interfacial area Vv Gas Wo Water 5 Core of phase References (i) Bais, Gs Scat, Chem Eng. Set 30 (1979p. 617. [2] Bie, R, Pesach der Pakula fr Maschinenbax, Rohr-Univesiat ‘Bochn 983) p24 [3] Bile, Fae Set Technol 9 (190) p36. (a) Bie, RSchues ML, Chem Eng. Techn 16 (199). [5] Chang, © S: Rochelle GT Ind Eng, Chem Pandan 24(1985)p. 7. Io) Dankworts PV, Gar Lipid Realfons McCraw Chemical Eogin- xing Series, 1970 (7) Doratwamy, I. K; Sharma, MM. Heterogenous Realions Anais ramps ond Reaktor Dslr, Yin Wiley Soas New Yor, VoL 2 984) (s] Harned, H, $; Owen, BB, The Physic! Chemistry ofElecoye Sob tin Editon; Reinbols, New Yor 1958. (o] Hin, H,; Ass, Tash T, AICHE Journal 23 (IST) No 4p. 538. (80) Hobie, fae Taner and Absorption Pergamon Press, Frankfurt 1966. {ut} Onde, KT; Kobayashi T; Fjng Ms Takabasti, M, Chem. Bg. Sok 25 1971p 208, 2027 12] Otunder D. Re, AICHE J 6 (1960) p. 238, 113} Panik Broakowska, W; Rodkinsky, K 3, Chem. Eng Se 46 (951) pz. (ig) Roberts. Ls Feedades, SK, AICHE J, 26 (1980) No 4p $98, [05] Schutes M, Diseraion an der Rabe Unverit Bochum 1990, Fo schrttBedchto VDERebe 3, Nr, 230, 4000 Duseldort 1, VDVerag m0. [06] Van Krevlen, DW: Hofer, PJ nt Congr Chim. Ind, 2s, Brus sel Special No. (198) 1658. {07} Vague, Gz Attrrena, Chen, F:Palo,E, Che Eng Teno 11 (OSH p56, [18] VD Wamendar VDLVerag Diseltoe (1991). 119} Ye erechnungsmatoden; Te 7, Sttésten, VCH: ering Weise 1985 (20) Vinograd, 1 R: MeBrian, 2. W.J-Am. Chem Soe. 63 (190) p, 2008, (21) Wong C; Hiamelbia,D.M, AICHE 10 (186) p14 [22] SchliogM, Abpurengung Springer Vea, Heldlbrg 196, (© WILEY-VCH Verlag GubH,D-65469 Weise, 1998 209