An effective utilization of the slag from acid leaching of coal-waste: Preparation of water glass with a low-temperature co-melting reaction

Abstract

This paper presents an effective utilization of slag from acid leaching of coal-waste with a novel approach, namely low-temperature co-melting method, for preparation of sodium silicate (Na₂O·nSiO₂) using slag from acid leaching of coal-waste as feedstock. It is very interesting that the co-melting reaction temperature of the mixture of Na₂CO₃ and the feedstock (50˜100 μm) was as low as 850 °C, which was significantly lower than the temperature used in traditional sodium silicate production (1400 °C). The optimum SiO₂/Na₂O ratio was identified as 7:3 according to the results of thermogravimetry–differential scanning calorimetry (TGA-DSC), ICP-AES, and X-ray diffraction (XRD) analyses. In this condition, the main product was sodium disilicate (Na₂O·2SiO₂), with water solubility of 85.0%. More importantly, the impurities such as aluminum in the feedstock, which had adverse effect on subsequent treatment, were concentrated almost completely in the filter residue as insoluble sodium alumunosilicates, i.e., Na(Si₂Al)O₆·H₂O. The lower co-melting temperature of this process demonstrates a significant energy-saving opportunity and thus a promising approach for highly effective utilization of coal-waste.

Implications:

Recently, alumina extraction from coal-waste has been extensively investigated and industrial applied in China. However, the slag-containing silica generated from the acid leaching process of coal-waste led to a secondary pollution, which hindered large-scale production. The proposed low-temperature co-melting method for preparation of sodium silicate (Na₂O·nSiO₂) using slag from acid leaching of coal-waste as feedstock indicated that it is an efficient approach for the recovery of silica from the acid-leached slag of coal-waste with minimal environmental impact.

Introduction

Coal-waste is the largest industrial solid wastes in China (Liu and Liu, 2010; Li et al., 2010). Nowadays, how to recycle and reuse coal-waste becomes a critical issue for effective resource utilization and environment protection. However, coal-waste has no way to be made use of except for some low-value-added products in China, such as constructive materials and road building materials. And there is still a great amount of coal-waste stack-stored in the open air, leading to land occupation and a series of environmental problems. Nevertheless, coal-waste is a useful resource, since coal-waste consists of about 20˜30% carbon, 15˜40% alumina, and 40˜60% silica. Therefore, the recovery of valuable elements from coal-waste to produce high-value-added products with minimal environmental impact is desired. In this regard, the extraction of alumina and silica from coal-waste can produce high-value chemicals and this kind of technology is in high demand. In our previously research, alumina was successfully extracted using an acid leaching process (Cheng et al., 2012); however, after this process, there are still a large amount of slag left, which contains a significant amount of SiO₂ and would lead to a secondary pollution if discharged into environment. Therefore, this research will focus on extracting silica using the slag from the acid leaching of coal-waste as feedstock to produce sodium silicate (Na₂O·nSiO₂; commercially named water glass).

Typically, water glass is prepared by fusing the mixture of Na₂CO₃ and silica sand (SiO₂) at 1100–1450 °C, followed by dissolving the liquid melts in water at 220–245 °C and 27–32 bar (Ismail et al., 2009; Carroll et al., 1998). Sometimes a hydrothermal method is also used. In this case, water glass aqueous solution is synthesized directly by the reaction of silica sand and an aqueous solution of sodium or potassium hydrate in an autoclave at 200 °C and 20 bar (Trabzuni et al., 2011). However, both the above methods are not suitable for sodium silicate production using the feedstock from the acid leaching of coal-waste due to a remainder of 5–10% Al₂O₃ after the acid leaching process. Fusing a mixture of the feedstock containing aluminum with Na₂CO₃ at a temperature of 1100˜1450 °C would lead to production of massive insoluble sodium aluminosilicates with high Si/Al ratios (Izumi, 2003), resulting in a dramatic decrease of both the solubility and the yield of the product. The other disadvantage was the high energy cost at this temperature. As for the latter method, hydrothermal reaction may lead to the difficulties on the separation of impurities and consequently influence the quality of the product.

