Effects of CO₂ Bubble Size, CO₂ Flow Rate and Calcium Source on the Size and Specific Surface Area of CaCO₃ Particles

Abstract

The effects of CO₂ flow rate and calcium source on the particle size (PS) and specific surface area (SSA) of CaCO₃ particles were evaluated using microbubble generator (MBG) and air diffuser (AD) systems. The carbonate mineralization (CM) and precipitated calcium carbonate (PCC) methods were employed to produce CaCO₃ using gypsum and Ca(OH)₂ as the calcium sources, respectively. The CaCO₃ particles prepared using the MBG were smaller (with larger specific surface area) than those obtained using the conventional AD, regardless of the calcium source. The average PSs were 2–3 and 7–9 μm for the MBG and AD systems, respectively. Moreover, the PS and SSA of the particles prepared using the MBG were not greatly affected by the CO₂ injection rate. This study clearly demonstrates that the use of an MBG ensures the stable production of fine CaCO₃ particles using various calcium sources and a wider range of CO₂ flow rates.

1. Introduction

Microbubbles (MBs) are very tiny bubbles in liquids, with diameters of several tens of micrometers [1]. One typical use of MBs is water purification, achieved by increasing the dissolved oxygen concentration in polluted water. As an extension of the use of MBs to a wider range of applications, the recovery of fine particles [2] and the precipitation of CaCO₃ using Ca(OH)₂ [3] have been reported. This application is apparently related to a physical property of the MBs not found in ordinary bubbles in liquids, namely that the concentration of a specific gas in a liquid remains higher for longer if the gas is contained in MBs.

Equation (1) explains the stability of the size of MBs:

$$\text{σ }={\left(\frac{\partial G}{\partial A}\right)}_{T, P}$$

(1)

where, σ, G, and A are the surface tension, Gibbs free energy, and surface area, respectively, of a CO₂ bubble:

$$J_{{\text{CO}}_{2}Z}=-D_{{\text{CO}}_{2}A}\frac{d{C}_{{\text{CO}}_2}}{d_z}$$

(2)

where, $J_{{\text{CO}}_{2}Z}$ is the concentration rate of CO₂ diffused in the z direction (mol/m²·s), $D_{{\text{CO}}_{2}A}$ is the diffusion coefficient of CO₂ through the suspension (m²/s), $d{C}_{{\text{CO}}_2}$ is the CO₂ concentration difference (mol/m³), and $d_z$ is the length difference in the z direction (m).

The surface tension (σ) increases as the surface area (A) of the bubble decreases, if the energy (G) in the two states is identical. Additionally, decreasing the surface area means decreasing the volume and floatation rate of the bubble. Therefore, MBs can remain suspended for longer than ordinary bubbles. The increased concentration of a gas in an aqueous phase enhances the diffusion rate. Fick’s first law of diffusion, presented in Equation (2), shows that the diffusion rate of CO₂ is enhanced in the z direction, if the concentration of CO₂ increases. The impact of these qualities of MBs in increasing the concentration of a gas is different from that of ordinary bubbles (as is evident on breathing into water through a straw). We have reported that a microbubble generator (MBG) precipitated smaller calcium carbonate particles faster than an ordinary air diffuser with 3 cm diameter did [3]. Feng et al. [4] also obtained smaller CaCO₃ particles by reducing the frit pore size used for CO₂ bubbling in a Ca(OH)₂ suspension: The carbonate mineralization (CM) method is used to fix CO₂ as a carbonate mineral (i.e., CaCO₃). With this method, carbonates are formed as precipitates from the chemical reaction between CO₂ and alkali earth metals such as calcium [5,6,7]. This method originated from a process for producing precipitated CaCO₃ [8]. Although this chemical reaction is simple and well known, as pointed out by Haywood et al. [9], some consideration of the reaction rate and efficiency of metal extraction is needed to realize a successful CO₂ fixation method. Ca(OH)₂ or CaO (the dehydrated form of Ca(OH)₂) is the calcined form of CaCO₃ (formed by dissociation of CO₂ when heated above 900 K). Therefore, for the purpose of CO₂ fixation, Ca(OH)₂ is not a good choice as a calcium source. Naturally obtainable gypsum is a better calcium-supply material for CM than Ca(OH)₂ because gypsum does not emit CO₂ as it does not involve thermal decomposition. Previously, we examined CM using CaSO₄·2H₂O (gypsum) as the calcium source with a NaOH solution, and using CO₂ MBs. Our results confirmed the enhanced efficiency of the reaction with this calcium source [10]. The crystal phases of gypsum, calcite, and portlandite were affected by the experimental conditions, such as differences in the concentration ratio of OH⁻ to Ca²⁺. Herein, the particle size (PS) and specific surface area (SSA) of CaCO₃ (calcite) produced by CM from gypsum are compared with those of particles precipitated using the precipitated calcium carbonate (PCC) method with Ca(OH)₂ as the calcium source. Additionally, the effect of CO₂ bubble size of larger bubbles from an air diffuser (AD) and tiny bubbles from MBG on the PS and SSA was estimated by the results from the experiments and the reference [4]. The PS and SSA results presented here for the PCC method are from our previous research [11]. The results from the current study should be useful for designing a CM process to achieve two aims: effective CO₂ fixation, and production of CaCO₃ with properties useful for industrial use.

