Introduction
The demand for lithium-ion batteries has grown with the popularity of electric cars and mobile devices [1]. The cathode of lithium-ion batteries uses lithium metal oxide, which is made from lithium compounds extracted from lithium deposits, such as high purity lithium hydroxide or lithium carbonate. Apart from brines, rock deposits are the most common form of lithium particularly in Australia, Canada, and China [2]. The conventional way to extract lithium from pegmatite containing spodumene [3] includes calcination at above 1050 °C to loosen the lattice to form b-spodumene, followed by leaching with concentrated sulfuric acid [4] to extract the lithium. The leaching process results in a lithium sulfate concentrate. This is then purified and reacted with sodium carbonate or calcium hydroxide to produce lithium carbonate [5] or lithium hydroxide [6]. A sulfate content below 0.01 wt-% must be obtained within the solid product, which corresponds to a sulfate concentration of less than 20 mg·L−1 in a saturated solution of 10.9 wt-% lithium hydroxide at 20 °C for battery grade lithium production [7].
There are a few existing methods, including membrane separation and chemical precipitation [8] for such lithium solution purification. Membrane separation uses charged membranes and Donnan exclusion theory to separate the differently charged ions [9], but is often prone to fouling and is incapable of producing ultra-high purity products [10]. Chemical precipitation normally uses barium ions to precipitate sulfate ions to form insoluble barium sulfate, which has a solubility equilibrium constant (Ksp) of 1.08 × 10−10 [11] with the following reaction [12].
This reaction makes chemical precipitation very selective and relatively inexpensive given the raw material used. However, barium sulfate can form very fine particles that often need to be separated using microfiltration [13]. Furthermore, the quantity of barium ions added needs to be stoichiometric, or the excess barium ions require an extra step to be removed, which substantially increases the operation complexity and operating costs.
This research focuses on the development of a polymeric matrix material with entrapped barium to selectively adsorb sulfate. Barium ions are encapsulated within microcapsules to resolve the aforementioned challenges in the conventional precipitation process while effectively removing the sulfate ions from the solution. Alginate was selected as the ligand material for the preparation of the polymeric microcapsules, due to its known ability to gel in the presence of multivalent cations such as Ba2+ and Ca2+. Conversely, the carboxyl groups on the alginate, with a single negative charge do not induce gelation with monovalent cations such as lithium [14]. Other advantages of such polymers include easy accessibility, low cost, biodegradability, and non-toxicity [15,16].
In this study, we aim to develop a novel barium@alginate adsorbent for sulfate removal in lithium refining. The research goal is to reduce the sulfate concentration to less than 20 mg·L−1 in the solution. A range of microcapsules prepared under different reaction conditions were studied in terms of Ba2+ loading, morphology, integrity, effectiveness and kinetics of sulfate removal. The findings of this work will contribute to future development of more efficient and environmentally friendly separation processes for lithium refinery.
Experimental
Chemicals
Ba(OH)2·8H2O (98.0 wt-%), Li2SO4·H2O (99.0 wt-%) and alginic acid sodium salt were purchased from Sigma-Aldrich. CaCl2·2H2O (99.0 wt-%) and BaCl2·2H2O (99.0 wt-%) were purchased from Chem-supply.
Synthesis of microcapsules
Three methods were developed to produce the barium@alginate microcapsules. The first method uses pure barium hydroxide as the gelation reagent to produce the Ba@alginate microcapsules as shown in Fig. 1(a). In a typical procedure, 0.5 g of alginic acid sodium salt was dissolved into 10 mL of deionised water; next, 0.5 g of Ba(OH)2·8H2O was dissolved in 50 mL of deionised water; then, the alginate was added to the barium hydroxide solution at room temperature drop by drop, using a syringe with gauge of 21G × 1.5′′ from a height of at least 15 cm, to allow for spherical gel beads to form. The alginate microcapsules were left to stand in the barium solution for 60 min and were then collected, rinsed with deionised water and dried by patting with a paper towel. The microencapsulation process was repeated with 2, 3, 4, 5, and 6 wt-% alginate solutions to study the impact of alginate concentration.