In this work, a high-purity Na₂O·nSiO₂ was prepared via a novel low-temperature co-melting reaction using slag from the acid leaching of coal-waste as the feedstock (referred as “feedstock” thereafter). The co-melting reaction temperature of the mixture of Na₂CO₃ and the feedstock was determined by thermogravimetry–differential scanning calorimetry (TGA-DSC) combined with inductively coupled plasma atomic emission spectroscopy (ICP-AES) and X-ray diffraction (XRD) analyses. The results indicate that this new method is a promising approach for highly effective utilization of coal-waste.

Materials and Experiment

Feedstock

The coal-waste, kindly supplied by Lu’an Group (Shanxi, China), was grinded to particles with average diameter of about 200 μm and then activated in muffle furnace at 750 °C for 2 hr. The feedstock was obtained from the acidic leaching process of the activated coal-waste according to the method described in our previously published paper (Cheng et al., 2012). Aluminum oxide, ferric oxide, and other impurities were extracted using hydrochloric acid and poly-aluminum chloride was obtained consequently. The main composition of the feedstock measured by a PANalytical X-ray fluorescence (XRF) spectrometer (Almelo, The Netherlands) is shown in Table 1. The Cl⁻ ions probably came mainly from the remainder of hydrochloric acid in the feedstock.

Table 1. The composition of the feedstock

Content (wt.%)
SiO2	Al2O3	Fe2O3	K2O	ZrO2	TiO2	Cl
80.49	8.37	0.53	1.14	0.15	2.31	6.83

Co-melting reaction mechanism

To overcome the disadvantages of the traditional water glass preparation mentioned above, a low-temperature co-melting approach was proposed for preparation of sodium silicate. It is clearly seen from the SiO₂-Na₂O phase diagram (Figure 1) (Paul, 1990) that the eutectic temperatures between sodium metasilicate (Na₂O·SiO₂) and sodium disilicate (Na₂O·2SiO₂) and between Na₂O·2SiO₂ and quartz are 840 and 789 °C, respectively. Therefore, in order to form soluble Na₂O·nSiO₂ (1 ≤ n ≤ 2.7) at the temperature range of 789˜840 °C, a suitable molar ratio of SiO₂ and Na₂CO₃ is necessary to meet SiO₂/Na₂O ratio within range of 1:1˜7.3:2.7 (the latter ratio corresponds to the composition of the eutectic mixture of Na₂O·2SiO₂ and quartz).

Figure 1. SiO₂-Na₂O phase diagram.

In our case, the possible reactions are reactions 1 and 2 shown below, whereas reaction 3 could not happen because of the limitation of the SiO₂/Na₂O ratio (Paul, 1990; Kham, 2005; Kosminski et al., 2006).

1

2

3

In particular, a small amount of Al₂O₃ in the feedstock will benefit the co-melting of Na₂CO₃-quartz due to its strong capability of lowering the refractoriness of quartz (Kham, 2005). Moreover, minimal additives such as NaCl will also decrease the eutectic temperature of the above co-melting system (Paul, 1990; Kham, 2005). All these factors make it possible to prepare soluble sodium silicate by low-temperature co-melting of the mixture of the feedstock and Na₂CO₃.

Sodium silicate preparation

Desired amounts of anhydrous Na₂CO₃ and the feedstock with different SiO₂/Na₂O molar ratios were mixed and well grinded to fine powder (diameter ca. 50˜100 μm) in a ball mill and then put in a muffle furnace. The temperature was raised to 850 °C and then kept for 3 hr. Afterwards, the liquid melt was quenched in boiled water and became small solid fragments. Then, the small fragments and water were transferred into a high-pressure reactor and heated to 200 °C with pressure of 13 bar for 3 hr. The obtained liquid mixture was filtered; thus, a Na₂O·nSiO₂ solution and a filter residue were obtained, respectively.