2. Materials and Methods

2.1. Raw Materials

Gypsum (96%, Junsei Chemicals, Tokyo, Japan) was selected as the calcium source for CM (concentration of 0.3 M). NaOH (97%, Samchun Chemicals, Seoul, Korea) was dissolved in water to prepare a 6 M stock solution, 100 mL of which was added to the gypsum suspension. Microbubbles of 99.9% CO₂ were injected from an MBG (at 0.4, 0.8 and 1.0 L/min) into the 1 L of NaOH-dosed gypsum suspension. A schematic of the MBG process was previously published [10]. All experiments were repeated at least twice.

For the PCC method, Ca(OH)₂ (96%, concentration 0.3 M, Junsei Chemicals) was selected as the calcium source of CaCO₃ precipitate. The CO₂ MBs were injected into the 1 L of the suspension at the same flow rates used for the CM method. A detailed experimental scheme for this can be found in another report [3].

Table 1 shows the specifications of the calcium supply materials, and Figure 1 indicates the 70,000× SEM images of the raw materials. All the solutions and suspensions were prepared using deionized water (Milli-Q Advantage A10, Millipore, Billerica, MA, USA).

Table 1. Specifications, particle size (PS), and specific surface area (SSA) of the raw materials.

Reagent	Assay (%)	PS (μm)	SSA (m2/g)	Concentration (M)	Volume (L)
Gypsum	96	42.9	9.9	0.3	1
Ca(OH)2	96	8 *	-	0.3	1

* The PS results for Ca(OH)₂ are from our previous paper [11].

Table 1. Specifications, particle size (PS), and specific surface area (SSA) of the raw materials.

Reagent	Assay (%)	PS (μm)	SSA (m2/g)	Concentration (M)	Volume (L)
Gypsum	96	42.9	9.9	0.3	1
Ca(OH)2	96	8 *	-	0.3	1

* The PS results for Ca(OH)₂ are from our previous paper [11].

Figure 1. SEM images of the calcium source materials: (a) Gypsum and (b) Ca(OH)₂.

Figure 1. SEM images of the calcium source materials: (a) Gypsum and (b) Ca(OH)₂.

2.2. Characterization

Circulation of the suspension, and injection of CO₂, were stopped when the pH (measured every 30 s, Orion 3 Star, Thermo Scientific, Waltham, MA, USA) of the suspension did not vary, and was ~7–8. The particles were collected using a nylon membrane (0.2 μm pore size, Whatman, Brentford, UK), dried at 303 K for at least 48 h; then, the PS was measured (Mastersize 2000, Malvern Instruments, Malvern, UK). The SSA was measured using Tristar 3000 (Micromeritics, Norcross, GA, USA) and Quadrasorb SI 4S (Quantachrome Instruments, Boynton Beach, FL, USA) analyzers after pre-treatment at 523 K to remove all moisture from the pores. The nitrogen gas provided was assumed to be adsorbed in multiple layers on the surfaces of CaCO₃ particles. The images of the particles were obtained using a Field Emission Scanning Electron Microscope (FE-SEM) (S4700, Hitachi, Japan) located at Korea Basic Science Institute (KBSI) Jeonju Center. The PS and SSA of the particles presented here for the PCC method were from previous research [11].

2.3. Control Experiments

A conventionally used ceramic air diffuser (AD) was employed to inject CO₂ into the gypsum and Ca(OH)₂ suspension to compare the effect of CO₂ MB on the characteristics of CaCO₃ particles produced by both methods: CM and PCC. A 3-cm diameter AD designed for use in a fishbowl was placed in a 2000-mL beaker. A 0.3 M gypsum suspension was used as calcium supplement. For the MBG experiments, 100 mL of 6 M NaOH solution was added to the 1 L of gypsum suspension, and the CO₂ flow rate was 0.8 L/min. Magnetic stirring at 300 rpm was used to compensate for the weak mixing of the suspensions compared with that during the MBG process.