Fig.1 Methods for preparing Ba@alginate microcapsules: (a) Method 1, pure barium hydroxide Ba@alginate microcapsules; (b) Method 2, two-step Ca/Ba@alginate microcapsules, where the polymerisation reaction first took place by using calcium as the firming agent, followed by barium; (c) Method 3, one step dual-ion Ca-Ba@alginate microcapsules, where both calcium and barium ions were used for the polymerisation reaction at the same time. |
The second method attempted to produce Ca/Ba@alginate microcapsules in two gelation steps as shown in Fig. 1(b). The alginate solution was injected dropwise into 50 mL of a 0.1 mol∙L–1 of CaCl2 solution, and crosslinked for 5 min to form calcium firmed microcapsules. The microcapsules were transferred onto a sieve and rinsed with deionised water immediately to terminate the crosslinking process. Next, the microcapsules were immersed into 50 mL of 0.1 mol∙L–1 barium chloride solution and further crosslinked with Ba2+ for an hour. Finally, the microcapsules were collected, rinsed with deionised water and dried with a paper towel.
In the third method two firming agents, namely Ca2+ and Ba2+, were involved at the same time for the preparation of the one step dual-ion Ca-Ba@alginate microcapsules as shown in Fig. 1(c). First, 50 mL of 0.1 mol∙L–1 barium chloride solution and 50 mL of 0.1 mol∙L–1 calcium chloride solution were well mixed at room temperature. The alginate solution was injected dropwise into the mixed-cation solution using a syringe with gauge of 21G × 1.5′′, and crosslinked for an hour. The microcapsules were collected, rinsed and dried as before.
Characterisation of microcapsules
The shapes of the microcapsules were visually observed. Their sizes were measured using a ruler and the average of the measurements for four microcapsules was reported.
A physical strength test was carried out by using an electronic balance. Increasing pressure was applied to the microcapsule against the balance under a glass plate until it ruptured, while the force was recorded. This process was repeated three times for the same type of microcapsules and the average value was calculated.
The structural integrity of the Ba@alginate microcapsules was tested with excess sulfate. Typically, ten microcapsules from the same batch were soaked in 10 mL of 800 mg·L−1 lithium sulfate solution for 4 h, and the broken microcapsules were visually observed. The integrity was calculated as the ratio of the number of microcapsules that survived to that of microcapsules tested.
Effectiveness of sulfate removal
The effectiveness of Ba@alginate microcapsules for sulfate removal was tested as follows. In a typical scenario, 0.2 g of microcapsules were immersed in 10 mL of 100 mg·L−1 lithium sulfate solution in an enclosed container that was left to stand for 4 h. The sulfate concentration in the solution after adsorption was tested to calculate sulfate removal efficiency using Eq. (2).
where C0 and Ce (mg·L−1) is the concentration of sulfate in the solution before and after adsorption, respectively.
Kinetics of sulfate adsorption
The kinetics of sulfate removal by adsorption was determined by submerging 0.2 g of microcapsules in 10 mL of 100 mg·L−1 lithium sulfate solution in an enclosed container for periods of 5, 10, 20, 30 min, and 1, 2, 3, 4, 5 h. The concentration of the solution after adsorption was tested. The amount of sulfate (Q) removed at time t, was calculated by the following equation:
where C0 (mg·L−1) is the initial concentration of the sulfate solution, Ct (mg·L−1) is the sulfate concentration at time t (s), m (g) is the mass of the microcapsules submerged, and V (L) is the volume of the sulfate solution.
Analytical methods
The sulfate solution after adsorption was filtered with 0.45 µm membrane filters, and tested with ion chromatography (Thermo Scientific Dionex ICS 1100 and Autosampler AS40). Scanning electron microscopy (SEM, Quanta) with an energy dispersive X-ray spectrometer (EDS) was conducted before and after sulfate adsorption on the microcapsules to observe the morphology as well as the distribution of the adsorption sites inside the microcapsules. The SEM samples were prepared by drying the microcapsules in a vacuum oven at 80 °C overnight, cutting into half, and mounting with carbon tape. An accelerating voltage of 10.00 kV was used by the SEM/EDS. with results smoothed with a Savitzky-Golay filter by a 2-point quadratic polynomial and 10 points of window.
