Thin-film versus slurry-phase carbonation of steel slag: CO2 uptake and effects on mineralogy

Journal of Hazardous Materials 283 (2015) 302–313

Contents lists available at ScienceDirect

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

Thin-film versus slurry-phase carbonation of steel slag: CO2 uptake and effects on mineralogy R. Baciocchi a , G. Costa a , M. Di Gianfilippo a , A. Polettini b,∗ , R. Pomi b , A. Stramazzo b a Laboratory of Environmental Engineering, Department of Civil Engineering and Computer Science Engineering, University of Rome “Tor Vergata”, Via del Politecnico 1, 00133 Rome, Italy b Department of Civil and Environmental Engineering, University of Rome “La Sapienza”, Via Eudossiana 18, 00184 Rome, Italy

h i g h l i g h t s • • • •

Batch thin-film and slurry-phase carbonation tests conducted on steelmaking slags. Max. CO2 uptakes: 280–403 g CO2 / kg slag (24-h slurry-phase carbonation). Theoretical expected sequestration potential: 2.2–4.2 Mt CO2 /y at a European level. Conversion of the original minerals into carbonates involved Ca, Mg, Fe and Mn.

a r t i c l e

i n f o

Article history: Received 10 April 2014 Received in revised form 8 September 2014 Accepted 11 September 2014 Available online 19 September 2014 Keywords: Steel slag Accelerated carbonation CO2 uptake Conversion yield Mineralogical characterization

a b s t r a c t The results of direct aqueous accelerated carbonation of three types of steel manufacturing residues, including an electric arc furnace (EAF) slag and two basic oxygen furnace (BOF) slags, are reported. Batch accelerated carbonation tests were conducted at different temperatures and CO2 pressures applying the thin-film route (liquid to solid, L/S, ratio = 0.3 L/kg) or the slurry-phase route (L/S ratio = 5 L/kg). The CO2 uptake strongly depended on both the slag characteristics and the process route; maximum yields of 280 (EAF), 325 (BOF1) and 403 (BOF2) g CO2 /kg slag were achieved in slurry phase at T = 100 ◦ C and pCO2 = 10 bar. Differently from previous studies, additional carbonates (other than Ca-based phases) were retrieved in the carbonated BOF slags, indicating that also Mg-, Fe- and Mn-containing phases partially reacted with CO2 under the tested conditions. The results hence show that the effects of accelerated carbonation in terms of CO2 uptake capacity, yield of mineral conversion into carbonates and mineralogy of the treated product, strongly rely on several factors. These include, above all, the mineralogy of the original material and the operating conditions adopted, which thus need specific case-by-case optimization to maximize the CO2 sequestration yield. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Among the CO2 capture and storage techniques, ex situ accelerated carbonation is the only option allowing permanent CO2 sequestration in short timeframes. In this process, a material rich in alkaline earth metal oxides and/or silicates is contacted with CO2 at optimized operating conditions, resulting in the formation of stable carbonate phases [1]. To maximize the CO2 uptake achievable by minerals such as olivine, serpentine and wollastonite, several reaction routes (e.g., indirect, direct gas–solid or direct aqueous) have been tested (see e.g., [2]). The CO2 sequestration capacity of several

∗ Corresponding author. Tel.: +39 06 44585037; fax: +39 06 44585037. E-mail address: [email protected] (A. Polettini). http://dx.doi.org/10.1016/j.jhazmat.2014.09.016 0304-3894/© 2014 Elsevier B.V. All rights reserved.

types of alkaline industrial residues (alkaline ashes from combustion processes [3–8], cement-based materials [9,10], cement kiln dust [7,11], steel manufacturing slag [12–22]) has also been recently evaluated. Such residues, although quantitatively limited, prove readily reactive toward CO2 at significantly milder operating conditions than Mg or Ca silicate minerals, and could hence be used to sequester part of the CO2 emitted from specific industrial facilities. The most common carbonation route for alkaline residues is the direct aqueous route (see e.g., [6,8,23]), in which the dissolution of reactive phases and the precipitation of Ca carbonates occur in a single stage. Direct aqueous carbonation studies have been performed either in slurry phase at liquid to solid (L/S) ratios of 5–50 L/kg (especially for waste materials with high silicate contents [6,13,15]) or via the thin-film route, i.e., at L/S ratios <1.5 L/ kg [3–5,7,8,11,12,23].

