Hydrothermal synthesis of mixtures of NaA zeolite and sodalite from Ti-bearing electric arc furnace slag

Abstract

Ti-bearing electric arc furnace slag (Ti-bearing EAF slag) is the main solid waste generated in the direct reduction iron making process. During the extraction process of Ti from Ti-bearing EAF slag, a by-product containing abundant Si and Al was seldom utilized, thus leading to waste of valuable elements and secondary pollution. In this paper, NaA zeolite (6Na₂O·6Al₂O₃·12SiO₂) and sodalite (SOD, 4Na₂O·3Al₂O₃·6SiO₂) were synthesized successfully using the by-product as a precursor. The effects of the SiO₂/Al₂O₃ molar ratio (n(SiO₂)/n(Al₂O₃)), H₂O/Na₂O molar ratio (n(H₂O)/n(Na₂O)), hydrothermal temperature and time on the crystal phase and microstructure of the prepared zeolites were systematically investigated. The results indicated that NaA zeolite with good crystallinity and cubic morphology was obtained at 140 °C for 3 h with n(SiO₂)/n(Al₂O₃) and n(H₂O)/n(Na₂O) fixed at 2.0 : 1 and 100 : 1, respectively. Decreasing n(H₂O)/n(Na₂O), raising the hydrothermal temperature and prolonging the hydrothermal time were beneficial for the formation of spherical SOD zeolite. The phase transformation between NaA and SOD zeolite was discussed and a mechanism was proposed to explain the phenomenon of these zeolites coexisting in the obtained samples. In addition, the removal performances of Cu²⁺ in aqueous solutions using zeolites synthesized at different temperatures were studied. The maximum removal capacity of the prepared zeolite can reach 1.346 mmol g⁻¹ for 180 min.

---

1 Introduction

Zeolite molecular sieves are crystalline porous solids with intricate pore and channel systems in the molecular size range of 0.3 to 3 nm.¹ NaA zeolite and sodalite (SOD) are kinds of low silica zeolites which have the general formulas 6Na₂O·6Al₂O₃·12SiO₂ (NaA) and 4Na₂O·3Al₂O₃·6SiO₂ (SOD), respectively. The molar ratios of SiO₂/Al₂O₃ in NaA and SOD zeolites are equal to 2.0 : 1. Their frameworks are composed of [SiO₄]⁴⁻ and [AlO₄]⁵⁻ tetrahedra which are linked at their corners to form channels and cages. Generally, zeolite molecular sieves are synthesized from chemical reagents or minerals with rich Al and Si,^2–5 leading to high cost, huge discharge of waste and high energy consumption. Thus, extensive efforts have been made to search cheaper raw materials and economic methods for synthesis of zeolite molecular sieves.^6–11 Qian et al.¹¹ prepared highly crystalline NaA zeolite with high Ca²⁺ exchange capacity (CEC) by using coal gangue as raw materials via the in situ crystallization technique. Bukhari et al.¹² compared direct hydrothermal method and indirect fusion method of converting coal fly ash to zeolite and using prepared zeolites as adsorbents for the removal of heavy metal ions from aqueous solutions. They found that the zeolites prepared by hydrothermal method showed better crystalline structure and performed better at immobilizing heavy metal ions. Therefore, the synthesis of zeolite from waste materials by hydrothermal method would have wide application prospects in solid waste reclamation utilization.

Ti-bearing electric arc furnace slag (Ti-bearing EAF slag) which contains Ti, Mg, Al and Si contents is the main solid waste generated in direct reduction ironmaking process, and at present, much more attention has been paid only on extraction of Ti by using various approaches such as acid leaching method,^13,14 selective enrichment method^15,16 and alkali fusion method^17,18 etc. However, few efforts have been taken to recycle other valuable elements comprehensively, which result in the potential secondary waste generated during this utilization. In order to make full use of valuable elements existed in Ti-bearing EAF slag, a novel process based on alkali fusion-water leaching-acidolysis process has been proposed by our group.^19–21 During the alkali fusion process, the elements existed in Ti-bearing EAF slag were converted to their corresponding sodium salts. Then the alkaline solution containing sodium aluminate and sodium silicate were obtained in the water leaching process, which was an ideal precursor to prepare NaA zeolite.²² However, until now, none of the work has been reported yet.

In this paper, NaA and sodalite (SOD) zeolites were synthesized from the leaching solution of the fused Ti-bearing EAF slag by hydrothermal method. The effects of SiO₂/Al₂O₃ molar ratio ((n(SiO₂)/n(Al₂O₃)), H₂O/Na₂O molar ratio (n(H₂O)/n(Na₂O)), hydrothermal time and temperature on the preparation of zeolites were systematically investigated. The transformation mechanism from NaA zeolite to SOD zeolite was also discussed. In addition, the removal capacities of the prepared zeolites with respect to Cu²⁺ in solution were studied. This paper may explore a novel pathway for efficient and comprehensive utilization of the Ti-bearing EAF slag.

