Business & Economics

Absorption of sulfur dioxide into aqueous reactive slurries of calcium and magnesium hydroxide in a stirred cell

Description
Chemical absorption of pure SO2 into aqueous slurries of fine and reactive Ca(OH)2 and Mg(OH)2 was studied in a stirred vessel at at realistically high mass transfer coefficients. The absorption process was theoretically analyzed using two different
Published
of 7
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
Share
Transcript
  * Corresponding author.Chemical Engineering Science 56 (2001) 1095 } 1101 Absorption of sulfur dioxide into aqueous reactive slurries of calciumand magnesium hydroxide in a stirred cell Manoj V. Dagaonkar   * , Antonie A. C. M. Beenackers  , Vishwas G. Pangarkar    Department of Chemical Engineering, Uni v ersity of Groningen, Nijenborgh 4, 9747 AG Groningen, Netherlands   Department of Chemical Engineering, Uni v ersity of Mumbai, Matunga, Mumbai 400019, India Abstract Chemicalabsorptionof pure SO   intoaqueousslurries of  " ne and reactiveCa(OH)   and Mg(OH)   wasstudiedin a stirredvesselat298 K at realistically high mass transfer coe $ cients. The absorption process was theoretically analyzed using two di !  erent models.Forthe SO  } Ca(OH)   system,a single-reactionplanemodel was usedand for the SO  } Mg(OH)   system, a two-reactionplane modelincorporating the solids dissolution promoted by the reactions with the absorbed SO   in the liquid  " lm was employed. A correctprocedure was adopted to estimate the contribution of the suspended particles to the enhancement of gas absorption. Theoreticalenhancement factors thus obtained compared well with the experimental data. The extra enhancement observed for the SO  } Mg(OH)   system could be explained from the reaction between SO   and the dissolved [SO  ]  .    2001 Elsevier Science Ltd. Allrights reserved.  Keywords:  Reactive solids; Enhancement; Solid dissolution; Calcium hydroxide; Magnesium hydroxide; Sulfur dioxide 1. Introduction Slurryreactorshaveawidespreadapplicationin chem-ical and bio-chemical industries. The problem of gasabsorption with reaction in a slurry containing  " ne par-ticles has become important in the development of pro-cesses for the removal of acidic pollutants. Mg(OH)   assuspended solids may yield a high scrubbing capacity asa result of the presence of the more soluble reactionproduct magnesium sul " te, relative to the correspondingcalcium salt (Sada, Kumazawa & Butt, 1977). The pres-ent work focuses on the enhancement of the absorptionrate of a gas into a slurry of small reactive particles. Theelementary processes involved in chemical absorptioninto the slurry are: (i) di !  usion of the solute gas in the " lm, (ii) chemical reaction and (iii) dissolution of solid.Applying the so-called  " lm theory for mass transfer, thechemical absorption and the solids dissolution are trans-fer processes either in series or in parallel, dependingupon whether the suspended particles size is signi " cantlysmalleror larger than the thickness of the liquid " lm ( " lmmodel,  " lm thickness " D /  k  ). The solids dissolution inthe liquid  " lm enhances the absorption rate and furtherthe rate of solute dissolution is enhanced by the reactionbetween the dissolved gas and the dissolved solid in theliquid  " lm when the particle size is signi " cantly smallerthan the  " lm thickness. As a result, the rate of gasabsorption is a !  ected by the solid dissolution rate as wellas the chemical reaction rate. The hydroxide particlesbeing reactive, help to increase the rate of absorption of SO   in the slurry.This problem has been discussed on the basis of the " lm model (Ramchandran & Sharma, 1969) consideringthe e !  ect of solids dissolution in the liquid  " lm to beboth, important and not important depending on theconditions. Later, Uchida, Koide and Shindo (1975)modi " ed the model proposed by Ramchandran andSharma and pointed out that the rate of solids dissolu-tion is enhanced by the reaction between the absorbedgas and the dissolved solid in the liquid  " lm. Sada et al.