Characterizing conditions

The chemical composition of the samples was determined by ICP using a PerkinElmer Optima 5300V ICP-AES instrument (Optima 5300V, PerkinElmer, Inc., USA). TGA and DSC analyses were carried out with a simultaneous TGA-DSC thermal analyzer (Netzsch, STA449C; Germany) in Ar atmosphere at a rising rate of 5 °C/min from room temperature to 1000 °C. Fourier transform infrared (FTIR) measurements were carried on with a Nicolet 380 FTIR spectrometer (Thermo Electron Co., USA) from 4000 to 400 cm⁻¹ with KBr pellets. X-ray diffraction (XRD) analysis was performed using a D/max-RB diffractometer (Rigaku, Japan) with graphite crystal monochromized Cu Kα radiation. The X-ray diffraction patterns were obtained at a scan rate of 5°/min.

Results and Discussion

TGA-DSC testing and analysis

To determine the co-melting reaction temperature of the powder mixture of anhydrous Na₂CO₃ and the feedstock, TGA-DSC analysis of the mixed powder with SiO₂/Na₂O molar ratios of 1:1, 3:2, 7:3, and 7.3:2.7 was carried out. For comparison, two other samples, SiO₂/Na₂O 3:2 with 2% NaCl and SiO₂/Na₂O 7:3 with pure silica instead of the feedstock were tested as well. The samples were labeled as S-1:1, S-3:2, S-7:3, S-7.3:2.7, S-3:2+NaCl, and S-7:3 (pure), respectively. The results are shown in Figures 2–4.

Figure 2. The comparative TGA-DSC curves of samples S-7:3 and S-7:3 (pure).

Figure 3. The TGA curves of the samples with different SiO₂/Na₂O molar ratios.

Figure 4. The comparative TGA-DSC curves of samples S-3:2 and S-3:2+NaCl.

Figure 2 shows the TGA-DSC curves of samples S-7:3 and S-7:3 (pure). It is obviously seen from the TGA curves that the first weight loss of both samples occurred at about 120 °C, corresponding to the weak endothermic peaks on the DSC curves. This was ascribed to the evaporation of the residual crystalline water of Na₂CO₃ (Kham, 2005).

In the case of S-7:3 (pure), the significant weight loss did not happen until 770 °C and the corresponding endothermic peaks presented at 580 and 854 °C, which were related to quartz inversion and the melting of Na₂CO₃, respectively (Kosminski et al., 2006). In addition, a solid-state reaction with release of CO₂ occurred between pure SiO₂ and Na₂CO₃ before the co-melting happened according to reaction 1 mentioned above. This was in accordance with Kosminski and the co-workers’ results (Kosminski et al., 2006).

It can also be seen from Figure 2 that the co-melting reaction temperature of S-7:3 (pure) was remarkably lower in comparison with the temperature (1100–1450 °C) used in the traditional manufacture of sodium silicate (Ismail et al., 2009; Carroll et al., 1998). This indicates that the fine particle size (50–100 μm) plays a key role in co-melting reaction at low temperatures, which might be attributed to the sufficient contact between the reactants (Paul, 1990). On the contrary, large particles could hardly react with each other before melting at a higher temperature.

Differed from that of S-7:3 (pure), there is no sudden weight loss observed for S-7:3; instead, the weight was lost slowly from 350 to 840 °C. Accordingly, no apparent endothermic or exothermic peak appeared during the weight loss process. The peaks corresponding to either quartz inversion at 573 °C or melting of Na₂CO₃ at 855 ± 5 °C were also not presented (Kham, 2005). This evidently indicates that the existence of the impurities promotes the decrease of both co-melting and solid reaction temperatures. It is believed that the solid-state reaction and co-melting occurred simultaneously for S-7:3. The weak endothermic and exothermic peaks observed at about 830 °C imply that a small melting process and a further reaction occurred subsequently, with no weight loss. This was probably due to the fact that reaction 2 occurred preferably at a lower temperature, whereas reaction 3 might not happen due to the SiO₂/Na₂O molar ratio at the range of 1˜2.7.