2 Experimental

2.1 Materials

The Ti-bearing EAF slag was obtained from Panzhihua Steel (Sichuan Province, China) and the main chemical compositions of the slag was shown in Table 1. All the chemical regents employed were analytical grade (Sinopharm Chemical Reagent Co. Ltd) and the distilled water was used throughout the experiment.

Table 1 Main chemical compositions of the Ti-bearing EAF slag (wt%)

Composition	TiO2	Al2O3	MgO	SiO2	CaO	Fe2O3
Content	50.9	19.4	12.9	8.0	5.4	2.9

---

According to our previous research,²² Si and Al could be converted to Na₂SiO₃ and NaAlO₂ during the molten NaOH treatment, respectively. Then, they were separated from other elements by water leaching and dissolved in water to form the precursor, which was the raw material to prepare zeolite.

2.2 Sample characterization

The contents of elements existed in the precursor was analyzed by inductively coupled plasma emission spectroscopy (ICP, Optima 7000 DV, PerkinElmer). The phase structure of the obtained sample was investigated by X-ray diffraction using Cu Kα radiation (λ = 0.154056 nm) with 40 kV, 200 mA and a speed of 10° min⁻¹ (XRD, M21X, MAC SCIENCE Co. Ltd, Japan). The morphology and microstructure of the prepared samples were examined by field emission scanning electron microscope (SEM, Zeiss, Supra-55). The Fourier transformed infrared (FT-IR) spectra of the prepared samples were measured by an Avatar 360 FT-IR ESP Spectrometer in the range of 2000–450 cm⁻¹. The Cu²⁺ concentrations in the filtrates were determined by a UV-Vis spectrophotometer (Persee, TU-1901) according to Lambert–Beer theory.

2.3 Synthesis of NaA and SOD zeolites from Ti-bearing EAF slag

2.3.1 Separation of Na₂SiO₃ and NaAlO₂ precursor from Ti-bearing EAF slag. The Ti-bearing EAF slag was mixed with NaOH homogeneously with Ti-bearing EAF slag/NaOH mass ratio of 1 : 1.5 and then roasting at 700 °C for 1 h to obtain alkali fusion slag. After that, the alkali fusion slag was dissolved in distilled water with mass ratio at 1 : 10 and stirred by electromagnetic stirring for 0.5 h. Finally, the obtained suspension was filtered and the obtained filtrate containing mainly Na₂SiO₃ and NaAlO₂ was used as precursor to synthesize zeolite.

2.3.2 Preparation of NaA and SOD zeolites. According to Table 1, the content of Al₂O₃ is larger than that of SiO₂ in Ti-bearing EAF slag, indicating that the concentration of Na₂SiO₃ is less than that of NaAlO₂ in the obtained leaching solution. In order to use this solution as precursor to synthesize zeolite, Na₂SiO₃·9H₂O should be added to adjust SiO₂/Al₂O₃ molar ratio (n(SiO₂)/n(Al₂O₃)) from 1.4 : 1 to 2.6 : 1. Then the precursor was well stirred and controlled H₂O/Na₂O molar ratio (n(H₂O)/n(Na₂O)) varied from 40 : 1 to 130 : 1. After that, 80 mL precursor was transported into a 100 mL Teflon lined stainless steel autoclave for hydrothermal treatment at a desired temperature for several hours to synthesize zeolite. Finally, the prepared zeolite was filtered, washed, and dried at 90 °C for 12 h before further measurement and characterization. The whole procedure can be depicted in detail as shown in Fig. 1.

Fig. 1 Flow sheet of the process for hydrothermal synthesis of zeolites from Ti-bearing EAF slag.

2.3.3 Removal performance measurement. The removal performance for heavy metal ions by the prepared zeolites under different hydrothermal temperatures were investigated by measuring removal capacity of Cu²⁺ in CuSO₄ solutions. Batch removal experiments were conducted using 0.5 g of prepared samples in 60 mL 40 mmol L⁻¹ CuSO₄ solutions at constant temperatures (25 °C). At the end of each specified reaction time, about 3 mL solutions were taken from the reaction vessel by graduated syringe and filtered through 0.45 μm polycarbonate filters. Since the exact concentration of CuSO₄ solution at the beginning of the process was known as 40 mmol L⁻¹, the concentration of CuSO₄ solution after specified time can be calculated using eqn (1) according to Lambert–Beer theory.

C_t = C₀A_t/A₀ (1)

where C₀ and C_t (mmol L⁻¹) are the concentrations of CuSO₄ solution at the beginning and after adsorption for t (min) time, respectively. A₀ and A_t are the absorbance of CuSO₄ solution at the beginning and after specified time period, respectively.

