Red mud as a carbon sink: Variability, affecting factors and environmental significance

Journal of Hazardous Materials 244–245 (2013) 54–59

Contents lists available at SciVerse ScienceDirect

Journal of Hazardous Materials journal homepage: www.elsevier.com/locate/jhazmat

Red mud as a carbon sink: Variability, affecting factors and environmental significance Chunhua Si a , Yingqun Ma a,c , Chuxia Lin b,∗ a b c

Centre for Ecological and Environmental Technologies, South China Agricultural University, Guangzhou 510642, China Australian Centre for Sustainable Catchments, University of Southern Queensland, Toowoomba, QLD 4350, Australia Chinese Research Academy of Environmental Sciences, Beijing 100012, China

h i g h l i g h t s     

Red mud is an abundant hazardous material that has the potential for carbon sequestration. Different red mud types were investigated to determine the factors affecting red mud carbonation. A set of parameters were established for evaluation of red mud’s carbon sequestration capacity. Global-scale estimation of carbon stock and carbon sequestration potential in red mud was made. Methods for maximizing red mud carbonation and their environmental significance were discussed.

a r t i c l e

i n f o

Article history: Received 15 June 2012 Received in revised form 9 November 2012 Accepted 10 November 2012 Available online 20 November 2012 Keywords: Red mud Carbon sequestration Alkalinity Carbonation Bauxite Calcite

a b s t r a c t The capacity of red mud to sequester CO2 varied markedly due to differences in bauxite type, processing and disposal methods. Calcium carbonates were the dominant mineral phases responsible for the carbon sequestration in the investigated red mud types. The carbon sequestration capacity of red mud was not fully exploited due to shortages of soluble divalent cations for formation of stable carbonate minerals. Titanate and silicate ions were the two major oxyanions that appeared to strongly compete with carbonate ions for the available soluble Ca. Supply of additional soluble Ca and Mg could be a viable pathway for maximizing carbon sequestration in red mud and simultaneously reducing the causticity of red mud. It is roughly estimated that over 100 million tonnes of CO2 have been unintentionally sequestered in red mud around the world to date through the natural weathering of historically produced red mud. Based on the current production rate of red mud, it is likely that some 6 million tonnes of CO2 will be sequestered annually through atmospheric carbonation. If appropriate technologies are in place for incorporating binding cations into red mud, approximately 6 million tonnes of additional CO2 can be captured and stored in the red mud while the hazardousness of red mud is simultaneously reduced. © 2012 Elsevier B.V. All rights reserved.

1. Introdution Global alumina production from bauxite processing generated over 2.7 billion tonnes of hazardous residues (red mud) during the period from the late 19th century to 2008, and the current annual production rate of red mud has been estimated at 120 million tonnes [1]. Bauxite is a heterogeneous material consisting primarily of one or more aluminium hydroxide minerals (mainly gibbsite, boehmite and diaspore) and impurities (mainly silicon-bearing clay minerals and iron oxides). To extract alumina from bauxite ores, sodium hydroxide (NaOH) is used to dissolve aluminium hydroxides at elevated temperature and pressure. Due to the presence

∗ Corresponding author. Tel.: +61 7 46312429; fax: +61 7 46315581. E-mail address: [email protected] (C. Lin). 0304-3894/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jhazmat.2012.11.024

of residual NaOH, red mud is highly alkaline with a pH usually ranging from 10 to 13 [2–6]. Consequently, red mud has the potential to sequester carbon dioxide (CO2 ), according to the following chemical equations: OH− (aq) + CO2(g) → HCO3 − (aq)

(1)

OH− (aq) + HCO3 − (aq) → H2 O + CO3 2− (aq)

(2)

The above reactions eventually reach a state of dynamic equilibrium unless the reaction products (i.e. bicarbonate and carbonate ions) are removed from the solution. Carbonate ions can react with a range of divalent metal cations to form stable carbonate minerals, as shown below: CO3 2− (aq) + M2+ (aq) → MCO3(s)

(3)

