J. Chem. Soc., Faraday Trans. I, 1984,80, 217-224 Crystal Growth of Strontium Fluoride from Aqueous Solution BY ROBERT A. BOCHNER, ABBAS ABDUL-RAHMAN AND GEORGE H. NANCOLLAS* Chemistry Department, State University of New York at Buffalo, Buffalo, New York 14214, U.S.A. Received 6th June, 1983 The kinetics of growth of strontium fluoride crystals has been studied in aqueous solution at 25 OC using a constant-composition method in which the supersaturation and ionic strength were maintained constant by the addition of titrants consisting of strontium nitrate and potassium fluoride solutions. The rate of crystallization is independent of ionic strength and, at low supersaturation, is best represented in terms of a spiral growth mechanism with a reaction order, n = 2, with respect to relative supersaturation.At higher supersaturations, a greater change in rate of growth with increasing concentration (n x 4.3) suggests a polynuclear mechanism. Inhibition of crystallization by the presence of phosphonates can be interpreted in terms of a Langmuir isotherm. The crystal growth of the alkaline-earth fluorides is of importance in view of their involvement in spectroscopy, electronics, lasers, glass manufacture and in the fluoridation of drinking waters. In the case of strontium fluoride, the crystallization as strontium-90 fluoride offers an attractive method for removing this element from highly radioactive fission products. The development of methods for growing large, easily filtered crystals of this sparingly soluble salt would significantly improve this process.The irreproducibility of spontaneous precipitation studies is well known since both nucleation and crystal growth may be involved concomitantly. These problems were overcome with the development of highly reproducible seeded-growth techniquesly which allowed factors such as temperature, supersaturation, ionic strength and solid: solution ratios to be studied quantitatively. In the present work a constant- composition method3 has been used to investigate the crystallization of strontium fluoride under conditions of constant ionic strength and supersaturation. The method yields kinetics data for the crystallization reaction even at very low supersaturation. The application of the technique to the industrial preparation of large crystals of strontium fluoride has attractive possibilities.EXPERIMENTAL Supersaturated solutions of strontium fluoride were prepared using both ultrapure (Alfa Chemicals) and reagent-grade (J. T. Baker) chemicals with triply distilled deionized water. The concentration of strontium ions was determined (0.2%) by exchanging the metal ion for hydrogen ion on a Dowex-50 ion-exchange resin and titrating the liberated acid with standard base. All fluoride solutions were prepared and stored in polyethylene or polypropylene bottles in order to prevent fluoride attack on glass surfaces. Seed crystals were prepared by simultaneously adding 200 cm3 of 0.956 mol dmP3 potassium fluoride and 190 cm3 of 0.504 mol dmP3 strontium chloride to 250 cm3 of triply distilled water. The slow mixing was made in a nitrogen atmosphere over a period of 2.5 h at 25 OC.The crystals were washed free of chloride ions with triply distilled water by decantation, and the suspension (seed A) was stored 217218 CRYSTAL GROWTH FROM AQUEOUS SOLUTION in polyethylene containers for up to one year before use. Seed suspension B was prepared by adding 1 dm3 of potassium fluoride (1 .O mol dmP3) and 1 dm3 of strontium nitrate (0.50 mol dm-3) to 500 cm3 of triply distilled water over a period of 12 h at 25 "C. The particles were kept in suspension by