- Email: [email protected]
Calcium sulfate precipitation in the presence of water-soluble polymers
Journal of Colloid and Interface Science 303 (2006) 164–170 www.elsevier.com/locate/jcis
Calcium sulfate precipitation in the presence of water-soluble polymers Maria G. Lioliou, Christakis A. Paraskeva, Petros G. Koutsoukos ∗ , Alkiviades C. Payatakes Institute of Chemical Engineering and High-Temperature Chemical Processes—Foundation for Research and Technology, P.O. Box 1414, GR-26500, Patras, Greece Department of Chemical Engineering, University of Patras, GR-26500, Patras, Greece Received 21 June 2006; accepted 22 July 2006 Available online 26 July 2006
Abstract The effect of four different polymers on the precipitation of calcium sulfate was investigated in the present work. The degree of inhibition was estimated from measurements of the calcium ion activity and from specific solution conductivity measurements in the supersaturated solutions during the course of the precipitation process. The effects of polyacrylic acid (PAA, three different polymers with average molecular weight 2000, 50,000, and 240,000, respectively) and of a co-polymer of PAA with polystyrene sulfonic acid (PSA, average molecular weight <20,000) were investigated with respect to their effect on the kinetics of spontaneous precipitation of calcium sulfate salts. The results of the kinetics experiments suggested that the spontaneous precipitation from supersaturated calcium sulfate solutions at 25 ◦ C yielded exclusively calcium sulfate dihydrate (gypsum) both in the absence and in the presence of the polymeric additives. The induction times, preceding the formation of the solid increased in all cases in the presence of the polymeric additives. Polymer concentrations as low as 2.0 ppm increased induction time from practically zero to 10 min. The rates of precipitation were reduced according to the solutions content in the polymers added and precipitation was completely suppressed in the presence of 6.0 ppm of the polymers tested, depending on their molecular weight. The lower the molecular weight of PAA, the more efficient was the threshold inhibition and the stronger the reduction of the rates of spontaneous precipitation. PSA yielded the poorest inhibition efficiency in comparison with the PAA, possibly because of the relatively lower affinity of the sulfonate groups for the calcium ions of the surface of the solid forming. The kinetics results analysis assuming Langmuir-type adsorption of the polymeric molecules on the growing supercritical gypsum nuclei showed different affinity for the polymers tested in agreement with the respective inhibition efficiency, in the order: PAA1 > PAA2 > PSA > PAA3. The presence of the polymers in the supersaturated solutions resulted in modification of the precipitated gypsum crystals morphology. © 2006 Elsevier Inc. All rights reserved. Keywords: Calcium sulfate dihydrate; Gypsum; Spontaneous precipitation; Inhibition; Polyacrylic acid; Polystyrene sulfonate
1. Introduction The formation of tenaciously adhering calcium sulfate scale in a number of processes from water desalination to heat exchangers and processes involving heating of water is a persistent problem [1]. Although six different calcium sulfate crystal forms are known to exist [2], three different salts are usually encountered in natural formations and scale precipitates: calcium sulfate dihydrate (CaSO4 ·2H2 O, CSD), calcium sulfate hemihydrate (CaSO4 ·1/2H2 O, CSH) and anhydrous calcium sulfate * Corresponding author. Fax: +30 2610997579.
E-mail address: [email protected] (P.G. Koutsoukos). 0021-9797/$ – see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2006.07.054
(CaSO4 , CSA). Both the CSA and the CSH salts may undergo further dehydration via phase transformation to the anhydrous form [3]. Despite the fact that considerable research has been going on during the past decades on the formation of calcium sulfate in aqueous media there is still large uncertainty concerning the mechanism of formation of this salt because of the largely variable conditions of the solutions in which the salt formation takes place, including temperature, pH, ionic strength and composition and the presence of foreign ions and or water soluble compounds. A large number of the published studies agree on the fact that the formation of the calcium sulfate nuclei is initiated on solid substrates. These substrates may be either metallic surfaces of heat exchangers or crystals of the same or different substrates [4–9].
M.G. Lioliou et al. / Journal of Colloid and Interface Science 303 (2006) 164–170
165
Among the most important objectives of the mechanistic investigations has been the possibility to control the formation of the various forms of calcium sulfate. The main effort to this end has been focused in the use of water soluble inhibitors which may act either as threshold inhibitors which block the development of the supercritical nuclei [10], or as retarders of the growth of the calcium sulfate crystals [11–16]. Several investigations have been carried out on the influence of trace amounts of polymeric scale inhibitors on both the precipitation and crystal habit modification of calcium sulfate forms [12,17–19]. Polymers containing carboxylic groups such as carboxymethyl cellulose (CMC), polymethacrylic acid (PMA), and polyacrylic acid (PAA) were found to be particularly effective as CaSO4 ·2H2 O growth inhibitors [12]. In the case of polymers in solution, there is general agreement that inhibition of the formation or the growth of the salt nuclei is effected by adsorption of these molecules on the active growth sites. Polymers tend to adsorb on solids from solutions due both to van der Waals [20] and/or electrostatic interactions [21,22]. The length of the polymer chains therefore as well as the functional groups present, which through ionization regulate the electrostatic charges of the polymers, are of primary importance for the investigation of the role of the respective polymeric additives in the crystal growth of calcium sulfate. In the present work, we have addressed the problem of the effect of PAA on the scale formation of gypsum (CSD) using polymers of markedly different molecular weight. Moreover, in order to compare the relative importance of the presence of the carboxylic groups, a co-polymer of PAA with sulfonated polystyrene was tested. The sulfonate groups are more strongly ionized in comparison to the carboxylic groups and are expected to promote stronger electrostatic interactions between the polymer and the surface of the calcium containing crystals which form in the supersaturated solutions. The effect of the water soluble polymeric additives was investigated in experiments in which calcium sulfate precipitation took place spontaneously from unstable supersaturated solutions, past the lapse of measurable induction time characteristic of the time frame needed for the formation of the supercritical nuclei and the subsequent initiation of the precipitation process. In order to accentuate the effect of the additives, the supersaturated solutions conditions selected for the test experiments in the absence of polymeric additives, yielded spontaneous precipitation with practically zero induction time.
