Microsoft Word - szolcs_eloszo.doc HUNGARIAN JOURNAL OF INDUSTRIAL CHEMISTRY VESZPRÉM Vol. 32. pp. 65-71 (2004) SEPARATION BY ION EXCHANGE AND ADSORPTION PARAMETRIC PUMPING V. Cycle Efficiency for Batch, Thermal Ion Exchange Parametric Pumping L. HANÁK, T. SZÁNYA and P. SZOLCSÁNYI Department of Chemical Process Engineering, University of Veszprém, Veszprém, HUNGARY The theoretically best attainable separation can easily be determined in the function of the theoretical cycle number. In the reality, due to the non-equilibrium ion exchange, the mixing phenomena and the “run out” of characteristics, the performance of the actual process, especially in the production of high purity materials, a cycle efficiency value can be ordered to the parametric pumping processes. The authors investigated the relationships between the more important process and operation parameters and the cycle efficiency in aqueous media and in the presence of VARION-KSM ion exchange resin, Ca2+, K+ cations and Cl– anions. Introduction Production of high purity materials accounts for the major production cost in fine chemical industries. There has been a growing demand for economical and energy-efficient separation processes. The increasing interest in this area is also reflected in the numerous research articles published in the field of ion exchange and adsorption separations [1-5]. Among these processes parametric pumping is a technique which has been studied extensively [6-8]. The most important property of this process is the fact that it is a reagent-free separation process in which no regeneration chemicals are needed which may pollute the environment, and in which no regenerant can contaminate either product. Emphasis is placed on the following areas: • Development of adsorbents and ion exchange materials [3-4, 9-10]. • Investigation of possible thermodynamic parameters [11-13]. • An improved theoretical understanding of the process [14-15]. • Development of new process configurations [16-17]. • Development of highly efficient processes [18]. In this paper experimental data were compared with the best attainable separation achieved by the application of theoretical cycles. Cycle efficiency (ηi) was defined as the ratio of the theoretical and actual cycle numbers. It presents an interesting possibility in the evaluation of experimental results, especially in the production of high purity ionic compounds. Experimental part An aqueous solution of calcium and potassium cations in the presence of chloride anion on a highly acidic cation exchanger (Varion-KSM, Nitrochemical Works of Balatonfűzfő, Hungary) was selected as a model system. 66 The ion exchange isotherms were determined by simple batch equilibration of a known mass of resin with a given volume of solution of known concentration and ionic composition over 36 hours at temperatures 288K and 333K. The solution was then separated from the resin and the concentration of calcium and potassium in solution were measured. The ion exchanger phase concentration was calculated by mass balance. A preliminary summary of these data have been reported elsewhere [19]. The experimentally determined ion exchange isotherms and the ion exchange selectivity coefficient ( CaKT ) as a function of resin phase composition (yCa) are shown in Figs. 1-3. 0 20 40 60 80 0 0.2 0.4 0.6 0.8 1yCa++ TC a+ + / K + 288K 333K Ca++-K+ -Cl- Varion KSM Co= 0.10 equ/dm 3 0 20 40 60 80 0 0.2 0.4 0.6 0.8 1yCa++ TC a+ + / K + 288K 333K Ca++-K+ -Cl- Varion KSM Co= 0.10 equ/dm 3 Figure 1. Ion exchange equilibrium of Ca2+ -K+ cations at temperatures of 288K and 333K, with a total cation content of 0.10 equ/dm3 Knowledge of ion exchange isotherms at two temperatures is required numerically for the mathematical modelling of the process. In our case, the ion exchange selectivity coefficient ( CaKT ) as a function of resin phase composition (yCa) was used for the definition of ion exchange equilibrium. The equilibrium data were fitted to a third degree polynome of the following form: dycybyaT Ca 2 Ca 3 Ca Ca K +++= (1) A least square optimization of the equilibrium data gave the constants of Eq.