Chemical regeneration of magnesium oxide used as a sorbent for fluoride

Chemical regeneration of magnesium oxide used as a sorbent for fluoride

Separation and Purification Technology 98 (2012) 24–30 Contents lists available at SciVerse ScienceDirect Separation and Purification Technology journ...

1MB Sizes 1 Downloads 9 Views

Separation and Purification Technology 98 (2012) 24–30

Contents lists available at SciVerse ScienceDirect

Separation and Purification Technology journal homepage: www.elsevier.com/locate/seppur

Chemical regeneration of magnesium oxide used as a sorbent for fluoride Keiko Sasaki ⇑, Naoyuki Fukumoto, Sayo Moriyama, Qianqian Yu, Tsuyoshi Hirajima Department of Earth Resources Engineering, Kyushu University, Fukuoka 819-0395, Japan

a r t i c l e

i n f o

Article history: Received 1 June 2012 Received in revised form 31 July 2012 Accepted 31 July 2012 Available online 10 August 2012 Keywords: Chemical regeneration Magnesium oxides Fluoride Sorption isotherm Temperature programmed desorption of CO2

a b s t r a c t Use of MgO for repeated calcination with regard to its sorption density for F and chemical stability was examined. Magnesium oxide was produced by the calcination of MgCO3 at 1273 K for 1 h. The sorption of 9.82 mM F on the calcined product and the calcinations of solid residues were carried out five times to evaluate trends in sorption density for F and the stability of the chemically regenerated sorbents. The order of sorption density of F (Q/mol g1) apparently seems to depend on the specific surface area. However, Q0 values after normalization of the sorption density for the specific surface area (Q0 /mmol m2) were found to be correlated with the solid basicities of the calcined products which were derived from CO2-TPD curves for the calcined products. The number of weak base sites, calculated from the peak intensity at 373 K in CO2-TPD, was considered to be responsible for the removal of F through hydration. Large quantities of NaMgF3 and elemental Mg were evaporated by the calcination at 1273 K of the solid residues formed after sorption of F. Ó 2012 Elsevier B.V. All rights reserved.

1. Introduction Fluoride (F) in drinking water has a narrow beneficial concentration range in relation to human health [1]. Small amounts in ingested water are usually considered to have a beneficial effect on the rate of occurrence of dental caries, particularly among children [2]. In contrast, excess intake of F leads to various diseases such as osteoporosis, arthritis, brittle bones, cancer, infertility, brain damage, Alzheimer’s syndrome, and thyroid disorder [3]. The World Health Organization (WHO) has classified F is as one of the contaminants of water and regulated the maximum contamination limits of fluoride to 1.0 mg L1 for drinking water and 10.0 mg L1 for industrial discharge [4]. Groundwater pollution is the result of glass and ceramic production, semiconductor manufacturing, electroplating, coal fired power stations, beryllium extraction plants, brick and iron works, aluminum smelters, and surface impoundments among many other sources [5,6]. Defluoridation is normally accomplished by sorption and precipitation processes [7–9]. The most commonly used sorbent for F removal from drinking water is activated alumina. However, the most serious disadvantage of this process is the formation of bulky sludge after the treatment of large amounts of water. Magnesium oxide (MgO) is a superior sorbent for F near the region of the maximum contamination limit of F, compared with other metallic oxides such as alumina, lime and bimetallic oxides and conventional sorbents such as carbon based sorbents and modified zeolite [8] and can be easily obtained by the simple ⇑ Corresponding author. Tel./fax: +81 92 802 3338. E-mail address: [email protected] (K. Sasaki). 1383-5866/$ - see front matter Ó 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.seppur.2012.07.028

modification of natural magnesite and/or hydromagnesite, both of which are low cost geomaterials [4,10,11]. The authors have reported that magnesia (MgO), which is derived from the calcination of magnesite (MgCO3), removes F ions effectively by destructive sorption. The sources of crystalline magnesite are limited and localized mainly in China (27%), Russia (22%) and North Korea (24%) [12]. In many reports, other Mg-bearing minerals and materials have been investigated for use as sorbents of F instead of MgO [13–17]. In particular, the Mg–Al–CO3 layered double hydroxides (LDHs) have been intensively studied as possible reusable sorbents for F [18,19]. However, calcination of LDHs sometimes lead the segregation of one metal on the surface to affect the surface reactivity. Therefore, it is important to understand the factors affecting the sorption efficiency in the chemical regeneration of MgO derived from magnesite, which is the simplest and the most common Mg resource. In the present work, repeated measurements of the sorption of F on MgO and calcined solid residues were carried out to evaluate their sorption density for F and the stability of the chemically regenerated sorbents. 2. Experimental MgO-rich phases were produced by calcination of MgCO3 (special grade, Sigma–Aldrich, St. Louis, MO, USA) for 1 h at 1273 K, and by calcination of the solid residues after sorption of 9.82 mM F at five times, designated as MgO-I, MgO-II, MgO-III, MgO-IV and MgO-V respectively, as shown in Fig. 1. Details of sorption experiments are described in the next section. The solid residue from each step was designated as Residue-I, Residue-II, Residue-III, Residue-IV and Residue-V, respectively. The calcined products were