R. Baciocchi et al. / Journal of Hazardous Materials 283 (2015) 302–313

Among the various alkaline industrial residues available, steel manufacturing slags are particularly interesting for mineral carbonation, on account of their widespread availability as well as chemical and mineralogical composition. Steel slags are produced either during the conversion of hot iron ores into crude steel in a basic oxygen furnace (BOF slag) in plants adopting the integrated manufacturing cycle, or during the melting of iron scrap in an electric arc furnace (slag from carbon steel – EAF-C – or stainless/high alloy steel production – EAF-S) [24–26]. Secondary metallurgical slags may also be generated in further secondary processing of crude steel [24–25]. A total crude steel production of 1.51 × 106 Mt was estimated worldwide in 2012, resulting in ∼125.8 kg of BOF slag and ∼168.6 kg of EAF slag per t of crude steel produced [27] The iron and steel industry is also recognized to be one of the largest industrial CO2 emitters, with a total of 115.6 Mt CO2 emitted by the EU-27 countries in 2007 [28], and a global annual emission of 1500–1600 Mt CO2 [29]. Steel slag carbonation may contribute to the sequestration of CO2 emissions while at the same time improving the technical and environmental behavior of the material, due to the mineralogical changes of thermodynamically unstable minerals produced by the process. Undesired alterations of steel slag minerals (e.g., hydration of CaO and MgO; C2 S transformation from the ␣- to the ␥-polymorph) cause significant volume instability (see e.g., [26,30,31]) with swelling and cracking phenomena that may hinder slag utilization. Moreover, some types of steel slag may release non-negligible amounts of toxic metals (e.g., Cr, V, Mo [32]). Mineralogical alterations of steel slag resulting from carbonation lead to improved mineralogical stability and lower contaminant mobility. Despite the fairly large amount of work that has been recently conducted in the field of accelerated carbonation of steelmaking residues, the individual mechanisms involved in the process and the resulting changes in the main physical, chemical and mineralogical characteristics of the materials have not been fully elucidated yet. With the present work we aimed at filling some of the existing gaps in the knowledge of accelerated carbonation of EAF and BOF slags applying the direct aqueous route by performing both slurry-phase and thin-film tests. Specifically, CO2 sequestration yields and related mineralogical changes were investigated in detail to gain insights into the transformations occurring during the carbonation process.

2. Materials and methods Slag samples were collected from a stainless steel plant after metals removal downstream the electric arc furnace unit (EAF slag), and from a steel plant employing the integrated steelmaking process both at the outlet of the basic oxygen furnace (BOF1 slag) and at the storage site after metals recovery (BOF2 slag). The samples were milled to pass the 150 ␮m sieve. The size distribution after milling was measured through laser particle size analysis (BT-9300S analyser). Slag characterization included the determination of the elemental composition, carbonate content and mineralogy of the tested samples. The elemental composition was determined by alkaline digestion of triplicate slag samples with Li2 B4 O7 in platinum melting pots at 1050 ◦ C for 2 h, followed by dissolution of the molten material with 10% HNO3 and analysis of element concentrations by atomic absorption spectrometry. Carbonate measurements were made in triplicate with a Shimadzu inorganic carbon (IC) analyser. Mineralogy was analyzed by powder X-ray diffraction (XRD) with Cu K␣ radiation using a Philips Expert Pro diffractometer equipped with a copper tube operated at 40 kV and 40 mA. Fourier-transform infrared spectroscopy (FT-IR) analyses were performed using an

303

Agilent 640 IR instrument operating in the mid-IR range. Scanning electron microscopy (SEM) analyses of polished thin sections were conducted with a GEMINI Supra 40 (Carl Zeiss) instrument. Direct aqueous accelerated carbonation tests were performed at L/S = 0.3 L/kg (thin-film process: humidified material in contact with the gaseous phase) and L/S = 5 L/kg (slurry-phase treatment: aqueous suspension of the slag contacted by the gaseous phase). While the thin-film process was operated under relatively mild conditions to assess the applicability of a low energy-intensive carbonation treatment, the slurry-phase route was intended as a more severe treatment to assess the maximum practicable CO2 sequestration yield. The values of pressure, temperature and L/S ratio adopted during the experiments were based on previous investigations [3–5,12,15]. The treatment was performed in a pressurized stainless steel reactor at 50 ◦ C (thin-film route) or 100 ◦ C (slurry-phase route), 100% CO2 atmosphere and 10 bar pressure for 0.5–24 h. After carbonation the samples were oven-dried at 105 ◦ C and analyzed to determine their carbonate content and mineralogy (through XRD, FT-IR and SEM analyses), as specified above. The process performance in terms of CO2 uptake was calculated through Eq. (1): CO2 uptake (%) =