C. Si et al. / Journal of Hazardous Materials 244–245 (2013) 54–59

It has been demonstrated that carbon sequestration by red mud took place rapidly under elevated CO2 pressure [4,7]. It has also been observed that calcium carbonates were present in large amounts in the red mud exposed to atmospheric CO2 under ambient environmental conditions [8–10]. Renforth et al. [6] recently found that atmospheric CO2 was rapidly sequestered by gypsumtreated red mud that was spread over the land surface following the Ajka (Hungary) red mud spill. Calcium (Ca) and magnesium (Mg) are common divalent metals in the Earth’s surface environments. These elements may be present in bauxite ores or introduced into red mud during bauxite processing [9,11–14]. Their abundance and reactivity are expected to have significant influences on the carbonation of red mud. On the other hand, the presence of elements such as silicon, titanium and phosphorus in red mud may also affect the CO2 sequestration by forming oxyanions that are likely to compete with carbonate ions for the available divalent cations in red mud. The chemical and mineralogical composition of individual red mud sources could vary markedly due to the differences in the bauxite ores and refining processes employed [9,15]. Consequently, the capacity of red mud to sequester CO2 is expected to be variable. Understanding the factors that affect red mud carbonation and the variability of carbon storage among different red mud types is important for assessing the role of red mud as a carbon sink. In this study, we examine carbon storage and a range of chemical parameters in the surface red mud samples collected from three selected alumina refineries. The objectives were to examine the chemical properties of red mud to (a) understand the magnitude of carbon sequestration in red mud using current disposal practices, (b) assess the maximum carbon capture potential of red mud, and (c) discuss on what practices may be employed to exploit this potential. 2. Materials and methods 2.1. The red mud samples Three different alumina refineries were selected for sample collection: (a) the Zhengzhou Alumina Refinery at Zhengzhou, central China (ZZ); (b) the Guizhou Alumina Refinery at Guiyang, Southwestern China (GZ), and (c) the Queensland Alumina Refinery at Gladstone, Australia (GS). Both the Zhengzhou and the Guizhou Alumina Refineries were designed to extract alumina from diaspore-dominated bauxite ore with a combined sintering and Bayer process, and a wet method for red mud disposal. The difference between the two refineries was that the Zhengzhou Alumina Refinery had an average Al/Si ratio of about 5.5 with illite and pyrophyllite being the main siliconbearing minerals [12,13] while the the Guizhou Alumina Refinery had a higher average Al/Si ratio (7.4) with kaolinite being the main silicon-bearing minerals [12]. The Queensland Alumina Refinery used the Bayer process method for alumina extraction from gibbsite-boehmite type bauxite ore mined from Weipa, northern Queensland [14]. The red mud was treated by seawater before being thickened by flocculants and deposited in the dam as slurry in the disposal facility. Surface red mud samples were collected from the red mud storage facilities of the above three refineries. In the laboratory, the red mud samples were air-dried, ground and sieved to prepare subsamples with different degrees of fineness (<2 mm, <0.25 mm and <0.15 mm) for various experiments and laboratory analyses. 2.2. Sample characterization and analytical methods Chemical analyses were performed to determine total carbon concentration (total C), pH, electrical conductivity (EC),