3. Results and Discussion

The average particle size of the raw gypsum was 42.9 μm (Table 1). The initial pH of the gypsum suspension without NaOH solution was around 8.5. While the gypsum suspension circulated with CO₂ MBs at 0.8 L/min through the MBG, the pH of the suspension decreased to ~7 and remained stable from the 4th minute of the process. After 16 min of circulation, the size of the gypsum particles was 25 µm. This reduction in size was not introduced by CM because NaOH was not added to the gypsum suspension; this was not a favorable condition for production of carbonate ions. The reduction in PS was due to frequent and random collisions between the suspended particles, CO₂ MBs, and parts of the MBG, owing to the turbulent flow resulting from the high flow rate [12]. The much smaller calcium-supply particles created by the MBs had much greater surface area and increased the reaction rates of both methods (CM and PCC). Ramachandran and Sharma [13] showed that, in theory, smaller particles should enhance chemical reactions between gas and liquid phases.

After 100 mL of NaOH solution was added, the gypsum suspension and CO₂ MBs were circulated through the MBG. The particles were collected when the suspension was around pH 7 and steady (after about 13 min circulation). XRD analysis revealed that all the collected particles were calcite [10]. The particle sizes obtained by CM with gypsum suspension were 2.4, 2.6 and 2.9 μm (black squares (□) in Figure 2), when the CO₂ injection rates were 0.4, 0.8 and 1.0 L/min, respectively. The average PS was 2.6 μm with a standard deviation of 0.3 μm.

Figure 2. Particle size (PS) of CaCO₃ produced by carbonate mineralization (CM) and precipitated calcium carbonate (PCC) from gypsum (squares □) and Ca(OH)₂ (circles ○). The effect of CO₂ injection methods (microbubble generator (MBG) and air diffuser (AD)) on the PS were marked additionally in black and red, respectively. The horizontal axis indicates the CO₂ injection rate into the calcium-source suspensions. The PS results using Ca(OH)₂ are from our previous paper [11]. The values in blue triangles are from reference [4] when 101–160 µm bubbles were used, and these are approximate values. The standard deviation was obtained from the repeated experimental results at least twice.

Figure 2. Particle size (PS) of CaCO₃ produced by carbonate mineralization (CM) and precipitated calcium carbonate (PCC) from gypsum (squares □) and Ca(OH)₂ (circles ○). The effect of CO₂ injection methods (microbubble generator (MBG) and air diffuser (AD)) on the PS were marked additionally in black and red, respectively. The horizontal axis indicates the CO₂ injection rate into the calcium-source suspensions. The PS results using Ca(OH)₂ are from our previous paper [11]. The values in blue triangles are from reference [4] when 101–160 µm bubbles were used, and these are approximate values. The standard deviation was obtained from the repeated experimental results at least twice.

For comparison, the size of the particles obtained by PCC with a suspension of Ca(OH)₂ were 2.1, 3.5 and 2.3 μm (black circles (○) in Figure 2) when the CO₂ injection rates were 0.4, 0.8 and 1.0 L/min, respectively. The average PS was 2.6 ± 0.8 μm [11]. It is interesting that the size of the CaCO₃ particles was almost the same regardless of the type of calcium source, when the MBG was used for either the CM or the PCC method.

However, the PS was not the same when an AD was used to inject CO₂ into the calcium-source suspension, for CM (9.2 ± 0.1 μm) or PCC (7.1 ± 0.6 μm) (red marks in Figure 2). Much smaller CaCO₃ particles were produced using the MBG than with the AD, regardless of the calcium source. Under the same CO₂ injection rate of 0.8 L/min into the calcium source suspension, tiny CO₂ bubbles by MBG produced the calcite particles sized in from 2.6 to 3.5 μm, and larger CO₂ bubbles by AD did the calcite particles sized in from 7.1 to 9.2 μm. MBG makes the CO₂ concentration higher even when the flow rate of CO₂ into the calcium-source suspension was low. For this reason, supersaturation of carbonate ions can be induced in the suspension, which will precipitate smaller CaCO₃ particles because of the higher nucleation rate [14]. It was observed that the PS of CaCO₃ was not sensitive to the CO₂ flow rate when a MBG was used. In contrast, when an AD or a gas distributor was used, the PS was significantly affected by the gas flow rate. Feng et al. [4] investigated the effect of CO₂ flow rate and bubble size on the PS of CaCO₃. A flow rate increment from 3.5 to 4 L/min, reduced the PS from 4 to 2.5 μm. For comparison, their results are also included as blue triangles (∆) in Figure 2.