CO2final (%) − CO2initial (%) × 100 100 − CO2final (%)

(1)

where CO2 uptake is expressed as g CO2 sequestered by 100 g of as-received dry slag, while CO2final and CO2initial are the CO2 weight percent contents of carbonated and untreated samples as derived from their respective IC contents. The conversion yield of reactive elements into carbonate forms, , was also evaluated according to Eq. (2): (%) =



CO2 uptake(%)/44



Mein,k (%)/AWk − CO3in (%)/60

× 100

(2)

k

where Mein (Me generically indicating Ca, Fe, Mg or Mn) and CO3in are the initial weight percent contents of metals and carbonate in the as-received untreated dry slag and AW is the atomic weight of Ca, Fe, Mg or Mn. In the present work, differently from our previous research [4,5] and other literature studies (e.g., [14,15,22,33,34]), the assumption that Ca only contributed to the formation of carbonate minerals appeared unreasonable, and explicit evidence was gained from XRD analyses (see below) of the conversion of Fe-, Mg- and Mn-bearing phases into carbonates. Eq. (2) is based on the implicit assumption that all the formed carbonate minerals had an equimolar Me/CO3 ratio. 3. Results and discussion 3.1. Materials characterization The main statistical parameters (10th, 50th and 90th percentiles, volume and surface mean diameters) of the grain size distribution of the milled materials are reported in Table 1. Although using identical milling conditions, the final grain size distributions were notably different, due to different initial dimensional characteristics and mechanical properties of the particles. The BOF2 slag displayed the largest amount of fine particles, with a final median diameter of 14.3 ␮m and a D90 of 50.2 ␮m. Conversely, the BOF1 slag contained significant proportions of larger particles (D90 = 208 ␮m); while physical aggregation may be excluded since a dispersant was used in the analysis, this finding may likely arise from particle hydration or cementation upon contact with the liquid medium used in the analysis. The actual grain size of the material when it comes in contact with water is important with a

304

R. Baciocchi et al. / Journal of Hazardous Materials 283 (2015) 302–313

Table 1 Representative parameters of the size distribution curves for the three slag samples. Slag sample

D10 (␮m)

D50 (␮m)

D90 (␮m)

Volume mean diameter, D [4,3] (␮m)

Surface mean diameter, D [3,2] (␮m)

BOF1 BOF2 EAF

7.3 2.2 6.1

78.6 14.3 33.9

208.0 50.2 108.1

99.4 21.5 46.5

15.4 4.7 10.0

Fig. 1. Elemental composition of the slags.