55

water-extractable K, Na, Ca and Mg and NH4 Cl-extractable K, Na, Ca and Mg. Total C was determined by a LECO CNS-2000 Analyzer. Since the organic carbon in all the three red mud types was negligible (determined by the Walkley-Black method), the total C can be used as an estimate of inorganic carbon (inorganic C). For each red mud type, 1:5 (red mud:water) and 1:5 (red mud:1 mol L−1 ammonium chloride (NH4 Cl)) extracts were prepared by shaking 5 g of red mud in 25 mL of a relevant solution for 1 h, followed by centrifugation to obtain the supernatant. pH and EC in the water extract were determined by a calibrated pH meter and a EC meter, respectively. K, Na, Ca and Mg in both water and 1 mol L−1 NH4 Cl extracts were determined by atomic absorption spectrometry (AAS). Water soluble K, Na, Ca and Mg were estimated by water-extractable K, Na, Ca and Mg, respectively; exchangeable K, Na, Ca and Mg were estimated by the difference between the 1 mol L−1 NH4 Cl extractable K, Na, Ca and Mg and the water-extractable K, Na, Ca and Mg, respectively. Since no exchangeable acidic cations are expected to exist in alkaline conditions, the sum of exchangeable K, Na, Ca and Mg can be used as an estimate of cation exchange capacity (CEC). Total metals were extracted by digesting a sample with a mixed HCl/HNO3 /HClO4 /HF solution. The concentration of Ca and Mg in all the extracts was determined by AAS. Mineral composition was determined using a Bruker D8 ADVANCE X-ray diffractometer. The Materials Data Inc. software Jade 5.0 was used for phase identification. Semi-quantitative phase analysis was performed using the computer program PCPDFWIN (I/Icor reference intensity ratio method). The samples were also used for examination of micromorphological characteristics by A FEI-XL30 environmental scanning electron microscope coupled with energy dispersive X-ray spectrometer (ESEM/EDS) and for thermogravimetric analysis (TGA) using a Netzsch TG 209 F1. TGA was conducted in nitrogen atmosphere from 20 to 900 ◦ C with a heating rate of 10 ◦ C min−1 . A successive water extraction procedure was used to estimate the water-extractable alkalinity in the red mud samples. Five grams of oven-dried red mud were repeatedly extracted by 25 mL of deionized water (shaking for 1 h in a rotary shaker) until the pH of the extract became stable. For this study, a total of 113 extractions were performed for each red mud type. Each extract was titrated with a standardized HCl solution to pH 5.5 in order to determine the water-extractable alkalinity. The total water-extractable alkalinity of the sample was then obtained by summing the values of the 113 extracts. The red mud residues from the successive water extraction were then further extracted successively by 25 mL of 0.2 mol L−1 HCl solution until no pH increase was observed upon addition of the HCl solution. For this study, 133, 126 and 60 acid extractions were required for the Zhengzhou (ZZ) red mud, Guizhou red mud (GZ) and Gladstone (GS) red mud, respectively. Each extract was titrated with either a standardized 0.2 mol L−1 HCl or a standardized 0.2 mol L−1 NaOH solution to pH 5.5, depending on the pH of the extract. The amount of alkali in each extract was then calculated based on the actual volume of the 0.2 mol L−1 HCl solution that was consumed during the titration. The summed value of all the extracts for each red mud type was defined here as total retained alkalinity. The sum of total water-extractable alkalinity and total retained alkalinity is defined as total alkalinity.

2.3. QC/QA procedure The extraction experiments were performed in triplicate. Repeatability analysis showed that the mean relative standard deviation (RSD) of replicate samples was <5% for both the successive water extraction and the successive acid extraction.

56

C. Si et al. / Journal of Hazardous Materials 244–245 (2013) 54–59

Maximum carbon sequestering capacity (Cmax)

Potential additional carbon sequestration capacity (Cpot)

Quick Cpot

Slow Cpot

Exploited carbonation percentage (ECP)

Actual carbon storage (Cact)

Fig. 1. A schematic flow chart showing the relationships among the parameters established for evaluation of carbon sequestration in red mud.

2.4. Evaluation parameters for carbon sequestration For assessing various carbon sequestration characteristics, the following parameters are established: 1. Maximum carbon sequestering capacity (Cmax ), which was calculated from the total alkalinity (At ) of a red mud based on the relationship that 2 moles of OH− have the potential to capture 1 mole of CO2 in the form of CO3 2− (Equations (1) and (2)) Cmax (mol/kg) =

At 2

2. Actual carbon storage (Cact ), which was estimated from the total carbon concentration of red mud, as measured by LECO CNS Analyzer. 3. Exploited carbonation percentage (ECP), which was calculated using the following formula: ECP(%) =

Table 1 A comparison of total carbon concentration and other chemical parameters among the three selected red mud samples. Parameter

ZZ

GZ

GS

pH EC (dS m−1 ) Total carbon (mmol kg−1 ) Retained alkalinity (mmol kg−1 ) Total Ca (mmol kg−1 ) Soluble K (mmol kg−1 ) Soluble Na (mmol kg−1 ) Soluble Ca (mmol kg−1 ) Soluble Mg (mmol kg−1 ) Exchangeable K (mmol kg−1 ) Exchangeable Na (mmol kg−1 ) Exchangeable Ca (mmol kg−1 ) Exchangeable Mg (mmol kg−1 )

12.45 8.03 1933 6910 4780 19.0 89.2 1.50 0.10 70.2 346 464 2.30

11.00 0.89 3441 8920 4596 6.24 70.2 1.19 0.06 83.7 219 341 4.01

9.40 1.26 870 2060 1536 16.9 71.9 0.36 0.11 74.0 236 23.7 4.16

3.2. Variations of pH and EC in the extracts The variation of extract pH during the successive water extraction is shown in Fig. 2a. There was a general trend showing that pH decreased with increasing number of extraction. The pH variation of GZ and ZZ were similar to each other, exhibiting a sharp drop from the 1st to the 9th extract. The pH of the extract was always lower in GS than in GZ and ZZ. The initial EC value in the extract was much higher in ZZ than in GZ and GS. The extract EC for all the three red mud types rapidly decreased within the first 10 extractions and then remained at very low levels throughout the rest of the experiment (Fig. 2b).