Results and discussion
Alginate can undergo gelation reactions with multivalent cations such as Ca2+ and Ba2+, forming ionic bonds with the carboxyl groups present [14], as shown in Scheme 1.
Calcium is selected for the reinforcement of the Ba@alginate microcapsules, since calcium ions are able to crosslink with alginate to form a stronger calcium@alginate hydrogel, which is non-toxic, low in cost and is widely used [17]. These properties greatly increase the versatility of the microcapsules.
Morphology and optimisation
The Ba@alginate microcapsules prepared with different weight percentages of alginate are shown in Fig. 2. The optimum concentration of alginate was selected by balancing the shape of the product microcapsules and the amount of alginate used. More uniform spheres enable adsorption to take place more evenly and offer better packing inside the adsorption column. Moreover, a higher ratio of alginate results in more carboxyl functional groups (R-COOH) available, leading to higher loading of barium and thus greater sulfate adsorption capacity, which is a desirable characteristic.
Fig.3 Optical images of Ba@alginate microcapsules prepared using three methods at different alginate concentrations: (a-e) Ba@alginate microcapsules produced from 2, 3, 4, 5, and 6 wt-% alginate, respectively; (f) Ca/Ba@alginate from 5 wt-% alginate; (g) Ca-Ba@alginate from 5 wt-% alginate. Illustrations of the alginate solution dropwise injection process (h) under the desirable range of viscosity and (i) under high viscosity. |
The 2 wt-% Ba@alginate was not able to form spherical microcapsules. The low viscosity of the alginate solution meant that the droplets were unable to keep their sphericity from the minimum height (at least 15 cm) [18] as they hit the surface of the solution. Instead, the droplets flattened as they hit the liquid surface. In order to avoid this deformation, we decreased the dropping height to as low as 1 cm above the liquid surface. However, at these shorter distances, teardrop shaped microcapsules were formed as shown in Fig. 2(a), which is not suitable for packed columns.
Similarly, we were unable to obtain spherical microcapsules using 6 wt-% alginate (Fig. 2(e)). The alginate liquid transforms from a teardrop shape to an egg shape, and then a spherical shape (Fig. 2(h)). However, due to the high viscosity, the surface tension force is reduced, which fails to compete against the air drag force pushing up against the bottom of the droplet [19]. Hence the droplet was deformed to an oval shape as illustrated in Fig. 2(i).
As shown in Figs. 2(b)–2(d), 3, 4 and 5 wt-% alginates are able to form spherical microcapsules. The sizes of the microcapsules are very similar, with diameters around 3 mm. Five wt-% alginate was selected for the synthesis of the microcapsules, as it provides a greater loading of barium in the microcapsules and thus a more compact packing in a column.
Since the preparation step for the alginate solution was the same for all production methods of Ba@alginate microcapsules, 5 wt-% alginate was also selected for synthesis of Ca/Ba@alginate and Ca-Ba@alginate microcapsules. The optical images of the microcapsules prepared as Ca-Ba@alginate are shown in Figs. 2(f) and 2(g), from which we can see the microcapsules are of high sphericity with diameters also around 3 mm, in a semi-transparent white colour similar to Ba@alginate microcapsules.
Structure integrity of the microcapsules
As shown in Fig. 3(a) Inset, it was found that the microcapsules tended to rupture into flakes in excessively high concentration of lithium sulfate solutions (i.e., 800 mg·L−1). Figure 3(a) illustrates a positive relationship between the mass fraction of alginate and the microcapsule integrity. As the alginate percentage increases, more microcapsules survive in the solution within a sulfate concentration of 800 mg·L−1. It is well known that sulfate reacts with barium cations to form insoluble precipitates [12]. The excessive sulfate depletes the Ba2+ from the Ba@alginate crosslinks, disrupting the gel and leading to the rupture of the hydrogel microcapsules. Since the 5 wt-% alginate was able to provide the highest integrity of 90% in an excess amount of sulfate, it was selected for the preparation of the microcapsules in the subsequent tests. Note that the integrity needs to be further increased for applications under extreme conditions to prevent contamination in the Li+ solution.
