- Email: [email protected]
Synthesis of nanosized BaSO4 and CaCO3 particles with a membrane reactor: effects of additives on particles
Journal of Colloid and Interface Science 266 (2003) 322–327 www.elsevier.com/locate/jcis
Synthesis of nanosized BaSO4 and CaCO3 particles with a membrane reactor: effects of additives on particles Zhiqian Jia,1 Zhongzhou Liu,∗ and Fei He Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences, PO Box 2871, Beijing, People’s Republic of China Received 16 May 2002; accepted 31 January 2003
Abstract Nanosized BaSO4 particles, about 15 nm in size, were synthesized successfully by a membrane reactor at the aids of additives, in which Na2 SO4 solutions were added into BaCl2 solutions gradually through the micropores of ultrafiltration membranes to control the saturation ratio, subsequently the nucleation and growth rates. The effects of additives species, additives concentrations, and membranes molecular weight cut-off (MWCO) on the particle morphology, along with the formation processes of particles, were investigated. CaCO3 nanoparticles of 30–60 nm in size were also prepared by the reactor. The results revealed that the addition of methyl alcohol, ethanol etc. favor the synthesis of nanoparticles with small size. The particles size decreases with the increase in ethanol concentrations. With the increase in membrane MWCO, the products tend from nanoparticles towards aggregates. 2003 Elsevier Inc. All rights reserved. Keywords: Synthesis; BaSO4 ; Nanoparticles; Membrane reactor; Additives
1. Introduction Liquid-reaction-induced precipitation is often used to prepare nanosized particles, in which process nucleation is usually dominated by homogeneous nucleation. Nevertheless, if the supersaturation and nucleation rates are too high, agglomeration will become an important growth mechanism, leading to irregular aggregates [1]. On the other hand, the distribution of supersaturation in a reaction system is another important factor in the dependence of particle size distribution on it. For an ideal process, the mixing of reactive ions should attain homogeneity at the molecular level before the establishment of a steady-state nucleation rate. Moderate supersaturation with good distribution can be gotten by means of gradual chemical reactions [2], e.g., hydrolysis of metal alkoxides, conversion of complexes, and decomposition of compounds, to control the nucleation and growth rate effectively. To control the particle size, the reaction, nucleation, and growth environment can be con* Corresponding author.
E-mail address: [email protected] (Z. Liu). 1 Present affiliation: Department of Photochemistry, Institute of Chem-
istry, Chinese Academy of Sciences, Zhongguancun 1st Street, Haidian District, Beijing, China. 0021-9797/$ – see front matter 2003 Elsevier Inc. All rights reserved. doi:10.1016/S0021-9797(03)00187-5
strained to the mesoscale by means of microemulsions, ion exchange resins, sol–gels, etc. Special additives can also be employed to inhibit the growth rates by blocking the kink positions because the growth rate is dominantly determined by the numbers of free kinks [3]. For example, the effects of EDTA [4], phosphates, acrylate [5], aminoethanol [6], etc. on the growth of BaSO4 crystals were studied. Intensifying mixing of the reactants by mixers, e.g., stirrers, jets, Tee mixers, static mixers, and rotating packed beds, can improve the micromixing conditions. But in these reactors, liquids can only be dispersed into eddies of 10–100 µm in average diameter, the subsequent micromixing should be completed by molecular diffusion. So new methods and reactors are still being sought for. In our former paper [7], a membrane reactor was proposed for the synthesis of nanosized BaSO4 particles, in which hollow fiber ultrafiltration (HF UF) membranes were employed to add Na2 SO4 solutions into BaCl2 solutions gradually through the membrane micropores, several or tens of nanometers in mean diameter, under the transmembrane pressure, to control the supersaturation ratio and the nucleation and growth rates. Meanwhile, the permeate size was dramatically reduced to nanoscale, resulting in a decrease in characteristic diffusion time and an improvement of mi-
Z.Q. Jia et al. / Journal of Colloid and Interface Science 266 (2003) 322–327
cromixing conditions. And particles about 70 nm in size were obtained. To further reduce the particle size, in this paper, BaSO4 and CaCO3 nanoparticles were synthesized by the membrane reactor with the aid of additives. The effects of additive species, additive concentrations, and membrane molecular weight cut-off (MWCO) on the particles, morphology, along with the formation of particles, were investigated.
323
was stopped. The equipment was rinsed four times with distilled water to remove impurities prior to each reaction. The morphology of the particles was analyzed with a Hitachi H800 transmission electron microscope (TEM).
3. Results and discussion 3.1. Calculation of saturation ratio
2. Experimental 2.1. Experimental materials and equipment The membrane module was fitted with 10 pieces of HF UF membranes, 1.0 mm in inner diameter, 0.130 m in effective length, produced by the Zhongke Membrane R&D Center of Beijing. The membrane material and MWCO were as follows: PS/PDC (30,000 Da), PS/PDC (10,000 Da), and PES/PDC (1,000 Da). The peristaltic pump (Cole–Parmer Instrument Company) was employed. The reagents, e.g., Na2 SO4 , BaCl2 , Na2 CO3 , CaCl2 , methyl alcohol, ethanol, n-butyl alcohol, acetone, isoamyl alcohol, and acetic acid, were all analytically pure.
The concentration of BaCl2 , CA , during the reaction can be expressed as t VA,0 CA,0 − CB,0 0 JB dt CA = (1) , t VA,0 + 0 JB dt where VA,0 is the initial volume of BaCl2 solution, CA,0 is the initial concentration of BaCl2 , CB,0 is the initial concentration of Na2 SO4 , and JB is the permeation rate of Na2 SO4 solutions for the membrane module. Supposing complete mixing and reaction of the reactants, the average saturation ratio, S, in the membrane lumina can be expressed as γA [Ba2+ ]γB [SO2− γA CA FA γB CB JB 1 4 ] = , Ksp FA + JB FA + JB Ksp
2.2. Experimental procedure
S=
The reaction was operated in semibatch made as shown in Fig. 1. Na2 SO4 solutions (0.025 mol l−1 ) in the graduated tube permeated into the lumina of the HF UF membrane through the micropores under the transmembrane pressure (P ). At the same time, BaCl2 solutions (0.0050 mol l−1 , 100.0 ml) in the stirring tank was drawn into the lumina of the HF membrane, reacted with permeated Na2 SO4 to form BaSO4 , and then returned into the stirring tank for recycling. The additives were added into the BaCl2 solutions and mixed sufficiently prior to each reaction. The liquid levels of Na2 SO4 solutions in the graduated tube were recorded at intervals and the permeation fluxes were calculated. The flow rate of BaCl2 solutions in the membrane lumina was 0.92 m s−1 . When the total molarity of Na2 SO4 solutions permeated equaled the initial molarity of BaCl2 solutions, the reaction
where Ksp is the solubility product of BaSO4 at equilibrium, which equals 1.0 × 10−9.2 at 25 ◦ C; FA is the volumetric flow rate of BaCl2 solutions; [Ba2+ ] and [SO2− 4 ] are the average concentrations in the membrane lumina; and γA and γB are the activity coefficients for respective ions. It should be noted that the real local saturation ratio in the boundary layer is probably higher than that calculated because of the instantaneous reaction, whereas the average saturation ratio provides a relative measure for evaluating the effects of operating parameters.
Fig. 1. Schematic of experimental apparatus.
(2)
3.2. Effects of additive species and concentrations The effects of ethanol, acetone, and acetic acid with the same concentration (0.0571 mol l−1 ) on the particles morphology were explored. The membrane with MWCO 1000 was employed and P was controlled to be 0.0050 MPa. Figure 2a shows that the particles are about 70 nm in diameter when no additives are added. Ethanol and acetic acid have better effects on the control of particle size, in which case particles about 15 nm in size were obtained (Figs. 2b, 2c, 2d). The effects of additives may include three main aspects as follows. Additives can adsorb on the steps and kinks of the particle surface and decrease the numbers of free steps and kinks, as well as the particle surface energy, leading to the inhibition of particle growth and agglomeration. Furthermore, the additives may also affect the nucleation process. As we know, the activation energy for homogeneous nucleation,
324
Z.Q. Jia et al. / Journal of Colloid and Interface Science 266 (2003) 322–327
(a)
(b)
(c)
(d)
(e)
(f)
(g)
(h)
Fig. 2. Effects of additive species and concentrations on the particles: (a) no additive; (b) ethanol, 0.0571 mol l−1 ; (c) acetone, 0.0571 mol l−1 ; (d) acetic acid, 0.0571 mol l−1 ; (e) methyl alcohol, 0.0571 mol l−1 ; (f) n-butyl alcohol, 0.0571 mol l−1 ; (g) ethanol, 0.1142 mol l−1 ; (h) ethanol, 0.1713 mol l−1 . The particles are prepared under the conditions of MWCO = 1000, t = 30 ◦ C, CA,0 = 0.0050 mol l−1 , CB,0 = 0.025 mol l−1 , P = 0.0050 MPa, and ripening for 24 h.
Z.Q. Jia et al. / Journal of Colloid and Interface Science 266 (2003) 322–327
Gc , is expressed as [1] 3 2 Gc = 16σLS v 3h2 T 2 (ln S)2 ,
325
(3)
where σLS is the solid–liquid interfacial tension, v is the molecular volume of the embryo, h is the Planck constant, and T is the temperature. The nucleation rate, I , can be expressed in terms of Gc and the diffusion activity energy, Ga , as I = N exp(−Gc / hT ) · B exp(−Ga / hT ),
(4)
where N is the solute molecular number per unit volume; B = av0 ; a is the vibration ratio of atoms in the direction of the embryo; and v0 is the vibration frequency of atoms. The critical nuclear size, r ∗ , is expressed as r ∗ = 2βa σLS v/(3βv kB T ln S),
Fig. 3. Effects of membrane MWCO on the permeation flux: (2) MWCO = 1000; (1) MWCO = 10,000; (Q) MWCO = 30,000. The particles are prepared under the conditions of P = 0.0050 MPa, t = 30 ◦ C, CA,0 = 0.0050 mol l−1 , CB,0 = 0.025 mol l−1 . The ethanol concentration is 0.1142 mol l−1 .
(5)
where βv is the volume conversion factor, βa is the area conversion factor, and kB is the Boltzmann constant. The addition of additives will result in a new liquid phase (L ) and subsequently a different solid–liquid interfacial tension, σL S . If σL S < σLS , the additives can reduce Gc and r ∗ and then boost the nucleation rate, I . The solubility of BaSO4 in the mixed solutions is also expected to decrease due to the decrease in the polarity of solvents, which can also boost the nucleation rate. As ethanol shows excellent effects on the control of particle size, the effects of methyl alcohol and n-butyl alcohol with the same concentration, 0.0571 mol l−1 , were also studied. Figures 2e, 2f show that methyl alcohol also has better inhibition effects on the particle size. The results suggest that the inhibition effects may be mainly attributed to the adsorption of additives on the particles surface because, according to the Traube rule [8], n-butyl alcohol will reduce the solution tension most effectively due to its longer carbon chain. The adsorption of additives on the particles is a competitive process with the water molecules and is determined by the molecular interaction between the BaSO4 surface and the additive, the additives and the water, and BaSO4 and the water. As we know, BaSO4 is a polar compound and tends to adsorb higher polar additives from the solutions. The sequence of adsorption amount can be related qualitatively to that of the additives polarities, i.e., methyl alcohol > ethanol > n-butyl alcohol. So methyl alcohol and ethanol have better growth inhibition effects. Figures 2b, 2g, 2h show that the particle sizes decrease with increasing ethanol concentrations. When the concentration is 0.1713 mol l−1 , the products seem to be loose small aggregates of minute particles. 3.3. Effects of membrane MWCO Membranes with different MWCO (1000, 10,000, and 30,000) were employed under the conditions of T = 303 K and P = 0.0050 MPa. The ethanol concentration was 0.1142 mol l−1 .
With the increase in membrane MWCO, the permeation rates of Na2 SO4 increase from 0.11 to 0.5 and 1.65 ml min−1 . Consequently, the initial saturation ratios increase from 29 to 145 and 505, and the reaction time decreases from 460 min to 65 and 10 min, respectively (Fig. 3). When MWCO = 1000, the sol was sampled at reaction degrees, ξ , of 0.5 and 1.0, with corresponding reaction times of 100 and 340 min, respectively. When ξ = 0.5, the products seem to be loose small aggregates of minute particles (Fig. 4a). As ξ = 1.0, particles with average size of 15 nm are obtained (Fig. 4b). When MWCO = 10,000, the sols were sampled at reaction degrees of 0.25 and 1.0, with corresponding reaction times of 7 and 111 min. The products sol is transparent and colorless, and the particle size is about 60 nm (Fig. 4d). When MWCO = 30,000, many white precipitates appear during the reaction, and the products appear as irregular agglomerations (Fig. 4e). The growth of particles shows an aggregation mechanism; that is, nuclei aggregate quickly after formation due to high surface energy. Then the nuclei in contact are cemented together by deposition of growth units from the bulk solutions and finally form amorphous particles. The aggregation rates are determined by the colloid stability factor and the number density of growth units in suspension [1]. When membranes with lower MWCO, e.g., 1000 and 10,000, are employed, the permeation fluxes, saturation ratios, nucleation and aggregation growth rates are low. The growth units deposit on the aggregates gradually, resulting in higher arrangement degree and regular nanoparticles. Higher MWCO (e.g., 30,000) will lead to higher S, greater nucleation and aggregation rate, and lower arrangement degree. So only irregular agglomerations are gained. Under the experimental conditions, membrane with MWCO 1000 seems to be more favorable for the synthesis of nanosized particles. To compare the membrane reactor with a classical stirring reactor, 20.0 ml BaCl2 solutions (0.0050 mol l−1 ) were added into a 50-ml beaker, and the ethanol concentration in the BaCl2 solutions was 0.1142 mol l−1 . Then 0.025 mol l−1
326
Z.Q. Jia et al. / Journal of Colloid and Interface Science 266 (2003) 322–327
(a)
(b)
(c)
(d)
(e) Fig. 4. Forming processes of the particles: (a) ξ = 0.50, MWCO = 1000; (b) ξ = 1.0, MWCO = 1000; (c) ξ = 0.25, MWCO = 10,000; (d) ξ = 1.0, MWCO = 10,000; (e) ξ = 1.0, MWCO = 30,000. P = 0.0050 MPa, t = 30 ◦ C, CA,0 = 0.0050 mol l−1 , CB,0 = 0.025 mol l−1 . The ethanol concentration is 0.1142 mol l−1 and without ripening.
Na2 SO4 solutions were added into the BaCl2 solutions by an acidic buret at a rate of 0.20 ml min−1 under 30 ◦ C and vigorous stirring (magnetic stirrer of ∅ 9 × 27 mm, 600 rpm). Although the adding rate (0.20 ml min−1 ) of Na2 SO4 solutions was lower than the permeation rate (∼0.35 ml min−1 ) in the membrane reactor with MWCO 10,000, lots of precipitates were generated during the addition because the droplet size (∼1 mm) was much larger than the permeate size (∼10 nm). 3.4. Synthesis of nanosized CaCO3 Nanosized CaCO3 particles were also synthesized by the membrane reactor with MWCO 1000, in which CaCl2
solutions with initial concentration of 0.020 mol l−1 were drawn into the membrane lumina, while Na2 CO3 solutions (0.10 mol l−1 ) permeated into the lumina from the shell side of the module. The reaction temperature was 30 ◦ C and the transmembrane pressure was 0.0050 MPa. The initial permeation flux of Na2 CO3 solutions was 2.5 × 10−3 ml min−1 cm−2 , and the initial saturation ratio calculated was about 10 (the solubility product of CaCO3 is 3 × 10−8 at 25 ◦ C. The particles prepared are about 60 nm in diameter (Fig. 5a). When isoamyl alcohol (0.110 mol l−1 ) was added into CaCl2 solutions prior to the reaction, the particles obtained are about 30 nm in diameter (Fig. 5b), showing that alcohol with low molecular weight also has good inhibition effects on the growth of CaCO3 particles.
Z.Q. Jia et al. / Journal of Colloid and Interface Science 266 (2003) 322–327
(a)
327
(b)
Fig. 5. CaCO3 nanoparticles: (a) with no additive; (b) with the addition of isoamyl alcohol (0.110 mol l−1 ). The concentrations of CaCl2 and Na2 CO3 solutions are 0.020 mol l−1 and 0.10 mol l−1 , respectively. MWCO = 1000, P = 0.0050 MPa, t = 30 ◦ C and without ripening.
4. Conclusions
JB
Nanosized BaSO4 and CaCO3 particles were synthesized successfully by a membrane reactor at the aids of additives. It was found that high polar additives, e.g., methyl alcohol and ethanol, have better inhibition effects on the particle growth. The inhibition effects are probably mainly related to the amounts of additives adsorpted on the particle surface. The particle size decreases with the increasing ethanol concentration. With the increase in membrane MWCO, the products tend from nanoparticles toward aggregates.
kB Ksp N r∗ S v VA,0
Acknowledgments The authors are grateful for financial support from the National Natural Sciences Foundation of China (Project 50072042) and the Chinese Academy of Sciences (Projects KZ951-A1-201-02, KZ95T-05).
permeation rate of Na2 SO4 of total membrane area (ml min−1 ) Boltzmann constant solubility product of BaSO4 at equilibrium solute molecular number per unit volume critical nucleus radius, m average saturation ratio in membrane lumina molecular volume of an embryo initial volume of BaCl2 solutions (ml)
Greek symbols ξ γA γB βv βa Ga σLS
reaction degree activity coefficient of BaCl2 activity coefficient of Na2 SO4 the volume conversion factor the area conversion factor diffusion activation energy solid–liquid interfacial tension, N m−1
References Appendix A. Nomenclature a CA CB CA,0 CB,0 FA h jB
atom vibration odds on the embryo direction concentration of BaCl2 (mol l−3 ) concentration of Na2 SO4 (mol l−3 ) initial concentration of BaCl2 (mol l−3 ) initial concentration of Na2 SO4 (mol l−3 ) volumetric flow rate of BaCl2 solution (ml min−1 ) Plank’s constant permeation flux of Na2 SO4 (ml cm2 min−1 )
[1] [2] [3] [4] [5]
J.A. Dirkksen, T.A. Ring, Chem. Eng. Sci. 46 (1991) 2389. T. Sugimoto, Adv. Colloid Interface Sci. 28 (1987) 65. G.M. Bliznakov, Sov. Phys. Crystallogr. 4 (1959) 134–140. K. Takiyama, Bull. Chem. Soc. Jpn. 31 (1958) 950–953. M.C. Van der Leeden, G.M. Van Rosmalen, J. Cryst. Growth 100 (1990) 109–116. [6] S. Kipp, R. Lacmann, M. Reichelt, Chem. Eng. Tech. 19 (1996) 543– 549. [7] Z. Jia, Z. Liu, J. Membrane Sci. 209 (2002) 153–161. [8] M.J. Jaycock, G.D. Parfitt, Chemistry of Interfaces, Wiley, New York, 1981.
Synthesis of nanosized BaSO4 and CaCO3 particles with a membrane reactor: effects of additives on particles Zhiqian Jia,1 Zhongzhou Liu,∗ and Fei He Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences, PO Box 2871, Beijing, People’s Republic of China Received 16 May 2002; accepted 31 January 2003
Abstract Nanosized BaSO4 particles, about 15 nm in size, were synthesized successfully by a membrane reactor at the aids of additives, in which Na2 SO4 solutions were added into BaCl2 solutions gradually through the micropores of ultrafiltration membranes to control the saturation ratio, subsequently the nucleation and growth rates. The effects of additives species, additives concentrations, and membranes molecular weight cut-off (MWCO) on the particle morphology, along with the formation processes of particles, were investigated. CaCO3 nanoparticles of 30–60 nm in size were also prepared by the reactor. The results revealed that the addition of methyl alcohol, ethanol etc. favor the synthesis of nanoparticles with small size. The particles size decreases with the increase in ethanol concentrations. With the increase in membrane MWCO, the products tend from nanoparticles towards aggregates. 2003 Elsevier Inc. All rights reserved. Keywords: Synthesis; BaSO4 ; Nanoparticles; Membrane reactor; Additives
1. Introduction Liquid-reaction-induced precipitation is often used to prepare nanosized particles, in which process nucleation is usually dominated by homogeneous nucleation. Nevertheless, if the supersaturation and nucleation rates are too high, agglomeration will become an important growth mechanism, leading to irregular aggregates [1]. On the other hand, the distribution of supersaturation in a reaction system is another important factor in the dependence of particle size distribution on it. For an ideal process, the mixing of reactive ions should attain homogeneity at the molecular level before the establishment of a steady-state nucleation rate. Moderate supersaturation with good distribution can be gotten by means of gradual chemical reactions [2], e.g., hydrolysis of metal alkoxides, conversion of complexes, and decomposition of compounds, to control the nucleation and growth rate effectively. To control the particle size, the reaction, nucleation, and growth environment can be con* Corresponding author.
E-mail address: [email protected] (Z. Liu). 1 Present affiliation: Department of Photochemistry, Institute of Chem-
istry, Chinese Academy of Sciences, Zhongguancun 1st Street, Haidian District, Beijing, China. 0021-9797/$ – see front matter 2003 Elsevier Inc. All rights reserved. doi:10.1016/S0021-9797(03)00187-5
strained to the mesoscale by means of microemulsions, ion exchange resins, sol–gels, etc. Special additives can also be employed to inhibit the growth rates by blocking the kink positions because the growth rate is dominantly determined by the numbers of free kinks [3]. For example, the effects of EDTA [4], phosphates, acrylate [5], aminoethanol [6], etc. on the growth of BaSO4 crystals were studied. Intensifying mixing of the reactants by mixers, e.g., stirrers, jets, Tee mixers, static mixers, and rotating packed beds, can improve the micromixing conditions. But in these reactors, liquids can only be dispersed into eddies of 10–100 µm in average diameter, the subsequent micromixing should be completed by molecular diffusion. So new methods and reactors are still being sought for. In our former paper [7], a membrane reactor was proposed for the synthesis of nanosized BaSO4 particles, in which hollow fiber ultrafiltration (HF UF) membranes were employed to add Na2 SO4 solutions into BaCl2 solutions gradually through the membrane micropores, several or tens of nanometers in mean diameter, under the transmembrane pressure, to control the supersaturation ratio and the nucleation and growth rates. Meanwhile, the permeate size was dramatically reduced to nanoscale, resulting in a decrease in characteristic diffusion time and an improvement of mi-
Z.Q. Jia et al. / Journal of Colloid and Interface Science 266 (2003) 322–327
cromixing conditions. And particles about 70 nm in size were obtained. To further reduce the particle size, in this paper, BaSO4 and CaCO3 nanoparticles were synthesized by the membrane reactor with the aid of additives. The effects of additive species, additive concentrations, and membrane molecular weight cut-off (MWCO) on the particles, morphology, along with the formation of particles, were investigated.
323
was stopped. The equipment was rinsed four times with distilled water to remove impurities prior to each reaction. The morphology of the particles was analyzed with a Hitachi H800 transmission electron microscope (TEM).
3. Results and discussion 3.1. Calculation of saturation ratio
2. Experimental 2.1. Experimental materials and equipment The membrane module was fitted with 10 pieces of HF UF membranes, 1.0 mm in inner diameter, 0.130 m in effective length, produced by the Zhongke Membrane R&D Center of Beijing. The membrane material and MWCO were as follows: PS/PDC (30,000 Da), PS/PDC (10,000 Da), and PES/PDC (1,000 Da). The peristaltic pump (Cole–Parmer Instrument Company) was employed. The reagents, e.g., Na2 SO4 , BaCl2 , Na2 CO3 , CaCl2 , methyl alcohol, ethanol, n-butyl alcohol, acetone, isoamyl alcohol, and acetic acid, were all analytically pure.
The concentration of BaCl2 , CA , during the reaction can be expressed as t VA,0 CA,0 − CB,0 0 JB dt CA = (1) , t VA,0 + 0 JB dt where VA,0 is the initial volume of BaCl2 solution, CA,0 is the initial concentration of BaCl2 , CB,0 is the initial concentration of Na2 SO4 , and JB is the permeation rate of Na2 SO4 solutions for the membrane module. Supposing complete mixing and reaction of the reactants, the average saturation ratio, S, in the membrane lumina can be expressed as γA [Ba2+ ]γB [SO2− γA CA FA γB CB JB 1 4 ] = , Ksp FA + JB FA + JB Ksp
2.2. Experimental procedure
S=
The reaction was operated in semibatch made as shown in Fig. 1. Na2 SO4 solutions (0.025 mol l−1 ) in the graduated tube permeated into the lumina of the HF UF membrane through the micropores under the transmembrane pressure (P ). At the same time, BaCl2 solutions (0.0050 mol l−1 , 100.0 ml) in the stirring tank was drawn into the lumina of the HF membrane, reacted with permeated Na2 SO4 to form BaSO4 , and then returned into the stirring tank for recycling. The additives were added into the BaCl2 solutions and mixed sufficiently prior to each reaction. The liquid levels of Na2 SO4 solutions in the graduated tube were recorded at intervals and the permeation fluxes were calculated. The flow rate of BaCl2 solutions in the membrane lumina was 0.92 m s−1 . When the total molarity of Na2 SO4 solutions permeated equaled the initial molarity of BaCl2 solutions, the reaction
where Ksp is the solubility product of BaSO4 at equilibrium, which equals 1.0 × 10−9.2 at 25 ◦ C; FA is the volumetric flow rate of BaCl2 solutions; [Ba2+ ] and [SO2− 4 ] are the average concentrations in the membrane lumina; and γA and γB are the activity coefficients for respective ions. It should be noted that the real local saturation ratio in the boundary layer is probably higher than that calculated because of the instantaneous reaction, whereas the average saturation ratio provides a relative measure for evaluating the effects of operating parameters.
Fig. 1. Schematic of experimental apparatus.
(2)
3.2. Effects of additive species and concentrations The effects of ethanol, acetone, and acetic acid with the same concentration (0.0571 mol l−1 ) on the particles morphology were explored. The membrane with MWCO 1000 was employed and P was controlled to be 0.0050 MPa. Figure 2a shows that the particles are about 70 nm in diameter when no additives are added. Ethanol and acetic acid have better effects on the control of particle size, in which case particles about 15 nm in size were obtained (Figs. 2b, 2c, 2d). The effects of additives may include three main aspects as follows. Additives can adsorb on the steps and kinks of the particle surface and decrease the numbers of free steps and kinks, as well as the particle surface energy, leading to the inhibition of particle growth and agglomeration. Furthermore, the additives may also affect the nucleation process. As we know, the activation energy for homogeneous nucleation,
324
Z.Q. Jia et al. / Journal of Colloid and Interface Science 266 (2003) 322–327
(a)
(b)
(c)
(d)
(e)
(f)
(g)
(h)
Fig. 2. Effects of additive species and concentrations on the particles: (a) no additive; (b) ethanol, 0.0571 mol l−1 ; (c) acetone, 0.0571 mol l−1 ; (d) acetic acid, 0.0571 mol l−1 ; (e) methyl alcohol, 0.0571 mol l−1 ; (f) n-butyl alcohol, 0.0571 mol l−1 ; (g) ethanol, 0.1142 mol l−1 ; (h) ethanol, 0.1713 mol l−1 . The particles are prepared under the conditions of MWCO = 1000, t = 30 ◦ C, CA,0 = 0.0050 mol l−1 , CB,0 = 0.025 mol l−1 , P = 0.0050 MPa, and ripening for 24 h.
Z.Q. Jia et al. / Journal of Colloid and Interface Science 266 (2003) 322–327
Gc , is expressed as [1] 3 2 Gc = 16σLS v 3h2 T 2 (ln S)2 ,
325
(3)
where σLS is the solid–liquid interfacial tension, v is the molecular volume of the embryo, h is the Planck constant, and T is the temperature. The nucleation rate, I , can be expressed in terms of Gc and the diffusion activity energy, Ga , as I = N exp(−Gc / hT ) · B exp(−Ga / hT ),
(4)
where N is the solute molecular number per unit volume; B = av0 ; a is the vibration ratio of atoms in the direction of the embryo; and v0 is the vibration frequency of atoms. The critical nuclear size, r ∗ , is expressed as r ∗ = 2βa σLS v/(3βv kB T ln S),
Fig. 3. Effects of membrane MWCO on the permeation flux: (2) MWCO = 1000; (1) MWCO = 10,000; (Q) MWCO = 30,000. The particles are prepared under the conditions of P = 0.0050 MPa, t = 30 ◦ C, CA,0 = 0.0050 mol l−1 , CB,0 = 0.025 mol l−1 . The ethanol concentration is 0.1142 mol l−1 .
(5)
where βv is the volume conversion factor, βa is the area conversion factor, and kB is the Boltzmann constant. The addition of additives will result in a new liquid phase (L ) and subsequently a different solid–liquid interfacial tension, σL S . If σL S < σLS , the additives can reduce Gc and r ∗ and then boost the nucleation rate, I . The solubility of BaSO4 in the mixed solutions is also expected to decrease due to the decrease in the polarity of solvents, which can also boost the nucleation rate. As ethanol shows excellent effects on the control of particle size, the effects of methyl alcohol and n-butyl alcohol with the same concentration, 0.0571 mol l−1 , were also studied. Figures 2e, 2f show that methyl alcohol also has better inhibition effects on the particle size. The results suggest that the inhibition effects may be mainly attributed to the adsorption of additives on the particles surface because, according to the Traube rule [8], n-butyl alcohol will reduce the solution tension most effectively due to its longer carbon chain. The adsorption of additives on the particles is a competitive process with the water molecules and is determined by the molecular interaction between the BaSO4 surface and the additive, the additives and the water, and BaSO4 and the water. As we know, BaSO4 is a polar compound and tends to adsorb higher polar additives from the solutions. The sequence of adsorption amount can be related qualitatively to that of the additives polarities, i.e., methyl alcohol > ethanol > n-butyl alcohol. So methyl alcohol and ethanol have better growth inhibition effects. Figures 2b, 2g, 2h show that the particle sizes decrease with increasing ethanol concentrations. When the concentration is 0.1713 mol l−1 , the products seem to be loose small aggregates of minute particles. 3.3. Effects of membrane MWCO Membranes with different MWCO (1000, 10,000, and 30,000) were employed under the conditions of T = 303 K and P = 0.0050 MPa. The ethanol concentration was 0.1142 mol l−1 .
With the increase in membrane MWCO, the permeation rates of Na2 SO4 increase from 0.11 to 0.5 and 1.65 ml min−1 . Consequently, the initial saturation ratios increase from 29 to 145 and 505, and the reaction time decreases from 460 min to 65 and 10 min, respectively (Fig. 3). When MWCO = 1000, the sol was sampled at reaction degrees, ξ , of 0.5 and 1.0, with corresponding reaction times of 100 and 340 min, respectively. When ξ = 0.5, the products seem to be loose small aggregates of minute particles (Fig. 4a). As ξ = 1.0, particles with average size of 15 nm are obtained (Fig. 4b). When MWCO = 10,000, the sols were sampled at reaction degrees of 0.25 and 1.0, with corresponding reaction times of 7 and 111 min. The products sol is transparent and colorless, and the particle size is about 60 nm (Fig. 4d). When MWCO = 30,000, many white precipitates appear during the reaction, and the products appear as irregular agglomerations (Fig. 4e). The growth of particles shows an aggregation mechanism; that is, nuclei aggregate quickly after formation due to high surface energy. Then the nuclei in contact are cemented together by deposition of growth units from the bulk solutions and finally form amorphous particles. The aggregation rates are determined by the colloid stability factor and the number density of growth units in suspension [1]. When membranes with lower MWCO, e.g., 1000 and 10,000, are employed, the permeation fluxes, saturation ratios, nucleation and aggregation growth rates are low. The growth units deposit on the aggregates gradually, resulting in higher arrangement degree and regular nanoparticles. Higher MWCO (e.g., 30,000) will lead to higher S, greater nucleation and aggregation rate, and lower arrangement degree. So only irregular agglomerations are gained. Under the experimental conditions, membrane with MWCO 1000 seems to be more favorable for the synthesis of nanosized particles. To compare the membrane reactor with a classical stirring reactor, 20.0 ml BaCl2 solutions (0.0050 mol l−1 ) were added into a 50-ml beaker, and the ethanol concentration in the BaCl2 solutions was 0.1142 mol l−1 . Then 0.025 mol l−1
326
Z.Q. Jia et al. / Journal of Colloid and Interface Science 266 (2003) 322–327
(a)
(b)
(c)
(d)
(e) Fig. 4. Forming processes of the particles: (a) ξ = 0.50, MWCO = 1000; (b) ξ = 1.0, MWCO = 1000; (c) ξ = 0.25, MWCO = 10,000; (d) ξ = 1.0, MWCO = 10,000; (e) ξ = 1.0, MWCO = 30,000. P = 0.0050 MPa, t = 30 ◦ C, CA,0 = 0.0050 mol l−1 , CB,0 = 0.025 mol l−1 . The ethanol concentration is 0.1142 mol l−1 and without ripening.
Na2 SO4 solutions were added into the BaCl2 solutions by an acidic buret at a rate of 0.20 ml min−1 under 30 ◦ C and vigorous stirring (magnetic stirrer of ∅ 9 × 27 mm, 600 rpm). Although the adding rate (0.20 ml min−1 ) of Na2 SO4 solutions was lower than the permeation rate (∼0.35 ml min−1 ) in the membrane reactor with MWCO 10,000, lots of precipitates were generated during the addition because the droplet size (∼1 mm) was much larger than the permeate size (∼10 nm). 3.4. Synthesis of nanosized CaCO3 Nanosized CaCO3 particles were also synthesized by the membrane reactor with MWCO 1000, in which CaCl2
solutions with initial concentration of 0.020 mol l−1 were drawn into the membrane lumina, while Na2 CO3 solutions (0.10 mol l−1 ) permeated into the lumina from the shell side of the module. The reaction temperature was 30 ◦ C and the transmembrane pressure was 0.0050 MPa. The initial permeation flux of Na2 CO3 solutions was 2.5 × 10−3 ml min−1 cm−2 , and the initial saturation ratio calculated was about 10 (the solubility product of CaCO3 is 3 × 10−8 at 25 ◦ C. The particles prepared are about 60 nm in diameter (Fig. 5a). When isoamyl alcohol (0.110 mol l−1 ) was added into CaCl2 solutions prior to the reaction, the particles obtained are about 30 nm in diameter (Fig. 5b), showing that alcohol with low molecular weight also has good inhibition effects on the growth of CaCO3 particles.
Z.Q. Jia et al. / Journal of Colloid and Interface Science 266 (2003) 322–327
(a)
327
(b)
Fig. 5. CaCO3 nanoparticles: (a) with no additive; (b) with the addition of isoamyl alcohol (0.110 mol l−1 ). The concentrations of CaCl2 and Na2 CO3 solutions are 0.020 mol l−1 and 0.10 mol l−1 , respectively. MWCO = 1000, P = 0.0050 MPa, t = 30 ◦ C and without ripening.
4. Conclusions
JB
Nanosized BaSO4 and CaCO3 particles were synthesized successfully by a membrane reactor at the aids of additives. It was found that high polar additives, e.g., methyl alcohol and ethanol, have better inhibition effects on the particle growth. The inhibition effects are probably mainly related to the amounts of additives adsorpted on the particle surface. The particle size decreases with the increasing ethanol concentration. With the increase in membrane MWCO, the products tend from nanoparticles toward aggregates.
kB Ksp N r∗ S v VA,0
Acknowledgments The authors are grateful for financial support from the National Natural Sciences Foundation of China (Project 50072042) and the Chinese Academy of Sciences (Projects KZ951-A1-201-02, KZ95T-05).
permeation rate of Na2 SO4 of total membrane area (ml min−1 ) Boltzmann constant solubility product of BaSO4 at equilibrium solute molecular number per unit volume critical nucleus radius, m average saturation ratio in membrane lumina molecular volume of an embryo initial volume of BaCl2 solutions (ml)
Greek symbols ξ γA γB βv βa Ga σLS
reaction degree activity coefficient of BaCl2 activity coefficient of Na2 SO4 the volume conversion factor the area conversion factor diffusion activation energy solid–liquid interfacial tension, N m−1
References Appendix A. Nomenclature a CA CB CA,0 CB,0 FA h jB
atom vibration odds on the embryo direction concentration of BaCl2 (mol l−3 ) concentration of Na2 SO4 (mol l−3 ) initial concentration of BaCl2 (mol l−3 ) initial concentration of Na2 SO4 (mol l−3 ) volumetric flow rate of BaCl2 solution (ml min−1 ) Plank’s constant permeation flux of Na2 SO4 (ml cm2 min−1 )
[1] [2] [3] [4] [5]
J.A. Dirkksen, T.A. Ring, Chem. Eng. Sci. 46 (1991) 2389. T. Sugimoto, Adv. Colloid Interface Sci. 28 (1987) 65. G.M. Bliznakov, Sov. Phys. Crystallogr. 4 (1959) 134–140. K. Takiyama, Bull. Chem. Soc. Jpn. 31 (1958) 950–953. M.C. Van der Leeden, G.M. Van Rosmalen, J. Cryst. Growth 100 (1990) 109–116. [6] S. Kipp, R. Lacmann, M. Reichelt, Chem. Eng. Tech. 19 (1996) 543– 549. [7] Z. Jia, Z. Liu, J. Membrane Sci. 209 (2002) 153–161. [8] M.J. Jaycock, G.D. Parfitt, Chemistry of Interfaces, Wiley, New York, 1981.