The supersaturated solutions were prepared directly in the reactor by mixing equal volumes of equimolar calcium chloride and sodium sulfate solutions. The master variable used to monitor the process of CSD spontaneous precipitation was the specific conductivity of the solution. The conductivity and temperature of the reacting solution were monitored during the crystallization process by a D/A converter attached to a computer unit with the appropriate software. The progress and the extent of the crystal growth process both in the presence and in the absence of the inhibitors tested, was characterized by the decrease of the solution conductivity as a function of time. In the experiments done in the presence of polymeric inhibitors, the additives were introduced in the sulfate solution to avoid complexation [23], conformational changes [24], and even precipitation of calcium-polymer salts [25]. The polymers tested are summarized in Table 1. The homogeneity of the supersaturated solutions was ensured by magnetic stirring at ca. 250 rpm. Past the establishment of the solution supersaturation the recording of the solution conductivity was initiated. The formation of calcium sulfate crystals was detected by a decrease of the solution conductivity due to the reduction of calcium concentration. The time lapse between the preparation of the supersaturated solutions and the appearance of the inflection point in the conductivity– time profile, was defined as the induction time, τ . The point of inflection was considered as the point in which the slope of the curve changed and was determined by the intersection of the two tangents drawn: one at the initial period (flat) and one at the dropping part of the curve. During the course of the precipitation process samples were withdrawn from the reactor and were filtered through membrane filters (0.22 µm, Millipore). The filtrates were analyzed for calcium by EDTA complexometric titrations [26]. The combination of the chemical analyses of the solution composition with the solution conductivity allowed for the construction of calibration curves. The recordings of the solution conductivity as a function of time were thus converted into calcium vs time profiles. These curves showed a profile similar to that corresponding to decreasing conductivity as a function of time, which allowed for the calculation of the initial rates of calcium sulfate precipitation. The calcium– time profiles were fitted according to fourth-order polynomial and the [d[Ca2+ ]/dt]t→0 was taken as the value for the initial rate in each experiment. This correlation allowed for the quantitative measurements of the rates of crystallization. The values reported are the mean of three different measurements.
2. Experimental
3. Results and discussion
Crystal growth experiments were carried out in a 0.250dm3 double-walled Pyrex vessel thermostated at 25.0 ± 0.2 ◦ C by water circulation from a constant-temperature bath. Stock calcium chloride and sodium sulfate solutions were prepared from the respective crystalline solids (Merck, pro analisi). The solutions were filtered through membrane filters (0.22 µm, Millipore) and standardized by atomic absorption spectrometry (Perkin Elmer A Analyst 300) and ion chromatography (Dionex) for calcium and sulfate respectively.
In all experiments of the present work, the pH of the supersaturated solutions was about 5.0 and it was not adjusted. It is established that pH over a wide range (3.0–8.0) does not affect the kinetics of spontaneous precipitation of CSD [9,27,28]. Three types of PAA polymers of different molecular weights and one polysulfonic acid polymer were tested as inhibitors of the calcium sulfate precipitation. The first three polymers are characterized by the same binding capacity with respect to calcium but different conformation and adsorption properties with
166
M.G. Lioliou et al. / Journal of Colloid and Interface Science 303 (2006) 164–170
Table 1 Polymeric inhibitors tested in the present work Polymer
K752
PAA
K702
K798
Acronym Chemical type M.W.
PAA1 Poly(acrylic acid)
PAA2 Poly(acrylic acid)
PAA3 Poly(acrylic acid)
PSA Poly(acrylic acid/sulfonic acid/sulfonated styrene)
2000
50,000
240,000
<20,000
Goodrich, M/SGoodrite
Polysciences Inc.
Goodrich, M/SGoodrite
Goodrich, M/SGoodrite
Chemical formula of repeating unit
Manufacturer
respect to the forming crystalline substrate. The polyacrylic polymer with the highest MW is expected to form stronger complexes with calcium and possibly binds more strongly with the surface of the CSD supercritical nuclei. It has been suggested that larger number of functional groups of negatively charged species increases the polar attraction between the adsorbate and the positive sites at the solution interface [29]. The proton dissociation constant, log K, for polyacrylates is in the order of 4.4 [30]. The presence of hydrophobic aromatic nucleus on the molecule may change the mode of adsorption and the respective efficiency. Indeed it has been reported that the presence of sulfonic or phenylsulfonic acid groups which substitute carboxylic groups in a polymer reduce the inhibitory function of the polyelectrolytes on the crystal growth of gypsum [31]. Moreover, it has been shown from adsorption competition experiments, that sulfonic groups form weaker complexes with surface calcium ions in comparison with carboxylic groups [32]. The concentration of the additives tested in the supersaturated solutions varied between 1 to 100 ppm. At these concentration levels the influence on the solution supersaturation was negligible. In all cases, the only form of calcium sulfate precipitating was CSD. The solution speciation and supersaturation were calculated using the MINEQL+ speciation code [33] and all relevant equilibria [3]. It should be noted that complexation of calcium with the PAA was taken into consideration in our calculations [34–36]. The solution supersaturation with respect to CSD is given by equation: Ω=
(Ca2+ )(SO2− 4 ) . 0 Ks
(1)
The onset of CSD formation from the supersaturated solutions was accompanied by a decrease of the free calcium ions in the solution with a concomitant decrease of the solution conductivity. Typical profiles of the solution conductivity as a function of
Fig. 1. Conductivity–time profiles for the spontaneous precipitation of calcium sulfate in the absence and in the presence of PAA (symbols as in Table 1); total calcium, Cat = 120 mM, total sulfate, St = 120 mM, 25 ◦ C. PAA1 6 ppm; PAA2 6 ppm; PAA3 6 ppm.
time in the absence and in the presence of PAA are shown in Fig. 1. The inhibition of CSD formation by PAA may be explained by the Cabrera and Vermileya model [37]. According to this model, growth inhibition is explained in terms of inhibitor ions adsorption upon the crystal surface. During the induction period, most of the active growing sites may be poisoned by the additive molecules/ions. However, some of the growth sites of lower energy may still be free to grow; thus, the reaction proceeds at a very low rate. The induction times preceding precipitation for different amounts of the polymers tested in the present work are summarized in Table 2, while the plots of the induction times measured as a function of the concentration of the polymers tested is shown in Fig. 2. As may be seen, PAA1, i.e., PAA with the least molecular weight, yielded the most efficient threshold effect on the spontaneous precipitation of CSD. This may be ascribed to the most efficient interaction of the low MW polymer with the CSD crys-
M.G. Lioliou et al. / Journal of Colloid and Interface Science 303 (2006) 164–170
167
Table 2 Induction times measured for different amounts of the polymers tested for the precipitation of CSD from supersaturated solutions PAA1
Concentration (ppm) Ind. time (min)
2 10
5 45
6 110
20 200
PAA2
Concentration (ppm) Ind. time (min)
6 50
20 90
40 135
50 >300
PAA3
Concentration (ppm) Ind. time (min)
6 10
20 30
30 36
40 50
PSA
Concentration (ppm) Ind. time (min)
6 27
20 60
40 125
Note. Total calcium, Cat = 120 mM; total sulfate, St = 120 mM, 25 ◦ C.