Cact × 100 Cmax

13

(a)

12 11 10

pH

4. Potential additional carbon sequestration capacity (Cpot ), which is defined as the difference between Cmax and Cact i.e. Cpot (mol/kg) = Cmax − Cact

9

Cpot can be further divided into quick Cpot (quick Cpot ) and slow Cpot (slow Cpot ). Quick Cpot is estimated from the total waterextractable alkalinity (Aw ) i.e.

7

It must be realized that the quick Cpot obtained using the above formula is based on the assumption that the supply of divalent cations is not limiting. The theoretical basis for and limitations of this method are discussed in the supplementary method section of the Supporting Information document. The slow Cpot is defined as the difference between Cpot and quick Cpot . The relationships among these parameters are illustrated in Fig. 1. 3. Results 3.1. Total carbon and other chemical characteristics Comparison of the total carbon concentration and other chemical parameters between the three selected red mud types is provided in Table 1. There was marked difference in the total carbon, pH, EC, retained alkalinity, total Ca and exchangeable Ca between the three types of red mud. Total carbon and retained alkalinity showed the following decreasing order GZ > ZZ > GS while the pH, total Ca and exchangeable Ca all displayed the following decreasing order: ZZ > GZ > GS. Either water-soluble Ca or Mg only accounted for a very small proportion of the total water-soluble basic cations.

GZ

GS

1

16

31

46

61

76

91

106

10 8

EC (dS/m)

Aw 2

ZZ

6

ZZ

GZ

(b)

GS

6 4 2 0 1

16

31

46

61

76

91

106

3 -extractable Cumulative water alkalinity (mol/kg)

Quick Cpot (mol/kg) =

8

ZZ

GZ

(c)

GS

2

1

0 1

16

31

46

61

76

91

106

Number of water extraction Fig. 2. Diagrams showing the variation of (a) pH and (b) EC in the extracts obtained in the successive water extraction experiment, and (c) the cumulative curves of the water-extractable alkalinity for the three red mud types.

C. Si et al. / Journal of Hazardous Materials 244–245 (2013) 54–59 Table 2 Major mineral types (identified by XRD) in the three red mud types. Minerals identified (in decreasing order)

ZZ

Perovskite (CaTiO3 ), Calcite (CaCO3 ), Vaterite (␮-CaCO3 ); Magnetite (FeO·Fe2 O3 ), Gibbsite (␥-Al(OH)3 ) Calcite (CaCO3 ), Perovskite (CaTiO3 ), Monohydrocalcite (CaCO3 ·H2 O), Magnetite (FeO·Fe2 O3 ) Boehmite (␥-AlO(OH), Quartz (SiO2 ), Larnite (Ca2 SiO4 ), Calcite (CaCO3 ), Magnetite (FeO·Fe2 O3 ), Perovskite (CaTiO3 ), Gibbsite (␥-Al(OH)3 ), Sodalite (Na8 Al6 Si6 O24 Cl2 ), Anatase (TiO2 )