(1). These functions and constants were used in all calculations (Table 1.). Table 1. Constants of the ion exchange selectivity coefficient at two temperatures C0=0.10equ/dm 3 C0=0.50equ/dm 3 C0=1.0equ/dm 3 288K 333K 288K 333K 288K 333K a -38.17 26.55 -0.727 -14.9 5.485 5.99 b 99.12 7.25 12 33.28 -8.063 -8.748 c -100.4 -100.2 -22.89 -38.9 0.596 -0.936 d 40.65 70.03 12.15 21.2 2.186 4.02 The parametric pumping experiments were carried out in a laboratory scale, automatic apparatus, described in detail in the previous paper [20]. Different size glass columns with jacket and reservoirs are included in the system for direct heat transfer experiments. 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1xCa++ y C a+ + 288K 333K Ca++-K+ -Cl- Varion KSM Co= 0.50 equ/dm 3 0 5 10 15 20 25 0 0.2 0.4 0.6 0.8 1yCa++ TC a+ + / K + 288K 333K Ca++-K+ -Cl- Varion KSM Co= 0.5 equ/dm 3 Figure 2. Ion exchange equilibrium of Ca2+ -K+ cations at temperatures of 288K and 333K, with a total cation content of 0.50 equ/dm3 The effect of the following more important operation and process parameters were investigated: • Liquid flow rate (vo, from 0. 1 to 2. 3 m/s) • Transfer of ion fraction during a half cycle(α’, from 0.026 to 1.0 (dimensionless)) • Initial ion composition (xCa,0, from 0.2 to 0.8) • Length of ion exchanger column (L, from 0.5 to 1.5m) • Total solution concentration (C0, from 0.1 to 1.0 equ/dm3). 67 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1xCa++ y C a+ + 288K 333K Ca++-K+ -Cl- Varion KSM Co= 1.0 equ/dm 3 0 1 2 3 4 5 0 0.2 0.4 0.6 0.8 1yCa++ TC a+ + / K + 288K 333K Ca++-K+ -Cl- Varion KSM Co= 1.0 equ/dm 3 Figure 3. Ion exchange equilibrium of Ca2+ -K+ cations at temperatures of 288K and 333K, with a total cation content of 1.0 equ/dm3 The following variables and operational conditions were fixed: • The temperatures (288K and 333K) in cold and hot half cycles • Particle size of ion exchanger, dp = 0,5-0,7 mm • Initial condition of the experiments, the ion exchanger column was always brought into equilibrium at the lower temperature (288K) with the solution to be separated • Location of the cold reservoir and the hot reservoir, in our case, the hot reservoir was always connected to the upper part of the column • Length of thermostating time is 15 minutes in case of a column diameter of 28 mm and, 8 minutes in case of a 10 mm column diameter • Dead-volume of the equipment. According to the temperature dependence of the ion exchange equilibrium, the concentration of calcium ions decreased in the cooled reservoir (288K). In the other case the concentration of potassium ions decreased in the heated reservoir (333K). The total cationic concentration of the solution in the mentioned reservoirs was constant. A part of the experimental results and data was published in the previous paper [20]. Table 2-4 summarizes the investigated parameters and results for selected experiments at three solution concentrations. Publication of about 1000 determined reservoir concentration from 45 measuring series is not possible here, only the hot and cold reservoir concentrations of the 20th cycle are given. The Comparison of the Calculated and Experimental Results Using relationships presented previously for a given parametric pumping operation, the theoretically best attainable separation can easily be determined in the function of the cycle number. Knowing this, it presents an interesting possibility in the evaluation of experimental results as it is shown below. Among the experimental data, the