The TGA curves of the samples with SiO₂/Na₂O molar ratios at the range of 1:1 to 7.3:2.7 are shown in Figure 3. It is seen that the co-melting reactions of all the samples started at much lower temperature in comparison with that of S-7:3 (pure) in Figure 2, which is in well accordance to the analysis results from SiO₂-Na₂O phase diagram. This could also be attributed to the lower eutectic temperature at higher SiO₂/Na₂O ratios, apart from the influence of the impurities in the feedstock.

Among these samples, S-7:3 shows the lowest reaction temperature compared with those of S-1:1 and S-3:2. On the contrary, S-7.3:2.7 shows a higher co-melting reaction temperature than that of S-7:3, although it is the co-melting point according to the SiO₂-Na₂O diagram. Ideally, the co-melting of the eutectic mixture with SiO₂/Na₂O molar ratio of 7.3:2.7 at the eutectic point of 789 °C requires a molecular-level mixing, for instance, the mixture formed by slow-cooling of the corresponding fused liquid. However, it is difficult to satisfy the requirements in the actual system we investigated. Besides, the total weight losses of S-1:1, S-3:2, S-7:3, and S-7.3:2.7 by release of CO₂ during the heating were about 23.8%, 19.9%, 16.0%, and 14.3%, respectively, which correspond to the amount decrease of Na₂CO₃ in the mixture.

The effect of minor addition of NaCl on the co-melting temperature of the sample with SiO₂/Na₂O molar ratio of 3:2 is shown in Figure 4. For the sample S-3:2+NaCl, the complete weight loss occurred at 780 °C, which was 70 °C lower than that for sample S-3:2. Although NaCl does not react with silica, it could be used as the melting agent during the reaction. By addition of NaCl, the melting temperature of Na₂CO₃ would be dramatically decreased due to the existence of the low eutectic temperature (about 640 °C) between NaCl and Na₂CO₃ (Paul, 1990; Kham, 2005).

Based on the results of TGA-DSC analysis, the co-melting reaction temperature was set to rise from room temperature to 850 °C by rate of 10 °C/min and then kept for 3 hr to make sure the reaction was complete. Meanwhile, in terms of practical application, it is worth to point out that the amount of the feedstock should be well controlled in practice to satisfy the requirement of the SiO₂/Na₂O molar ratio in the mixture prior to the co-melting process, since the amount of silica in the feedstock may vary according to various sources of coal-waste.

The chemical composition analysis of the products

Table 2 lists the chemical composition of the prepared sodium silicate and filter residue from S-3:2. The amounts of Al₂O₃, Fe₂O₃, and TiO₂ in the filter residue were much higher than those in the feedstock (Table 1), and they were all below detection limits in sodium silicate solution. This illustrates that nearly all the impurities were removed and enriched in the filter residue.

Table 2. The composition of the filtrate and residue

		Content (wt.%)
Sample	Fraction	Na2O	SiO2	Al2O3	Fe2O3	TiO2
S-3:2	Filtrate	36.28	60.97	—	—	—
	Residue	13.38	54.78	22.38	1.35	5.02

Notes: Detection limits: Al, ≥0.05 μg/mL; Fe, ≥0.07 μg/mL; Ti, ≥0.08 μg/mL.

More notably, the contents of SiO₂ and Na₂O in water glass are 60.97 and 36.28, respectively, indicating that Na₂O·2SiO₂ rather than Na₂O·1.5SiO₂ was formed during the low-temperature co-melting process of S-3:2. This probably was due to the formation of Na₂O·2SiO₂ at a lower temperature, as shown in the SiO₂-Na₂O phase diagram (Paul, 1990). Meyer et al. also reported that a glass transition temperature of 741 K existed for Na₂Si₂O₅ synthesized from ultrapure Na₂CO₃ and SiO₂ powders (Meyer et al., 2002). Another interesting finding is that titanium in the feedstock was also enriched in the filter residue (˜5%), indicating the possibility of extracting titanium as well.