Fig.4 (a) The relationship between the integrity and the concentration of alginate in an 800 mg·L−1 lithium sulfate solution (red line: a line of best fit; inset: optical image of Ba@alginate flakes, scale: mm); (b) the physical strength of calcium@alginate microcapsules under 0.1 and 0.05 mol·L–1 Ca2+ and 60, 3000, 1200 s crosslink time. |
It is well documented that calcium also crosslinks with alginate [20], and it is readily accessible and low in price [17]. Calcium@alginate microcapsules prepared from the same method as the pure barium hydroxide Ba@alginate microcapsules were able to achieve an integrity rate of 100%, higher than the Ba@alginate counterpart. This is because CaSO4 is slightly soluble in water with a solubility constant Ksp of 1.9 × 10−4, whereas BaSO4 is insoluble with a Ksp of 6 orders smaller in magnitude [11]. Thus, Ca2+ is not as easily drained from the Ca@alginate gel in the presence of excessive sulfate. The greater stability of the calcium gel is also reflected in the use of CaSO4 as a firming agent for alginate based gelation reactions in tissue engineering [21]. Therefore, calcium was selected as another crosslinking agent for the reinforcement of Ba@alginate in the two additional methods to produce Ca/Ba@alginate and Ca-Ba@alginate.
To prepare the Ca/Ba@alginate microcapsules, the concentration of the Ca2+ solution and the crosslinking time were optimised using the physical strength test. As shown in Fig. 3(b), microcapsules produced from 0.1 mol∙L–1 Ca2+ and 5 min gelation time was able to withstand 2 N of force before crushing, implying the microcapsules can stack within a packed column for approximately 1 m. For 0.05 mol∙L–1 Ca2+, it takes as long as 20 min to obtain a similar strength. The reaction was too fast to be controlled in solutions of Ca2+ concentrations higher than 0.1 mol∙L–1. It is undesirable if the alginate inside a microcapsule is fully crosslinked by Ca2+, as it loses the capacity for barium crosslinking.
Distribution of SO42− adsorption sites on microcapsules
To investigate the distribution of the sulfate adsorbed by the Ba@alginate microcapsules, SEM/EDS was conducted to locate the key elements inside the microcapsules. The content of Ba and S in the Ba@alginate microcapsules before and after sulfate adsorption is presented in Figs. 4(a)–4(d), respectively. The SEM/EDS analyses require the sample to be completely dried, the microcapsules deformed after dehydration (Figs. 4(b) and 4(d)). The barium contents in both samples are similar, around 5%-10% along the chord of the cross-section under the EDS scan. However, the distribution of barium along the line is not even. This is very likely due to the geometric differences in height, when the Ba@alginate gel is closer to the electron receptor more signals were detected, resulting in a higher spectrum. Another reason could be the shrinking core reaction process [22] leading to higher barium in the outer layer than that in the core of the microcapsules. Before adsorption, the background sulfate concentration is very low, less than half an atomic percentage, as shown in Fig. 4(a). After adsorption, the sulfur content inside the microcapsule peaks at 3% (Fig. 4(c)), which is significantly higher than before. This observation clearly shows that the Ba@alginate microcapsules served as the sulfate adsorbent and successfully entrapped barium sulfate.
Effectiveness of sulfate removal
As discussed in section 3.3, the capsules contain 5%-10% barium, 0.2 g of microcapsules equal to 7.3 × 10−5- 1.5 × 10−4 mol of barium which is well in excess of the 10 mL of 100 mg·L−1 lithium sulfate present, which equals to 1.0 × 10−5 mol of sulfate ions. The efficiencies of sulfate removal and the resulting sulfate concentration achieved by Ba@alginate microcapsules are shown in Fig. 5.