means of a Teflon overhead stirrer, and the filtered solid was washed with saturated strontium fluoride solution and allowed to age before use for up to six months. A third seed suspension (seed C) was obtained by the rapid mixing (ca.4 min) of 1 dmP3 volumes of potassium fluoride (0.780 mol dm-3) and strontium nitrate (0.390 mol dm-3) solutions. The crystals were washed with distilled water and allowed to age at 25 O C . Seed crystals were confirmed as strontium fluoride by X-ray powder diffraction (Phillips XRG 3000 X-ray diffractometer, Ni filter and Cu Ka radiation). Specific surface areas (s.s.a.) of the seed and grown phases were measured (& 1 %) by a single-point B.E.T. nitrogen adsorption using a 30% nitrogen + helium mixture (Quantasorb 11, Quantachrome Greenvale, N.Y .). The values were l9.7,2.68 and 33 m2 g-l for seeds A, B and C, respectively. Scanning electron micrographs (IS1 model I1 scanning electron microscope) and transmission electron micrographs (Lietz electron microscope) showed the seed materials to consist of small aggregates of particles having cube-like morphology. Crystal-growth experiments were made in a stream of pre-saturated nitrogen gas using a cell consisting of a double-walled Pyrex glass vessel of 500 cm3 capacity fitted with a Teflon lid and polyethylene liner.The cell was maintained at 25 OC by circulating thermostatted water through the outer jacket. Stirring was effected using a Teflon stirrer with an overhead variable-speed motor (model I, Eastern Engineering Co., Conn.). Supersaturated solutions were prepared by slowly mixing strontium nitrate and potassium (or sodium) fluorides in the range of super- saturation (ti = O.OC1.8) in which the supersaturation, ti, is defined by 0 = (([Sr2+] [F-]2)1'3 - ([Sr2+]o[F-]f)1~3})/([Sr2fl,[F-]~)1~3 where [Sr2+], [F-] and [Sr2+],,, [F-1, are the concentrations of lattice ions at time t and at equilibrium, respectively.The latter values were calculated from conditional solubility products at the ionic strengths of the experiments. In the crystallization experiments, following the addition of a known volume of suspension of seed crystals to the metastable supersaturated solutions, the activities of ion species were maintained constant by the addition of titrant solutions consisting of strontium nitrate and potassium fluoride from mechanically coupled burets. The addition was controlled by means of a pH-stat (Metrohm Combitrator model 3D, Brinkmann Instrument Co.) using a fluoride- specific electrode (Orion model 94-09), coupled with a thermal-electrolytic silver, silver chloride reference electrode.This reference electrode was designed to include an intermediate liquid junction containing background electrolyte so as to eliminate errors due to leakage of salt bridge solution into the crystallization cell during the reactions. The reference electrode limb which was immersed in the cell solution was constructed of Teflon. During the crystallization experiments, aliquots were withdrawn from time to time, filtered (0.2 pm Millipore filtration) and analysed for strontium ion by spectrophotometric titration with EDTA and by atomic absorption (Perkin-Elmer model 503). The data confirmed the constancy of the lattice-ion concentrations to within 1 %. The solid phases, collected during the experiments, were also investigated by X-ray diffraction, specific-surface-area analysis and scanning electron microscopy. RESULTS AND DISCUSSION Concentrations of ionic species in the supersaturated solutions were calculated as described previously4 using expressions for mass balance, electroneutrality and the thermodynamic equilibrium constants, K , for the various associated species in the equilibria : K H F e H++ F- 6.61 x ref.