Fig. 3. Spontaneous precipitation of CSD in the presence of 6 ppm of PAA. Total calcium, Cat = 120 mM, total sulfate, St = 120 mM, 25 ◦ C; (2) blank; (!) PAA1 (M.W. 2000); (P) PAA2 (M.W. 50,000); (1) PAA3 (M.W. 240,000).
Fig. 2. Variation of the induction times preceding the spontaneous precipitation of CSD as a function of the concentration of polymers; total calcium, Cat = 120 mM, total sulfate, St = 120 mM, 25 ◦ C. ( ) PAA1; (1) PAA2; (!) PAA3; (P) PSA.
talline substrate. It is thus possible that the low MW chains adsorb flat on the surface blocking the active sites of the critical nuclei which cannot grow further. The higher the MW the mode of adsorption may be such that the presence of loops and trains on the solid substrate [22] do allow for a larger number of unblocked active sites. The efficiency of the PAAs tested with respect to the threshold inhibition of CSD, decreased with increasing MW. The order found was PAA1 > PAA2 > PAA3 in agreement with previous reports [12,29]. In all cases, it was found that an increase in polymer concentration results in an increase of the induction time values. This means that the duration of the induction period is increased by increasing the amount of the additive in the reacting solution. Induction times for CSD spontaneous precipitation in the presence of 6 ppm of PAA increased from ∼20 to more than 110 min for MW decrease from 240,000 (PAA3) to 2000 (PAA1). The effect of the polymers on the spontaneous precipitation of CSD may be seen from the desupersaturation curves shown in Fig. 3. PSA, inhibited also the spontaneous precipitation of CSD at concentration levels exceeding 20 ppm, as may be seen in the concentration–time profile shown in Fig. 4. The precipitation was significantly suppressed in the presence of 60 and 100 ppm of the polymer. The PSA inhibitor contained hydrophobic phenyl groups attached to the polymer chain. In
Fig. 4. Spontaneous precipitation of CSD in the presence of PSA: (2) blank; (1) 6 ppm; (!) 20 ppm; () 60 ppm; () 100 ppm; total calcium, Cat = 120 mM, total sulfate, St = 120 mM, 25 ◦ C.
general, hydrophobic groups have been reported to be detrimental for the performance of an inhibitor. The presence of the phenyl group may, therefore, explain the relatively lower inhibition capacity of the PSA co-polymer in comparison to the PAA. It should be noted however that sulfonic acid groups are present in the PSA molecules. Sulfonic anions may replace the sulfate ions in the calcium sulfate lattice, enhancing, thus, the inhibiting capacity of the additive. In the present work however, it was found that PSA did not perform better than the PAA polymers, containing –COOH functional groups. Superior performance of PSA may have been anticipated because of the more acidic character of the sulfonic acid group compared to the carboxylic acid group. It may therefore be suggested that the presence of hydrophobic aromatic nuclei in PSA prevailed affecting the conformation of the polymeric chains at the gypsum/water interface yielding weaker interactions and therefore lower affinity of the polymer for the crystalline material. These results are in agreement with earlier reports which have attributed the weaker interactions between the crystal surfaces and the aromatic sulfonate containing polyelectrolytes to the fact that they do not favor the flat on conformation of the ad-
168
M.G. Lioliou et al. / Journal of Colloid and Interface Science 303 (2006) 164–170
(a)
(b)
(c)
(d)
(e) Fig. 5. Scanning electron micrographs of CSD crystals precipitated spontaneously; total calcium, Cat = 120 mM, total sulfate, St = 120 mM, 25 ◦ C; (a) no additive, bar = 40 µm; (b) 6 ppm PAA1, bar = 30 µm; (c) 6 ppm PAA2, bar = 30 µm; (d) 6 ppm PAA3, bar = 30 µm; (e) 6 ppm PSA, bar = 60 µm.
sorbed molecules, which is more likely for the carboxyl group containing PAAs [31]. The crystal habit of the precipitated CSD crystals was affected by presence of the polymers, as may be seen in the scanning electron photographs shown in Fig. 5. The habit of the CSD crystals precipitated in the absence of additives (Fig. 5a) was the well-known thin, elongated calcium sulfate crystals as a result of rapid growth on the [111] faces. The presence of
polyacrylic acid polymers in the reacting solutions enhanced the agglomeration of the crystals, retarded the growth of the [111] faces and as a result more plate-like crystals were obtained. In general, due to their adsorption on the crystal surface, PAAs do change the surface charge of the crystals, enhancing agglomeration. The presence of high molecular weight PAA (PAA3) resulted in even higher agglomeration as may be seen in Fig. 5d. In this case, the precipitated crystals consisted of
M.G. Lioliou et al. / Journal of Colloid and Interface Science 303 (2006) 164–170
169
Table 3 Affinity constants as calculated for the polymers tested for calcium sulfate crystal growth
Fig. 6. Plot of ratio R0 /(R0 − Ri ) as a function of the inverse of inhibitor concentration for (1) PAA1; (!) PAA2; (P) PAA3; () PSA.
rather deformed crystals, covered by sponge-like mass of tiny crystallites. The kinetic data obtained in the presence of the polymers were fitted to a kinetic model based on the assumption of the Langmuir isotherm [38]. Assuming that the rate of crystal growth in the absence of an additive is R0 and the limiting rate in its presence is bR0 and defining the ratio Cm /Cads as the fraction of the occupied sites, θ , the overall rate measured in the presence of an inhibitor, Ri , is given by Ri = R0 − θ R0 (1 − b).