25

K. Sasaki et al. / Separation and Purification Technology 98 (2012) 24–30

MgO(331) MgO(222)

MgO-V

MgO(220)

5 kcps

MgO(200)

MgO(111)

(a)

Intensity

MgO-IV

MgO-III

MgO-II

MgO-I 10

20

30

40

50

Diffraction angle, 2

60

Cu K

70

80

degree

Fig. 1. Sequential calcination process to synthesize and regenerate MgO after use for sorption of fluoride.

Intensity

MgO (222) Mg(OH)2 (202)

Mg(OH)2 (103)

Mg(OH)2 (201) MgO (331)

MgO (220)

Mg(OH)2 (110)

MgO (200)

Mg(OH)2 (102)

Residue-V

Mg(OH)2 (101)

MgO (111)

5 kcps

Mg(OH)2 (001)

(b)

Residue-IV

Residue-III

10

MgO-I MgO-II MgO-III MgO-IV MgO-V

45.8 8

Q / mmol·g

-1

23.3

Residue-II

Residue-I

6

10

9.82

20

30

40

Diffraction angle, 2 4

0

0

10

70

80

degree

45.8

23.3 23.3 23.3

9.82 9.82 6.82 1.24

2

60

Fig. 3. XRD patterns of (a) calcined products and (b) solid residues after sorption of F. Symbols: s, MgO (JCPDS 45-946); d, Mg(OH)2 (JCPDS 44-1482).

23.3

6.82 9.82

50

Cu K

45.8 45.8 45.8

20

30

40

50

-1

Ceq / mmol·L

Fig. 2. Sorption isotherm of F on MgO-rich phases produced by repeating calcinations of magnesium carbonates and solid residues. Sorption density is expressed as Q (mmol g1). Numbers in the figure indicate the initial F concentration (mM). Data were fit to Freundlich type of isotherm.

characterized by X-ray diffraction with Cu Ka, 40 kV, 20 mA (XRD, Multi Flex, Rigaku, Akishima, Japan), scanning electron microscopy (SEM, VE-9800, KEYENCE, Osaka, Japan) and transmission electron

microscopy (TEM, FEI TECNAI-20, JEOL). The specific surface areas were determined by the seven-point N2-adsorption BET method (AUTOSORB-1, YUASA, Osaka, Japan) and temperature programmed desorption curves (TPD, BELCAT-B, BEL JAPAN Inc., Toyonaka, Japan) using CO2 as a probe gas. CO2-TPD curves were measured using a positive temperature gradient of 3.3 K min1 after degassing by heating the samples at 773 K. The data were analyzed with a BELCAT Chem Master Ver. 2.3.9 (BEL JAPAN Inc., Toyonaka, Japan) to evaluate the basicity and the number of base sites of the calcined products, which consisted mainly of MgO. To evaluate the sorption density of F on the calcined products for the calcination cycles shown in Fig. 1, the sorption isotherms were determined. 1.24–45.77 mmol L1 F solutions at pH 6.09 were prepared using NaF (special grade, Wako, Osaka, Japan). For the sorption experiments, 0.100 g of the calcined products MgO-I,

26

K. Sasaki et al. / Separation and Purification Technology 98 (2012) 24–30

Fig. 4. SEM images of (a) MgO-I, (b) Residue-I, (c) MgO-II, (d) Residue-II, (e) MgO-III, (f) Residue-III, (g) MgO-IV, (h) Residue-IV, (i) MgO-V and (j) Residue-V. Horizontal bars indicate 1 lm.