(5) HF+ F- s HF; 0.295, ref. (5) SrF+ e Sr2+ + F- 0.147, ref. (6) SrOH+ + Sr2+ +OH- 0.150, ref. (7) H,O e H+ + OH- 1.002 x ref. (8).R. A. BOCHNER, A. ABDUL-RAHMAN AND G. H. NANCOLLAS 219 The computations were made by successive approximations for the ionic strength, I, as described previouslyg using activity coefficients calculated from the extended form of the Debye-Hiickel equation proposed by Davies.l0 For the experiments which were made at constant ionic strength (0.15 mol dm-3), values of the conditional equilibrium constants were also estimated by using activity-coefficient data calculated from the Davies equation.The solubility of strontium fluoride was obtained by allowing growth experiments to proceed to equilibrium; the value at 25 "C was 1.343 x mol dm-3 at an ionic strength of 0.046 mol dm-3. The corresponding thermodynamic solubility product, K,, = 2.82 x m0l3 dm-9. The value compares well with that reported by Talipov and Khadeev, 2.45 x m013 dm-9.11 In the crystal-growth experiments, the rate of reaction was calculated from the rate of addition of titrant solutions.It has been shown that errors introduced by the limited response time of the Orion fluoride electrode were negligible. The results of the crystal-growth experiments are summarized in table 1, in which 7&. and TF are the molar strontium and fluoride concentrations. Typical plots of the Table 1. Crystallization experiments at 25 OC, &.ITF = 0.5 T,r I seed ~/10-7 / 1 0-3 mol / 1 O-* mol mol dmP3 dm-3 ff amount/mg type min-' mP2 10, 11, 12a 20 21 16 18 19 17 107 109 106 111 110 108, 112a 23 29 31 33 35 39 40 49 50 52 54 55 57 59 60 63 65 3.75 3.50 3.30 3.20 3.00 2.70 2.50 3.75 3.50 3.30 3.00 2.75 2.50 1.60 1.80 2.00 2.20 2.30 2.40 2.60 2.80 3.00 1.90 2.00 2.30 . 2.60 2.90 3.30 3.50 3.70 6.86 6.74 6.74 6.59 6.49 6.37 6.24 6.86 6.74 6.64 6.49 6.37 6.24 15.0 15.0 15.0 15.0 15.0 15.0 15.0 15.0 15.0 6.80 6.80 6.80 6.80 6.80 6.80 6.80 6.80 1.81 1.63 1.49 1.42 1.27 1.09 0.9 I 1.81 1.63 1.49 1.27 1.09 0.9 1 0.09 0.22 0.36 0.49 0.56 0.63 0.76 0.90 1.03 0.44 0.52 0.74 0.97 1.20 1 S O 1.65 1.80 48.3 48.3 48.3 48.3 48.3 48.3 48.3 28 1 28 1 28 1 28 1 28 1 28 1 3120 1800 390 2000 780 780 2000 800 800 266 266 133 266 266 266 133 133 A A A A A A A B B B B B B C C C C C C C C C C C C C C C C C 41 30 20 18 11 7.2 4.0 75 51 34 18 14 5.3 0.0 1 0.03 0.10 0.18 0.25 0.45 0.64 0.83 1.15 0.13 0.21 0.38 0.70 2.20 4.10 6.19 8.74 a Effect of stirring rate (in r.p.m.) expt 10 (60); expt 11 (100); expt 12 (80); expt 108 (100); expt 112 (60 and 300); all others 100 r.p.m.220 CRYSTAL GROWTH FROM AQUEOUS SOLUTION amount of strontium fluoride precipitated as a function of time are shown in fig.1 . It can be seen that for given values of CT, the rate of growth is constant for at least 60 min, at which time the amount of precipitation corresponds to as much as 25% of the initial mass of inoculating seed. The rates of crystallization, R, in table 1 were calculated from the slopes of linear plots such as those in fig. 1 . It can be seen that 36 20 40 tlmin 60 Fig. 1. Crystal growth of strontium fluoride. Plots of amount precipitated against time. The numbers refer to the experiment numbers in table 1 . Expt 107, 109, 21 and 16, left-hand ordinate; expt 40 and 50, right-hand ordinate. R, normalized for the initial surface area of each inoculating seed, is constant, confirming that crystallization takes place on the seed crystals without additional nucleation or spontaneous precipitation. The rates of growth of strontium fluoride seed fall in the order