The Ca/Ba@alginate microcapsules prepared by the two-step gelation method obtained the highest percentage of sulfate removal (99.9%), resulting in a sulfate concentration of 0.2 mg·L−1 in the lithium sulfate solution. This is significantly below the 20 mg·L−1 target. Furthermore, the Ca-Ba@alginate microcapsules achieved 9.4 mg·L−1 sulfate concentration for the treated solution after adsorption, which is still well below the 20 mg·L−1 allowable value. Finally, the Ba@alginate microcapsules achieved 16 mg·L−1 sulfate concentration after adsorption, which is sufficient for the target concentration, with 84% of the sulfate removed from the solution. During the production of battery grade lithium, sulfate is the major impurity that needs to be thoroughly removed. Other undesired species are very minor can include Cl−, Mg2+, Ca2+, Fe3+, K+, Na+ [23]. However, they have no interference with the sulfate adsorption process and are normally removed by the re-crystallization process.
The addition of Ca2+ resulted in greater sulfate removal effectiveness in both types of the reinforced microcapsules compared to the pure barium hydroxide Ba@alginate microcapsules. This is because the presence of barium in the gelling solution results in a lower permeability for the transfer of external ions through the microcapsules [24]. In this case, the Ba@alginate microcapsules were prepared from a pure barium hydroxide gelling bath, leading to the lowest permeability. Hence it is more difficult for the SO42− ions to penetrate through the alginate gel and react with the Ba2+ ions.
On the other hand, the Ca/Ba@alginate microcapsules prepared by the two-step gelation method have the smallest amount of barium compared with other samples, as the calcium crosslinking reaction takes place before barium. In this method, intermediate Ca@alginate microcapsules were fabricated first, and a larger amount of calcium (compared with the Ca-Ba@alginate prepared by the dual-ion method) reacted with alginate to give the desired physical strength for the intermediate microcapsules. In contrast, for Ca-Ba@alginate microcapsules in which the Ca2+ and Ba2+ undergo gelation reaction at the same time, the Ba2+ ions have a higher affinity compared to Ca2+, and thus a greater tendency to react with the free carboxyl functional groups. Therefore, the ion permeability of Ca/Ba@alginate is greater than the Ca-Ba@alginate, resulting in a higher sulfate removal efficiency. This result suggests that other multivalent cations can also be considered to reinforce the Ba@alginate microcapsules which can be explored in future work.
Kinetics analysis
The kinetics of the sulfate removal process for the Ba@alginate microcapsules are shown in Fig. 6. It can be seen that the initial adsorption of sulfate Q increases quickly during the first 30 min, reaching 85% of the maximum capacity, then the adsorption reactions slow down and plateau to an equilibrium value at about 1 h. With the maximum capacity of over 1.5 mg·g−1, the Ba@alginate microcapsules showed a much better adsorption performance than other potential adsorbents, such as the polysryrene divinylbenzene (PS-DVB) copolymer beads with the sulfate adsorption capacity of 0.318 mg·g−1 [25], which is about 5 times lower. In optimum condition, the sulfate removal efficiency from saline water of PS-DVB was 65.64%, whereas our Ba@alginate microcapsules show removal efficiency of 84%, and our other adsorbent Ca/Ba@alginate microcapsules show removal efficiency of over 99% (Fig. 5(a)). The amount of sulfate removed by the Ba@alginate microcapsules decreases from the initial maximum capacity of 2 mg·g−1 to a new equilibrium value of 1.5 mg·g−1 after 1 h of adsorption. This unexpected result may be because a small fraction of the sulfate that was physically adsorbed initially in the microcapsules was released back to the solution as the Ba@alginate gel de-crosslinked due to the competitive adsorption of sulfate. The data points after the 1 h mark for the Ba@alginate microcapsules were not included in the fitting of kinetic models.
The time dependent removal rate of sulfate per gram of microcapsules over time in Fig. 6 was fitted with the pseudo second order (Fig. 6(a)), pseudo first order (Fig. 6(b)), and Elovich models (Fig. 6(c)) [26] using Eqs. (4), (5) and (6), respectively:
where k1 (s−1) and k2 (g·mg−1·s−1) are adsorption rates for pseudo first and second order, respectively; Q is the amount of sulfate removed at time t (s); Qe (mg·g−1) is the maximum amount of sulfate adsorbed by the microcapsules at equilibrium. For the Elovich model, the adsorption rate is described by kE = a∙exp(‒bQ).