view to interpreting the behavior observed in aqueous carbonation tests. The chemical composition of the EAF and BOF slags is reported in Fig. 1. Ca displayed significantly higher concentrations (21–23% for the BOF samples and 35% for the EAF slag) than Mg (3.3–3.8% for the BOF samples and 2.4% for the EAF slag). Other elements at appreciable concentrations in the BOF slags were (in descending order): Fe (20–25%), Si (4–6%), Mn (2.5–3.8%), Al (0.7–1.1%), V (7800–9300 mg/kg) and Cr (1300–1800 mg/kg). In the EAF slag such elements were present at different concentrations, namely: Si (13%), Cr (3.7%), Fe (3.2%), Al (2%) and V (970 mg/kg). In addition to Cr and V, other metals of potential environmental concern in the BOF slags included Zn (35–88 mg/kg), Pb (∼60 mg/kg) and Cu (8–16 mg/kg). For the EAF slag, Ni was the most abundant element (480 mg/kg), followed by Mo (287 mg/kg), Zn (260 mg/kg), Cu (171 mg/kg) and Pb (91 mg/kg). The initial carbonate content of the material was 0.8%, 6.7% and 3.5% for the BOF1, BOF2 and EAF slag, respectively. The XRD results (Fig. 2) show that the main phases in the BOF1 sample included silicate and oxide phases: tricalcium (hatrurite) and dicalcium (larnite) silicate, a mixed Ca–Al–Fe oxide (Ca2 Fe1.4 Al0.6 O5 ), a mixed Ca–Cr–Fe oxide (Ca2 Cr0.5 Fe1.5 O5 ), wüstite (FeO), magnetite (FeO·Fe2 O3 ), corundum (Al2 O3 ) and traces of portlandite. The BOF2 sample displayed a marked presence of portlandite and smaller amounts of silicate and oxide minerals, including Mg silicate, wüstite, mixed oxides ((MgO)0.725 (MnO)0.275 ; Ca2 Fe1.4 Al0.6 O5 ; Ca2 Cr0.5 Fe1.5 O5 ), calcite and magnetite. The presence of portlandite was also indicated (see below) by a sharp FT-IR band at ∼3640 cm−1 , commonly associated to hydroxyl stretching [35]. The presence of portlandite and the higher initial carbonate content in BOF2 indicate the occurrence of hydration and natural carbonation during storage at the disposal site. For the EAF slag, silicates (dicalcium silicate, gehlenite [Ca2 Al(AlSi)O7 ], akermanite [Ca2 MgSi2 O7 ], cuspidine [Ca4 Si2 O7 (F,OH)2 ]), oxide phases (periclase [MgO], Ca–Al oxide, Cr oxides [CrO, MgCr2 O4 ] and magnetite) and fluorite (CaF2 ) were detected.

Fig. 2. XRD spectra for the three untreated slag samples. Legend: (a) portlandite [Ca(OH)2 ]; (b) brucite [Mg(OH)2 ]; (c) calcite [CaCO3 ]; (d) larnite [Ca2 SiO4 ]; (e) magnetite [FeO·Fe2 O3 ]; (f) Ca12 Al14 O33 ; (g) hatrurite [Ca3 SiO5 ]; (h) Ca–Fe–Al oxide [Ca2 Fe1.4 Al0.6 O5 ]; (i) Mg–Cr oxide [MgCr2 O4 ]; (j) Cr oxide [CrO]; (k) corundum [Al2 O3 ]; (l) Ca–Cr–Fe oxide [Ca2 Cr0.5 Fe1.5 O5 ]; (m) wüstite [FeO]; (s) Mg silicate [MgSiO3 ]; (t) Mg–Mn oxide [(MgO)0.725 (MnO)0.275 ]; (u) periclase [MgO]; (v) akermanite [Ca2 MgSi2 O7 ]; (w) gehlenite [Ca2 Al(AlSi)O7 ]; (x) cuspidine [Ca4 Si2 O7 (F,OH)2 ]; (y) fluorite [CaF2 ].

3.2. Carbonation performance The CO2 sequestration and the conversion yields attained in the carbonation experiments, calculated through Eqs. (1) and (2) as illustrated in Section 2, are summarized in Fig. 3(a)–(d). The amount of sequestered CO2 largely differed between the two process routes, with uptakes after 24 h of 17.6 (EAF) – 20.9% (BOF2) and 28.0 (EAF) – 40.3% (BOF2) for the thin-film and slurry processes, respectively. The slurry-phase conditions likely favored CO2 solvation into the liquid phase, dissolution of reactive metals from the

R. Baciocchi et al. / Journal of Hazardous Materials 283 (2015) 302–313

305

Fig. 3. CO2 uptake (a and b) and conversion yield (c and d) in thin-film and slurry-phase carbonation. Continuous lines indicate the theoretical curves represented by Eq. (4).