The concentration of the heavy metal ions in the solid phase (Q_t, mmol g⁻¹) was calculated using integral mass balance for the heavy metal ions in the vessel according to eqn (2).

Q_t = V(C₀ − C_t)/m (2)

where V (L) represents the volume of CuSO₄ solution and m (g) is the sorbent mass used in the reaction. Q_t (mmol g⁻¹) is the relative removal capacity of zeolite for Cu²⁺.

3 Results and discussion

Considering the content of Al and Si existed in Ti-bearing EAF slag, they were suitable to synthesize the low silica zeolite such as NaA and sodalite (SOD) zeolites. The preparation of NaA zeolite and SOD zeolite can be illustrated by the following generalized equations:

12Na₂SiO₃ + 12NaAlO₂ + 12H₂O = 6Na₂O·6Al₂O₃·12SiO₂(NaA) + 24NaOH (3)

6Na₂SiO₃ + 6NaAlO₂ + 5H₂O = 4Na₂O·3Al₂O₃·6SiO₂(SOD) + 10NaOH (4)

According to eqn (3) and (4), SiO₂/Al₂O₃ molar ratio (n(SiO₂)/n(Al₂O₃)), H₂O/Na₂O molar ratio (n(H₂O)/n(Na₂O)), hydrothermal time and temperature may play important roles in synthesis of NaA and SOD zeolites from Ti-bearing EAF slag.

3.1 The effect of n(SiO₂)/n(Al₂O₃)

Fig. 2 gave the XRD patterns of prepared zeolites with different n(SiO₂)/n(Al₂O₃). It can be seen that NaA zeolites could be obtained when n(SiO₂)/n(Al₂O₃) was fixed at 2.6 : 1, but the crystallinity was poor. Moreover, an obvious diffraction peak assigned to SOD zeolite could be detected at 2θ = 14°, suggesting that a few SOD zeolite coexisted with poor crystallized NaA zeolite. When n(SiO₂)/n(Al₂O₃) was adjusted to consist with the stoichiometry of reactions (2.0 : 1), the intensities of diffraction peaks of NaA zeolite became stronger while that of diffraction peak at 2θ = 14° assigned to SOD zeolite became weaker, indicating that the crystallinity of prepared NaA zeolite became better. Further decreasing n(SiO₂)/n(Al₂O₃) to 1.4 : 1, SOD zeolite became the predominant phase in the obtained sample, suggesting that NaA zeolite would be converted to SOD zeolite with the amount of NaAlO₂ increasing. This phenomenon may be ascribed to the formed SOD zeolite with high structure stability. According to ref. 23, it is known that cubic NaA and SOD zeolites can be built using the β-cage consisting of 24 T-atoms [T = Si⁴⁺ or Al³⁺] (Fig. 3a) where primary units is SiO₄ and AlO₄ tetrahedra. In NaA zeolite, each β-cage is connected to the six nearest neighboring β-cages through double T4-rings ([D4R]) as shown in Fig. 3b. In SOD zeolite, each β-cage is connected to the six nearest neighboring β-cages through common simple T4-rings ([S4R]) (Fig. 3c). Therefore, the mass density of SOD zeolite (2.29 g cm⁻³) is slightly higher than that of NaA zeolite (2.00 g m⁻³), resulting in SOD zeolite showed a higher framework density and structure stability compared to NaA zeolite.

Fig. 2 XRD patterns of prepared samples with different n(SiO₂)/n(Al₂O₃). (n(H₂O)/n(Na₂O) = 100 : 1, T = 140 °C, t = 3 h).

Fig. 3 The connection types of β-cage. (a) β-Cage, (b) β-cage connected through a double T4-ring ([D4R]) to one of its nearest neighbors, (c) β-cage connected through a common simple T4-ring ([S4R]) to one of its nearest neighbors.

Since NaAlO₂ is necessary for the formation of zeolite framework, the presence of aluminum atom on neighboring T (Si, Al) atom site distorts silicon tetrahedron by shortening one of the Si–O bonds.^24,25 Thus, increasing the amount of NaAlO₂ in solution would result in more serious distortion of [SiO₄] tetrahedral, leading to the zeolite framework unstable and easy conversion to SOD zeolite with higher stable structure.

The morphologies of obtained zeolites with different n(SiO₂)/n(Al₂O₃) were shown in Fig. 4. From SEM images analysis, it can be seen that cubic grains and spherical particles were formed and agglomerated in all the prepared samples, which were respectively identified as NaA zeolite (cubic) and SOD zeolite (spherical). Notably, when n(SiO₂)/n(Al₂O₃) was 1.4 : 1, spherical particles with less imperfect cubic grains appeared, suggesting that the main phase of the obtained sample was SOD zeolite as shown in Fig. 4a. With n(SiO₂)/n(Al₂O₃) increasing from 2.0 : 1 (Fig. 4b) to 2.6 : 1 (Fig. 4c), highly crystallized cubic grains with larger size and intact crystal plane were formed. Meanwhile, the amount of spherical particles decreased obviously, indicating that higher n(SiO₂)/n(Al₂O₃) was beneficial for crystallization of NaA zeolite.