II, III, IV and V were added to 40 mL fluoride solutions, followed by shaking at 100 rpm at 298 K. In preliminary experiments, it was confirmed that equilibrium was achieved within 72 h under these conditions. After 72 h, the supernatants were filtered (0.45 lm) for

the determination of the total concentrations of F by ion chromatography (ICS-90, DIONEX, Osaka, Japan). Solid residues after sorption were also characterized by XRD and SEM in the same manner as for the sorbents before sorption. Thermogravimetric

27

K. Sasaki et al. / Separation and Purification Technology 98 (2012) 24–30

Fig. 5. TEM images of (a) MgO-I, (b) MgO-II, (c) MgO-III, (d) MgO-IV and (e) MgO-V. Horizontal bars indicate 200 nm in (b) and 100 nm in others.

measurements were performed using a TG–DTA 2000SA (RUKER, Yokohama, Japan) under a flow of 100 mL min1 N2 with Residue-I after sorption of 9.82 mM F. Samples were heated from room temperature to 1373 K at 10 K min1.

Table 1 Changes in specific surface areas of calcined products and sorbed mass of CO2 molecules per unit surface area derived from component (a) in CO2-TPD curves.

SBET (m2/g) CO2 (lmol/m2)a

3. Results and discussion Fig. 2 shows the sorption isotherm of F on the calcined products at 298 K. The sorption density of F on the calcined product (Q in mmol g1) was the greatest for MgO-I and the smallest for MgO-II. It increased with the number of calcination cycles. Fig. 3 shows the XRD patterns for MgO-I, II, III, IV and V, and Residue-I, II, III, IV and V. In the calcined products, all peaks are assigned to MgO (JCPDS 45-946). For the solid residues, the major peaks are assigned to Mg(OH)2 (JCPDS 44-1482) and MgO. The mechanism of F removal is explained by the destructive sorption of MgO [20]. During the hydration of MgO, F is co-precipitated. Peak positions assigned to Mg(OH)2 were identical to data JCPDS

a

MgO-I

MgO-II

MgO-III

MgO-IV

MgO-V

29.1 0.86

1.5 15.3

2.3 8.70

5.1 3.73

7.1 2.96

Calculated from component (a) in Fig. 7.

44-1482 in Residue-V, but most had shifted to lower angles in XRD pattern for Residue-I. This might be caused by structural strain due to the sorption of large amounts of F in Residue-I (Fig. 2). The relative peak intensities for Mg(OH)2 were the largest in Residue-I, gradually decreased with increase in regeneration times and were the smallest in Residue-V. The remaining peaks assigned to MgO are particularly prominent in Residue-V. The relative intensities of MgO in the solid residues increased with the number of regeneration cycles.

28

K. Sasaki et al. / Separation and Purification Technology 98 (2012) 24–30

Because of the relationship with SBET, the maximum value of Q0 appeared with MgO-II and the minimum with MgO-I. This calculation suggests that the sorbent, which has strong basicity, showed a high affinity for F. To clarify these findings, the basicity and the number of base sites per SBET were measured. The CO2-TPD curves were collected for CO2-desorption temperatures in the range from room temperature to 500 °C and analyzed so as to separate the results into three components using the previously reported peak positions for MgO [21,22], as shown in Fig. 7. Although the CO2-TPD curves in Fig. 7 were obtained using a detection method based on measurement of electro-conductivity, the same samples were independently identified using a mass spectrometer (Q-mass). From the Q-mass results, it was confirmed that the first component (a) at 373 K is from CO2 and the second and third components are from H2O. The quantities of sorbed CO2 for component (a) in Fig. 7 are also summarized in Table 1. There is an association between the basicity per SBET of the calcined products and the sorption density of F obtained in Fig. 6, as depicted in Fig. 8. This exemplifies the findings in our previous report, in which different calcinations temperatures produced different basicities, affecting the sorption density of F per unit surface area [24]. After being normalized for surface area, the larger basicity leads to higher reactivity with H2O, as shown in the following equation.

Time / sec 0

500

-2

70 µV·m

1000

1500

2000

(b)

(a)