GZ GS

3.3. Cumulative water-extractable alkalinity Fig. 2c compares the cumulative curves of water-extractable alkalinity among the three selected red mud samples. The total water-extractable alkalinity was in the following decreasing order: ZZ (2.28 mol kg−1 ) > GZ (1.39 mol kg−1 ) > GS (0.50 mol kg−1 ). 3.4. Thermal gravimetric analysis The TGA diagrams of the three red mud types all showed the presence of a weight loss band occurred in the range of 630–770 ◦ C, corresponding to release of CO2 during calcination of calcite (Supplementary Fig. S1a-c). The temperature required for the maximum weight loss was in the following decreasing order: GZ > ZZ > GS. 3.5. SEM observation The SEM images of the three selected red mud types are provided in Supplementary Fig. S2a-c. Different micromorphological characteristics were observed. In particular, the GZ red mud contained substantial amounts of poorly crystallized materials. 3.6. XRD analysis The mineralogical composition of the three red mud types differed markedly from each other. Calcium carbonates were the only carbon-bearing minerals detected by the XRD method and the content of the calcium carbonate was in the following decreasing order GZ > ZZ > GS. The GS red mud only contained calcite while the other two red mud types showed the presence of other mineral forms of calcium carbonates in addition to calcite; the GZ red mud contained monohydrocalcite and the ZZ red mud contained vaterite. Perovskite was present for all the three red mud types with the following decreasing order: ZZ > GZ > GS. The GS red mud contained substantial amounts of larnite (Ca2 SiO4 ) and quartz (SiO2 ), while these two minerals were not detected from either the ZZ red mud or the GZ red mud (Table 2). 4. Discussion The maximum carbon sequestering capacity (Cmax ) based on the calculation from the total alkalinity was much greater than the amount of carbon that was actually stored (i.e. actual carbon storage or Cact ) in all the three investigated red mud types (Fig. 3). The exploited carbonation percentage (ECP) was 41.9%, 66.7% and 68.1% for the ZZ red mud, GZ red mud and GS red mud, respectively. The carbonate minerals consisted of calcite, vaterite and monohydrocalcite, suggesting that the carbonate was primarily bound to Ca. Magnesium carbonates, if any, must be present at a small amount since no MgCO3 type minerals were identified from the XRD analysis. The red mud samples had very low concentrations of soluble Ca and Mg (Table 1), which limited further formation of any Ca or Mg carbonates through the chemical reaction described in Equation 3.

6000

5000

Carbon (mmol/kg)

Red mud type

57

4000 Cmax 3000

Cact Slow Cpot Quick Cpot

2000

1000

0 ZZ

GZ

GS

Fig. 3. Comparison of maximum carbon sequestering capacity (Cmax ), actual carbon storage (Cact ), slow potential carbon sequestration capacity (slow Cpot ), and quick potential carbon sequestration capacity (quick Cpot ) among the three different red mud types.

Substantial amounts of Ca-containing perovskite were detected in all the three red mud types. In the GS red mud, larnite was also present. This suggests that titanate ion (TiO3 2− ) and silicate ion (SiO4 2− ) in the red mud could strongly compete with carbonate ion for the available Ca to form calcium minerals. The occurrence of larnite in the GS red mud can be attributed to its high-silicon content, as indicated by the presence of significant amounts of quartz in this red mud sample. The combined sintering-Bayer process used in both the Zhengzhou and Guizhou refineries removes silicon from the solid residue prior to the discharge of red mud into the storage facility. Consequently, the ZZ red mud and the GZ red mud contained much less silicon, as compared to the GS red mud. Magnetite was present in all the three red mud types. Since magnetite is not common in bauxite ores, it is likely that the red mud-borne magnetite was generated during the process of alumina refining. Grafe et al. [15] attributed the formation of magnetite to the sintering process for extraction of diaspora- and boehmite-type bauxite ores. However, magnetite was also detected from the GS red mud that was produced by the Bayer process. Whatever the exact cause might be, the presence of magnetite suggests that part of the ferric Fe was reduced to ferrous Fe during the course of bauxite processing. However, such Fe(II) was not available for the formation of carbonate-bearing siderite (FeCO3 ), which was probably only metastable to magnetite [16]. The much lower Cmax in the GS red mud than in the other two red mud types was attributable to the seawater treatment process, which removes substantial amounts of soluble alkalis from the red mud prior to the discharge of red mud into the disposal sites [17]. Larnite in the GS red mud may act as a source of Ca for future carbonation [18] through carbonic acid-driven weathering: Ca2 SiO4(s) + 2H2 CO3(aq) → 2CaCO3(s) + SiO2(s) + 2H2 O G◦ )

(4) 25 ◦ C

The free energy of the above reaction (r at is −150 kJ mol−1 , suggesting that the reaction is likely to proceed. It has been shown that artificial silicates (including cement materials containing larnite) could undergo carbonation under ambient temperature and pressure [19,20]. However, such chemical reactions may be kinetically very slow under ambient conditions [21], making this carbonation pathway insignificant since, within a red mud storage facility, a given volume of red mud is usually only exposed to the atmosphere for a limited period of time before it is covered by a new layer of red mud. Therefore, any attempt to use red mud-borne calcium and/or magnesium silicates as a source of Ca and Mg for carbonation requires a purpose-built processing facility to catalyze the reaction [4,22–26]. Yang and Xiao [27]