Figure 3 shows SEM images of Figure 3a raw gypsum and Figure 3b gypsum circulated through the MBG for 16 min without a dose of NaOH. Evidently, the gypsum particles become significantly smaller during circulation. Therefore, the SSA of the particles was expected to increase owing to this refinement. The raw gypsum crystals were covered by numerous fine particles (Figure 3a), which greatly increased the surface area. Therefore, the surface area would be decreased if these surface particles were dissolved during circulation. The very flat surface of gypsum after circulation (Figure 3b) supports this explanation. The expected increase in the measured SSA was not observed. Instead, the SSA of the gypsum decreased from 9.9 to 6.1 m²/g during circulation, despite significant reduction in the PS. The reason for this discrepancy can be observed in the SEM images. The high flow rate provided by the MBG is expected to readily dissolve these tiny particles on the surface of gypsum crystals.

Figure 3. Scanning Electron Microscope (SEM) images of gypsum: (a) Raw gypsum with tiny particles (<2 µm) on the surface of larger particles; (b) Gypsum after circulation for 16 min through the MBG: The tiny particles disappeared and the large particles were broken into smaller ones.

Figure 3. Scanning Electron Microscope (SEM) images of gypsum: (a) Raw gypsum with tiny particles (<2 µm) on the surface of larger particles; (b) Gypsum after circulation for 16 min through the MBG: The tiny particles disappeared and the large particles were broken into smaller ones.

The SSAs of the CaCO₃ particles prepared by CM or PCC with CO₂ MBs were similar whether the calcium source was Ca(OH)₂ or gypsum. Figure 4 shows the effect of the CO₂ injection rate on the SSA for both the MBG and AD. The SSA of the particles from the gypsum suspension after the MBG process (black squares (□) in Figure 4) was 23.7, 22.0 and 17.5 m²/g at the CO₂ rates of 0.4, 0.8 and 1.0 L/min, respectively. The average value was 21.1 ± 3.20 m²/g: the standard deviation of 15% being rather high. In comparison, the SSA of particles produced by PCC with Ca(OH)₂ was 21.7, 22.2 and 22.4 m²/g at CO₂ rates of 0.4, 0.8 and 1.0 L/min, respectively (black circles (○) in Figure 4), with an average SSA of 22.1 ± 0.36 m²/g [11]. This result agrees with the PS data in the sense that the SSA values are almost constant when using the MBG, regardless of the calcium source. The SSA of particles prepared at a CO₂ flow rate of 0.8 L/min, using the AD (red marks in Figure 4), were significantly less than for those obtained using the MBG. In this case, the particles prepared by CM with gypsum had a SSA of 10.7 ± 3.3 m²/g, whereas those prepared with Ca(OH)₂ had a SSA of 17.2 ± 6.1 m²/g as AD supplied larger CO₂ into the calcium source suspension; MBG with tiny CO₂ bubbles produce the calcite particles with larger SSA (22.0–22.2 m²/g).

Figure 4. SSA of CaCO₃ produced by CM and PCC from gypsum (squares □) and Ca(OH)₂ (circles ○) respectively. The effect of the CO₂ injection methods (MBG and AD) on the SSA were marked in black and red, respectively. The horizontal axis indicates the CO₂ injection rate into the calcium-source suspensions. The SSA results using Ca(OH)₂ are from our previous paper [11]. The standard deviation was obtained from the repeated experimental results at least twice.

Figure 4. SSA of CaCO₃ produced by CM and PCC from gypsum (squares □) and Ca(OH)₂ (circles ○) respectively. The effect of the CO₂ injection methods (MBG and AD) on the SSA were marked in black and red, respectively. The horizontal axis indicates the CO₂ injection rate into the calcium-source suspensions. The SSA results using Ca(OH)₂ are from our previous paper [11]. The standard deviation was obtained from the repeated experimental results at least twice.