Among the three models, the pseudo second order model gives the best fit with the experimental data for all three types of microcapsules. As summarised in Table 1, the correlation coefficient, R2, is 0.990, 0.991, and 0.984 for Ba@alginate, Ca/Ba@alginate, and Ca-Ba@alginate microcapsules, respectively. For the pseudo second order model, the maximum capacity of sulfate at equilibrium for the two-step Ca/Ba@alginate is the highest at 2.76 mg·g−1. The equilibrium adsorption capacity of sulfate for pure barium hydroxide Ba@alginate and the one step dual-ion Ca-Ba@alginate microcapsules are similar at 2.10 mg·g−1 and 2.18 mg·g−1, accordingly, with Ca-Ba@alginate slightly higher. The adsorption rate (k2) for the two-step Ca/Ba@alginate microcapsules is 1.71 × 10−3 g·mg−1·s−1 being the highest of the three types which implies that the Ca/Ba@alginate microcapsule adsorb sulfate in a solution fastest.
Tab.1 Kinetic parameters of second order pseudo, first order pseudo, and Elovich for Ba@alginate microcapsules in 100 mg·L−1 lithium sulfate solution |
| Parameters | Secondorder pseudo | Firstorder pseudo | Elovich | ||||||
|---|---|---|---|---|---|---|---|---|---|
| Qe/(mg·g−1) | k2× 10−3/(g·mg−1·s−1) | R2 | Qe /(mg·g−1) | k1× 10−3/s−1 | R2 | α/s−1 | β | R2 | |
| Ca/Ba@alginate | 2.76 | 1.71 | 0.991 | 2.64 | 2.68 | 0.988 | 2.08 | 4.19 | 0.952 |
| Ca-Ba@alginate | 2.18 | 1.32 | 0.984 | 2.04 | 1.91 | 0.953 | 8.25 × 10−3 | 3.80 | 0.971 |
| Ba@alginate | 2.10 | 1.56 | 0.990 | 2.01 | 1.83 | 0.990 | 1.07 × 10−2 | 2.02 | 0.955 |
The results of these adsorption kinetics experiments are consistent with the sulfate removal efficiency test where the two-step Ca/Ba@alginate microcapsules also gave the best sulfate adsorption capacity. On the other hand, the Ba@alginate and Ca-Ba@alginate microcapsules have k2 coefficients of 1.56×10−3 g·mg−1·s−1 and 1.32×10−3 g·mg−1·s−1, respectively. This means that the Ba@alginate microcapsules adsorb sulfate faster than the Ca-Ba@alginate microcapsules, although the permeability in the Ca-Ba@alginate microcapsules is slightly higher than that in the Ba@alginateas microcapsules, as discussed before (Section 3.4). This is because the adsorption reaction is governed by both the permeability of the sulfate ions through the microcapsules as well as the adsorption site available in the outer layers. The Ba@alginate microcapsules have a higher barium content compared to the Ca-Ba@alginate microcapsules, and hence the rate of adsorption is faster at the beginning. However, due to the limitation of the sulfate permeability through the microcapsules the adsorption capacity of the Ca-Ba@alginate microcapsules plateaus at a higher equilibrium value. The other two models with slightly lower R2 values also suggest that the two-step Ca/Ba@alginate microcapsules are the best among the three samples.
Conclusions
We have prepared Ba@alginate microcapsules as a novel adsorbent for lithium purification in lithium refining to remove sulfate in solution. Three methods were developed namely pure barium hydroxide Ba@alginate, two-step Ca/Ba@alginate, and the one step dual-ion Ca-Ba@alginate for preparing three varieties of microcapsules. The Ba@alginate microcapsules could successfully remove sulfate from the lithium sulfate solution, reducing its concentration to 16 mg·L−1. Although the addition of calcium did not further improve the integrity of the microcapsules, it increased the effectiveness of sulfate removal. The two-step Ca/Ba@alginate microcapsules showed the highest rate of adsorption and the greatest sulfate removal rate, achieving a sulfate concentration (as low as 0.02 mg·L−1) that is three orders of magnitude below the impurity limit of sulfate for producing battery-grade lithium. Furthermore, the Ba@alginate microcapsules have the advantages of ease of preparation, high efficiency, low cost, and adaptability to commercialise adsorption technologies, demonstrating great potential for future applications in lithium refining processes.