UDK 546.46:54-36:551.464 Izvirni znanstveni članek ISSN 1318-0010 KZLTET 33(6)473(1999) N. PETRIC ET AL.: EFFECT OF 818A AND 827N FLOCCULANTS ON SEAWATER… EF FECT OF 818A AND 827N FLOCCULANTS ON SEAWATER MAG NE SIA PRO CESS VPLIV 818A IN 827N FLOKULANTA NA PROCES PRIDOBIVANJA MAGNEZIJE IZ MORSKE VODE Nedjeljka Petric, Vanja Martinac, Miroslav La bor, Ol i ver Jurin Faculty of Chemical Technology, Department of Thermodynamics, Teslina 10/V, 21000 Split, Croatia Prejem rokopisa - received: 1999-09-08; sprejem za objavo - accepted for publication: 1999-11-17 Pos si bil ities for the ap pli ca tion of the an ionic 818A flocculant (polyacrylamide) pro duced by the Dutch firm Her cu les have been stud ied in or der to in crease the set tling rate of mag ne sium hy drox ide from sea wa ter and t o de fine the op ti mum con di tions dur ing the pre cip i ta tion pro cess. The ef fi ciency of the 818A flocculant has been com pared to the nonionic 827N flocculant. Ex am i na tions were car ried out with dif fer ent de grees of com plete ness of pre cip i ta tion and with dif fer ent quan ti ties of the 818A flocculant. The de pend ence ob tained was de scribed by the ap pro pri ate an a lyt i cal ex pres sion. The study has shown that the set tling rate is higher dur ing nonstoichiometric pre cip i ta tion, i.e. pre cip i ta tion when t he quan tity added is lower than the stoichiometrically re quired quan tity of the pre cip i ta tion agent. This has been ex plained by the Mg(OH) 2 par ti cle model. In or der to in ter pret the ex per i men tal data ob tained by batch-settling tests a cal cu la tion of a co n tin u ous thick ener was made ap ply ing the Kynch the ory. It was ob served that at 100% pre cip i ta tion the quan tity of the pre cip i tate ob tained was 23.05 kgh -1 , while that ob tained at 80% pre cip i ta tion was 52.69 kgh -1 , which in di cates a con sid er able in crease in the thick ener pro duc tive ca pac ity. Thus, it could be seen that the set tling rate of the mag ne sium hy drox ide slurry was very im po r tant in com put ing the thick en ing area re quired. Key words: mag ne sium hy drox ide from sea wa ter, set tling rate, nonstoichiometric pre cip i ta tion, ef fect of flocculant Preiskovana je bila uporaba anionskega 818A flokulanta (poliakriamida) nizozemskega proizvajalca He rkules z namenom povečanja hitrosti usedanja magnezijevega hidroksida iz morske vode in definiranja optimalnih pogoj ev usedanja. Učinkovitost flokulanta 818A smo primerjali z učinkovitostjo 827N flokulanta. Preiskave so bile narejene pri raz ličnih stopnjah usedanja in različnih količinah dodanega 818A flokulanta. Dobljene odvisnosti so opisane z odgovarjajočimi anal itičnimi izrazi. Rezultati preiskav kažejo na znatno povečanje hitrosti usedanja delcev pri nestehiometričnem usedanju, to je, usedanju pri uporabi manjše količine usedalnega sredstva od stehiometrično potrebne količine. Navedeno se pojasnjuje z modelom delca Mg(OH) 2. Na osnovi ugotovljenih usedalnih karakteristik suspenziji mahnezijevega hidroksida v mezuri pri pogoji h diskontinuiranega usedanja, je bil narejen preračun za kontinuirano usedanje z uporabo Kynchove teorije. Pri 100% use danju se pridobi 23,05, pri 80% usedanju pa 52,69 kgh -1 magnezijevega hidroksida, kar predstavlja znatno povečanje proizvodnosti usedalnika. Iz ugotovljenega je razvidno, da je hitrost usedanja suspenzije magnezijevega hidroksida zelo važna ko ličina v proračunu potrebne površine usedalnika. Ključne besede: magnezijev hidroksid, morska voda, hitrost usedanja, nestehiometrično usedanje, vpl iv flokulanta 1 IN TRO DUC TION The seawater magnesia process 1-6 is chemically very simple in principle, requiring only the addition of an alkaline base such as dolomite lime to precipitate the magnesium salts present in seawater as magnesium hydroxide. This reaction is the crucial part of the process. The central prob lem is how to increase the magnesium hydroxide settling rate 7-10 since the magnesium hydroxide suspensions are characterized by low settling rates; also the precipitate is very difficult to filtrate. The settling rate is the "bottleneck" of this technology and it is one of the most important unit operations