reservoir concentrations of the 20th cycle were taken as basic data. (The sensitivity of the used ZEISS-AAS1 atomic absorption spectrophotometer was Table 2. Conditions and results for selected batch experiments at 0.10equ/dm3 solution concentration, L=1.0m, Db=28mm No. Xca,0 ∆V,dm 3 B,dm3/h C20 ,mequ/dm 3 NT ηi P20 ,m/m% α' vo,mm/s Ca 2+ Ca2+ Ca2+ K+ K+ K+ K+ Ca2+ 1 0.20 0.4 1.0 0.038 9.8 0.49 99.981 0.026 0.45 6.07 6.6 0.33 88.555 2 0.20 0.8 1.0 0.034 10.4 0.52 99.983 0.052 0.46 2.05 8.6 0.43 95.982 3 0.20 2 1.0 0.026 11.6 0.58 99.987 0.13 0.45 1.12 9.8 0.49 97.785 4 0.20 4 1.0 0.022 12.2 0.61 99.989 0.26 0.43 7.76 6.2 0.31 85.598 5 0.20 6 1 0.019 12.6 0.63 99.990 0.39 0.45 22.7 4.4 0.22 62.999 6 0.20 8 1 0.018 12.8 0.64 99.991 0.52 0.44 45.1 2.6 0.15 37.836 7 0.50 0.4 1.0 0.1 10.4 0.5 99.971 0.026 0.44 0.025 13.6 0.68 99.950 8 0.50 0.8 1 0.052 10.6 0.53 99.974 0.052 0.45 0.02 13.8 0.69 99.960 9 0.50 2 1 0.045 11.1 0.55 99.977 0.13 0.45 0.015 14.4 0.72 99.970 10 0.50 4 1 0.033 11.4 0.57 99.983 0.26 0.47 1.28 7.2 0.36 97.472 11 0.50 6 1 0.88 6.4 0.32 99.558 0.39 0.45 5.95 5 0.25 88.768 12 0.50 8 1 8.55 3.6 0.18 95.534 0.52 0.45 13.8 3 0.15 75.747 13 0.80 0.8 1 0.68 9 0.45 99.659 0.052 0.46 0.0033 14.4 0.72 99.993 14 0.80 2 1 13.4 5.4 0.27 92.819 0.13 0.45 0.0025 15.4 0.77 99.995 15 0.80 4 1 38 3.2 0.16 76.543 0.26 0.45 0.002 15.8 0.79 99.996 16 0.80 6 1 50.2 2.4 0.12 66.489 0.39 0.44 0.27 7.2 0.36 99.461 17 0.80 8 1 58 1.8 0.09 59.155 0.52 0.45 1.86 3.8 0.19 96.348 68 inadequate for the accurate determination of the low concentration of given component, after the 20th cycle). Using the calculated data, the theoretical cycle number of the former reservoir concentrations was determined, and the cycle efficiency (ηi) was defined as the ratio of the theoretical and actual cycle numbers. It is shown in the function of process parameters for Ca2+–K+ ion pairs, in the presence of Cl– anions, and in the case of VARION KSM cation exchanger resin. Table 3. Conditions and results for selected batch experiments at 0.50equ/dm3 solution concentration, L=1.53m, Db=28mm, ∆V=800cm 3, α’=0.17 No. XCa,0 B,dm 3/h C20 ,mequ/dm 3 NT hi P20 ,m/m% vo,mm/s Ca 2+ Ca2+ Ca2+ K+ K+ K+ K+ Ca2+ 1 0.20 0.2 0.014 15.6 0.78 99.999 0.09 0.92 9.2 0.46 99.908 2 0.20 0.4 0.018 15 0.75 99.998 0.18 0.88 9.4 0.47 99.912 3 0.20 0.7 0.02 14.8 0.74 99.998 0.32 14.9 5.6 0.28 98.487 4 0.20 1.1 0.025 14.4 0.72 99.997 0.51 20.7 4.2 0.21 97.886 5 0.20 1.6 0.032 13.2 0.66 99.997 0.71 23.2 3.6 0.18 97.625 6 0.20 2.8 0.74 7.6 0.38 99.926 1.24 25.1 3.2 0.16 97.425 7 0.50 0.4 0.11 13.6 0.68 99.989 0.18 0.07 12.2 0.61 99.993 8 0.50 0.7 0.24 12.2 0.61 99.976 0.31 0.11 11.2 0.56 99.989 9 0.50 1.1 0.56 8.6 0.43 99.944 0.50 0.83 8.2 0.41 99.917 10 0.50 1.6 0.76 6.4 0.32 99.924 0.71 2.1 7.6 0.38 99.790 11 0.50 2.8 0.9 4.8 0.24 99.910 1.24 6.8 4.8 0.24 99.315 12 0.80 0.4 4.2 9.6 0.48 99.578 0.18 0.004 16.6 0.83 99.999 13 0.80 0.7 14.9 8 0.4 98.487 0.31 0.005 16.2 0.81 99.999 14 0.80 1.1 38.7 5.4 0.27 95.974 0.48 0.01 15.5 0.77 99.999 15 0.80 1.6 44.2 4.8 0.24 95.376 0.71 0.025 13 0.65 99.997 16 0.8 2.8 51.5 4.2 0.21 94.570 1.22 0.23 8.2 0.41 99.977 In Figure 4 is shown the effect of flow rate, in case of 0.5 equ/dm3 solution concentrations, if various initial ion compositions are used. The cycle efficiency was increasing with the decrease of flow rate but, the measure of growth and the values of cycle efficiencies varied widely. If the initial ion composition is low, then cycle efficiency generally increases for a given ion and the effect of flow rate is small. In Figure 5 are shown the values of cycle efficiency as a function of the relative ion amount (α’) correspond to a half cycle, in case if various initial ion compositions and 0.1 equ/dm3 solution concentration. The curves reach maxima, indicating the lack of equilibrium. The