The XRD analyses for the feedstock, the sodium silicate, and the filter residue of S-3:2 are shown in Figure 5. The main XRD characteristic peaks for the feedstock (Figure 5a) were attributed to α-silica, whereas the XRD patterns for the filtered slag were attributed to sodium hydroaluminosilicate (Na(Si₂Al)O₆·H₂O; Figure 5c) according to the PDF2 XRD (2004) database. However, no peak corresponding to sodium silicate was observed in Figure 5b, which may be due to the formation of glassy state silicate during the evaporative drying of the sodium silicate aqueous solution.

Figure 5. XRD patterns of the feedstock (a), obtained sodium silicate (b), and filter residue (c) from sample S-3:2.

The weak diffraction peaks of Na₂CO₃ observed in Figure 5b indicate that Na₂CO₃ was in surplus during the co-melting reaction, which maybe because of the formation of Na₂O·2SiO₂ at low temperature. Besides, CO₂ in ambient air absorbed during the evaporative drying of the sodium silicate aqueous solution could also produce a very small amount of Na₂CO₃. This is in good accordance with the ICP results in Table 2. More importantly, the impurities existed in the feedstock were concentrated in the filter residue; thus, a very pure sodium silicate could be prepared by the co-melting reaction method.

FTIR measurement and analysis

The comparative FTIR spectra of the obtained sodium silicate and filter residue from sample S-3:2 are shown as Figure 6A and B. The broad bands at about 3450 and 1647 cm⁻¹ for both spectra are ascribed to absorption of O–H stretching (Giannopoulou and Panias, 2010; Panias et al., 2007; Barbosa et al., 2000). The bands at around 1450 cm⁻¹ indicate the presence of Na₂CO₃ in both the sodium silicate and residue (Barbosa et al., 2000). The sharp elimination of this band in Figure 6B indicates the existence of some Al₂O₃ in the filtered residue (Giannopoulou and Panias, 2010; Panias et al., 2007). There are three main bands associated with different vibrations in the –Si–O–Si– units (Parker and Frost, 1996): (a) a Si–O bending vibration at about 450 cm⁻¹ (Barbosa et al., 2000; Parker and Frost, 1996), (b) 750-800 cm⁻¹ due to symmetric stretching vibration, and (c) the main absorption band at 1200 and 930 cm⁻¹ assigned to the asymmetric stretching vibration. The broadness of band c of Figure 6A indicates the overlapping of the individual absorption bands of the SiQⁿ structural units (where Q represents the bridging oxygen atom in the first coordination sphere of silicon atom, and n denotes the number of bridging oxygen atoms), according to the analysis by Panias and his colleagues (Giannopoulou and Panias, 2010; Panias et al., 2007). It is in good agreement with his point of view that the maximum absorption band is in the range 980–1020 cm⁻¹ due to the predominance of SiQ² and SiQ³ structural units over SiQ°, SiQ¹, and SiQ⁴ ones. Besides, the band at the range 980–1020 cm⁻¹ of Figure 6B shifts toward higher wavenumber as a result of the formation of insoluble Na(Si₂Al)O₆·H₂O during the co-melting process and concentration in the filter residue. With Al incorporating in silicate gels, it improved the cross-linking in between silicate chains and favored the structural change from SiQ² to SiQ³ units. These results are consistent with the ICP and XRD results as well.

Figure 6. Comparative FTIR spectra of sodium silicate (A) and the filter residue (B) obtained from sample S-3:2.