slag matrix, or prevented coating of the unreacted slag particles by the precipitated carbonate product (see e.g., [11,15,20,36,37]). Despite the highest Ca content, the EAF sample exhibited the lowest CO2 sequestration capacity, likely due to its different mineralogy. It should be noted that the CO2 uptake resulting for the EAF slag in the slurry phase-treatment was significantly higher (23 versus 14% after 4 h) than the one previously obtained [5]. It is hypothesized that at the lower L/S ratio adopted in the present study the ionic strength of the solution increased and so did the solubility of reactive phases [15]. On the basis of the physical, chemical and mineralogical properties of the tested materials, it is also inferred that the CO2 sequestration yields attained are, for given carbonation conditions, the combined result of their mineralogical composition and grain size characteristics, however the available data does not allow to separate their individual effects. As for the conversion yield (see Fig. 3(c) and (d)), the maximum observed values were 38.6 (BOF1), 46.7 (BOF2), 39.8% (EAF) for the thin-film process, and 62.3 (BOF1), 90.0 (BOF2), 63.5% (EAF) for the slurry-phase route. For comparison purposes only, the conversion yield was calculated with Eq. (2) also for the EAF slag, although for this slag sample there was no evidence of Fe, Mg and Mn conversion into carbonates. If the Ca content only were considered for the EAF slag, conversion yields of 47% and 75% would result for the wet and slurry processes, respectively. The high conversion yields attained in the slurry-phase process, especially for the BOF2 sample, indicate, for long treatment times, virtually complete carbonation even under relatively mild conditions and also suggest the opportunity of exploring the process performance at lower temperatures and/or CO2 pressures. For comparison purposes, Table 2 shows the results of steel slag carbonation tests obtained in the present work and in

previous studies, reported in terms of the operating conditions that showed to maximize the CO2 uptake yield. The data show that the CO2 sequestration yields obtained in the present study for both the EAF and the BOF slags are comparable or even considerably higher than the best results documented by previous literature studies: in particular, for the two slag types our best CO2 uptake results (EAF: 28.0%; BOF: 40.3%) should be compared with the maximum reported performances of 30.7 and 33.1%, respectively. While the maximum carbonation yields attained in the present study were obtained with a residence time of 24 h which may be judged to be hardly practicable in view of full-scale implementation of the process, our results also show that lower process durations can still ensure appreciable CO2 uptake yields; for instance, the 1-h sequestration capacity was shown to account for 46–82% (thin-film carbonation) and 60–73% (slurry-phase carbonation) of the 24-h yield. It should also be mentioned that the results reported in Table 2 were obtained under notably different process configurations and operating conditions, which may result in different technical implications, energy requirements and related costs. Hence, in order to judge the practical feasibility of the proposed CO2 sequestration process based on accelerated carbonation of steel slag, an energetic and economic assessment of each process configuration, taking into account also the characteristics of the specific type of tested slag, as well as the operating conditions applied in each treatment step, should be conducted. The assessment of the energy and material requirements of the two process routes examined in this study applied to different types of industrial residues was specifically addressed in another paper [41]. The total energy requirement resulting for the slurryphase process, considering the operating conditions discussed in the present study (T = 100 ◦ C; P = 10 bar; L/S = 5 L/kg), a carbonation time of 1 h, and applying the approach described in detail in

306

R. Baciocchi et al. / Journal of Hazardous Materials 283 (2015) 302–313

Table 2 Summary of the best carbonation conditions obtained in this and previous studies. Material type

Material properties

Carbonation route

Reactor type

Gas properties

Operating conditions maximizing CO2 uptake

Carbonation performance

Ref.

BOF

CaO = 31.7% MgO = 6.0% d < 38 ␮m

Direct aqueous carbonation (slurry)

Pressurized stirred batch reactor

100% vol. CO2

Ca = 74% CO2uptake = 15.6%

[15]

EAF

CaO = 32.8% MgO = 10% d = 38–106 ␮m CaO = 58.1% MgO = 6.20% d = 38–106 ␮m CaO = 32.1% MgO = 9.4% d = 150–250 ␮m

Direct aqueous carbonation (slurry) Direct aqueous carbonation (slurry) Direct aqueous carbonation (slurry)

Stirred batch reactor

15% vol. CO2 Q = 5 mL/min

Ca = 6.6%a CO2uptake = 1.7%

[13]

Stirred batch reactor

15% vol. CO2 Q = 5 mL/min

Ca = 54.1%a CO2uptake = 24.7%

[13]

Batch reactor

100% vol. CO2

Ca = 12% CO2uptake = 3.0%a

[17]

LSb (Al-killed)

CaO = 49.9% MgO = 4.3% d = 45–75 ␮m

Direct aqueous carbonation (slurry)

Batch reactor

100% vol. CO2

Ca = 10% CO2uptake =3.9%a

[17]

EAF + AOD

CaO = 70.0% MgO = 7.4% d < 105 ␮m CaO = 49.3% MgO = 4.1% d < 105 ␮m

Direct aqueous carbonation (thin-film) Direct aqueous carbonation (thin-film)