place of the maxima shows the amount of relative ions (α’), which should be the load for the best separation. The role of the initial ion composition is clear: The maximum place of the given component is shifted to the larger loads, proportionally to the decrease of the initial ion composition. Table 4. Conditions and results for selected batch experiments at 1.0equ/dm3 solution concentration, L=1.50m, Db=10mm, B=0.087dm 3/h, v0=0.31mm/s N o . X ca ,0 ∆ V ,d m 3 C 2 0 ,m eq u /d m 3 N T η i P 2 0 ,m /m % α' C a 2+ C a 2 + C a 2 + K + K + K + K + C a2 + 1 0.20 0.1 0.035 16.1 0.80 99.998 0.17 6.01 10.6 0.53 99.699 2 0.20 0.1 0.01 18 0.9 99.999 0.34 10 9.6 0.48 99.497 3 0.20 0.2 0.01 18 0.9 99.999 0.68 426 3.4 0.17 72.935 4 0.20 0.3 0.01 18 0.9 99.999 1 552 2.4 0.12 61.878 5 0.50 0.1 0.65 13.8 0.7 99.967 0.17 3.96 9.4 0.47 99.802 6 0.50 0.1 0.13 16.4 0.82 99.993 0.34 0.17 15.6 0.78 99.991 7 0.50 0.2 0.18 15.8 0.79 99.991 0.68 0.08 16 0.8 99.996 8 0.50 0.3 76.80 5 0.25 96.007 1 131.00 3.2 0.16 92.991 9 0.80 0.1 2.72 13.2 0.66 99.864 0.17 0.65 11.6 0.58 99.967 10 0.80 0.1 1.12 15 0.75 99.944 0.34 0.01 17.2 0.86 99.999 11 0.80 0.2 354 3.6 0.18 78.493 0.68 0.006 18.2 0.91 99.999 12 0.80 0.3 500.3 2.4 0.12 66.640 1 0.61 11.8 0.59 99.969 In Figure 6 are shown the effects of the above discussed components in case of 1 equ/dm3 solution concentration. The results are the same as above, besides the considerably greater loads. In Figure 7 the effects of the total concentration of solutions are presented in the function of relative ion amounts, flown through during a half cycle. With the increase of solutions total concentration, markedly increases the cycle efficiency belonging to the maxima of the curves on one hand, while on the other hand the maximum cycle efficiencies appear always at higher loads of relative ion amounts. The equilibria related to Ca2+–K+ ion pairs depends strongly on the total concentration of solutions (Fig. 1-3.). Increasing the concentration, the selectivity values decrease and the isotherms of the so called sigmoid shape form. Figure 7 proves that the decrease of selectivity values improves the load of the column. Due to the sigmoid type isotherm, at the concentration of 1 equ/dm3, the shape and the 69 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 1.2 1.4 vo, mm/s η C a+ + Xca++,o=0.80 Xca++,o=0.50 Xca++,o=0.20 Co=0.50 equ/dm 3, L=1.0 m, Db=28 mm, α'=0.17 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 1.2 1.4 vo, mm/s η K + Xca++,o=0.80 Xca++,o=0.50 Xca++,o=0.20 Figure 4. The changes of cycle efficiencies in the function of the flow rate 0 0.2 0.4 0.6 0.8 1 0 0.1 0.2 0.3 0.4 0.5 0.6α' η C a+ + Xca++,o=0.80 Xca++,o=0.50 Xca++,o=0.20 Co=0.10 equ/dm 3, L=1.0 m, Db=28mm, vo=0.45 mm/s 0 0.2 0.4 0.6 0.8 1 0 0.1 0.2 0.3 0.4 0.5 0.6α' η K + Xca++,o=0.80 Xca++,o=0.50 Xca++,o=0.20 Figure 5. The changes of cycle efficiencies with α’ at different initial ion composition and at 0.1 equ/dm3 solution concentration 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 1.2α' η C a+ + Xca++,o=0.80 Xca++,o=0.50 Xca++,0=0.20 Co=1.0 equ/dm 3 vo=0.31 mm/s L=1.5 m, Db=10 mm 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 1.2 α' η Κ + Xca++,o=0.80 Xca++,o=0.50 Xca++,o=0.20 Figure 6. The changes of cycle efficiencies with α’ at different initial ion composition at 1 equ/dm3 solution concentration 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 1.2 α' η C a+ + Co=0.10 equ/dm3 Co=0.50 equ/dm3 Co=1.0 equ/dm3 vo=0.31 mm/s, L=1.5 m, Db=10 mm, Xca++,o =0.80 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 1.2 α' η K + Co=0.10 equ/dm3 Co=0.50 equ/dm3 Co=1.0 equ/dm3 Figure 7. The changes of cycle efficiencies with different solution concentration 70 0 0.2 0.4 0.6 0.8 1 0.5 1 1.5 L, m η C a+ + Xca++,o=0.80 Xca++,o=0.50 Xca++,o=0.20 Co=0.50 equ/dm 3, vo=0.18 