(2)
Substituting θ and rearranging: R0 1 kdes 1 (3) = + . R0 − Ri 1 − b kads (1 − b) C The ratio kads /kdes is the affinity constant of the additive for the particular substrate and the parameter b is a measure of the effectiveness of the inhibitor present at infinite concentration in the case of monolayer coverage or at concentrations lower than the one corresponding to monolayer coverage (0 < b < 1). 0 as a function of C1 yielded satisfacPlots of the ratio R0R−R i tory linear fits for all the polymers used in the present work, as shown in Fig. 6. The intercept of the fitted lines for PAA1 and PAA2 were found to be <1, implying that these additives may cause complete inhibition of the precipitation process at concentrations lower than the one corresponding to monolayer coverage. The fact that the intercept was found to be >1 for the other two polymers (PAA3 and PSA) may be interpreted as the inability of these additives to completely inhibit crystal growth of calcium sulfate. The inverse of the slopes of the fitted lines yield the affinity constants of the polymers used for calcium sulfate crystal surfaces, which are given in Table 3. The affinities calculated for PAA1 and PAA2 were 1.96 × 106 and 1.19 × 106 , respectively, while the affinities for PAA3 and PSA were 1.33 × 105 and 2.48 × 105 . It is interesting to point out the small difference between the values for the first two PAAs in spite of the difference in their molecular weights, while PAA3 gave a value of affinity constant lower by one order of magnitude. This difference may be ascribed to the chain
Polymer
Acronym
Affinity constant
K752 PAA K702 K798
PAA1 PAA2 PAA3 PSA
1.96 × 106 1.19 × 106 1.33 × 105 2.48 × 105
length of the polymer which affected the respective surface conformation. It should also be mentioned that the experiments were conducted at pH approximately 5.0. At this pH value the higher molecular weight polymer is probably ionized to a lower extent, in comparison to the lower MW PAAs. Lower extent of ionization would in turn affect the conformation and/or the extent of adsorption of PAA on the newly formed gypsum nuclei. The presence of the aromatic ring in PSA structure apparently influenced its hydrophobic character and the conformation of the molecule at the solid–solution interface, resulting in reduced inhibition activity. Finally it should be noted that despite the fact that the present work was done at acidic pH, sufficiently high to have the polyelectrolytes in ionized form, increasing the solution pH is expected to increase the efficiency of the inhibitors tested because of the increase of the degree of ionization of the soluble polyelectrolytes [28,31,39]. Acknowledgments The authors wish to acknowledge financial support by the General Secretariat for Research and Technology, Ministry of Development, through PENED Program Contract M413/2002. References [1] Z. Amjad, J. Colloid Interface Sci. 123 (1988) 523. [2] J. Glater, J.L. York, K.S. Campbell, in: K.S. Spiegler, A.D.K. Laird (Eds.), Principles of Desalination, Part B, second ed., Academic Press, New York, 1980, pp. 627–678. [3] P.G. Klepetsanis, P.G. Koutsoukos, J. Colloid Interface Sci. 143 (2) (1991) 299. [4] D. Hasson, J. Zahavi, Ind. Eng. Chem. Fundam. 9 (1) (1970) 26. [5] O.D. Linnikov, Desalination 128 (2000) 35. [6] J.S. Gill, G.H. Nancollas, J. Cryst. Growth 48 (1980) 34. [7] G.H. Nancollas, W.P. Klima, Mater. Performance 21 (1982) 9. [8] G.H. Nancollas, W.F. Klima, Paper 81, Corrosion/81, National Association of Corrosion Engineers Conference, Toronto, Ontario, 1981. [9] S.T. Liu, G.H. Nancollas, J. Colloid Interface Sci. 44 (1973) 422. [10] S. He, J.E. Oddo, M.B. Tomson, J. Colloid Interface Sci. 162 (2) (1994) 297. [11] B.R. Smith, A.E. Alexander, J. Colloid Interface Sci. 34 (1970) 81. [12] E.R. McCartney, A.E. Alexander, J. Colloid Interface Sci. 13 (1958) 383. [13] G.H. Nancollas, W. White, F. Tsai, L. Maslow, Corrosion 35 (1979) 304. [14] M.E. Tadros, I. Mayes, J. Colloid Interface Sci. 72 (1979) 245. [15] M.P.C. Weijnen, G.M. van Rosmalen, J. Cryst. Growth 79 (1986) 157. [16] M.P.C. Weijnen, G.M. van Rosmalen, P. Bennema, J.J.M. Rijpkema, J. Cryst. Growth 82 (1987) 509. [17] R.A. Kuntz, Nature 211 (1966) 406. [18] L.W. Jones, Corrosion 17 (1961) 232. [19] Z. Amjad, Desalination 54 (1985) 263. [20] P. Somasundaran, T.W. Healy, D.W. Fuerstenau, J. Phys. Chem. 68 (1964) 3562.
170
M.G. Lioliou et al. / Journal of Colloid and Interface Science 303 (2006) 164–170
[21] J.T.G. Overbeek, Pure Appl. Chem. 46 (1976) 91. [22] J. Lyklema, in: B.M. Moudgil, P. Somasundaran (Eds.), Proceedings of the Engineering Foundation Conference, Sea Island, Georgia, 1985, pp. 3–21. [23] T. Miyajima, M. Mori, S.-I. Ishiguro, J. Colloid Interface Sci. 187 (1997) 259. [24] K. Huber, J. Phys. Chem. 97 (1993) 9825. [25] B.L. Rivasa, E.D. Pereiraa, I. Moreno-Villoslada, Prog. Polym. Sci. 28 (2003) 173. [26] APHA, Standard Methods for the Examination of Water and Wastewater, 15th ed., American Public Health Association, Washington, DC, 1981. [27] P.G. Klepetsanis, P.G. Koutsoukos, J. Cryst. Growth 98 (1989) 480. [28] M.C. van der Leeden, G.M. van Rosmalen, Desalination 66 (1987) 35. [29] M. Öner, Ö. Do˘gan, G. Öner, J. Cryst. Growth 186 (1998) 427. [30] G. Battaglia, F. Crea, P. Crea, S. Sammartano, Ann. Chim. 95 (2005) 643. [31] M.P.C. Weijnen, G.M. van Rosmalen, Desalination 54 (1985) 239.
[32] U.S. Adam, I.D. Rob, J. Chem. Soc. Faraday Trans. I 79 (1983) 2745. [33] MINEQL+ v.4.06, A Chemical Equilibrium Modeling System, Environmental Research Software, Hallowell, ME, 1998. [34] C. De Stefano, A. Gianguzza, D. Piazzese, S. Sammartano, Talanta 61 (2003) 181. [35] C. De Stefano, A. Gianguzza, D. Piazzese, S. Sammartano, J. Chem. Eng. Data 45 (2000) 876. [36] R.-m. Wen, S.-q. Deng, Z.-l. Zhu, W. Fan, Y.-f. Zhang, Chem. Res. Chinese Univ. 20 (1) (2004) 36. [37] N. Cabrera, D.A. Vermileya, in: R.H. Doremus, B.W. Roberts, D. Turnbull (Eds.), Growth and Perfection of Crystals, Wiley, New York, 1958, p. 393. [38] K. Bouropoulos, N. Bouropoulos, M. Melekos, P.G. Koutsoukos, G.C. Chitanu, A.G. Anghelescu-Dogaru, A.A. Carpov, J. Urol. 159 (5) (1998) 1755. [39] Z. Amjad, Can. J. Chem. 66 (1988) 1529.