Figure 5 shows SEM images of the CaCO₃ particles obtained from the CM process (gypsum source) and PCC process (Ca(OH)₂ source) using either MBG or AD. The particles prepared using the MBG appear smaller than those prepared using AD, regardless of calcium source. In addition, the particles produced using the MBG had a narrower size distribution (Figure 5a,b), than those using AD (Figure 5c,d). This was probably because of comparatively larger clusters in the latter. The SSA is affected by the size of primary particles that are agglomerate to form larger (secondary) particles. The size of primary particles can be calculated by Equation (3) when the particles are granular [15]. Table 2 shows the results for primary particle size calculated using this equation. These values were in accordance with the SEM images in Figure 5 when the MBG was used. However, the PS of the primary CaCO₃ particles produced using the AD were not matched with the SEM results in Figure 5 due to the irregular shapes of the particles and their agglomerates. In this case, the contacted surfaces between the irregular shaped primary particles were larger than the granular particles in order to the SSA would be reduced:

$$d=6/{\left(S_{\text{g}}·{\text{ρ}}_{\text{g}}\right)}_{ }$$

(3)

where, d is primary particle size, $S_{\text{g}}$ is SSA, and ${\text{ρ}}_{\text{g}}$ is density of the particles (in this case 2.71 g/cm³ for calcite of CaCO₃).

The novelty of this study is that the size of the CaCO₃ particles prepared using the MBG were finer and almost identical in size (~2.6 μm and 22 m²/g) regardless of the calcium source, whereas the use of an AD produced coarser CaCO₃ particles (~8 μm and less than 17 m²/g). Generally, when such CaCO₃ particles are used in paper making, it is said that their finer PS and greater SSA enhance smoothness and printing quality. A rationale for the homogeneities in (primary) PS and SSA may be that the MBG causes changes in the fluid dynamics, such as the surface tension and buoyancy of bubbles in the calcium-source suspension, which are quite different from those obtained with the AD.

Figure 5. SEM images of CaCO₃ produced by AD and MBG using gypsum and Ca(OH)₂: (a) gypsum by MBG; (b) Ca(OH)₂ by MBG; (c) gypsum by AD and (d) Ca(OH)₂ by AD. The particles in this panel were obtained CO₂ flow at 0.8 L/min, with AD and MBG.

Figure 5. SEM images of CaCO₃ produced by AD and MBG using gypsum and Ca(OH)₂: (a) gypsum by MBG; (b) Ca(OH)₂ by MBG; (c) gypsum by AD and (d) Ca(OH)₂ by AD. The particles in this panel were obtained CO₂ flow at 0.8 L/min, with AD and MBG.

Table 2. Calculated size of the primary particles based on specific surface area.

CO2 Rate (L/min)	Size of the Primary Particles (nm)
	CM by MBG	PCC by MBG	CM by AD	PCC by AD
0.4	94	102	-	-
0.8	101	100	207	129
1.0	127	99	-	-

Table 2. Calculated size of the primary particles based on specific surface area.

CO2 Rate (L/min)	Size of the Primary Particles (nm)
	CM by MBG	PCC by MBG	CM by AD	PCC by AD
0.4	94	102	-	-
0.8	101	100	207	129
1.0	127	99	-	-

According to Henry’s law, the degree of solubility of a gas in aqueous media depends on the pressure and temperature. The addition of chemicals such as amines [16,17,18] and ammonia [19,20,21] could enhance the solubility of CO₂ by forming complexes. Moreover, changes in the physical properties of a gas bubble, such as decreasing the bubble size, could also increase their solubility under given conditions of atmospheric pressure and temperature [3,10,22]. The strong surface tension of MBs described in Equation (1) causes these results. This will improve the ability to build rather small, unpressurized CM facilities, to use less and thus reduce the cost of CO₂ sorbent chemicals, and to reduce the loss of whiteness of CaCO₃ particles due to those chemicals. The use of the MBG accelerates the nucleation rate of CaCO₃ due to the enhanced diffusion rate indicated in Equation (2), which decreases the PS. In addition, Gibbs-Thomson and Ostwald ripening effects exert an influence on the PS, with a limit for CaCO₃ near 2 μm [23]. This appears to agree with our result of PS around 2 μm, regardless of the calcium source, and of the process (CM or PCC), when the MBG was used.

4. Conclusions

The effect of the MBG on the PS and SSA of CaCO₃ particles prepared using the CM process from gypsum was evaluated. The properties of the particles obtained by the PCC process from Ca(OH)₂ were compared. The PS and SSA were almost identical regardless of calcium source, and CO₂ flow rates if MBG used. Moreover, the particles produced by MBG had smaller size (around 2 μm) and greater SSA (around 22 m²/g) than those obtained in the control experiments by AD (around 8 μm and less than 17 m²/g). This result was caused by innate qualities of the MBs: strong surface tension and enhanced diffusion rate. The results from this study are expected to help predict the qualities of the particles produced by CM from various calcium sources using MBG.