mm/s, Db=28 mm, α'=0.13 0 0.2 0.4 0.6 0.8 1 0.5 1 1.5 L, m η K + Xca++,o=0.80 Xca++,o=0.50 Xca++,o=0.20 Figure 8. The changes of cycle efficiency with the length of ion exchange column position of cycle efficiency curves of Ca2+ and K+ are very similar, this is very advantageous, because the column can be operated with optimum load regarding both the ions. At last, in Figure 8 are shown the change of cycle efficiencies as a function of column height, in case of 0.5 equ/dm3 initial concentration and different initial ion composition. The cycle efficiencies increase in different extent with the column height. In given circumstances, the number of equilibrium units formed in the column is proportional with the length of the ion exchanger charge, filled in the column. This basically determines the level of the component separation. Summary The results of these calculations were used in the evaluation of results determined in the experiments with Ca2+–K+ ion pair, the idea of cycle efficiency was defined. This is the ratio of the theoretical and practical cycle numbers belonging to a given reservoir concentration. With the help of the cycle efficiency, diagrams were constructed to elucidate the effect of some operation parameters on the component separation. With the selection of appropriate parameters and starting concentration in favourable composition range, the cycle efficiencies remain high, and the products purity reach the 99.99 or 99.999 m/m% values. SYMBOLS v0 superficial velocity of liquid phase xj ion ratio of component “j” in the liquid phase xj,0 initial ion ratio of component “j” in the liquid phase yj ion ratio of component “j” in resin phase C0 total cationic concentration of liquid phase C20,j reservoir concentration of component “j” after the 20th cycle L the length of ion exchange column NT theoretical cycle number P20,j product purity of component “j” after the 20th cycle i jT ion exchange selectivity coefficient of ion “i”, related to ion “j” ∆V volume flow through in half cycle α’ dimensionless ion proportion, the ratio of cation amount flowing through during half cycle and ion exchange capacity of the column ηi cycle efficiency related to component “i” 71 REFERENCES 1. HUMPHREY J.: Chem. Eng. Prog.,1995, 91(10), 31-41 2. KELLER G.: Chem. Eng. Prog.,1995, 91(10), 56-67 3. SHENG P. Z. and COSTA C. A. V.: Ind. Eng. Chem.Res.,1998, 37(12), 4808-4815 4. DIAZ M., RENDUELES M., FERNANDEZ A. and SUAREZ C.: Sep. Sci. Techn.,1998, 33(7), 1025-1042 5. AGRAWAL A. and BURNS M. A.: AIChE J.,1996, 42(1), 131-146 6. DIEZ S, LEITAO A., FERREIRA L. M. and RODRIGUES A. E.:Sep. Purif. Techn., 1998, 13, 25-35 7. SIMON G., HANÁK L., GREVILLOT G., SZÁNYA T. and MARTON G.: J. Chrom. B, 1995, 664, 17-31 8. SIMON G., HANÁK L., GREVILLOT G., SZÁNYA T. and MARTON G.: Chem. Eng. Sci., 997, 52, 467-480 9. IVANOV V. A., TIMOFEEVSKAJA V. D., GARLINA O. T. and GORSHKOV V. I.: Microporous and Mesoporous Materials, 2003, 65, 257-265 10. IVANOV V. A., TIMOFEEVSKAJA V. D. and GORSHKOV V. I.: React. Polym., 1992, 17(1), 101-107 11. AHMED Z. M.: React. Polym., 1987, 5(3), 227- 235 12. KIEFER R. and HÖLL W. H.: React. Funct. Polym., 2000, 45, 197-210 13. FERREIRA L. M. and RODRIGUES A. E.: Adsorption, 1995, 1(3), 233-252 14. GREVILLOT G., BAILLY M. and TONDEUR D.: Int. Chem. Eng., 1982, 22, 440-455 15. FERREIRA L. M. and RODRIGUES A. E.: Adsorption, 1995, 1(3), 213-231 16. NIKOLAEV N. P., MURAVIEV D. N. and MUHAMMED M.: Sep. Sci. Techn., 1997, 32(1-4), 849-866 17. SHENG P. Z. and COSTA C. A. V.: Sep. Sci. Techn., 1997, 12(1), 81-95 18. WANKAT P. C.: Large-Scale Adsorption and Chromatography, CRC Press, Boca Raton, 1986 19. HANÁK L., SZÁNYA T. and SZOLCSÁNYI P.: Hung. J. Ind. Chem., 1988,16, 253-260. 20. SZÁNYA T., HANÁK L., MOHILLA R. and SZOLCSÁNYI P.: Hung. J. Ind. Chem., 1988, 16, 261-271.