58

C. Si et al. / Journal of Hazardous Materials 244–245 (2013) 54–59

showed the formation of portlandite from partial hydrolysis of larnite during the process of red mud brick making. The portlandite may then react with CO2 to form calcite. Perovskite is thermodynamically metastable with respect to TiO2 [28] and therefore can react with carbonic acid to form calcium carbonate minerals according to the following equation: CaTiO3(s) + H2 CO3(aq) → CaCO3(s) + TiO2(s) + H2 O(r G◦ = −56 kJ mol−1 )

(5)

For the same reason as the larnite, industrial processes are required to catalyze the above-mentioned reaction in order to make it technically feasible. Apart from chemical reactions shown in Equations (4) and (5), the protonation of Al(OH)4 − may also contribute to the slow potential additional carbon sequestration capacity (slow Cpot ) according to the following chemical equations: 2Al(OH)4 − (aq) + H2 CO3(aq) → 2Al(OH)3 0 (s) + 2H2 O + CO3 2− (aq) (r G◦ = −44 kJ mol−1 )

(6)

Immobilization of CO3 2− generated from Equation 6 requires the presence of sufficient amount of soluble divalent cations in the red mud. Besides, carbonic acid tends to preferentially react with free OH− that constitutes the quick potential carbon sequestration capacity (quick Cpot ). Therefore, the reaction in Equation (6) may only proceed after the quick Cpot is depleted. Enhancement of CO2 sequestration by using red mud’s quick Cpot is relatively straightforward. It only requires sufficient supply of divalent binding metals to the reaction (Equation (3)). The consumption of quick Cpot simultaneously reduces the causticity of the red mud, a major hazardous issue that requires costly management. Where the soluble Ca- or/and Mg-bearing substances such as gypsum and concentrated brines (e.g. bitterns) are readily available at low cost, mixing such materials with red mud prior to its disposal to allow carbonation to take place (Equation (3)) could be a cost-effective approach for simultaneous CO2 sequestration and causticity reduction of the red mud. Apart from the soluble Caor/and Mg-bearing substances, ferrous sulfate (FeSO4 ) may also be used to form siderite: FeSO4(aq) + H2 CO3(aq) + 2OH− (aq) → FeCO3(s) + SO4 2− (aq) + H2 O (7) However, ferrous ion is subject to rapid oxidation under high pH conditions [29]. Therefore, anaerobic conditions need to be created to promote the reaction (Equation (7)). Water inundation, coupled with addition of organic matter to red mud may be a possible approach for achieving such a goal. It has been demonstrated that some microorganisms were able to colonize in red mud by decomposing the added organic matter [30]. The microbial processes can drive the reduction of Fe(III) in the red mud [31], which may enhance the formation of siderite. Due to the removal of soluble alkalinity by seawater washing, the GS red mud is likely to represent a red mud type with less additional carbon sequestration capacity. This was in contrast with the GZ red mud and the ZZ red red, which contained 4.1% and 2.3% of inorganic carbon, respectively. The higher carbon content in the Chinese red mud can be attributed to the abundant presence of calcium in the red mud as a result of addition of lime during bauxite processing. The total storage of red mud in China was estimated to be over 100 million tonnes as in 2008 and the current annual production of red mud in China is around 30 million tonnes [13]. It is therefore likely that, to date, over 19 million tonnes of CO2 have been sequestered by the Chinese red mud alone if we assume that the red mud has an average carbon content of 3%. This could largely