seed B > seed A > seed C , suggesting different densities of growth sites on the crystal surfaces.During the reactions, the crystals (ca. 2pm in size) maintained their cubic morphology, as observed by scanning electron microscopy, and were present largely as aggregates ca. 10 pm in size. For many sparingly soluble salts M,A,, the rate of crystallization can be expressed by12 where K is a constant (= ks Kson’(a+b)), k is the rate constant for crystal growth, s is proportional to the number of growth sites available on the seed crystals, K,, is the solubility product at the ionic strength of the experiment and n is the apparent order of reaction.Analysis of the growth data for strontium fluoride is shown in fig. 2, in which -log R is plotted against -log CT. The slopes of the linear plots indicate a change in the value of n as the supersaturation of the solutions is reduced. In the concentration range 0.9 < CT < 2.0 the value of n is 3.5k0.1, while at lower supersaturation (0.09 < CT < 1.03) the apparent reaction order is 2.0 & 0.1. At all supersaturations, the growth rate is insensitive to changes in stirring rate, as can be seen from the results of experiments 10-12 in table 1. l4 hydrated monolayer15 and polynuclear/birth and spread.l6-l* In general, it has been shown13 that for the spiral growth and hydrated-monolayer models, a value of n M 2 would rate = R = d[M,A,]/dt = Kan (1) Crystallization models for solution growth include spiralR. A.BOCHNER, A. ABDUL-RAHMAN AND G. H. NANCOLLAS 22 1 be expected in eqn (1) while for polynuclear crystallization n > 2. The slopes of the linear plots in fig. 2 indicate spiral growth or dehydration mechanisms at lower supersaturation and a polynuclear crystallization at higher concentrations. Similar results have been obtained in the crystallization of other alkaline-earth flu~rides.'~ A typical plot of the specific surface area (s.s.a.) of the grown phases as a function of the extent of reaction is shown in fig. 3. It can be seen that the decrease in s.s.a. is more rapid than that calculated on the basis of an isometric three-dimensional crystal growth.The reduction in s.s.a. during the first 100% of crystal growth probably -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 -log cr Fig. 2. Plots of -log R against -log 0 for the crystallization of strontium fluoride. Seed A: 0 ; seed B: V; seed C: a, 0 and 0. 0 and 0, Right-hand ordinate, all others left-hand ordinate. 20 16 - I M n E 12 9 > 2 8 4 500 1000 1500 2000 growth (%) Fig. 3. Plots of s.s.a. against growth with respect to original seed. Seed A, 10, 1 1 and 12 : 0 ; calculated values: 7. a = 1.810. Expt222 CRYSTAL GROWTH FROM AQUEOUS SOLUTION reflects crystal-lattice perfection and surface annealing processes. Indeed, the decrease in s.s.a. during this period over-compensates for the general increase in surface area accompanying macroscopic crystal growth. Scanning electron micrographs of the crystals grown to an extent of more than twenty times the mass of inoculating seed showed that although the crystals maintained their orthorhombic form, aggregation into larger particles occurred.These compensating effects of increasing and decreasing available surface area for growth appeared to account for the negligible change in overall surface area during the linear growth plots in fig. 1. Crystallization experiments made in the presence of phosphonate inhibitors, TENTMP (triethylenediaminetetramethylene phosphonic acid, Dequest 205 1, Mon- santo Chemical Co.) and NTA (nitrilotrimethylene phosphonic acid, Dequest 200 1, Monsanto Chemical Co.) are summarized in table 2. The marked reduction in rate Table 2. Crystal growth in the presence of phosphonates.T,, = OST, = 3.30 x mol dm-3, [KNO,] = 5.66 x mol dm-3, seed A expt no. concentration R/ 1 0-6 inhibitor (PPm) mol m i x 1 m-2 21 22 24 26 25 28 27 47 43 44 45 46 - NTMP NTMP NTMP NTMP NTMP NTMP TENTMP TENTMP TENTMP TENTMP TENTMP - 0.02 0.08 0.23 0.463 0.650 0.93 0.084 0.232 0.463 0.650 0.79 1 1.88 1.64 1.51 0.66 0.36 0.1 1 0 1.52 1 .oo 0.72 0.24 0.21 12 ' N E 7 ' 8 2 --. U m a J aJ a * .y .