(1977), Sada, Kumazawa and Butt (1979) and Sada,Kumazawa, Sawada and Hashizume (1980) formulatedthe process of gas absorption in the slurry on the basis of the  " lm model incorporating instantaneous reactionsbetween the absorbed gas and the dissolved solid in theliquid of the  " lm. Their model assumes that solids dis-solution in the " lm for mass transfer is one of the elemen-tary steps. This is the case when the average size of the 0009-2509/01/$-see front matter    2001 Elsevier Science Ltd. All rights reserved.PII: S0 0 09 -2 5 0 9 (0 0 ) 0 0 3 26 - 2  Fig. 1. Stirred cell.Table 1Size distribution of the hydroxide particles% particles Particle size (  m)Ca(OH)   Mg(OH)  10 6.09 29.1625 4.711 24.4450 3.516 18.3875 3.339 16.5790 3.238 15.65Mean size: 4.352   m 21.204   m suspended particles is considerably smaller than thethickness of the  " lm. The reaction was interpretatedboth, by a single-reaction plane model (Sada et al., 1977)and a two-reaction plane model (Sada et al., 1979), re-spectively. However, the experimental results for highSO   concentrations could not be interpretated by theproposed models (Sada et al., 1980), possibly becausethese did not incorporate the fact that the solids dissolu-tion in the liquid  " lm can be enhanced by the chemicalreactions.Sada, Kumazawa, Sawada and Hashizume (1981) de-veloped a two-reaction plane model incorporating thesolids dissolution enhanced by the reactions in the liquid " lm. The theoretical enhancement factors compared wellwith the experimental data. Previous authors (Sada et al.,1977, 1979, 1980, 1981; Uchida et al., 1975) have mea-sured the enhancement factor at a speed range of 1 } 5 s  which gives a very low mass transfer coe $ cient(2 } 4  10   m/s) which is much lower than that appliedin industrial practice. The purpose of the present work isto present the absorption data for the removal of SO   bymicro-sized reactive particles of calcium and magnesiumhydroxide at stirring speeds of 3 } 15 s   ( Re ' 10  ) andto check whether the experimentally observed enhance-ment factors can be described by a suitable model. 2. Experimental The experiments were carried out in a thermostattedreactor (0.105 m dia., 1.8  10   m   capacity) made of glass and stainless steel as shown in Fig. 1. A six-bladedturbine stirrer was located centrally in the liquid ata height above the reactor bottom equal to half thereactor diameter. Four symmetrically mounted glassba % es increased the e !  ectiveness of stirring and pre-vented the formation of a vortex. The pressure and tem-perature transducers together with valves 1 and 2 wereconnected to an Olivetti M240 computer, thus enablingautomatic data collection and programmed reactoroperation. After " lling the reactor with the desired slurry,the liquid was degassed by closing valve 1 and openingvalve 2. Once the slurry was equilibrated under the va-pour pressure of water, N  O was fed to the reactor up toa  " xed pressure (8  10   Pa). Then, the stirrer was startedand the decrease of pressure due to the physical absorp-tion of N  O was recorded over time. These data wereused to estimate the solubility of the gas and the liquidside mass transfer coe $ cient.After the physical absorption experiments, the chem-ical absorption of pure SO   into aqueous slurries of Ca(OH)   and Mg(OH)   was carried out. The experi-ments were carried out in both, batch mode, with respectto gas phase and the slurry solution and semi-batchmode, where the gas was continuously supplied into thereactor. The volume of the slurry loaded in the reactorwas always kept at 10   m   and the slurry concentrationwas varied from 0 to 20 wt%. The reactive particles of Ca(OH)   and Mg(OH)   of size 4.35 and 21.20   m wereused for the experimentation. Table 1 gives the sizedistribution of the hydroxide particles used in theexperimentation.The rate of SO   absorption in the slurry follows from J  a " <  <  R ¹  ! d P  d t   " k  EC  . (1)The experimental enhancement factor was calculatedby taking the ratio of the initial rates in the presenceof solids and in the absence of suspended solidparticles i.e. the saturated solution of the hydroxideinvolved. 