MgO-V

MgO-IV

-2

70 µV·m Desorption of CO2 / µV·m

-2

SEM images for MgO-I, II, III, IV, V, and Residue-I, II, III, IV, V are shown in Fig. 4. The MgO-I particles (Fig. 4a) were rod shaped, 10– 20 lm in length and 5–6 lm in diameter. The surface of Residue-I (Fig. 4c) was totally covered with very fine flakes leading to an increase in the diameters of the rods, probably due to being covered with bulky Mg(OH)2 precipitates. The surface of MgO-II (Fig. 4b) was covered with a large quantities of granular particles 0.5– 1.0 lm in length. MgO-III, IV, V had similar morphologies to MgO-II. In the SEM images of Residue-II, III, IV, V (Fig. 4d,f,h,j), well-developed hexagonal flakes were observed on the surface consisting of Mg(OH)2. As shown in Fig. 3b, a phase of MgO still remained in Residue-V and peak shifts of Mg(OH)2 were surprisingly observed in Residue-I. Larger amounts of F were sorbed on MgO-I than MgO-V, as shown in Fig. 2a. Combining these observations, it is suggested that Mg(OH)2 crystals with large quantities of F show less developed flakes in the SEM images (Fig. 4b) and lead to an increase in lattice parameters for Residue-I (Fig. 3b). On the other hand, Residue-V has a smaller proportion of F ion impurities leading to well developed crystals of Mg(OH)2, as shown in the XRD and SEM. Analyzing the XRD data by using the Rietveld method, the fractions of remaining MgO were calculated 4% in Residue I, 13 % in Residues II, 38% in Residue III, 57 % in Residue IV and 52% in Residue V. This indicates that the repetition of calcination produces highly crystalline MgO in the cores of the grains which remains even after sorption of F in the same conditions. TEM images for all calcined products are shown in Fig. 5. The size of crystals observed in the calcined products increases with repeated calcination (Fig. 5e). These data indicate that the stability of MgO has increased with the number of calcination cycles. It was also observed that crystal sizes became more variable with the increase in calcinations times. The specific surface areas (SBET) of the calcined products are listed in Table 1. The largest SBET was for MgO-I. The second calcination resulted in a very low value of SBET, and after the third calcination the value increased gradually with the number of calcination cycles. The variation of the specific surface area might be caused by trace amounts of moisture produced by lyophilization before the calcination. The capacity Q is a function of the value of SBET (Fig. 2). Therefore, by converting the Q values to Q0 (mmol m2), a different trend can be seen, as shown in Fig. 6. 2.0

MgO-I MgO-II MgO-III MgO-IV MgO-V

Q' / mmol·m

-2

1.5

1.0

6.82

45.8

23.3

9.82

-2

200 µV·m

45.8

23.3

9.82

23.3

0.5

6.82 0

MgO-II

45.8

9.82

-2

20 µV·m

45.8 0.0

MgO-III

-2

200 µV·m

23.3

9.82 10

20

MgO-I

45.8 30

40

50

-1

Ceq / mmol·L 

Fig. 6. Sorption isotherm of F on MgO-rich phases produced by repeating calcinations of magnesium carbonates and solid residues, with conversion of sorption density into Q0 (mmol m2). Numbers in the figure are the same as in Fig. 2. Data were fit to Freundlich type of isotherm.

40

80

120

160

200

240

Temperature / Fig. 7. Plots of Q0 against sorbed mass of CO2 per unit surface area in component (a) in Fig. 6 for calcined products.

29

K. Sasaki et al. / Separation and Purification Technology 98 (2012) 24–30

I

1.2

V IV

0

II

III

0.00

Ce 20 mM 10 mM

1.0

MgO

-0.02

1 mM

0.8

Gravimetric loss / %

Q / mmol·m

-2

-0.04

0.6

0.4

-20

-0.06

Solid residue

-0.08

-1

-30

0.2

Differential gravimetric loss/ %·s

-10

-0.10

0.0

0

5

10

15

Mg(OH)2

20

-2

CO2 in component (a) / µmol·m

-40

-0.12 200

400

600

800

1000

o

Fig. 8. CO2-TPD curves of calcined products at different repeating times of calcination.

Temperature / C Fig. 9. DT–TGA curves of the solid residue after sorption of 9.82 mM F (Residue-I) and MgO and Mg(OH)2.

The elemental compositions in the sorbents before and after sorption of 9.82 mmol L1 F ions, and changes in their mass relative to MgO-I are summarized in Table 2. As the cycle number increased the mass of Mg in the sorbent decreased and the relative mass of MgO was reduced to 53% after the 5th calcination. The F content was not detected for the sorbent obtained after any calcination. It was found that calcination causes the volatilization of Mg and F. It is known that neighborite (NaMgF3) can be synthesized from MgF2 by mixing solutions of MgCl2 and NaF followed by aging for several hours [23]. The boiling points of MgF2 and NaMgF3 are reported to be 2533 K and less than 1296 K, respectively [24]. There is a possibility that NaMgF3 might begin to vaporize

Table 2 Changes of total weights in calcination and sorption cycles, and Mg (wt%) and F (ppm) contents in the solid residues before and after sorption of F, normalized for weight of sorbents before 1st sorption of F. 1st Total weight (wt%)

Mg contents (wt%)

F contents (mg/ kg)

Others (wt%)

n.d: not detected.