Calcium sulfate precipitation in the presence of water-soluble polymers Maria G. Lioliou, Christakis A. Paraskeva, Petros G. Koutsoukos ∗ , Alkiviades C. Payatakes Institute of Chemical Engineering and High-Temperature Chemical Processes—Foundation for Research and Technology, P.O. Box 1414, GR-26500, Patras, Greece Department of Chemical Engineering, University of Patras, GR-26500, Patras, Greece Received 21 June 2006; accepted 22 July 2006 Available online 26 July 2006
Abstract The effect of four different polymers on the precipitation of calcium sulfate was investigated in the present work. The degree of inhibition was estimated from measurements of the calcium ion activity and from specific solution conductivity measurements in the supersaturated solutions during the course of the precipitation process. The effects of polyacrylic acid (PAA, three different polymers with average molecular weight 2000, 50,000, and 240,000, respectively) and of a co-polymer of PAA with polystyrene sulfonic acid (PSA, average molecular weight <20,000) were investigated with respect to their effect on the kinetics of spontaneous precipitation of calcium sulfate salts. The results of the kinetics experiments suggested that the spontaneous precipitation from supersaturated calcium sulfate solutions at 25 ◦ C yielded exclusively calcium sulfate dihydrate (gypsum) both in the absence and in the presence of the polymeric additives. The induction times, preceding the formation of the solid increased in all cases in the presence of the polymeric additives. Polymer concentrations as low as 2.0 ppm increased induction time from practically zero to 10 min. The rates of precipitation were reduced according to the solutions content in the polymers added and precipitation was completely suppressed in the presence of 6.0 ppm of the polymers tested, depending on their molecular weight. The lower the molecular weight of PAA, the more efficient was the threshold inhibition and the stronger the reduction of the rates of spontaneous precipitation. PSA yielded the poorest inhibition efficiency in comparison with the PAA, possibly because of the relatively lower affinity of the sulfonate groups for the calcium ions of the surface of the solid forming. The kinetics results analysis assuming Langmuir-type adsorption of the polymeric molecules on the growing supercritical gypsum nuclei showed different affinity for the polymers tested in agreement with the respective inhibition efficiency, in the order: PAA1 > PAA2 > PSA > PAA3. The presence of the polymers in the supersaturated solutions resulted in modification of the precipitated gypsum crystals morphology. © 2006 Elsevier Inc. All rights reserved. Keywords: Calcium sulfate dihydrate; Gypsum; Spontaneous precipitation; Inhibition; Polyacrylic acid; Polystyrene sulfonate
1. Introduction The formation of tenaciously adhering calcium sulfate scale in a number of processes from water desalination to heat exchangers and processes involving heating of water is a persistent problem [1]. Although six different calcium sulfate crystal forms are known to exist [2], three different salts are usually encountered in natural formations and scale precipitates: calcium sulfate dihydrate (CaSO4 ·2H2 O, CSD), calcium sulfate hemihydrate (CaSO4 ·1/2H2 O, CSH) and anhydrous calcium sulfate * Corresponding author. Fax: +30 2610997579.
E-mail address: [email protected] (P.G. Koutsoukos). 0021-9797/$ – see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2006.07.054
(CaSO4 , CSA). Both the CSA and the CSH salts may undergo further dehydration via phase transformation to the anhydrous form [3]. Despite the fact that considerable research has been going on during the past decades on the formation of calcium sulfate in aqueous media there is still large uncertainty concerning the mechanism of formation of this salt because of the largely variable conditions of the solutions in which the salt formation takes place, including temperature, pH, ionic strength and composition and the presence of foreign ions and or water soluble compounds. A large number of the published studies agree on the fact that the formation of the calcium sulfate nuclei is initiated on solid substrates. These substrates may be either metallic surfaces of heat exchangers or crystals of the same or different substrates [4–9].
M.G. Lioliou et al. / Journal of Colloid and Interface Science 303 (2006) 164–170
165
Among the most important objectives of the mechanistic investigations has been the possibility to control the formation of the various forms of calcium sulfate. The main effort to this end has been focused in the use of water soluble inhibitors which may act either as threshold inhibitors which block the development of the supercritical nuclei [10], or as retarders of the growth of the calcium sulfate crystals [11–16]. Several investigations have been carried out on the influence of trace amounts of polymeric scale inhibitors on both the precipitation and crystal habit modification of calcium sulfate forms [12,17–19]. Polymers containing carboxylic groups such as carboxymethyl cellulose (CMC), polymethacrylic acid (PMA), and polyacrylic acid (PAA) were found to be particularly effective as CaSO4 ·2H2 O growth inhibitors [12]. In the case of polymers in solution, there is general agreement that inhibition of the formation or the growth of the salt nuclei is effected by adsorption of these molecules on the active growth sites. Polymers tend to adsorb on solids from solutions due both to van der Waals [20] and/or electrostatic interactions [21,22]. The length of the polymer chains therefore as well as the functional groups present, which through ionization regulate the electrostatic charges of the polymers, are of primary importance for the investigation of the role of the respective polymeric additives in the crystal growth of calcium sulfate. In the present work, we have addressed the problem of the effect of PAA on the scale formation of gypsum (CSD) using polymers of markedly different molecular weight. Moreover, in order to compare the relative importance of the presence of the carboxylic groups, a co-polymer of PAA with sulfonated polystyrene was tested. The sulfonate groups are more strongly ionized in comparison to the carboxylic groups and are expected to promote stronger electrostatic interactions between the polymer and the surface of the calcium containing crystals which form in the supersaturated solutions. The effect of the water soluble polymeric additives was investigated in experiments in which calcium sulfate precipitation took place spontaneously from unstable supersaturated solutions, past the lapse of measurable induction time characteristic of the time frame needed for the formation of the supercritical nuclei and the subsequent initiation of the precipitation process. In order to accentuate the effect of the additives, the supersaturated solutions conditions selected for the test experiments in the absence of polymeric additives, yielded spontaneous precipitation with practically zero induction time.