Fig. 4 SEM images of samples prepared with different n(SiO₂)/n(Al₂O₃). (a) 1.4 : 1, (b) 2.0 : 1, (c) 2.6 : 1 (n(H₂O)/n(Na₂O) = 100 : 1, T = 140 °C, t = 3 h).

3.2 The effect of n(H₂O)/n(Na₂O)

It is well known that the alkalinity of the precursor strongly influences the kinetics of nucleation, crystal growth and the transformation between zeolites.²⁶ In this paper, the alkalinity of the precursor mainly depended on n(H₂O)/n(Na₂O), and it increased with n(H₂O)/n(Na₂O) decreasing.

Fig. 5 was the XRD patterns of samples obtained with different n(H₂O)/n(Na₂O). It can be seen that no obvious diffraction peaks appeared with n(H₂O)/n(Na₂O) at 130 : 1, indicating that few crystalline structure could be formed from the precursor. Decreasing n(H₂O)/n(Na₂O) to 100 : 1 and 70 : 1, diffraction peaks with strong intensities assigned to NaA zeolite appeared, suggesting that well crystallized NaA zeolite could be prepared when n(H₂O)/n(Na₂O) ranged from 70 : 1 to 100 : 1. However, further decreasing n(H₂O)/n(Na₂O) to 40 : 1, NaA zeolite was converted to SOD zeolite, indicating that strong alkalinity was benefit for the formation of SOD zeolite rather than NaA zeolite. Based on the aforementioned analysis, it is believed that low alkalinity is unfavorable for the crystallinity, and high alkalinity can promote the transformation from a metastable phase (NaA zeolite) to a new phase with high framework density (SOD zeolite). Therefore, relatively pure NaA zeolite could be obtained when n(H₂O)/n(Na₂O) was fixed at 100 : 1.

Fig. 5 XRD patterns of samples prepared with different n(H₂O)/n(Na₂O). (n(SiO₂)/n(Al₂O₃) = 2.0 : 1, T = 140 °C, t = 3 h).

SEM images of samples prepared with different n(H₂O)/n(Na₂O) were illustrated in Fig. 6. It can be seen that when n(H₂O)/n(Na₂O) was 40 : 1, only uniform spherical particles appeared, indicating that obtained sample were mainly SOD zeolite (Fig. 6a). With n(H₂O)/n(Na₂O) increasing to 70 : 1, many small cubic NaA zeolite particles were clearly identified among the spherical SOD zeolites (Fig. 6b). Further increasing n(H₂O)/n(Na₂O) to 100 : 1, highly crystallized cubic grains with intact crystal plane and uniform size combined with some little spheres were obtained. Moreover, the size of the obtained cubic particles became larger and the surface of cubic particle appeared smoother, implying that the crystallinity of the prepared NaA zeolites became better. When n(H₂O)/n(Na₂O) was continuously up to 130 : 1, no crystallized product but amorphous gel was formed. This observation agreed well with the XRD results (Fig. 5).

Fig. 6 SEM images of samples prepared with different n(H₂O)/n(Na₂O). (a) 40 : 1, (b) 70 : 1, (c) 100 : 1, (d) 130 : 1 (n(SiO₂)/n(Al₂O₃) = 2.0 : 1, T = 140 °C, t = 3 h).

3.3 The effect of hydrothermal temperature

Fig. 7 illustrated XRD patterns of samples synthesized at different hydrothermal temperatures. When the hydrothermal temperature was controlled at 100 °C for 3 h, none of the crystalline structure was detected. Increasing the hydrothermal temperature to 120 °C, the diffraction peaks indexed as NaA with weak intensity appeared, suggesting that NaA could be formed at 120 °C. However, some strong diffraction peaks assigned to NaA and SOD zeolites appeared at 140 °C and 160 °C, respectively, indicating that the reaction temperature strongly affects the nucleation process and crystal growth process for the formation of zeolites at different temperatures.

Fig. 7 XRD patterns of samples obtained at different hydrothermal temperatures. (n(SiO₂)/n(Al₂O₃) = 2.0 : 1, n(H₂O)/n(Na₂O) = 100 : 1, t = 3 h).