Pressurized batch reactor

100% vol. CO2

Ca = 23.6%a CO2uptake = 13.0%

[12]

Pressurized batch reactor

100% vol. CO2

Ca = 47.0%a CO2uptake = 18.2%

[4]

CaO = 56.5% MgO = 2.8% d < 105 ␮m CaO = 49.3% MgO = 4.1% d < 105 ␮m CaO = 36.12% MgO = 18.98% d < 100 ␮m

Direct aqueous carbonation (thin-film) Direct aqueous carbonation (thin-film) Direct aqueous carbonation (slurry)

Pressurized batch reactor

100% vol. CO2

Ca = 69.2%a CO2uptake = 30.7%

[4]

Pressurized batch reactor

100% vol. CO2

Ca = 36.0%a CO2uptake = 13.9%

[5]

Stirred batch reactor

15% vol. CO2 Q = 50 L/h

Ca = 33.6%a CO2uptake = 30.7%

[22]

EAF

CaO = 26.91% MgO = 18.95% d < 100 ␮m

Direct aqueous carbonation (slurry)

Stirred batch reactor

15% vol. CO2 Q = 50 L/h

Ca = 9.0%a CO2uptake = 1.9%

[22]

LSb

CaO = 42.22% MgO = 14.99% d < 100 ␮m

Direct aqueous carbonation (slurry)

Stirred batch reactor

15%vol. CO2 Q = 50 L/h

Ca = 13.9%a CO2uptake = 4.6%

[22]

BOF

CaO = 51.11% MgO = 4.17% d < 44 ␮m

Direct aqueous carbonation (slurry)

Slurry reactor

100% vol. CO2 Q = 0.1 L/min

Ca = 72.2% CO2uptake = 29.0%a

[33]

BOF

CaO = 42.43% MgO = 9.15% d < 63 ␮m

Direct aqueous carbonation (slurry)

High gravity rotating packed bed reactor

100% vol. CO2 Q = 2.5 L/min

Ca = 93.5% CO2uptake =31.2%a

[14]

BOF

CaO = 57.8% MgO = 0.5% d < 80 ␮m

Pressurized batch reactor

60% vol. CO2

Ca = 30.2%a CO2uptake = 13.7%

[19]

BOF

CaO = 57.8% MgO = 0.5% d = 80–500 ␮m CaO = 56.8% MgO = 7.0% d = 200–63 ␮m

Direct aqueous carbonation (thin-film carbonation; 8 bar pressurized steam) Gas–solid carbonation

T = 100 ◦ C; P = 19 bar; t = 30 min; L/S = 10 T = 20 ◦ C; P = 1 atm; t = 24 h; L/S = 10 T = 20 ◦ C; P = 1 atm; t = 40 h; L/S = 10 T = 25 ◦ C; P = 1 atm; t = 70 h; L/S = 8.3 T = 25 ◦ C; P = 1 atm; t = 70 h; L/S = 8.3 T = 50 ◦ C; P = 3 bar; t = 8 h; L/S = 0.3 T = 50 ◦ C; P = 1 bar; t = 24 h; L/S = 0.4 T = 50 ◦ C; P = 10 bar; t = 8 h; L/S = 0.4 T = 100 ◦ C; P = 10 bar; t = 4 h; L/S = 0.4 T = 25 ◦ C; P = 1 atm; L/S = 100; t = 65 min T = 25 ◦ C; P = 1 atm; L/S = 100; t = 65 min T = 25 ◦ C; P = 1 atm; L/S = 100; t = 65 min T = 60 ◦ C; P = 1 atm; t = 60 min; L/S = 10 T = 65 ◦ C; P = 1 bar; t = 30 min; L/S = 20 T = 650 ◦ C; P = 20 bar; t = 30 min

Batch reactor

100% vol. CO2

Ca = 14.1%a CO2uptake = 6.4%

[19]

Direct aqueous carbonation (slurry)

Sonication stirred batch reactor

100% vol. CO2 Q = 0.24 L/min

Ca = 48.5% CO2uptake = 21.6%a

[20]

Direct aqueous carbonation (slurry)

Sonication stirred batch reactor

100% vol. CO2 Q = 0.24 L/min

Ca = 73.2% CO2uptake = 30.4%a

[20]