The supersaturated solutions were prepared directly in the reactor by mixing equal volumes of equimolar calcium chloride and sodium sulfate solutions. The master variable used to monitor the process of CSD spontaneous precipitation was the specific conductivity of the solution. The conductivity and temperature of the reacting solution were monitored during the crystallization process by a D/A converter attached to a computer unit with the appropriate software. The progress and the extent of the crystal growth process both in the presence and in the absence of the inhibitors tested, was characterized by the decrease of the solution conductivity as a function of time. In the experiments done in the presence of polymeric inhibitors, the additives were introduced in the sulfate solution to avoid complexation [23], conformational changes [24], and even precipitation of calcium-polymer salts [25]. The polymers tested are summarized in Table 1. The homogeneity of the supersaturated solutions was ensured by magnetic stirring at ca. 250 rpm. Past the establishment of the solution supersaturation the recording of the solution conductivity was initiated. The formation of calcium sulfate crystals was detected by a decrease of the solution conductivity due to the reduction of calcium concentration. The time lapse between the preparation of the supersaturated solutions and the appearance of the inflection point in the conductivity– time profile, was defined as the induction time, τ . The point of inflection was considered as the point in which the slope of the curve changed and was determined by the intersection of the two tangents drawn: one at the initial period (flat) and one at the dropping part of the curve. During the course of the precipitation process samples were withdrawn from the reactor and were filtered through membrane filters (0.22 µm, Millipore). The filtrates were analyzed for calcium by EDTA complexometric titrations [26]. The combination of the chemical analyses of the solution composition with the solution conductivity allowed for the construction of calibration curves. The recordings of the solution conductivity as a function of time were thus converted into calcium vs time profiles. These curves showed a profile similar to that corresponding to decreasing conductivity as a function of time, which allowed for the calculation of the initial rates of calcium sulfate precipitation. The calcium– time profiles were fitted according to fourth-order polynomial and the [d[Ca2+ ]/dt]t→0 was taken as the value for the initial rate in each experiment. This correlation allowed for the quantitative measurements of the rates of crystallization. The values reported are the mean of three different measurements.
2. Experimental
3. Results and discussion
Crystal growth experiments were carried out in a 0.250dm3 double-walled Pyrex vessel thermostated at 25.0 ± 0.2 ◦ C by water circulation from a constant-temperature bath. Stock calcium chloride and sodium sulfate solutions were prepared from the respective crystalline solids (Merck, pro analisi). The solutions were filtered through membrane filters (0.22 µm, Millipore) and standardized by atomic absorption spectrometry (Perkin Elmer A Analyst 300) and ion chromatography (Dionex) for calcium and sulfate respectively.
In all experiments of the present work, the pH of the supersaturated solutions was about 5.0 and it was not adjusted. It is established that pH over a wide range (3.0–8.0) does not affect the kinetics of spontaneous precipitation of CSD [9,27,28]. Three types of PAA polymers of different molecular weights and one polysulfonic acid polymer were tested as inhibitors of the calcium sulfate precipitation. The first three polymers are characterized by the same binding capacity with respect to calcium but different conformation and adsorption properties with
166
M.G. Lioliou et al. / Journal of Colloid and Interface Science 303 (2006) 164–170
Table 1 Polymeric inhibitors tested in the present work Polymer
K752
PAA
K702
K798
Acronym Chemical type M.W.
PAA1 Poly(acrylic acid)
PAA2 Poly(acrylic acid)
PAA3 Poly(acrylic acid)
PSA Poly(acrylic acid/sulfonic acid/sulfonated styrene)
2000
50,000
240,000
<20,000
Goodrich, M/SGoodrite
Polysciences Inc.
Goodrich, M/SGoodrite
Goodrich, M/SGoodrite
Chemical formula of repeating unit
Manufacturer
respect to the forming crystalline substrate. The polyacrylic polymer with the highest MW is expected to form stronger complexes with calcium and possibly binds more strongly with the surface of the CSD supercritical nuclei. It has been suggested that larger number of functional groups of negatively charged species increases the polar attraction between the adsorbate and the positive sites at the solution interface [29]. The proton dissociation constant, log K, for polyacrylates is in the order of 4.4 [30]. The presence of hydrophobic aromatic nucleus on the molecule may change the mode of adsorption and the respective efficiency. Indeed it has been reported that the presence of sulfonic or phenylsulfonic acid groups which substitute carboxylic groups in a polymer reduce the inhibitory function of the polyelectrolytes on the crystal growth of gypsum [31]. Moreover, it has been shown from adsorption competition experiments, that sulfonic groups form weaker complexes with surface calcium ions in comparison with carboxylic groups [32]. The concentration of the additives tested in the supersaturated solutions varied between 1 to 100 ppm. At these concentration levels the influence on the solution supersaturation was negligible. In all cases, the only form of calcium sulfate precipitating was CSD. The solution speciation and supersaturation were calculated using the MINEQL+ speciation code [33] and all relevant equilibria [3]. It should be noted that complexation of calcium with the PAA was taken into consideration in our calculations [34–36]. The solution supersaturation with respect to CSD is given by equation: Ω=
(Ca2+ )(SO2− 4 ) . 0 Ks
(1)
The onset of CSD formation from the supersaturated solutions was accompanied by a decrease of the free calcium ions in the solution with a concomitant decrease of the solution conductivity. Typical profiles of the solution conductivity as a function of
Fig. 1. Conductivity–time profiles for the spontaneous precipitation of calcium sulfate in the absence and in the presence of PAA (symbols as in Table 1); total calcium, Cat = 120 mM, total sulfate, St = 120 mM, 25 ◦ C. PAA1 6 ppm; PAA2 6 ppm; PAA3 6 ppm.
time in the absence and in the presence of PAA are shown in Fig. 1. The inhibition of CSD formation by PAA may be explained by the Cabrera and Vermileya model [37]. According to this model, growth inhibition is explained in terms of inhibitor ions adsorption upon the crystal surface. During the induction period, most of the active growing sites may be poisoned by the additive molecules/ions. However, some of the growth sites of lower energy may still be free to grow; thus, the reaction proceeds at a very low rate. The induction times preceding precipitation for different amounts of the polymers tested in the present work are summarized in Table 2, while the plots of the induction times measured as a function of the concentration of the polymers tested is shown in Fig. 2. As may be seen, PAA1, i.e., PAA with the least molecular weight, yielded the most efficient threshold effect on the spontaneous precipitation of CSD. This may be ascribed to the most efficient interaction of the low MW polymer with the CSD crys-
M.G. Lioliou et al. / Journal of Colloid and Interface Science 303 (2006) 164–170
167
Table 2 Induction times measured for different amounts of the polymers tested for the precipitation of CSD from supersaturated solutions PAA1
Concentration (ppm) Ind. time (min)
2 10
5 45
6 110
20 200
PAA2
Concentration (ppm) Ind. time (min)
6 50
20 90
40 135
50 >300
PAA3
Concentration (ppm) Ind. time (min)
6 10
20 30
30 36
40 50
PSA
Concentration (ppm) Ind. time (min)
6 27
20 60
40 125
Note. Total calcium, Cat = 120 mM; total sulfate, St = 120 mM, 25 ◦ C.