Fig. 8 gave the SEM images of samples obtained at different hydrothermal temperatures. The amorphous phase samples obtained at 100 °C for 3 h had predominant amorphous floccule morphologies (Fig. 8a), while cubic grains with uniform size covered by some amorphous substance were synthesized at 120 °C (Fig. 8b). With the temperature increasing to 140 °C, the amorphous floccule disappeared and majority of well crystallized cubic grains combined with some little spheres on their surface were formed (Fig. 8c). Further increasing the temperature to 160 °C, most of the cubic NaA zeolite was converted to SOD zeolite with spherical morphology (Fig. 8d), which was also evidenced by the XRD results shown in Fig. 7.

Fig. 8 SEM images of samples obtained at different hydrothermal temperatures. (a) 100 °C, (b) 120 °C, (c) 140 °C, (d) 160 °C (n(SiO₂)/n(Al₂O₃) = 2.0 : 1, n(H₂O)/n(Na₂O) = 100 : 1, t = 3 h).

3.4 The effect of hydrothermal time

In order to investigate the phase transformation process for NaA and SOD zeolites, the preparation of samples were conducted at 120 °C and 160 °C for different hydrothermal time, respectively. Fig. 9 was the XRD patterns of samples synthesized at 120 °C for 3 to 12 h. It was found that the intensity of diffraction peaks indexed to NaA zeolite became stronger with the reaction time from 3 to 9 h, indicating that the crystallinity of prepared NaA zeolite became better, correspondingly. However, further prolonging the hydrothermal time to 12 h, the obtained NaA zeolite disappeared and changed to SOD zeolite, suggesting that NaA zeolite would easily convert into the more stable SOD zeolite after long reaction time.²

Fig. 9 XRD patterns of samples synthesized at different hydrothermal time. (n(SiO₂)/n(Al₂O₃) = 2.0 : 1, n(H₂O)/n(Na₂O) = 100 : 1, T = 120 °C).

Fig. 10 showed the SEM images of samples synthesized at different hydrothermal time. As shown in Fig. 10a, it clearly revealed that 3 h crystallized sample were a lot of cubic particles covered by some amorphous floccule. With the hydrothermal time increasing to 6 h, the amorphous floccule disappeared and highly crystallized cubic grains combined with some little spheres on the surface could be obtained (Fig. 10b). In addition, the crystal plane of obtained cubic grains became more intact crystal plane and the sizes of the grains were uniform with about 10 μm (shown as inset in Fig. 10b). However, the aggregation degree of cubic particles became more serious, and the defined cubic structure started to collapse with time prolonging to 9 h (Fig. 10c). As shown in the inset of Fig. 10c, it can be seen that cubic particles with surface corrosion and some imperfect spherical particles were coexisted in the obtained sample. With the time up to 12 h, nearly all the cubic structure destroyed and the spherical crystal grains became the majority phase. This means that relatively long reaction time was beneficial for the conversion of metastable NaA zeolite into more thermodynamically stable SOD zeolite phase.²⁷

Fig. 10 SEM images of samples synthesized at different hydrothermal time. (a) 3 h, (b) 6 h, (c) 9 h, (d) 12 h (n(SiO₂)/n(Al₂O₃) = 2.0 : 1, n(H₂O)/n(Na₂O) = 100 : 1, T = 120 °C).

The parallel experiments were conducted at 160 °C for different time to investigate the phase transformation mechanism further. The XRD patterns and SEM images of samples obtained at 160 °C for different time were shown in Fig. S1 and S2,† respectively. According to the analysis of XRD and SEM results, the same conclusion can be drawn: the prepared zeolites could transform from cubic NaA zeolite to spherical SOD zeolite with the reaction time prolonging. Moreover, the higher temperature, the longer hydrothermal time would be beneficial for the formation of SOD zeolite.

3.5 The transformation mechanism between NaA and SOD zeolite

It is known that the framework of NaA zeolite is built up by β-cages connected by double 4-rings ([D4R]), whereas in SOD zeolite, the β-cages are connected by sharing single 4-rings ([S4R]).² So, the FT-IR spectra were used to analyze the arrangement of Si and Al tetrahedral units. Fig. 11a illustrated the FT-IR spectrum of sample obtained at 120 °C for 3 h. It can be seen that five obvious characteristic bands are at about 1440, 992, 882, 668 and 464 cm⁻¹. Compared to the FT-IR spectra of samples prepared at 140 °C (Fig. 11b) and 160 °C (Fig. 11c), the absorption bands at around 1440 cm⁻¹ weaken with temperature increasing, indicating that the amount of amorphous material in the synthesized samples gradually decreased.²⁸ According the ref. 2, 27, 29 and 30, the broad bands at about 1000 cm⁻¹ can be assigned to antisymmetric stretching of T–O bonds (T = Si or Al) in aluminosilicates. The characteristic bands at about 1654 cm⁻¹ are attributed to bound water existed in zeolites. Moreover, the absorption bands in the range of 500–420 cm⁻¹ are typical for the T–O bonds in the β-cage. The absorption bands in the range of 600–500 cm⁻¹ are corresponding to the double 4-rings ([D4R]) and the absorption bands in the range of 800–600 cm⁻¹ are corresponding to the simple 4-rings ([S4R]), which are typical for the SOD structure. Therefore, it is reasonable to assume that the weak band at about 464 cm⁻¹ could point to the beginning of the crystallization of a zeolite with β-cage. Meanwhile, the intensity of absorption band at about 555 cm⁻¹ was strong when the hydrothermal temperature was at 140 °C (Fig. 11b), indicating that NaA zeolite was the majority phase in the prepared samples since [D4R] only existed in NaA zeolite. However, it is important to note that after crystallization at 160 °C for 3 h, the absorption bands at about 616 and 662 cm⁻¹, which are in the range of 600–800 cm⁻¹ associated with [S4R] are absent (Fig. 11c), suggesting that NaA zeolite has converted to SOD zeolite. Similar results were obtained when the samples were prepared at 120 °C for different time (Fig. S3†).