compensate for the emission of CO2 from the calcination of limestone for production of lime that was used for bauxite processing. The low exploited carbonation percentage suggests that a large proportion of the carbon sequestration capacity was not used for the historical red mud. If we conservatively estimate that the red mud has an average Cpot of 15 kg of carbon per tonne, the annual amount of CO2 to be sequestered will be around 1.5 million tonnes if appropriate treatment methods are in place to ensure adequate supply of binding divalent metals for red mud carbonation. In contrast with the Chinese red mud, most of red mud types around the world have much lower calcium concentration [15]. This could limit the carbon sequestration potential in these red mud types. If we assume the average carbon content is, globally, 1% for red mud (excluding the Chinese one), the total amount of CO2 that has been sequestered by the red mud outside China could be over 90 million tonnes worldwide. The amount of CO2 sequestered in red mud could therefore exceed 100 million tonnes. It is worthwhile to note that, for the low-Ca red mud outside of China, the exploited carbonation percentage could be lower. Therefore, the red mud in other parts of the world is likely to have a larger potential additional carbon sequestration capacity, as compared to the Chinese red mud. If we apply the same rate of Cpot in the Chinese red mud for the global-scale estimation, approximately 6 million tonnes of CO2 may be sequestered through enhanced red mud carbonation in addition to another 6 million tonnes of CO2 that are captured and stored in the red mud through naturally occurring processes. The global production rate of alumina and therefore red mud is still increasing [1]. Development of treatment methods for enhanced red mud carbonation is a research priority considering the dual environmental benefits to be brought together through the applications of such technologies: minimization of environmental hazards and generation of carbon credits. 5. Conclusion The differences in bauxite type, alumina refining method and the red mud disposal method could markedly affect carbon sequestration in red mud. For the three investigated red mud types, calcium carbonates were the dominant mineral phases responsible for the carbon sequestration. The carbon sequestration capacity of the red mud was not fully used due to unavailability of soluble divalent cations for the formation of stable carbonate minerals. It was likely that titanate and silicate ions strongly competed with carbonate ions for the available soluble Ca in these red mud types. Therefore, supply of additional soluble Ca and Mg could be a viable pathway for maximizing carbon sequestration in red mud and simultaneously reducing the causticity of red mud. It is roughly estimated that globally over 100 million tonnes of CO2 have been sequestered in red mud from the late 19th century to date. Based on the current production rate of red mud, it is likely that some 6 million tonnes of CO2 will be sequestered annually through atmospheric carbonation. If appropriate technologies are in place for incorporating binding cations into red mud, at least 6 million tonnes of additional CO2 can be sequestered by red mud while the alkalinity of red mud is simultaneously reduced. Therefore, development of treatment methods for enhanced red mud carbonation is a research priority considering the dual environmental benefits to be brought together through the applications of such technologies: minimization of environmental hazards and generation of carbon credits. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/ j.jhazmat.2012.11.024.

C. Si et al. / Journal of Hazardous Materials 244–245 (2013) 54–59

References [1] G. Power, M. Gräfe, C. Klauber, Bauxite residue issues: I. Current management, disposal and storage practices, Hydrometallurgy 108 (2009) 33–45. [2] C. Lin, G. Maddocks, J. Lin, G. Lancaster, C. Chu, Acid neutralizing capacity of two bauxite residues and their potential applications for treating acid sulfate soil and water, Aust. J. Soil Res. 42 (2004) 649–657. [3] P. Castaldi, M. Silvetti, G. Garau, S. Deiana, Influence of the pH on the accumulation of phosphate by red mud (a bauxite ore processing waste), J. Hazard. Mater. 182 (2010) 266–272. [4] R.C. Sahu, R.K. Patel, B.C. Ray, Neutralization of red mud using CO2 sequestration cycle, J. Hazard. Mater. 179 (2010) 28–34. [5] T.C. Santini, C. Hinz, A.W. Rate, C.M. Carter, R.J. Gilkes, In situ neutralisation of uncarbonated bauxite residue mud by cross layer leaching with carbonated bauxite residue mud, J. Hazard. Mater. 194 (2011) 119–127. [6] P. Renforth, W.M. Mayes, A.P. Jarvis, I.T. Burke, D.A.C. Manning, K. Gruiz, Contaminant mobility and carbon sequestration downstream of the Ajka (Hungary) red mud spill: The effects of gypsum dosing, Sci. Total Environ. 421–422 (2012) 253–259. [7] V.S. Yadav, M. Prasad, J. Khan, S.S. Amritphale, M. Singh, C.B. Raju, Sequestration of carbon dioxide (CO2 ) using red mud, J. Hazard. Mater. 176 (2010) 1044–1050. [8] D. Bonenfant, L. Kharoune, S. Sauve, R. Hausler, P. Niquette, M. Mimeault, M. Kharoune, CO2 sequestration by aqueous red mud carbonation at ambient pressure and temperature, Ind. Eng. Chem. Res. 47 (2008) 7617–7622. [9] Y. Liu, C. Lin, Y. Wu, Characterization of red mud derived from a combined Bayer process and bauxite calcination method, J. Hazard. Mater. 146 (2007) 255–261. [10] K.S. Lackner, Capture of carbon dioxide from ambient air, Eur. Phys. J. Special Topics 176 (2009) 93–106. [11] P. Smith, The processing of high silica bauxites — Review of existing and potential processes, Hydrometallurgy 98 (2009) 162–176. [12] S. Gu, Z. Yin, Z. Yang, W. He, Improvements of the processes to produce alumina from Chinese diaspore bauxite, Light Met. (2003) 153–158. [13] W.C. Liu, J.K. Yang, B. Xiao, Review on treatment and utilization of bauxite residues in China, Int. J. Miner. Process. 93 (2009) 220–231. [14] L. McAuthur, C. Greensill, Mineralogical analysis of Weipa Bauxite using NIR spectroscopy. Australian Institute of Physics 17th National Congress 2006, Brisbane December 3-8, 2006. [15] M. Gräfe, G. Power, C. Klauber, Bauxite residue issues: III. Alkalinity and associated chemistry, Hydrometallurgy 108 (2011) 60–79. [16] S. Yur, Decomposition of siderite to magnetite at lower oxygen fugacities: a thermochemical interpretation and geological implications, Econ. Geol. 61 (1966) 768–776.