- .- 20 4 0 6 0 t /m in Fig. 4. Crystallization rate plots in the presence of phosphonate inhibitors.R. A. BOCHNER, A. ABDUL-RAHMAN AND G. H. NANCOLLAS 223 of crystallization is illustrated by the rate plots in fig. 4. In these experiments, the phosphonate additives were introduced into the strontium fluoride supersaturated solutions prior to inoculation with seed crystals.It is apparent that constant rates of growth were again observed for at least 60 min of reaction and that both NTMP and TENTMP were effective inhibitors of crystallization. Applying a simple Langmuir adsorption the influence of both inhibitors can be interpreted in terms of their selective adsorption at growth sites on the crystal surface. The decrease in crystallization rate can be related to the crystal surface area covered by the adsorbed inhibitor molecules. If R, and Ri are the rates of crystallization in the absence and presence of inhibitors, respectively, the Langmuir isotherm requires a linear relationship between the relative reduction in rate, Ro/(Ro-Ri), and the reciprocal of the inhibitor concentration.20 Fig. 5 confirms the applicability of this simple adsorption isotherm, with both phosphonates being approximately equally effective in reducing the rate of crystallization of strontium fluoride.6.0 '4 e l < 4.0 % v . 0 2.0 2.0 6.0 1 /[inhibitor] 10.0 Fig. 5. The influence of TENTMP (v) and NTMP (0) on the crystallization of strontium fluoride at 25 "C. We acknowledge financial support from the National Science Foundation in a University/Industry Cooperative research grant (no. CPE 8005345) between the State University of New York at Buffalo and Martin Marietta Laboratories. C . W. Davies and A. L. Jones, Discuss. Faraday SOC., 1949,5, 103. G. H. Nancollas and N . Purdie, Q. Rev. Chem. SOC., 1964, 18, 1. P. Koutsoukos, Z. Amjad, M. B. Tomson and G. H. Nancollas, J . Am. Chem. Soc., 1980,102, 1553. L. J. Shyu and G. H. Nancollas, Croat. Chem. Acta, 1980, 53, 281. A. J. Ellis, J . Chem. SOC., 1963, 4300. R. E. Connick and M. S. Tsao, J . Am. Chem. SOC., 1954,76, 5311. ' F. G. Gimbell and C. B. Monk, Trans. Faraday SOC., 1954, 50, 965. H. S. Harned and W. J. Hamer, J. Am. Chem. Soc., 1933,55, 2194. G. H. Nancollas, Interactions in Electrolyte Solutions (Elsevier, Amsterdam, 1966). lo C . W. Davies, Ion Association (Butterworths, London, 1960). l 1 Sh. T. Talipov and V. A. Khadeev, Zh. Obshch. Khim., 1950, 20, 783.224 CRYSTAL GROWTH FROM AQUEOUS SOLUTION l2 G. H. Nancollas, A h . Colloid Interface Sci., 1979, 10, 215. l3 A. E. Nielsen, Pure Appl. Chem., 1981, 53, 2025. l4 G. M. van Rosmalen, P. J. Dandey and W. G. J. Marchee, J. Cryst. Growth, 1981, 52, 801. l5 C. W. Davies and A. L. Jones, Trans. Faraday SOC., 1955,51, 812. l6 G. H. Gilmer and P. Bennema, J. Appl. Phys., 1972,43, 1347. l7 L. H. E. Madsen and R. J. Boistelle, J. Cryst. Growth, 1979, 46, 681. l9 G. H. Nancollas, R. A. Bochner, E. Liolios, L. J. Shyu, Y. Yoshikawa, J. P. Barone and D. Svrjek, 2o P. G. Koutsoukos, Z. Amjad and G. H. Nancollas, J. Colloid Interface Sci., 1981, 83, 599. W. B. Hillig, Acta Metall., 1966, 14, 1868. Am. Inst. Chem. Eng., Symp. Ser., 1982, 215, 26. (PAPER 3/921)