Before sorption After sorption Before sorption After sorption Before sorption After sorption Before sorption After sorption

3rd

4th

5th

100.0

95.12

80.77

70.69

52.73

135.6

105.97

88.45

77.46

68.24

55.80

57.06

45.57

39.19

29.82

54.14

51.08

42.15

35.18

33.10

n.d.

n.d.

n.d.

n.d.

2nd

n.d.

0.37

0.31

0.29

0.31

0.34

44.20

38.06

35.20

31.50

22.91

81.48

54.89

46.30

42.28

35.14

during calcination at 1273 K. The molar ratio of Mg/F volatilized from the solid residue was much higher than 0.33 so it suggests that the elemental Mg should have also evaporated, along with NaMgF3. The boiling point of the elemental magnesium is reported to be 1363 K [25]. To remove NaMgF3, while also preventing the evaporation of elemental Mg by calcinations, requires control within a very narrow range of calcination temperatures. To confirm the reason why the content of Mg and F decreased with increase in calcinations cycles, the TGA curves for the solid residue after sorption of 9.82 mmol L1 F ions (Residue-I) are summarized in Fig. 9, as well as those for the chemical reagents MgO and Mg(OH)2. Mg(OH)2 showed a large gravimetric loss, ten times greater than that of MgO. This arises from dehydration at around 700 K. Another gravimetric loss was confirmed for the solid residues at 1100 K. This might arise from volatilization of NaMgF3 around this temperature. 4. Conclusions The degree of destructive sorption density of F with MgO was investigated using sorption isotherms at 298 K in with multiple (five times fold) of repeat of calcination and sorption under the same conditions. Solid residues after sorption of much large amounts of F lead to an increase in the lattice parameters of Mg(OH)2 and look less developed flake-like structures in the SEM images. Repeated calcinations did not show a clear trend in either the values of SBET for the calcined products or the sorption efficiency of F per unit mass and surface area, but did appear to increase the crystal size of crystals observed by TEM observation. The sorption density of F was not clearly related with to the initial SBET value of the sorbents. However, a positive correlation was found between the sorbed mass of F per SBET (Q0 ) and the sorption capacity of CO2, used as an indicator of basicity which was obtained from CO2-TPD analysis. The basicity on the surface of the MgO affects to its hydration and the resulting in sorption density of F during co-precipitation of F with Mg(OH)2. Increase in calcinations times lead well to highly crystalline MgO in cores of the

30

K. Sasaki et al. / Separation and Purification Technology 98 (2012) 24–30

grains cores but it did not always increase the total basicity of the surface of the MgO particle. Considerable amounts of NaMgF3 and elemental Mg were evaporated by calcinations of the solid residues at 1273 K after each sorption of F. These results show that a sequential system of sorption and calcination needs to be optimized to maximize the overall efficiency of sorption of F. Acknowledgements Financial support was provided to KS by the Funding Program for the Next Generation of World-Leading Researchers (‘‘NEXT program’’ GR078) in the Japan Society for the Promotion of Science (JSPS). XRD patterns were collected at the Advanced Analytical Center, Kyushu University, and TEM observations were carried out in the Research Laboratory for High Voltage Electron Microscopy (HVEM), Kyushu University.

[11]

[12] [13]

[14]

[15]

[16]

[17]

References

[18]

[1] Meenakshi, R.C. Maheshwari, Fluoride in drinking water and its removal, J. Hazard. Mater. 137 (2006) 456–463. [2] M. Mahramanlioglu, I. Kizilcikli, I.O. Bicer, Adsorption of fluoride from aqueous solution by acid treated spent bleaching earth, J. Fluorine Chem. 115 (2002) 41–47. [3] N.J. Chinoy, Effects of fluoride on physiology of animals and human beings, Indian J. Environ. Toxicol. 1 (1991) 17–32. [4] M.G. Sujana, H.K. Pradhan, S. Anand, Studies on sorption of some geomaterials for fluoride removal from aqueous solutions, J. Hazard. Mater. 161 (2009) 120– 125. [5] F. Shen, X. Chen, P. Gao, G. Chen, Electrochemical removal of fluoride ions from industrial wastewater, Chem. Eng. Sci. 58 (2003) 987–993. [6] W. Nigussie, F. Zewge, B.S. Chandravanshi, Removal of excess fluoride from water using waste residue from alum manufacturing process, J. Hazard. Mater. 147 (2007) 954–963. [7] M. Young, T. Hashimoto, N. Hosho, H. Myoga, Fluoride removal in a fixed bed packed with granular calcite, Water Res. 33 (1999) 3395–3402. [8] A. Bhatnagar, E. Kumar, M. Sillanpa, Fluoride removal from water by adsorption – a review, Chem. Eng. J. 171 (2011) 811–840. [9] A.M. Raichur, M.J. Basu, Sorption of fluoride onto mixed rare earth oxides, Sep. Purif. Technol. 24 (2001) 121–127. [10] H. Bilinski, B. Matkovic, D. Kralj, D. Radulovic´, V. Vrankovic´, Model experiments with CaO, MgO and calcinated dolomite for fluoride removal in