Fig. 11 FT-IR spectra of samples synthesized at different hydrothermal temperatures. (a) 120 °C, (b) 140 °C, (c) 160 °C (n(SiO₂)/n(Al₂O₃) = 2.0 : 1, n(H₂O)/n(Na₂O) = 100 : 1, t = 3 h).

Based on the above analysis, it can be found that the samples synthesized from Ti-bearing EAF slag would transform from NaA zeolite to SOD zeolite with the temperature increasing or time prolonging. Generally, it is believe that the formation of two different phases was resulted from temperature gradients or compositional variations during the course of the reaction. However, it is important to find out the transformation mechanism of the existing cubic NaA zeolite converted to SOD zeolite step by step.

Fig. 12 illustrated the formation process of cubic NaA zeolite and the phase transformation mechanism from cubic NaA zeolite to spherical SOD zeolite. At the very beginning, amorphous aluminosilicates precursor aggregated into irregular particles by the arrangement of Si and Al tetrahedral units (primary building) (step 1). These particles further aggregated into secondary building units (SBUs) such as simple 4-rings ([S4R]) and simple 6-rings ([S6R]) by the association of TO₄ (T = Si, Al) tetrahedral units (step 2). Then some of β-cages were formed by further polymerization of simple 4-rings ([S4R]) and simple 6-rings ([S6R]). Surface crystallization then occurred to form many crystalline islands by connecting β-cages via double 4-rings ([D4R]) (step 3). When the crystalline islands extended and fused together, a cubic shell of NaA zeolite could be obtained (step 4 and 5). After the formation of the crystalline cubic shell, the crystallization process extended inward to the amorphous core to increase the thickness of the shell (step 6).

Fig. 12 Schematic diagram of NaA zeolite formation process and the phase transformation from NaA zeolite to SOD zeolite.

When the core–shell cubes formed, mass transportation across the crystalline shell became difficult. As the crystalline shell extended inward to the amorphous core, the pressure in the core may increase if the core was dense, because the process of crystallization of the core of zeolite A was a process of density decrease. Based on the aforementioned analysis, it is known that nucleation of SOD zeolite, having a higher density in comparison with NaA zeolite, took place in the amorphous cores of cubic particles where the pressure was built up (step 7). Moreover, high pressure induced phase transformation from a low density zeolite to a high density one was also reported in other research.³¹ The SOD nucleus expanded in size to break through the NaA zeolite cubic shells to form spherical structure, which can be confirmed by the SEM image (step 8). As shown in the SEM image, the well-developed surfaces of some cubic particles were corroded by the flakes developed within their centers. The broken cubes did not show regular edges, and the cubic morphology was partially maintained in the particle (part A). In other particles such as part B, the SOD zeolite with spherical morphologies developed further to break the cubic shells of NaA zeolite and left only one surface of cubic. Finally, the consumption of NaA zeolite occurred through an Ostwald ripening process and resulted in the cubic NaA zeolite convert to SOD zeolite with spherical morphology (step 9). The results obtained by Wang et al.³² and Radulović et al.³³ confirmed that the formation of more stable phase occurred following the Ostwald ripening rule. This process was characterized by a simultaneous growth and dissolution of particles separated in the same media. With time prolonging or temperature increasing, more stable phases like SOD or cancrinite will form at the expense of the metastable NaA zeolite. However, the appearance of these new phases requires the occurrence of a series of very complex processes mass (transference, nucleation and growth) by the rupture of the chemical equilibrium of the system. The stable phases reach the required equilibrium conditions to promote the formation of the final zeolite materials.

According to the above analysis, it is believed that the transformation from NaA zeolite to SOD probably caused by local high pressure rather than by nucleation and growth from residual inhomogeneities in the aluminosilicate gel. Fortunately, we found some intermediate phases (Fig. S4†) formed during the transformation from NaA zeolite to SOD following the Ostwald ripening rule. Therefore, under the present synthetic conditions, no pure NaA zeolite could be produced at any synthetic stage, although reasonably large cubic particles appeared and NaA zeolite became the major phase in the obtained samples as shown in Fig. 8c and 10b.