59

[17] N.W. Menzies, I.M. Fulton, W.J. Morrell, Seawater neutralization of alkaline bauxite residue and implications for revegetation, J. Environ. Qual. 33 (2004) 1877–1884. ˜ [18] A. Santos, M. Ajbary, V. Morales-Flórez, A. Kherbeche, M. Pinero, L. Esquivias, Larnite powders and larnite/silica aerogel composites as effective agents for CO2 sequestration by carbonation, J. Hazard. Mater. 168 (2009) 1397–1403. [19] P. Renforth, D.A.C. Manning, E. Lopez-Capel, Carbonate precipitation in artificial soils as a sink for atmospheric carbon dioxide, Appl. Geochem. 24 (2009) 1757–1764. [20] P. Renforth, D.A.C. Manning, Laboratory carbonation of artificial silicate gels enhanced by citrate: Implications for engineered pedogenic carbonate formation, Int. J. Greenh. Gas Con. 5 (2011) 1578–1586. [21] P.V. Brady, S.A. Carroll, Direct effects of CO2 and temperature on silicate weathering: Possible implications for climate control, Geochim. Cosmochim. Acta 58 (1994) 1853–1856. [22] F. Dufaud, I. Martinez, S. Shilobreeva, Experimental study of Mg-rich silicates carbonation at 400 and 500 ◦ C and 1 kbar, Chem. Geol. 265 (2009) 79–87. [23] M. Kakizawa, A. Yamasaki, Y. Yanagisawa, A new CO2 disposal process via artificial weathering of calcium silicate accelerated by acetic acid, Energy 26 (2001) 341–354. [24] M.F. Bertos, S.J.R. Simons, C.D. Hills, P.J. Carey, A review of accelerated carbonation technology in the treatment of cement-based materials and sequestration of CO2 , J. Hazard. Mater. 112 (2004) 193–205. [25] C. Shi, J. Xu, E. Beckman, R. Enick, Carbon dioxide sequestration via pH reduction of red mud using liquid CO2 , ACS Div. Fuel Chem. 45 (2005) 703–705. [26] S. Khaitan, D.A. Dzombak, V.L. Gregory, Mechanisms of neutralization of bauxite residue by carbon dioxide, J. Environ. Eng. 135 (2009) 433–438. [27] J. Yang, B. Xiao, Development of unsintered construction materials from red mud wastes produced in the sintering alumina process, Constr. Build. Mater. 22 (2008) 2299–2307. [28] H.W. Nesbitt, M.G. Bancroft, W.S. Fyfe, S.N. Karkhanis, A. Nishijima, S. Shin, Thermodynamic stability and kinetics of perovskite dissolution, Nature 289 (1981) 358–362. [29] W. Stumm, G.F. Lee, Oxygenation of Ferrous Iron, Ind. Eng. Chem. 53 (1961) 143–146. [30] M. Hamdy, F. Williams, Bacterial amelioration of bauxite residue waste of industrial alumina plants, J. Ind. Microbiol. Biotechnol. 27 (2001) 228–233. [31] K.L. Straub, M. Benz, B. Schink, Iron metabolism in anoxic environments at near neutral pH, FEMS Microbiol. Ecol. 34 (2001) 181–186.