3. Theory of gas absorption [OH]   ions are fed by the dissolution of the solidparticles in the liquid  " lm. In the case of theSO  } Ca(OH)   system the reaction between SO   and[OH]   is instantaneous and the product of the reactionCaSO   is insoluble in the medium (Sada et al., 1981). The 1096  M.V. Dagaonkar et al.  /   Chemical Engineering Science 56 (2001) 1095 } 1101  Fig. 2. Concentration pro " le for SO  /Mg(OH)   slurry [(a) no sus-pended solids, (b) in the presence of suspended solids]. reaction scheme for this process of gas absorption can berepresented as:SO    P SO    , (i)Ca(OH)    P [Ca]   # 2[OH]  , (ii)SO    # 2[OH]  P [SO  ]  # H  O. (iii)As the rate of solid dissolution is enhanced by the instan-taneous reaction of SO   and Ca(OH)  , the model pro-posed by Uchida et al. (1975) can be used to describe theabsorptionprocess.Therate ofgas absorptionis givenbythe expression: J  " mD  A H coth m  # mD  C  z   coth m  ! 1sinh m    .(2)The parameter    can be calculated by the equation: D  C  2   coth m  # coth m (  !  ) ! 1sinh m   ! D  A H sinh m  " 0 . (3)When the solution contains no suspended solids, theexpression becomes J  " k  A H  1 # D  B  D  A H   . (4)In the SO  } Mg(OH)   slurry process, however, the prod-uct of the reaction MgSO   has a much higher solubilityin water than that of Mg(OH)  . The MgSO   formedexists in a dissolved state. Thus the dissolved SO   alsoreacts with [SO  ]   and forms [HSO  ]   which in turnfurther enhances the rate of absorption. Thus, dissolvedSO   is consumed bySO  # 2[OH]  " [SO  ]  # H  O , (I)SO  # [SO  ]  # H  O " 2[HSO  ]   , (II)[HSO  ]  # [OH]  " [SO  ]  # H  O . (III)In the process of SO   absorption in Mg(OH)   slurrywith no suspended particles, [HSO  ]   cannot coexistwith [OH]  , so that reaction (I) never takes place dir-ectly [Fig. 2(a)]. The above consideration shows thatreactions (II) and (III) take place at two di !  erentlylocatedplanes in the two reaction plane model. However,in the slurry process, both dissolved SO   and the[HSO  ]   to be produced by reaction (II) can react with[OH]   which is fed by the dissolution of the solid par-ticles in the liquid  " lm. So, dissolved SO   can be con-sumed by reactions (I) and (II) simultaneously. Fora saturated solution of magnesium hydroxide, a plausiblesketch of the concentration pro " le is given in Fig. 2(a).When the particles are suspended in the liquid  " lm, theconcentration pro " les shift as shown in Fig. 2(b). Themass balances for the relevant species in regions I } IIIare as follows:  Region  I: D  d  C  d z  ! k  2   1 # 2 D  C  C  D    A  C  " 0 , (5) D  d  C  d z  ! k   1 # D  C  C  D   A  C  " 0 . (6)  Region  II: D  d  C  d z  ! k   1 # D  C  C  D   A  C  " 0 , (7) D  d  C  d z  ! k   1 # D  C  C  D   A  C  " 0 . (8)  Region  III: D  d  C  d z  # k  A  ( C  ! C  ) " 0 , (9) D  d  C  d z  " 0 . (10)The boundary conditions imposed are:At  z " 0,  C  " C  , d C  /d z " 0 , (11) z " z  ,  C  " C  " 0,  C  " C   , (12) ! D  (d C  /d z ) " D  (d C  /d z ) , (13)  M.V. Dagaonkar et al.  /   Chemical Engineering Science 56 (2001) 1095 } 1101  1097  Fig. 3. Enhancement factor ratio as a function of solid concentration: (a) SO   absorption in Ca(OH)   slurry [Lines 1 } 4: from theoretical values of enhancement factor predicted according to Uchida et al. (1975)]. (b) SO   absorption in Mg(OH)   slurry [   and   : experimental values of [ E /  E H   ( E /  E  )   ] at 10.8 and 5 s   respectively, Lines 1 and 2: Theoretical values (Sada et al., 1981) with correct  N /  w  values, Line 3: Theoreticalvalues predicted by Model II]. (c)   : SO   absorption in Mg(OH)   and   : Ca(OH)   slurry in semi-batch mode at 10.8 s  . [Lines 1 and 2 aretheoretical values of enhancement factors]. At  z " z  ,  C  " C  " 0,  C  " C   , (14) D  (d C  /d z ) "! D  (d C  /d z ) , (15)At  z " z  ,  C  " C  ,  C  " C    . (16)The expression for the enhancement factor as suggestedby Sada et al. (1981) is E "  1 # 12 r  q      N tan   Nx  !   12 r  q      N Sinh   Nx  ,(17) E   represents the enhancement factor for a clear solutionsaturated with the hydroxides and is de " ned by E  "  1 # 12 r  q    . (18) 4. Results and discussion To show the contribution of the presence of the solidsto the absorption rate in a batch mode, the ratio of enhancement factor into slurry to that into saturatedsolution ( E /  E  ) is plotted against wt% solids in Fig. 3(a),(b) and for a semi-batch mode in Fig. 3(c). The ratio  E /  E  represents the degree of enhancement owing to the pres-ence of solid particles in the slurry. For the case of SO  absorption in Ca(OH)   slurry, the concentration of [SO  ]   (in the bulk) to be produced by the reaction of SO   and Ca(OH)   is extremely low, because the solubil-ityofCaSO   in wateris about25 times lowerthanthat of Ca(OH)  . Consequently, the reaction between dissolvedSO   and [SO  ]   can be neglected. The solid curvesplotted in Fig. 3(a) (Curves 1 } 5) and Fig. 3(c) (Curve b)were obtained from the theoretical values of the enhance-ment factor. These theoretical values are calculated usingthe model proposed by Uchida et al. (1975), by usingEqs. (2) } (4).In order to compare the experimental results with thetheoretical predictions for the SO  } Mg(OH)   system, itis necessary to know the values of the dimensionlessparameters  r ,  q  and  N . For the evaluation of   r , thedi !  usivity of SO   in the slurry was assumed to be thesame as that in pure water. Thus, the value of   r  is 1.22 forthis system. The value of   q  was calculated from the ratioof concentration of SO   at the gas } liquid interface andthe solubility of the hydroxide in water.  x   and  x  were calculated computationally by using the equations 1098  M.V. Dagaonkar et al.  /   Chemical Engineering Science 56 (2001) 1095 } 1101  Table 2Values of the solid dissolution parameter at di !  erent solid loadings forSO  /Mg(OH)   slurry processSpeed (rpm) Solids  k   10   z " D /  k   N /  w (kg/m  ) (m/s) (  m)300 0.01 4.4 31.8 11.310.04 4.4 31.8 11.310.08 4.4 31.8 11.310.16 4.4 31.8 11.31650 0.01 5.6 25.0 7.580.04 5.1 27.45 9.140.08 4.9 28.57 9.910.16 4.9 28.57 9.91 suggested by Sada et al. (1981). The position of theprimary reaction plane was smaller than the averagediameter of the suspended particles; typically,  z  /  d   wasin the range of 0.4 } 0.45.Fig. 3(b) and (c) (Curve a) gives the variation of   E /  E  with the solid loading for the absorption of SO   inMg(OH)   slurry in a batch mode and in semi-batchmode, respectively. The observed enhancement in the gasabsorption during the semi-batch mode was found to bemore than that of the batch mode which could be due tothe constant driving force for the gas absorption in theearlier case.For the estimation of the parameter  N , Sada et al.(1981) compared their experimental results with the the-oretical prediction according to model II (Sada et al.,1979).However,as model II did not take into accounttheextra reaction between SO   and [SO  ]   (reaction (II))thevalues of experimentalenhancementfactors predictedby Sada et al. (1981) did not match with the theoreticalvalues predicted by Model II.The enhancementobserved in this systemis consideredto be due to both, the presence of solid particles (react-ants of reactions (I) and (III)) in the liquid  " lm and theextra reaction of SO   and [SO  ]   (reaction (II)). Theconcentration of [SO  ]   is allowed to be a function of time.