controlling the cost of production. The settling rate can be significantly increased if small quantities of organic long chain polymers 8-10 are used. In this study we have investigated the possibility of improving the settling rate by using the anionic (818A) and nonionic (827N) polyacrylamide flocculants for obtaining magnesium hydroxide from seawater with dolomite lime as the precipitation agent. KOVINE, ZLITINE, TEHNOLOGIJE 33 (1999) 6 2 EX PER I MEN TAL The content of magnesium oxide and calcium oxide in the seawater used for the precipitation of magnesium hydroxide was MgO = 2.2339 g dm -3 and CaO = 0.5810 g dm -3 . The composition of dolomite lime used was as follows: MgO = 40.90 mass.%, CaO = 57.89 mass.%, SiO 2 = 0.102 mass.%, Fe 2O 3 = 0.319 mass.% and Al 2O3 = 0.866 mass.%. The seawater was pretreated for precipitation of magnesium hydroxide to remove the bicarbonate and carbonate ions by adding a defined quantity of H 2SO 4 with on-line control through pH measurement (pH = 3.8-4.0), and degassing the acidified water to remove the released CO 2. Degassing is accomplished in a desorption tower by blowing air. Precipitation of magnesium hydroxide took place after pre-treatment of seawater. The experimental procedure was similar to that employed in previous investigations 8,11 . Experiments were carried out in graduated glass cylinders of the same diameter. 1mm of the height of magnesium hydroxide precipitate is equivalent to a 473 N. PETRIC ET AL.: EFFECT OF 818A AND 827N FLOCCULANTS ON SEAWATER… volume of 2.9268 cm 3 of the suspension. This method is suitable for tests with solid-liquid systems containing at least 0.1 mass.% of solid materials. The powdery 818A and 827N flocculants used in these examinations were synthetic high molecular weight polymers (M(818A) = 3-4x10 6 , M(827N)=2x10 6) which are soluble in water. When in solution they ionize according to their ionic nature. The nonionic type (827N) can be dispersed in water and has a weak anionic character in dispersion. They are both with 100 mass percentage of active matter (the actual activity of the specific component from which it was prepared). The active groups over the molecular chain should probably be acryl and/or amide groups, because Hercofloc 818A and 827N are high molecular polymers of the poly -acrylamide type. The functionality (the number of active groups in relation to the chain length) of these flocculants was very low for the nonionic 827N flocculant and low for the anionic 818A flocculant. The prepared solution contained 0.05 mass percentage of active matter, and was prepared by dissolving 0.1 g of the flocculants in 190 cm 3 of distilled water, with gentle mixing. 1 cm 3 of the solution contained 5 .10 -4 g of the flocculant 818A and/or 827N. All the results presented are an average of 5. The results obtained were within the experimental error of approx. ± 0.5 mm. All measurements showed good reproducibility. Measurements were carried out at a temperature of 18.5 ± 1°C. 3 RE SULTS AND DIS CUS SION Figures 1-4 show settling curves for different degrees of completeness of precipitation of magnesium hydroxide from seawater with the addition of 1.5, 1.7, Fig ure 2: The de pend ence of pre cip i tate level (Z) upon the set ting time (t), for dif fer ent de grees of pre cip i ta tion com plete ness, with ad di tion of 1.7 cm 3 of the 818A flocculant per 500 cm 3 of sea wa ter Slika 2: Odvisnost višine usedline (Z) v odvisnosti od časa usedanja (t) pri različnih stopnjah popolnosti usedanja magnezijevega hidroksida iz morske vode pri dodatku 1,7 cm 3 flukolanta 818A/500cm 3 morske vode 1.9, and 2.5 cm 3 of the 818A flocculant per 500 cm 3 o f seawater respectively. Figure 5 shows settling curves for different degrees of completeness of precipitation with the addition of 5.0 cm 3 of the 827N flocculant per 500 cm 3 seawater. The results obtained indicate a difference in the settling rate during complete (stoichiometric) and incomplete (nonstoichiometric) precipitation, i.e. when the quantity of the precipitation agent is lower than that required stoichiometrically. Different degrees of completion of precipitation indicate the difference of initial mass concentrations