3.6 Removal of Cu²⁺ by prepared zeolite

The removal capacities of Cu²⁺ ions by zeolites synthesized at different hydrothermal temperatures had been studied to elucidate removal of heavy metal ions from wastewaters. Fig. 13 illustrated the relationship between contact time and removal capacities of Cu²⁺ by zeolites synthesized at different temperatures. It is indicated that the removal process was consisted of two steps. The first step was rapid removal of metal ions within 30 min, followed by a slower second step, suggesting that the adsorption equilibrium between Cu²⁺ ions with the adsorption sites on the surface of zeolite might be attained in 30 min. This behavior, characterized by an initial rapid and quantitatively predominant removal followed by a slower quantitatively insignificant removal, has been extensively reported.^34–36 For shorter contact time, the amounts of adsorbed metal increased rapidly, due to the more active surface sites on zeolites and high driving force for the mass transfer. However, in order to immobilize Cu²⁺ ions, Cu²⁺ should not only move through the pores of the zeolite, but also through channels of the lattice and replace exchangeable cations (mainly sodium). The adsorption would became less efficient as the active surface sites became increasingly occupied and the removal efficiency of Cu²⁺ was retarded when Cu²⁺ moved through channels of the lattice to replace exchangeable cations in zeolite.

Fig. 13 Relationship between contact time and removal capacity of Cu²⁺ by zeolite synthesized at different temperatures.

According to eqn (1) and (2), it can be calculated that the removal capacity of Cu²⁺ by zeolite synthesized at 120 °C for 180 min was 1.346 mmol g⁻¹, much better than that of obtained at 140 °C or 160 °C, which were only 0.722 and 0.673 mmol g⁻¹, respectively. The removal of Cu²⁺ in wastewater was attributed to different mechanisms of ion exchange as well as adsorption process. During the ion exchange process, Cu²⁺ can move through the pores of zeolites and channels of the lattice to take place ion exchange with Na⁺ in zeolite. The diffusion is faster through the pores, whereas it is retarded when the ions move through the smaller channels.^9,37 Based on the aforementioned analysis, the pores and channels of SOD zeolites were much smaller than that of NaA zeolites, leading to difficult movement of Cu²⁺ to exchange with Na⁺ ions. Therefore, the removal performance of NaA zeolite for Cu²⁺ was better than that of SOD zeolite.

4 Conclusion

In this paper, NaA and sodalite (SOD) zeolites were synthesized successfully from Ti-bearing EAF slag for removing heavy metal ions in waste water. The effects of n(SiO₂)/n(Al₂O₃), n(H₂O)/n(Na₂O), hydrothermal temperature and hydrothermal time on the crystal phase and microstructure of synthesized zeolites were systematically investigated. The results indicated that NaA zeolite with well crystallinity and cubic morphology could be obtained at 140 °C for 3 h with n(SiO₂)/n(Al₂O₃) and n(H₂O)/n(Na₂O) fixed at 2.0 : 1 and 100 : 1, respectively. Decreasing n(H₂O)/n(Na₂O), raising the hydrothermal temperature and prolonging the hydrothermal time were beneficial for the formation of spherical SOD zeolite. Moreover, when n(SiO₂)/n(Al₂O₃) and n(H₂O)/n(Na₂O) were controlled at 2.0 : 1 and 100 : 1, respectively, hydrothermal conditions were fixed at 120 °C for 12 h or 160 °C for 3 h, single crystal of SOD zeolite could be synthesized. However, no matter which condition was chosen, NaA zeolite was coexisted with SOD zeolite. In addition, the removal performances of Cu²⁺ in Cu₂SO₄ solutions using zeolites synthesized at different temperatures were studied. The maximum removal capacity reached 1.346 mmol g⁻¹ by zeolite obtained at 120 °C for 3 h.

Acknowledgements

The work was financially supported by National Basic Research Program of China (No. 2014CB643401, No. 2013AA032003), National Science Foundation of China (No. 51372019, 51471122), China Postdoctoral Science Foundation (No. 2015M581581).