[19]

[20]

[21]

[22]

[23]

[24] [25]

a wet scrubbing system with sea water in recirculation, Water Res. 19 (1985) 163–168. B. Nagappa, G.T. Chandrappa, Mesoporous nanocrystalline magnesium oxide for environmental remediation, Micropor. Mesopor. Mater. 106 (2007) 212– 218. S.S. Fu, P.J. Li, Q. Feng, X.J. Li, P. Li, Y.B. Sun, Y. Chen, Soil quality degradation in a magnesite mining area, Pedosphere 21 (2011) 98–106. D. Thakre, S. Rayalu, R. Kawade, S. Meshram, J. Subrtc, N. Labhsetwar, Magnesium incorporated bentonite clay for defluoridation of drinking water, J. Hazard. Mater. 180 (2010) 122–130. M. Mohapatra, D. Hariprasad, L. Mohapatra, S. Anand, B.K. Mishra, Mg-doped nano ferrihydrite – a new adsorbent for fluoride removal from aqueous solutions, Appl. Surf. Sci. 258 (2012) 4228–4236. X. Xu, Q. Li, H. Cui, J. Pang, L. Sun, H. An, J. Zhai, Adsorption of fluoride from aqueous solution on magnesia-loaded fly ash cenospheres, Desalination 272 (2011) 233–239. S.M. Maliyekkal, S. Shukla, L. Philip, I.M. Nambi, Enhanced fluoride removal from drinking water by magnesia-amended activated alumina granules, Chem. Eng. J. 140 (2008) 183–192. L. Lv, J. He, M. Wei, D.G. Evans, X. Duan, Factors influencing the removal of fluoride from aqueous solution by calcined Mg–Al–CO3 layered double hydroxides, J. Hazard. Mater. B133 (2006) 119–128. J. Fan, Z. Xu, S. Zheng, Comment on ‘‘Factors influencing the removal of fluoride from aqueous solution by calcined Mg–Al–CO3 layered double hydroxides’’, J. Hazard. Mater. B139 (2007) 175–177. G. Wu, X.L. Wang, W. Wei, Y.H. Sun, Fluorine-modified Mg–Al mixed oxides: a solid base with variable basic sites and tunable basicity, Appl. Catal. A: Gen. 377 (2010) 107–113. M.E. Martin, R.M. Narske, K.J. Klabunde, Mesoporous metal oxides formed by aggregation of nanocrystals. Behavior of aluminum oxide and mixtures with magnesium oxide in destructive adsorption of the chemical warfare surrogate 2-chloroethylethyl sulfide, Micropor. Mesopor. Mater. 83 (2005) 47–50. S. Fan, N. Zhao, J. Li, F. Xiao, W. Wei, Y. Sun, Effective and green synthesis of methyl pyrrole-1-carboxylate with dimethyl carbonate over solid base, Catal. Lett. 120 (2008) 299–302. K. Sasaki, N. Fukumoto, S. Moriyama, T. Hirajima, Sorption characteristics of fluoride on to magnesium oxide-rich phases calcined at different temperatures, J. Hazard. Mater 191 (2011) 240–248. I. Sevonkaev, D.V. Goia, E. Matijevi´c, Formation and structure of cubic particles of sodium magnesium fluoride (neighborite), J. Colloid Interf. Sci. 317 (2008) 130–136. J.B. Holn, J.L. Holn, The enthalpy of mixing liquid NaF and NaMgF3 from drop calorimetry, Thermochim. Acta 10 (1974) 393–397. E.Y. Shafirovich, U.I. Goldshleger, The superheat phenomenon in the combustion of magnesium particles, Combust. Flame 88 (1992) 425–432.