in such studies. Fig ure 1: The de pend ence of the pre cip i tate level (Z) upon the set ting time (t), for dif fer ent de grees of pre cip i ta tion com plete ness, with ad di tion of 1.5 cm 3 of the 818A flocculant per 500 cm 3 of sea wa ter Slika 1: Odvisnost višine usedline (Z) v odvisnosti od časa usedanja (t) pri različnih stopnjah popolnosti usedanja magnezijevega hidroksida iz morske vode pri dodatku 1,5 cm 3 flokulanta 818A/500 cm 3 morske vode Fig ure 3: The de pend ence of the pre cip i tate level (Z) upon the set ting time (t), for dif fer ent de grees of pre cip i ta tion com plete ness, with ad di tion of 1.9 cm 3 of the818A flocculant per 500 cm 3 of sea wa ter Slika 3: Odvisnost višine usedline (Z) v odvisnosti od časa usedanja (t) pri različnih stopnjah popolnosti usedanja magnezijevega hidroksida iz morske vode pri dodatku 1,9 cm 3 flukolanta 818A/500cm 3 morske vode 47 KOVINE, ZLITINE, TEHNOLOGIJE 33 (1999) 6 N. PETRIC ET AL.: EFFECT OF 818A AND 827N FLOCCULANTS ON SEAWATER… Fig ure 4: The de pend ence of the pre cip i tate level (Z) upon the set ting time (t), for dif fer ent de grees of pre cip i ta tion com plete ness, with ad di tion of 2.5 cm 3 of the 818A flocculant per 500 cm 3 of sea wa ter Slika 4: Odvisnost višine usedline (Z) v odvisnosti od časa usedanja (t) pri različnih stopnjah popolnosti usedanja magnezijevega hidroksida iz morske vode pri dodatku 2,5 cm 3 flukolanta 818A/500cm 3 morske vode Figures 1 to 5 show that the settling rate is higher the lower the initial mass concentration of magnesium hydroxide. If the mass concentration of the solid phase in suspension increases, density and viscosity of the suspension increase, and thus the settling rate decreases. In concentrated suspensions porous aggregates are formed that significantly reduce the settling rate. Thus, the suspension properties are largely affected by mass concentration, and, by association, by the quantity of the reagent (precipitation agent) added. In other words, settling taking place with excess Mg 2+ ions is more efficient than that with surplus OH - ions, both as regards the settling rate and the filtration rate. Taking into account that the mass concentration of magnesium hydroxide, i.e. the mass of solids per unit volume of slurry varies for different degrees of completeness of the precipitation, concentrations were equaled mathematically in order to eliminate their effect on the settling rate. The material balance for the solid phase in the slurry, used in the comparison of stoichiometric and nonstoichiometric precipitation, has the form: Z o Z Io. Y * 1 (1) where Z is the height of the interface, and ? the mass concentration of the solid phase. Subscript "o" denotes 100% (stoichiometric) precipitation, and the subscript "1" incomplete (nonstoichiometric) precipitation. It then follows that Z 1 is the height which the slurry would occupy if all the solids phase present were at the mass concentration ?o. Starting from this value of Z 1 i.e. the value for which concentrations are equal for different degrees of completeness of precipitation, the change in the precipitate level, ?Z = Z 1 - Z 2, was determined for the KOVINE, ZLITINE, TEHNOLOGIJE 33 (1999) 6 t/ min Fig ure 5: The de pend ence of the pre cip i tate level (Z) upon the set ting time (t), for dif fer ent de grees of pre cip i ta tion com plete ness, with ad di tion of 5.0 cm of the 827N flocculant per 500 cm of sea wa ter Slika 5: Odvisnost višine usedline (Z) v odvisnosti od časa usedanja (t) pri različnih stopnjah popolnosti usedanja magnezijevega hidroksida iz morske vode pri dodatku 5,0 cm flukolanta 827N/500cm morske vode time interval of 20 minutes. Table 1 shows the values obtained for Z 1 and Z 2 where Z 1 is the precipitate level determined according to expression (1) and Z 2 is the precipitate level 20 minutes after Z 1. The relations obtained ?Z = f (% stoich.) (Figure 6) indicate the degree of completeness of precipitation for which the quantity of the flocculant added is the optimum one. For the maximum value of change in the precipitate level ?Z, that degree of completeness of precipitation is determined for which the settling rate is maximal for the flocculant