[SO  ]   is producedin the " lm bythe reactionof SO  and the hydroxide ions (reaction (I)) and also by thereaction of [HSO  ]   and the hydroxide ions (reaction(II)). The consumption of [SO  ]  , however, is only byits reaction with SO  . Hence, the rate of generation of [SO  ]   is greater than its consumption in the  " lm, dueto whichthe concentrationof [SO  ]   starts building upin the  " lm. The concentration of [SO  ]   in the  " lm islarger than its concentration in the bulk liquid due towhich [SO  ]   di !  uses in the liquid bulk.The enhancement caused by the reaction between SO  and the [SO  ]   formed was assessed by extrapolatingthe observed enhancement in the presence of solids tothat for a saturated solution of magnesium hydroxide,that is ( E /  E  )   . This value was found experimentallyto be 1.8 at 10.8 s   and 2.4 at 5 s   Fig. 3(b)) and 2.1 at10.8 s   for the semi-batch process Fig. 3(c)).To estimate the contribution of solids to the observedenhancement factor, the degree of enhancementwhich was also due to reaction (II) was avoided bySada et al. (1981) by converting the ratio  E /  E   to thequantity [( E /  E  ! ( E /  E  )   ! 1)], where the quantity[( E /  E  )   ! 1] correspondedto the extra enhancementdue to reaction (II). Model II predicts the data of e !  ect of addition of solids on enhancement factor at di !  erentvalues of   N /  w . Sada et al. (1981) thus, estimated the valueof   N /  w  to be 12.3 by comparing their experimental  " nd-ings with the theoretical values predicted by Model II.However, their procedure of calculating the contributionof solids to the enhancement factor does not seem to becorrect as the enhancement e !  ects are multiplicative andhence they cannot be subtracted. Hence, in our calcu-lations, the ratio of   E /  E   was converted to the quantity[ E /( E  .( E /  E  )   )], to estimate the e !  ect of solid on theenhancement factor where the quantity ( E /  E  )    ac-counts for the additional increase in the enhancementdue to the reaction (II) which is represented by  E  . Thevalues of [ E /( E  .( E /  E  )   )] are represented as darkpoints in Fig. 3(b) and (c). These dark points show thecontribution to the enhancement in the rate of absorp-tion due to the presence of the suspended solids only.For comparison, the values predicted by Model II for N /  w " 12.3 are shown in the form of a solid line inFig. 3(b). The di !  erence in the values between our experi-mental  " ndings and line 3 is more, especially at highersolid loadings, which is due to the more accurate proce-dure for the estimation of contribution of solids in theenhancement used in our calculations.For the slurry process, the Sherwood number( k  d  /  D  ) gives the value of solid dissolution parameter N  from the equation: N " Sh (6 w /   )( z  /  d  )  . (19)This indicates that the quantity  N /  w  is dependent on Sh  and  d  . For calculation of   N /  w ,  Sh " 2 was taken inour calculations with an assumption of spherical par-ticles, which resulted in the value of   N /  w " 11.31 for5 s  . The values of   N /  w  at the various solid loadings areshown in Table 2.This value of   N /  w  was used to calculate the theoreticalenhancement factors due to the presence of solids usingEq. (17), which are indicated by the three solid curves(1 } 3) as represented in Fig. 4. Fig. 4 shows the variationof the enhancement factor with time as observed forreactive SO   absorption in a Mg(OH)   slurry. It can beseen that the enhancement is constant at the start forsome time and then decreases and ultimately becomeszero. Initially, at higher partial pressures of SO  ,the consumption of [SO  ]   by reaction (II) isbalanced by its regeneration by reaction (III). Hence, the  M.V. Dagaonkar et al.  /   Chemical Engineering Science 56 (2001) 1095 } 1101  1099
Search
Tags
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks