Preparation of Calcium Carbonate Nanoparticles with a Continuous ...

MATERIALS AND PRODUCT ENGINEERING Chinese Journal of Chemical Engineering, 21(2) 121—126 (2013) DOI: 10.1016/S1004-9541(13)60449-8

Preparation of Calcium Carbonate Nanoparticles with a Continuous Gas-liquid Membrane Contactor: Particles Morphology and Membrane Fouling* JIA Zhiqian (贾志谦)**, CHANG Qing (常青), QIN Jin (秦晋) and MAMAT Aynur (阿伊努尔·买买提) Lab for Membrane Science and Technology, College of Chemistry, Beijing Normal University, Beijing 100875, China Abstract Nanosized calcium carbonate particles were prepared with a continuous gas-liquid membrane contactor. The effects of Ca(OH)2 concentration, CO2 pressure and liquid flow velocity on the particles morphology, pressure drop and membrane fouling were studied. With rising Ca(OH)2 concentrations, the average size of the particles increased. The effects of Ca(OH)2 concentration and CO2 pressure on particles were not apparent under the experimental conditions. When the Ca(OH)2 concentration and liquid flow velocity were high, or the CO2 pressure was low, the fouling on the membrane external surface at the contactor entrance was serious due to liquid leakage, whereas the fouling was slight at exit. The fouling on the membrane inner-surface at entrance was apparent due to adsorption of raw materials. The membrane can be recovered by washing with dilute hydrochloric acid and reused for at least 6 times without performance deterioration. Keywords membrane contactor, gas-liquid reaction, CaCO3 nanoparticles, two-phase reaction, membrane absorption

1

INTRODUCTION

Gas-liquid reactions are often employed in the preparation of nanoparticles, for examples, CaCO3 [1, 2], ZrO2 [3], CdS [4], Pt [5], etc. In the bi-phase reactions, the mass transfer of gas into the liquid phase often plays a crucial role in determining the absorption rate. In traditional gas-liquid reaction devices, the two phases are inter-dispersed, and the interfacial area is limited and varies with operation conditions. The interdependence of the two phases also leads to problems such as emulsion, foaming, unloading and flooding [6]. Wang et al. [7] used a membrane dispersion mini-reactor to introduce CO2 into Ca(OH)2 suspensions in the preparation of CaCO3 nanoparticles. The average diameter of gas bubbles was reduced to 0.1-2 mm and the mass transfer was enhanced. In our previous work [8, 9], a gas-liquid membrane contactor was employed in the preparation of nanosized calcium carbonate for the first time, and the effects of Ca(OH)2 concentration, CO2 partial pressure and liquid flow velocity on the absorption rate were studied theoretically and experimentally. The membrane contactor provides substantially large and fixed specific interfacial area. All the gas can be absorbed in the non-disperse contacting mode, which is especially important for the expensive or hazardous gases. The energy consumption is low and the process can be easily scaled-up. Therefore, this membrane contactor provides a green and efficient alternative for gas-liquid reaction process. In this paper, as a continuation of the study, the effects of Ca(OH)2 concentration, CO2 partial pressure and liquid flow rate on the particles morphology, pressure drop and membrane fouling along with the

membranes cleaning and reuse were investigated. These studies were crucial important for the process development and design. 2 2.1

EXPERIMENTAL Materials

The Ca(OH)2 powder was analytical grade and was dissolved with degassed distilled water prior to experiment. The membrane contactor was made up of 20 polypropylene hollow fiber membranes with pores diameter of 1-10 μm, porosity of 0.6, outer diameter of 0.5 mm, inner diameter of 0.4 mm and effective length of 90.0 mm (Tianjin Blue Cross Membrane Technology Co., Ltd.). The fibers were well-separated by two porous stainless steel sheets mounted at the two ends of membrane module. 2.2

Synthesis

The experimental set-up and procedure were given in detail in our previous work [8, 9]. Pure CO2 from cylinder flowed into the gas supply tank, the buffer tank and then the shell side of the membrane contactor (Fig. 1). CO2 pressure was monitored with pressure gauge (0-10 kPa, 2.5 grade). Ca(OH)2 suspension in the stirred tank flowed into the membranes lumens, contacted with CO2 gas permeated through membrane pores, reacted and then flowed into the product vessel. The flow rate of liquid was measured with a rotameter and the pH was monitored by a pH-meter. Liquid pressure was monitored with pressure gauge (0-0.16 MPa, 2.5 grade).

Received 2011-05-12, accepted 2012-07-12. * Supported by the National Natural Science Foundation of China (20676016, 21076024). ** To whom correspondence should be addressed. E-mail: [email protected]

122

Chin. J. Chem. Eng., Vol. 21, No. 2, February 2013

Figure 1 Schematic of experiment apparatus 1—stirred tank; 2—peristaltic pump; 3—liquid rotameter; 4—pressure gauge; 5—membrane contactor; 6—product tank; 7—CO2 cylinder; 8—gas supply tank; 9—buffer tank; 10—barometer; 11—gas chromatography

2.3

3

Analysis

For the observation of membrane fouling, after reaction pure water was pumped to flow through the membranes lumens at flow velocity of 0.053 m·s−1 for 1 min to wash the membrane inner surface. Then the membranes specimens at the contactor inlet and outlet were taken, dried at ambient conditions, sputtered with platinum and observed by cold-field scanning electron microscopy (SEM S-4800, Japan). The particle morphology was observed by SEM and transmission electron microscope (TEM, Hitachi). The particles mean diameter was obtained by measuring at least one hundred particles. The XRD pattern was analyzed by monochromatized Cu Kα incident radiation (Shimadzu XRD-6000). The membrane mass-transfer coefficient of the contactor, km, was calculated by Eq. (1) according to the absorption rate of pure CO2 with 100 mol·m−3 aqueous NaOH solution [8]. 1 ( pA − pA* ) 1 = − km N/A HEkL

(1)

where N/A is the absorption rate per interfacial area, H is the solubility coefficient of CO2 in NaOH solution, kL is the individual mass-transfer coefficients in the liquid film, E is the enhancement factor due to chemical reaction, pA is the CO2 pressure, and pA* is the CO2 pressure equilibrated with the bulk liquid. 2.4

Membrane cleaning and reuse

To remove the fouling materials on membrane inner surface, dilute hydrochloric acid was pumped to flow through the membranes lumens at flow velocity of 0.053 m·s−1 for 30 min after reaction. Then the membrane inner surface was kept contacting with the hydrochloric acid for 15 h to dissolve the residues. Lastly, the membrane was washed with distilled water to neutral. The mass-transfer coefficient of the recovered membrane was determined and the contactor was reused in the next reaction.

3.1

RESULTS AND DISCUSSION Effects of Ca(OH)2 concentration

To explore the influences of Ca(OH)2 concentrations, the experiments were conducted at 295 K, CO2 pressure of 101451 Pa, liquid flow rate of 0.033 m·s−1, Ca(OH)2 concentration of 2.7, 5.4, 8.1, 10.8 and 13.5 mol·m−3, respectively. With increasing Ca(OH)2 concentrations, the pH of inlet liquid increases from 11.67 to 12.38, and the pH of outlet liquid rises from 6.60 to 7.36, indicating that the reaction is completed even at the highest concentration. Meanwhile, the absorption rates of CO2 increase from 0.00048 to 0.0005, 0.00052, 0.00055 and 0.00057 m3 m−2·s−1, respectively [9]. The molar ratio of absorbed CO2 to Ca(OH)2 flowing though the lumens per unit time is described as: pA ( N / A)2nπrl 2lpA ( N / A) RT (2) k= = 2 RTruC nπr uC where n is the fibers number, r is the fiber radius, l is the fiber length, N/A is the absorption rate per interfacial area, T is reaction temperature, u is the liquid flow velocity, and C is the Ca(OH)2 concentration. With increasing Ca(OH)2 concentrations, k is found to decrease from 10 to 5.23, 3.62, 2.87 and 2.298 respectively, indicating that the CO32− concentrations in the liquid reduce correspondingly. Furthermore, the solubility product (Ksp, I = 0.1) of Ca(OH)2 is 10−4.9 [10], and the OH− concentrations rise from 10−2.33 mol·dm−3 to 10−1.62 mol·dm−3 for the three Ca(OH)2 concentrations. Thus, the equilibrium concentrations of Ca2+ decrease from 10−0.24 mol·dm−3 to 10−1.66 mol·dm−3 in the three Ca(OH)2 suspensions. The low Ca2+ and CO32− concentrations at high Ca(OH)2 concentration result in low supersaturation and nucleation rate. Therefore, with the rising Ca(OH)2 concentrations, the particles average size increases from 75 nm to 79 nm [Figs. 2 (a, b)]. The selected area electron diffraction [insert of Fig. 2 (b)] shows that the particles are polycrystalline

123


(a) C = 5.4 mol·m−3, pA = 101451 Pa, u = 0.033 m·s−1

(b) C = 13.5 mol·m−3, pA = 101451 Pa, u = 0.033 m·s−1

(c) C = 5.4 mol·m−3, pA = 101451 Pa, u = 0.013 m·s−1

Figure 2 Effects of parameters on particles morphology

and the diffraction rings coincide well with 104, 110, 202 and 300 planes of calcite, respectively. The powder X-ray diffraction (XRD) pattern (Fig. 3) is well consistent with calcite (PDF 05-0586, rhombohedral lattice). The above results confirm the generation of calcite particles.

To explore the mechanism of membrane fouling, the membrane samples were observed with SEM. Figs. 5 (a) and (b) show that at the entrance of membrane contactor, the fouling on the external surface is slight for Ca(OH)2 concentration of 2.7 mol·m−3 but becomes serious for concentration of 13.5 mol·m−3. The relative pressure of liquid phase is defined as,

p = pL − pG

(3)

where pL is the liquid pressure and pG is the gas phase pressure. When p exceeds the breakthrough pressure (Δpb) of membranes, the liquid leakage from the membrane pores occurs. Δpb is expressed as follows [11],

Δpb = −

Figure 3

Powder XRD pattern of CaCO3 nanoparticles

During the reaction, the liquid pressure at the entrance of membrane contactor rises gradually due to membrane fouling and fiber clogging, whereas the pressure at the exit keeps at zero. Thus, the pressure drop in the contactor increases with increasing reaction time (Fig. 4). Meanwhile, it is found that higher Ca(OH)2 concentration leads to faster increment of pressure drop mainly due to fiber clogging (Fig. 4).

(4)

where σ is the surface tension of liquid, θ is the contact angle, and dmax is the maximum diameter of membrane pores (dmax = 10 μm). The contact angle between polypropylene and water is about 117.7° [12], and the surface tension of water is 0.0723 N·m−1. Thus, the breakthrough pressure is found to be 13.4 kPa. According to the pressure drop curves (Fig. 4) and the gas pressure, it is found that for Ca(OH)2 concentrations of 5.4, 10.8 and 16.2 mol·m−3，the relative liquid pressure at entrance exceeds the breakthrough pressure after reaction time of 450 s, 240 s and 80 s, respectively. Thus, the external-surface fouling at entrance is serious for higher Ca(OH)2 concentration. Fig. 5 (c) shows that the external-surface fouling at exit is slight in spite of Ca(OH)2 concentration because the liquid pressure at exit nears zero. The inner-surface fouling at entrance is apparent due to adsorption of Ca(OH)2 particles [Fig. 5 (d)], but it is slight at exit where Ca(OH)2 has been consumed. Apparently, liquid leakage should be avoided in the membrane contactor. Narrow distribution of membrane pores size and low Ca(OH)2 concentration favor the process stability and inhibition of membrane fouling. 3.2

Figure 4 Effects of Ca(OH)2 concentration on pressure drop of membrane contactor C/mol·m−3: ■ 5.4; ○ 10.8; ▲ 16.2

4σ cos θ d max

Effects of liquid flow velocity

The effects of liquid flow velocity were studied

124


(a) External surface at entrance, C = 2.7 mol·m−3

(c) External surface at exit, C = 2.7 mol·m−3

(b) External surface at entrance, C = 13.5 mol·m−3

(d) Inner-surface at entrance, C = 5.4 mol·m−3

Figure 5 Effects of Ca(OH)2 concentration on membranes fouling

at 295 K, CO2 pressure of 101451 Pa, Ca(OH)2 concentration of 5.4 mol·m−3, and liquid flow velocity of 0.013, 0.033, 0.053, 0.066 and 0.080 m·s−1, respectively. With increasing flow velocity from 0.013 m·s−1 to 0.080 m·s−1, the pH of outlet liquid rises from 7.19 to 7.76, indicating that the reaction is completed. Meanwhile, the absorption rate increases from 0.0004 to 0.0005, 0.00058, 0.0006 and 0.00063 m3·m−2·s−1, respectively [9]. k is found to decrease from 10.6 to 5.22, 3.90, 3.13 and 2.71, respectively, which leads to the reduced CO32− concentrations and nucleation rate. On the other hand, at high flow velocity, the reaction time and nucleus growth time are short. As the effects of nucleation rate and growth time on particles size are opposite, the particles are not sensitive to the liquid flow velocity [Figs. 2 (a, c)]. The initial pressure drop of membrane contactor increases with increasing flow velocity [Fig. 6 (a)]. For laminar flow in membrane fibers (Re = 5.2-31.8), the pressure drop, Δp, is expressed by the HagenPoiseuille equation [13], 32 μ ul (5) Δp = d2 where d is the fiber inner diameter, l is the fiber length, μ is the liquid viscosity and u is the average flow velocity. Therefore, the initial pressure drop increases linearly with the flow velocity. Furthermore, with the rising flow velocity, the productivity of the continuous contactor increases linearly, leading to the rise of membrane fouling degree and then the significant increment of pressure drop [Fig. 6 (b)]. This is really different from the general membrane separation process where high velocity often leads to removal of the fouling on the membranes.

(a) Initial pressure drop

(b) Evolution of pressure drop u/m·s−1: ■ 0.013; ○ 0.033; ▲ 0.066 Figure 6 Effects of liquid flow velocity on pressure drop at Ca(OH)2 concentration of 5.4 mol·m−3

According to the pressure drop curves [Fig. 6 (b)] and gas pressure, it can be seen that for liquid flow rates of 0.013, 0.033 and 0.066 m·s−1, the relative pressure of liquid exceeds the breakthrough pressure


(a) u = 0.013 m·s−1

(b) u = 0.080 m·s−1

(c) pA = 101397 Pa

(d) pA = 102785 Pa

125

Figure 7 Parameters effects on membrane fouling of external surface at entrance

after reaction time of 1250, 400 and 0 s, respectively. Therefore, the fouling on the membrane external surface at entrance is serious for high flow velocity [Figs. 7 (a, b)]. Furthermore, the fouling on the external surface at exit is slight and that on the inner-surface at entrance is serious. It was found that, for low Ca(OH)2 concentration and low flow rate, the experiment can last at least 1 h without serious membrane fouling. 3.3

Effects of gas pressure

The effects of CO2 pressure were studied at 293 K, liquid flow velocity of 0.033 m·s−1, Ca(OH)2 concentration of 10.8 mol·m−3 ， CO2 pressure of 100851, 101322, 101645, 101843 and 102785 Pa，respectively. With increasing CO2 pressure from 100851 Pa to102785 Pa, the pH of outlet liquid reduces from 7.16 to 6.86, and the absorption rate increases from 0.0005 m3·m−2·s−1 to 0.00055 m3·m−2·s−1 [9]. According to Henry’s law, the interfacial CO2 concentration in the liquid at equilibrium, CA, is expressed as CA = pAH. Herein, H = 0.00036 mol·m−3·Pa−1 [9]. When pA increases from 100851 Pa to 102785 Pa, CA rises from 36.31 to 37.00 mol·m−3. It can be seen that the changes of absorption rates and CO2 concentrations at intersurface are not large. Thus, the influence of CO2 pressure on particles is negligible, and the as-obtained particles are about 80 nm in size. With increasing CO2 pressure, the relative pressure of liquid phase reduces and liquid leakage is inhibited. Therefore, the degree of membranes fouling on the external surface at entrance decreases [Figs. 7 (c, d)].

3.4

Recovery and reuse of membranes

In the preparation of nanoparticles, membrane fouling is caused by the absorption and clogging of Ca(OH)2 and CaCO3. After reaction, the membrane was washed with dilute hydrochloric acid, and then the mass-transfer coefficient of the recovered membrane, km, was measured. Fig. 8 shows the membrane mass-transfer coefficients versus the reuse times. It illustrates that km can be recovered and the membranes can be reused for at least 6 times without performance deterioration. The foulants on membrane external surface can be removed by dilute hydrochloric acid.

Figure 8 Membrane mass-transfer coefficients versus reuse times ▼ membrane 1; ○ membrane 2

4

CONCLUSIONS

Polycrystalline CaCO3 particles are prepared with

126


a continuous gas-liquid membrane contactor. With increasing Ca(OH)2 concentrations, the particles average size increases. The effects of Ca(OH)2 concentration and CO2 pressure on particles are negligible. At high Ca(OH)2 concentration, high liquid flow velocity or low CO2 pressure, the relative pressure of liquid phase may exceed the breakthrough pressure of membranes, resulting in liquid leakage especially at the contactor entrance. The fouling on the inner-surface at entrance is apparent due to absorption and clogging of reactants. Low reactant concentration, low flow velocity and high CO2 pressure favor the process stability and inhibition of membrane fouling. The membranes can be recovered by cleaning with dilute hydrochloric acid and reused for at least 6 times without apparent performance loss, indicating that the membrane reactor is convenient and prospective in the preparation of nanoparticles. REFERENCES

4

5

6

7

8

9

10 11

1

2

3

Wang, M.J., Sung, H.K., Kim, W., Chang, K.C., “Particle morphology of calcium carbonate precipitated by gas-liquid reaction in a Couette-Taylor reactor”, Chem. Eng. Sci., 55, 733-747 (2000). Gomez-Diaz, D., Navaza, J.M., Sanjurjo, B., “Analysis of mass transfer in the precipitation process of calcium carbonate using a gas/liquid reaction”, Chem. Eng. J., 116, 203-209 (2006). Kim, C.H., Ri, C.S., Lee, D.H., Choo, H.S., Lee, B.K., Chung, J.J.,

12

13

Kim, S.D., “Effect of pH and concentration on the synthesis of monodispersed spherical fine zirconia powders using gas-liquid phase reaction”, In: Symposium on Nanostructured Powders and Their Industrial Applications, San Francisco, California (1998). Wu, P.W., Gao, L., “Synthesis of CdS nanoparticles in reverse micelles by gas-liquid reaction technique”, J. Inorg. Mater., 18, 937-941 (2003). Narayanan, R., El-Sayed, M.A., “Changing catalytic activity during colloidal platinum nanocatalysis due to shape changes: electron-transfer reaction”, J. Am. Chem. Soc., 126, 7194-7195 (2004). Lia, J.L., Chen, B.H., “Review of CO2 absorption using chemical solvents in hollow fiber membrane contactors”, Sep. Purif. Tech., 41, 109-122 (2005). Wang, K., Wang, Y.J., Chen, G.G., Luo, G.S., Wang, J.D., “Enhancement of mixing and mass transfer performance with a microstructure minireactor for controllable preparation of nanoparticles”, Ind. Eng. Chem. Res., 46, 6092-6098 (2007). Jia, Z.Q., Chang, Q., Qin, J., Hong, X.M., “Preparation of nanoparticles with a gas-liquid membrane contactor”, J. Membr. Sci., 342, 1-5 (2009). Jia, Z.Q., Chang, Q., Qin, J., Sun, H.J., “Preparation of nanoparticles with a continuous gas-liquid membrane contactor: absorption process”, J. Membr. Sci., 352, 50-54 (2010). Peng, C.H., Feng, J.Z., Zhang, X.Y., Quantitative Analytical Chemistry, Beijing University Press, Beijing, 339-340 (1984). (in Chinese) Gabelman, A., Hwang, S.T., “Hollow fiber membrane contactors”, J. Membr. Sci., 159, 61-106 (1999). Dindore, V.Y., Brilman, D.W.F., Geuzebroek, F.H., Versteeg, G.F., “Membrane-solvent selection for CO2 removal using membrane gas-liquid contactors”, Sep. Purif. Technol., 40, 133-145 (2004). Bird, R.B., Stewart, W.E., Lightfoot, E.N., Thansport Phenomena, John Wiley & Sons, Inc., New York, 69-70 (1960).