addition examined. Ta ble 1: De pend ence of Z 1 and Z 2 on the de grees of com plete ness of pre cip i ta tion for dif fer ent ad di tions of flocculant 818A and 827N Tabela 1: Odvisnost Z 1 in Z 2 od stopnje popolnosti usedanja pri različnih količinah dodatka flokulantov 818A in 827N Flocculant cm 3 of the flocculant per 500 cm 3 o f sea water Degree of completeness of precipitation % Z1/mm Z2/mm 818A 1.5 60 12.00 4.00 70 14.00 5.50 80 16.00 9.50 100 20.00 17.00 1.7 70 13.00 8.50 80 15.00 9.50 100 18.50 15.00 1.9 80 15.00 10.00 90 17.00 11.00 100 19.00 15.00 2.5 80 18.00 12.50 100 22.50 16.00 120 27.00 22.00 827N 5.0 70 15.00 5.00 80 17.50 7.00 100 22.00 16.50 475 N. PETRIC ET AL.: EFFECT OF 818A AND 827N FLOCCULANTS ON SEAWATER… Figure 6: The dependence of the change in the precipitate level ( ?Z ) upon the degree of precipitation completeness of magnesium hydroxide from sea water, with different additions of the 818A flocculant per 500 cm 3 of seawater Slika 6: Odvisnost spremembe višine usedline ( ?Z) od stopnje popolnosti usedanja magnezijevega hidroksida iz morske vode pri različnih količinah dodanega flukolanta 818A/500cm 3 morske vode In this way, the optimum degree of completion of precipitation for the 818A flocculant addition examined is determined from the settling rate measurement results and vice versa. The experimental results obtained in examination of the settling rate of magnesium hydroxide from seawater indicate that there is a certain relation between the addition of the 818A flocculant and the degree of completion of precipitation under optimum settling conditions. Therefore, a mathematical analysis of the experimental results was performed in order to obtain a function that would best describe this dependence. The relations obtained have been described by the polynomials of the second and third degree: Y = 2 .10 -3 x2 - 0.276x + 12.56 r = 2.20% Y = 1.333 . 10 -4 x3 - 0.032x 2 + 2.587x - 67 r = 0.20% where: Y = the volume of the 818A flocculant in cm 3 per dm 3 of seawater, x = the degree of completeness of precipitation, and r = the mean relative error. The mean relative error, r, being much lower in the function with the form of the third degree polynomial, this function may be concluded to approximate the dependence shown. To explain this behavior of synthetic polymeric compounds, one could say that the affinity of the 8181A polyelectrolyte to the surface of the solid phase suspended in the suspension is one of its more important properties. These long straight polymeric chains adsorb on to particles suspended forming strong bonds between the polymers and solid particles. After one end of the molecule adsorbs on to the particle, the rest of the molecule is still free in the suspension to adsorb on to other particles, and this leads to fast agglomeration and flocculation of the particles suspended. Owing to this phenomenon (known as "bridging" 12 ) large agglomerates 476 Fig ure 7: The de pend ence of the change in the pre cip i tate level ( ?Z ) upon the de gree of com plete ness of pre cip i ta tion of mag ne sium hy drox ide from sea wa ter, with ad di tion of 5.0 cm 3 of the 827N flocculant per 500 cm 3 of sea wa ter Slika 7: Odvisnost spremembe višine usedline ( ?Z) od stopnje popolnosti usedanja magnezijevega hidroksida iz morske vode pri dodatku 5,0 cm 3 flukolanta 827N/500cm 3 morske vode of suspended particles and polyelectrolyte chains are formed, that, due to their large diameters and mass, greatly accelerate the settling process. Bearing in mind the significant advantages of the substoichiometric (80%) precipitation that has been established in previous studies 8,9,13 , applicability i.e. efficiency of the nonionic 827N flocculant was examined for this precipitation mode. The results obtained for the magnesium hydroxide settling rate for 80% precipitation (Figure 5) with the addition of the 827N flocculant indicate that much higher quantities of this flocculant are needed to obtain the optimum process conditions. The results obtained for the functional dependence ?Z = f (% stoich.), (Fig. 7) , have shown that the optimum quantity of nonionic 827N flocculant is 5.0 cm 3 per 500 cm 3 seawater, i.e. 10.0 cm 3 per dm 3 of seawater. The comparison with the previously examined 818A flocculant indicates that, although 1cm 3 of the solution of both examined flocculants contains 5 ?10 -4 g 818A (827N) of flocculant, the quantity of anionic flocculant needed is much lower (approx. 