Fig. 3. Spontaneous precipitation of CSD in the presence of 6 ppm of PAA. Total calcium, Cat = 120 mM, total sulfate, St = 120 mM, 25 ◦ C; (2) blank; (!) PAA1 (M.W. 2000); (P) PAA2 (M.W. 50,000); (1) PAA3 (M.W. 240,000).
Fig. 2. Variation of the induction times preceding the spontaneous precipitation of CSD as a function of the concentration of polymers; total calcium, Cat = 120 mM, total sulfate, St = 120 mM, 25 ◦ C. ( ) PAA1; (1) PAA2; (!) PAA3; (P) PSA.
talline substrate. It is thus possible that the low MW chains adsorb flat on the surface blocking the active sites of the critical nuclei which cannot grow further. The higher the MW the mode of adsorption may be such that the presence of loops and trains on the solid substrate [22] do allow for a larger number of unblocked active sites. The efficiency of the PAAs tested with respect to the threshold inhibition of CSD, decreased with increasing MW. The order found was PAA1 > PAA2 > PAA3 in agreement with previous reports [12,29]. In all cases, it was found that an increase in polymer concentration results in an increase of the induction time values. This means that the duration of the induction period is increased by increasing the amount of the additive in the reacting solution. Induction times for CSD spontaneous precipitation in the presence of 6 ppm of PAA increased from ∼20 to more than 110 min for MW decrease from 240,000 (PAA3) to 2000 (PAA1). The effect of the polymers on the spontaneous precipitation of CSD may be seen from the desupersaturation curves shown in Fig. 3. PSA, inhibited also the spontaneous precipitation of CSD at concentration levels exceeding 20 ppm, as may be seen in the concentration–time profile shown in Fig. 4. The precipitation was significantly suppressed in the presence of 60 and 100 ppm of the polymer. The PSA inhibitor contained hydrophobic phenyl groups attached to the polymer chain. In
Fig. 4. Spontaneous precipitation of CSD in the presence of PSA: (2) blank; (1) 6 ppm; (!) 20 ppm; () 60 ppm; () 100 ppm; total calcium, Cat = 120 mM, total sulfate, St = 120 mM, 25 ◦ C.
general, hydrophobic groups have been reported to be detrimental for the performance of an inhibitor. The presence of the phenyl group may, therefore, explain the relatively lower inhibition capacity of the PSA co-polymer in comparison to the PAA. It should be noted however that sulfonic acid groups are present in the PSA molecules. Sulfonic anions may replace the sulfate ions in the calcium sulfate lattice, enhancing, thus, the inhibiting capacity of the additive. In the present work however, it was found that PSA did not perform better than the PAA polymers, containing –COOH functional groups. Superior performance of PSA may have been anticipated because of the more acidic character of the sulfonic acid group compared to the carboxylic acid group. It may therefore be suggested that the presence of hydrophobic aromatic nuclei in PSA prevailed affecting the conformation of the polymeric chains at the gypsum/water interface yielding weaker interactions and therefore lower affinity of the polymer for the crystalline material. These results are in agreement with earlier reports which have attributed the weaker interactions between the crystal surfaces and the aromatic sulfonate containing polyelectrolytes to the fact that they do not favor the flat on conformation of the ad-
168
M.G. Lioliou et al. / Journal of Colloid and Interface Science 303 (2006) 164–170
(a)
(b)
(c)
(d)
(e) Fig. 5. Scanning electron micrographs of CSD crystals precipitated spontaneously; total calcium, Cat = 120 mM, total sulfate, St = 120 mM, 25 ◦ C; (a) no additive, bar = 40 µm; (b) 6 ppm PAA1, bar = 30 µm; (c) 6 ppm PAA2, bar = 30 µm; (d) 6 ppm PAA3, bar = 30 µm; (e) 6 ppm PSA, bar = 60 µm.
sorbed molecules, which is more likely for the carboxyl group containing PAAs [31]. The crystal habit of the precipitated CSD crystals was affected by presence of the polymers, as may be seen in the scanning electron photographs shown in Fig. 5. The habit of the CSD crystals precipitated in the absence of additives (Fig. 5a) was the well-known thin, elongated calcium sulfate crystals as a result of rapid growth on the [111] faces. The presence of
polyacrylic acid polymers in the reacting solutions enhanced the agglomeration of the crystals, retarded the growth of the [111] faces and as a result more plate-like crystals were obtained. In general, due to their adsorption on the crystal surface, PAAs do change the surface charge of the crystals, enhancing agglomeration. The presence of high molecular weight PAA (PAA3) resulted in even higher agglomeration as may be seen in Fig. 5d. In this case, the precipitated crystals consisted of
M.G. Lioliou et al. / Journal of Colloid and Interface Science 303 (2006) 164–170
169
Table 3 Affinity constants as calculated for the polymers tested for calcium sulfate crystal growth
Fig. 6. Plot of ratio R0 /(R0 − Ri ) as a function of the inverse of inhibitor concentration for (1) PAA1; (!) PAA2; (P) PAA3; () PSA.
rather deformed crystals, covered by sponge-like mass of tiny crystallites. The kinetic data obtained in the presence of the polymers were fitted to a kinetic model based on the assumption of the Langmuir isotherm [38]. Assuming that the rate of crystal growth in the absence of an additive is R0 and the limiting rate in its presence is bR0 and defining the ratio Cm /Cads as the fraction of the occupied sites, θ , the overall rate measured in the presence of an inhibitor, Ri , is given by Ri = R0 − θ R0 (1 − b).