References

1. A. C. Lopes, P. Martins and S. Lanceros-Mendez, Prog. Surf. Sci., 2014, 89, 239–277.
2. V. Y. Prokof'Ev and N. E. Gordina, Appl. Clay Sci., 2014, 101, 44–51.
3. T. S. Jamil, H. S. Ibrahim, I. H. Abd El-Maksoud and S. T. El-Wakeel, Desalination, 2010, 258, 34–40.
4. J. Wang, Y. Huang, Y. Pan and J. Mi, Microporous Mesoporous Mater., 2014, 199, 50–56.
5. W. Pichór, W. Mozgawa, M. Król and A. Adamczyk, Mater. Res. Bull., 2014, 49, 210–215.
6. C. Kosanovi, J. Broni, A. Mek, B. Suboti, I. Mit, M. Stubiar and A. Tonejc, Zeolites, 1995, 15, 247–252.
7. C. Wang, J. Li, L. Wang and X. Sun, J. Hazard. Mater., 2008, 155, 58–64.
8. S. Vichaphund, D. Aht-Ong, V. Sricharoenchaikul and D. Atong, Environ. Technol., 2014, 35, 2254–2261.
9. W. Bao, L. Liu, H. Zou, S. Gan, X. Xu, G. Ji, G. Gao and K. Zheng, Chin. J. Chem. Eng., 2013, 21, 974–982.
10. M. C. Barnes, J. Addai-Mensah and A. R. Gerson, Microporous Mesoporous Mater., 1999, 31, 287–302.
11. T. Qian and J. Li, Adv. Powder Technol., 2015, 26, 98–104.
12. S. S. Bukhari, J. Behin, H. Kazemian and S. Rohani, J. Mater. Sci., 2014, 49, 8261–8271.
13. X. Lei and X. Xue, Trans. Nonferrous Met. Soc. China, 2010, 20, 2294–2298.
14. L. Zhang, H. Hu, Z. Liao, Q. Chen and J. Tan, Hydrometallurgy, 2011, 107, 40–47.
15. L. Zhang, L. N. Zhang, M. Y. Wang, T. P. Lou, Z. T. Sui and J. S. Jang, J. Non-Cryst. Solids, 2006, 352, 123–129.
16. Z. Z. Guo, T. P. Lou, L. Zhang, L. N. Zhang and Z. T. Sui, Acta Metall. Sin., 2007, 20, 9–14.
17. Y. Han, T. Sun, J. Li, T. Qi, L. Wang and J. Qu, Int. J. Miner., Metall. Mater., 2012, 19, 205–211.
18. A. J. Manhique, W. W. Focke and C. Madivate, Hydrometallurgy, 2011, 109, 230–236.
19. Y. Li, Y. Yue, Z. Que, M. Zhang and M. Guo, Int. J. Miner., Metall. Mater., 2013, 20, 1012–1020.
20. Z. Que, Y. Li, H. Yu, F. Zheng, M. Zhang and M. Guo, Rare Met., 2015, 34, 267–275.
21. Y. Li, Y. Yang, M. Guo and M. Zhang, RSC Adv., 2015, 5, 13478–13487.
22. Y. Li, H. Yu, Z. Zhang, M. Zhang and M. Guo, ISIJ Int., 2015, 55, 134–141.
23. R. R. Xu, Chemistry-zeolite and porous materials, Science Press, Beijing, China, 2004.
24. L. A. Bugaev, J. A. Bokhoven and V. V. Khrapko, J. Phys. Chem. B, 2009, 113, 4614–4618.
25. C. Kosanović, T. A. Jelić, J. Bronić, D. Kralj and B. Subotić, Microporous Mesoporous Mater., 2011, 137, 72–82.
26. T. Antonic, B. Subotic and N. Stubicar, Zeolites, 1997, 18, 291–300.
27. X. Zhang, D. Tang, M. Zhang and R. Yang, Powder Technol., 2013, 235, 322–328.
28. Y. Huang, H. Wang, K. Wang, D. Dong, D. Li and H. Hill, Microporous Mesoporous Mater., 2010, 127, 167–175.
29. M. Alkan, I. D. Hopa, Z. Yilmaz and H. Güler, Microporous Mesoporous Mater., 2005, 86, 176–184.
30. F. Hasan, R. Singh, G. Li, D. Zhao and P. A. Webley, J. Colloid Interface Sci., 2012, 382, 1–12.
31. L. Xiaoyang, S. Wenhui, W. Yifeng and Z. Xudong, Chem. Commun., 1992, 902–903.
32. N. Wang, Y. Liu, A. Huang and J. Caro, Microporous Mesoporous Mater., 2015, 207, 33–38.
33. A. Radulović, V. Dondur, R. Dimitrijević and D. Arandjelović, Thermochim. Acta, 2010, 511, 37–42.
34. H. Song, H. Cheng, Z. Zhang and F. Cheng, J. Environ. Biol., 2014, 35, 983–988.
35. M. Šljivić Ivanović, I. Smičiklas and S. Pejanović, Chem. Eng. J., 2013, 223, 833–843.
36. R. Shawabkeh, Process Saf. Environ. Prot., 2009, 87, 261–266.
37. M. Jovanovic, Z. Grbavcic, N. Rajic and B. Obradovic, Chem. Eng. Sci., 2014, 117, 85–92.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra26881h

---