3 times) i.e. that it is more effective in improving the settling rate. The results obtained can lead to the conclusion that the 818A flocculant, being negatively charged, has a greater number of active groups in the molecule chain, while the effect of the nonionic 827N flocculant is most probably based on the polarity of its molecule. This leads to different ionization degrees in the solution. To interpret the experimental results obtained by analysis of the settling properties of the magnesium hydroxide suspension in a graduated glass cylinder in conditions of discontinuous precipitation (batch-settling tests), a calculation of the continuous thickener 14,15 h a s been made, applying the Kynch theory. It was interesting to see how the results obtained were important for the design of the thickener as its KOVINE, ZLITINE, TEHNOLOGIJE 33 (1999) 6 N. PETRIC ET AL.: EFFECT OF 818A AND 827N FLOCCULANTS ON SEAWATER… construction is the time-controlling factor in plants of this type. Table 2 shows the characteristic thickener sizes for different degrees of completion of precipitation (80%, 90%, and 100% precipitation) for the optimum addition of the 818A flocculant. The first part of table (I) shows the thickener dimension (diameter and area) values obtained for the constant volume flow of the suspension entering the thickener (Q in = 3.6 m 3 h-1 ). The density (mass of solid per unit volume of slurry) of the underflow after 30 minutes ( ?out ) was found to be greater for the lower degree of completion of precipitation, while the thickener area, A, and its diameter, d, are much smaller. The second part of the table (II) shows the values for the range of the volume flow of the suspension fed per unit time to the thickener (Q in ) and the thickener dimensions (A and d), based on the constant density of the underflow ( ?out = 45.0297 kg m -3 ) and the mass flow of the solid phase fed per unit time to the thickener (Q m = 23.0458 kg h -1 ). It has been observed that Q in increases, while the thickener dimensions (A and d) significantly decrease with the decrease in the degree of completion of precipitation. The third part of table (III) , shows the values for the range of the volume flow of the entering suspension (Q in ) and the quantity of precipitate obtained in the unit of time (Q m), based on the constant density of the underflow ( ?out = 45.0297 kg m -3 ) and thickener dimensions (A = 1.77 m 2 and d = 1.5 m). The quantity of Ta ble 2: Char ac ter is tic val ues of the thick ener for dif fer ent de grees of in com plete pre cip i ta tion (80%, 90% and 100% re spec tively) and with op ti mal ad di tion of the 818A flocculant Tabela 2: Karakteristične velikosti usedalnika pri različnih stopnjah usedanja magnezijevega hidroksida iz m orske vode pri optimalnem dodatku flokulanta 818A Part % stoich. flocculant 818A dm 3m-3 Pin kgm-3 ?out kgm -3 Q in m 3h-1 Qm kgh-1 Qout m3h-1 Qp m3h-1 q kgm2h-1 A m2 d m I 80 3.4 5.1212 46.2612 3.60 18.4363 0.40 3.20 27.61 0.67 0.9 90 3.8 5.7610 46.2077 3.60 20.7396 0.45 3.15 11.41 1.82 1.5 100 5.0 6.4016 45.0297 3.60 23.0458 0.51 3.09 13.01 1.77 1.5 II 80 3.4 5.1212 45.0297 4.48 23.0458 0.51 3.99 29.77 0.77 0.99 90 3.8 5.7610 45.0297 3.99 23.0458 0.51 3.49 12.02 1.92 1.56 100 5.0 6.4016 45.0297 3.60 23.0458 0.51 3.09 13.01 1.77 1.5 III 80 3.4 5.1212 45.0297 10.29 52.6929 1.17 9.12 29.77 1.77 1.5 90 3.8 5.7610 45.0297 3.69 21.2754 0.47 3.22 12.02 1.77 1.5 100 5.0 6.4016 45.0297 3.60 23.0458 0.51 3.09 13.01 1.77 1.5 Symbols: ?in - is the density (mass concentration) of feed, kg m -3 ?out - is the density (mass of solids per unit volume of slurry) of underflow, kg m -3 Qin - is the volume flow of the incoming suspension, m 3 h -1 Qm - is the mass of solids fed per unit time to thickener, kg h -1 Qout - is the volume flow of overflow, m 3 h -1 Qp - is the volume flow of overflow, m 3 h -1 q - is the capacity of the thickener, kg m 2 h -1 A - is the surface of the