(2)
Substituting θ and rearranging: R0 1 kdes 1 (3) = + . R0 − Ri 1 − b kads (1 − b) C The ratio kads /kdes is the affinity constant of the additive for the particular substrate and the parameter b is a measure of the effectiveness of the inhibitor present at infinite concentration in the case of monolayer coverage or at concentrations lower than the one corresponding to monolayer coverage (0 < b < 1). 0 as a function of C1 yielded satisfacPlots of the ratio R0R−R i tory linear fits for all the polymers used in the present work, as shown in Fig. 6. The intercept of the fitted lines for PAA1 and PAA2 were found to be <1, implying that these additives may cause complete inhibition of the precipitation process at concentrations lower than the one corresponding to monolayer coverage. The fact that the intercept was found to be >1 for the other two polymers (PAA3 and PSA) may be interpreted as the inability of these additives to completely inhibit crystal growth of calcium sulfate. The inverse of the slopes of the fitted lines yield the affinity constants of the polymers used for calcium sulfate crystal surfaces, which are given in Table 3. The affinities calculated for PAA1 and PAA2 were 1.96 × 106 and 1.19 × 106 , respectively, while the affinities for PAA3 and PSA were 1.33 × 105 and 2.48 × 105 . It is interesting to point out the small difference between the values for the first two PAAs in spite of the difference in their molecular weights, while PAA3 gave a value of affinity constant lower by one order of magnitude. This difference may be ascribed to the chain
Polymer
Acronym
Affinity constant
K752 PAA K702 K798
PAA1 PAA2 PAA3 PSA
1.96 × 106 1.19 × 106 1.33 × 105 2.48 × 105
length of the polymer which affected the respective surface conformation. It should also be mentioned that the experiments were conducted at pH approximately 5.0. At this pH value the higher molecular weight polymer is probably ionized to a lower extent, in comparison to the lower MW PAAs. Lower extent of ionization would in turn affect the conformation and/or the extent of adsorption of PAA on the newly formed gypsum nuclei. The presence of the aromatic ring in PSA structure apparently influenced its hydrophobic character and the conformation of the molecule at the solid–solution interface, resulting in reduced inhibition activity. Finally it should be noted that despite the fact that the present work was done at acidic pH, sufficiently high to have the polyelectrolytes in ionized form, increasing the solution pH is expected to increase the efficiency of the inhibitors tested because of the increase of the degree of ionization of the soluble polyelectrolytes [28,31,39]. Acknowledgments The authors wish to acknowledge financial support by the General Secretariat for Research and Technology, Ministry of Development, through PENED Program Contract M413/2002. References [1] Z. Amjad, J. Colloid Interface Sci. 123 (1988) 523. [2] J. Glater, J.L. York, K.S. Campbell, in: K.S. Spiegler, A.D.K. Laird (Eds.), Principles of Desalination, Part B, second ed., Academic Press, New York, 1980, pp. 627–678. [3] P.G. Klepetsanis, P.G. Koutsoukos, J. Colloid Interface Sci. 143 (2) (1991) 299. [4] D. Hasson, J. Zahavi, Ind. Eng. Chem. Fundam. 9 (1) (1970) 26. [5] O.D. Linnikov, Desalination 128 (2000) 35. [6] J.S. Gill, G.H. Nancollas, J. Cryst. Growth 48 (1980) 34. [7] G.H. Nancollas, W.P. Klima, Mater. Performance 21 (1982) 9. [8] G.H. Nancollas, W.F. Klima, Paper 81, Corrosion/81, National Association of Corrosion Engineers Conference, Toronto, Ontario, 1981. [9] S.T. Liu, G.H. Nancollas, J. Colloid Interface Sci. 44 (1973) 422. [10] S. He, J.E. Oddo, M.B. Tomson, J. Colloid Interface Sci. 162 (2) (1994) 297. [11] B.R. Smith, A.E. Alexander, J. Colloid Interface Sci. 34 (1970) 81. [12] E.R. McCartney, A.E. Alexander, J. Colloid Interface Sci. 13 (1958) 383. [13] G.H. Nancollas, W. White, F. Tsai, L. Maslow, Corrosion 35 (1979) 304. [14] M.E. Tadros, I. Mayes, J. Colloid Interface Sci. 72 (1979) 245. [15] M.P.C. Weijnen, G.M. van Rosmalen, J. Cryst. Growth 79 (1986) 157. [16] M.P.C. Weijnen, G.M. van Rosmalen, P. Bennema, J.J.M. Rijpkema, J. Cryst. Growth 82 (1987) 509. [17] R.A. Kuntz, Nature 211 (1966) 406. [18] L.W. Jones, Corrosion 17 (1961) 232. [19] Z. Amjad, Desalination 54 (1985) 263. [20] P. Somasundaran, T.W. Healy, D.W. Fuerstenau, J. Phys. Chem. 68 (1964) 3562.
170
M.G. Lioliou et al. / Journal of Colloid and Interface Science 303 (2006) 164–170
[21] J.T.G. Overbeek, Pure Appl. Chem. 46 (1976) 91. [22] J. Lyklema, in: B.M. Moudgil, P. Somasundaran (Eds.), Proceedings of the Engineering Foundation Conference, Sea Island, Georgia, 1985, pp. 3–21. [23] T. Miyajima, M. Mori, S.-I. Ishiguro, J. Colloid Interface Sci. 187 (1997) 259. [24] K. Huber, J. Phys. Chem. 97 (1993) 9825. [25] B.L. Rivasa, E.D. Pereiraa, I. Moreno-Villoslada, Prog. Polym. Sci. 28 (2003) 173. [26] APHA, Standard Methods for the Examination of Water and Wastewater, 15th ed., American Public Health Association, Washington, DC, 1981. [27] P.G. Klepetsanis, P.G. Koutsoukos, J. Cryst. Growth 98 (1989) 480. [28] M.C. van der Leeden, G.M. van Rosmalen, Desalination 66 (1987) 35. [29] M. Öner, Ö. Do˘gan, G. Öner, J. Cryst. Growth 186 (1998) 427. [30] G. Battaglia, F. Crea, P. Crea, S. Sammartano, Ann. Chim. 95 (2005) 643. [31] M.P.C. Weijnen, G.M. van Rosmalen, Desalination 54 (1985) 239.
[32] U.S. Adam, I.D. Rob, J. Chem. Soc. Faraday Trans. I 79 (1983) 2745. [33] MINEQL+ v.4.06, A Chemical Equilibrium Modeling System, Environmental Research Software, Hallowell, ME, 1998. [34] C. De Stefano, A. Gianguzza, D. Piazzese, S. Sammartano, Talanta 61 (2003) 181. [35] C. De Stefano, A. Gianguzza, D. Piazzese, S. Sammartano, J. Chem. Eng. Data 45 (2000) 876. [36] R.-m. Wen, S.-q. Deng, Z.-l. Zhu, W. Fan, Y.-f. Zhang, Chem. Res. Chinese Univ. 20 (1) (2004) 36. [37] N. Cabrera, D.A. Vermileya, in: R.H. Doremus, B.W. Roberts, D. Turnbull (Eds.), Growth and Perfection of Crystals, Wiley, New York, 1958, p. 393. [38] K. Bouropoulos, N. Bouropoulos, M. Melekos, P.G. Koutsoukos, G.C. Chitanu, A.G. Anghelescu-Dogaru, A.A. Carpov, J. Urol. 159 (5) (1998) 1755. [39] Z. Amjad, Can. J. Chem. 66 (1988) 1529.