To further investigate the influence of Al on the formation of sodium aluminosilicates, the FTIR spectra of the filter residues from the samples with different SiO₂/Na₂O molar ratios were characterized, which are shown in Figure 7. For comparison, the IR spectrum of the residue from calcined S-7:3 at 1400 °C is illustrated in Figure 7 as well. It is clearly seen that the peak denoting the absorption of Al–O–Si shifted from 1030 to 1054 cm⁻¹ with the increase of the SiO₂/Na₂O molar ratio in the feedstock (Giannopoulou and Panias, 2010). It is concluded that the higher Al in the feedstock, the more sodium aluminosilicates formed, and higher temperature would make that even worse.

Figure 7. FTIR spectra of the filter residue with different SiO₂/Na₂O ratios. (a) S-1:1, (b) S-3:2, (c) S-7:3, (d) S-7:3 (1400 °C).

The solubility of the obtained water glass

The solubility of the silicates produced by the mixture of the feedstock and Na₂CO₃ with different SiO₂/Na₂O molar ratios at 850 °C together with that of S-7:3 (1400 °C) are shown in Table 3. Apparently, the solubility of the silicates decreases with the increase of the SiO₂/Na₂O molar ratio. This could be attributed to the formation of more sodium aluminosilicates with the amount increase of Al in the feedstock. This was confirmed by the above-mentioned results.

Table 3. The solubilities (wt.%) of the produced sodium silicates

Sample
S-1:1	S-3:2	S-7:3	S-7:3 (1400 °C)
92.2	89.4	85.0	54.6

Furthermore, it can be seen that the fusing process at 1400 °C led to a remarkable drop of the solubility of the obtained product; for instance, it dropped to 54.6% for S-7:3 (1400 °C) in comparing with the high solubility (85%) of S-7:3 (850 °C) owning to the formation of more insoluble salts at 1400 °C. It is further clarified that the temperature for industrial production of sodium silicate from pure silica sand is not suitable for the slag from acid leaching of coal-waste.

Figure 8 presents the XRD pattern of the filtered slag of sample S-7:3 (1400 °C). By comparing with the PDF2 XRD (2004) database, the characteristic diffractions of the residue belong to an insoluble aluminosilicate, Na_3.552(Al_3.6Si_12.4O₃₂)·10.656H₂O, in which one Al atom almost binds four Si atoms, much higher than that of the residue obtained at 850 °C. This could be the main reason for the rapid decrease of the solubility for fusing sample S-7:3 at 1400 °C.

Figure 8. XRD pattern of the filter residue of the sample S-7:3 at 1400 °C.

Conclusion

A novel approach named low-temperature co-melting reaction was used for preparation of sodium silicate using the slag from acid leaching of coal-waste as the feedstock. The results showed that the co-melting reaction temperature was as low as 850 °C, which was significantly lower than the temperature of 1400 °C used in the traditional sodium silicate production process. The existence of Al in the feedstock decreased the co-melting temperature of the fine powder mixture of the feedstock and Na₂CO₃ but increased the amount of the insoluble aluminosilicate, Na(Si₂Al)O₆·H₂O. The proper SiO₂/Na₂O ratio of the mixture was identified to be 7:3, by which the main product was sodium disilicate (Na₂O·2SiO₂) and the solubility reached 85%. In particular, the impurities such as aluminum and titanium in the feedstock were concentrated in the filter residue; therefore, only trace amount of impurities was detected in the sodium silicate solution, which had no effect on the subsequent treatment. This new method is a promising approach for highly effective utilization of coal-waste.

Funding

The authors wish to thank the financial support from the National High Technology Research and Development Program of China (863 Program, grant no. 2011AA06A103).

Additional information

Notes on contributors

Li Fang

Li Fang is an associate professor, Fangqin Cheng is a professor, and Xiaofang Duan and Rongming Chen are postgraduate students at the State Environmental Protection Key Laboratory of Efficient Utilization Technology of Coal-waste Resources, Institute of Resource and Environment, Shanxi University, Taiyuan City, P.R. China.