thickener, m 2 d - is the diameter of the thickener, m KOVINE, ZLITINE, TEHNOLOGIJE 33 (1999) 6 477 magnesium hydroxide obtained (Q m), the volume flow at the entrance (Q in ), and the volume flow at the overflow (Q p) were found to be much higher for lower degrees of completion of precipitation. The quantity of precipitate obtained in the unit of time was 23.046 kg h -1 for 100% precipitation and 52.693 kg h -1 for 80% precipitation, which represents an increase in the thickener productive capacity by approximately 128%. The improvements observed in the settling rate of the magnesium hydroxide obtained by nonstoichiometric (80%) precipitation are due to lower viscosity and density of the magnesium hydroxide suspension, which results in faster settling of the precipitate formed. Also, taking into account the electrokinetic potential ? , that decreases in the presence of excess Mg 2+ ions, as well as different adsorption of individual ions on to the surface of the magnesium hydroxide particle (Mg 2+ ions are bound to the surface with stronger adsorption forces than Ca 2+ ions), the much improved precipitate settling rate in nonstoichiometric precipitation can be explained by the Mg(OH) 2 particle model. 4 CONCLUSIONS – Application of the powdery 818A flocculant (polyacrylamide) to improve the settling rate of magnesium hydroxide in seawater is much more efficient than that of the nonionic flocculant 827N – Based on the functional dependence ?Z = f (% stoich.), the optimum quantity of the nonionic 827N N. PETRIC ET AL.: EFFECT OF 818A AND 827N FLOCCULANTS ON SEAWATER… flocculant was found to be 10.0 cm 3 per dm 3 seawater for 80% precipitation – Based on the functional dependence ?Z = f (% stoich.), the optimum quantity of the anionic 818A flocculant was found to be 3.4 cm 3 per dm 3 seawater for 80% precipitation – The results obtained indicate that the 818A flocculant, being negatively charged, has a greater number of active groups along the molecular chain, while the action of the nonionic 827N flocculant is probably based on the polarity of its molecule. This results in different degrees of ionization in the solution – Mathematical analysis has yielded the most appropriate function for the dependence of the quantity of the 818A flocculant added (Y) on the degree of completeness of precipitation (x) under optimum settling conditions. The dependence obtained is described by the third degree polynomial: Y = 1.333 . 10 -4 x3 - 0.032x 2 + 2.587x -67 – When results obtained for 80% and 100% precipitation are compared, the calculation for the continuous thickener indicates that the thickener productive capacity is much higher (by approx. 128%) if the anionic flocculant 818A is used in 80% precipitation (nonstoichiometric) of magnesium hydroxide from seawater. 5 LIT ER A TURE 1 Tsuge H., Kotaki Y., Asano S., Seventh Symposium on Salt, Elsevier Science Publishers B. V., Amsterdam, Vol. II, (1993) 219 2 Gildersleeve M. J., Brook R. J., Br.Ceram.Trans. , 83 (1984) 181 3 Sims C., Indust.Minerals , July, 1997 , p. 21 4 Gilpin W. C., Heasman N., Chem.Ind. , 16 (1977) 567 5 Heasman N., Gas Wä rme International , 28 (1979) 392 6 Hicks J. C., Tangney S., Ceram. Bull. , 59 (1980) 7 1 1 7 Barba D., Brandani V., Di Giacomo G., Foscolo P. U., Desalination , 33 (1980) 2 4 1 8 Petric B., Petric N., Ind.Eng.Chem.Process Des.Dev. , 1 9 (1980) 329 9 Petric N., Petric B., Martinac V., J.Chem.Tech.Biotechnol. , 52 (1991) 519 10 Vohra R. N., Patel K. N., Shukla B. K., Chem. Age of India, 1 9 (1968) 441 11 Petric B., Petric N., J.Chem. Tech.Biotechnol. , 2 9 (1979) 642 12 Benefield L. D., Judkins J. F., Ewand B. L., Process Chemistry for Water and Wastewater treatment , Prentice Hall, Inc., Englewood Cliffs, New Jersey, 1982 , p. 218 13 Petric N., Martinac V., Labor M., Proceedings of the 16th Symposium on Technology developments and Ecological Solutions in the production of Cement and Fibre Cement products , Split, 1995 , p. I-13-I-17 14 Foust A. S., Wenzal L. A., Chemp C. W., in Principles of Unit Operations , John Wiley & Sons, Inc. New York, 1960 , p. 465-472 15 Mitrovi}-Kessler E., Žaneti} R., Vojnovi} I., Kem.Ind., 38 (1989) 17 47 KOVINE, ZLITINE, TEHNOLOGIJE 33 (1999) 6