LSb

EAF

EAF

AOD

EAF

EAF

AOD

AOD

CaO = 52.9% MgO = 9.0% d = 63–200 ␮m

T = 650 ◦ C; P = 1 bar; t = 10 min T = 50 ◦ C; P = 1 atm; L/S = 100; t = 4 h; US = 24 kHz T = 50 ◦ C; P = 1 atm; L/S = 100; t = 4 h; US = 24 kHz

R. Baciocchi et al. / Journal of Hazardous Materials 283 (2015) 302–313

307

Table 2 (Continued ) Material type

Material properties

Carbonation route

Reactor type

Gas properties

Operating conditions maximizing CO2 uptake

Carbonation performance

Ref.

BOF

CaO = 41.15% MgO = 9.21% d < 44 ␮m

Column reactor

100% vol. CO2 Q = 1 L/min

T = 25 ◦ C; P = 1 atm; t = 120 min; L/S = 20

Ca = 89.4% CO2uptake = 28.3%

[34]

BOF

CaO = 36.37% MgO = 7% d < 62 ␮m

Direct aqueous carbonation (slurry with metalworking wastewater) Direct aqueous carbonation (slurry)

Rotating packed bed reactor

98.9% vol. CO2

Ca = 36.1% CO2uptake = 10.3%a

[38]

BOF

CaO = 46.4% MgO = 6.5% d < 125 ␮m

High gravity rotating packed bed reactor

30% vol. CO2

Ca = 90.7% CO2uptake = 33.1%a

[39]

AOD

CaO =54.8% MgO = 9% d < 500 ␮m

Direct aqueous carbonation (slurry with metalworking wastewater) Direct aqueous carbonation (thin-film)

T = 25 ◦ C; P = 1 atm; t = 1 min; L/S = 20 T = 25 ◦ C; P = 1 atm; t = 20 min; L/S = 20

Batch reactor

20% vol. CO2

Ca = 56.2%a CO2uptake = 24.2%

[40]

AOD

CaO =54.8% MgO = 9% d < 500 ␮m

Direct aqueous carbonation (slurry)

Stirred batch reactor

100% vol. CO2

Ca = 93.6%a CO2uptake = 40.3%

[40]

AOD

CaO =54.8% MgO = 9% d < 500 ␮m

Direct aqueous carbonation (slurry)

Stirred batch reactor

100% vol. CO2

Ca = 94.3%a CO2uptake = 40.6%

[40]

CCc

CaO = 50.0% MgO = 10.9% d < 500 ␮m

Direct aqueous carbonation (thin-film)

Batch reactor

20% vol. CO2

Ca = 94.2%a CO2uptake = 37.0%

[40]

CCc

CaO = 50.0% MgO = 10.9% d < 500 ␮m

Direct aqueous carbonation (slurry)

Stirred batch reactor

100% vol. CO2

Ca = 56.3%a CO2uptake = 24.3%

[40]

EAF

CaO = 49.3% MgO = 4.1% d < 150 ␮m

Direct aqueous carbonation (thin-film)

Pressurized batch reactor

100% vol. CO2

Ca = 45.4%a CO2uptake = 17.6%

This study

EAF

CaO = 49.3% MgO = 4.1% d < 150 ␮m CaO = 29.9% MgO = 6.3% d < 150 ␮m

Direct aqueous carbonation (slurry) Direct aqueous carbonation (thin-film)

Pressurized stirred batch reactor Pressurized batch reactor

100% vol. CO2

Ca = 72.3% CO2uptake = 28.0%

This study

Ca = 89.0%a , d CO2uptake = 20.9%

This study

CaO = 29.9% MgO = 6.3% d < 150 ␮m

Direct aqueous carbonation (slurry)

Pressurized stirred batch reactor

100% vol. CO2

T = 30 ◦ C; P = 1 atm; t = 144 h; L/S = 0.33 T = 90 ◦ C; P = 30 bar; t = 60 min; L/S = 16 T = 90 ◦ C; P = 6 bar; t = 120 min; L/S = 16 T = 30 ◦ C; P = 1 atm; t = 144 h; L/S = 0.33 T = 90 ◦ C; P = 9 bar; t = 60 min; L/S = 16 T = 50 ◦ C; P = 10 bar; t = 24 h; L/S = 0.3 T = 100 ◦ C; P = 10 bar; t = 24 h; L/S = 5 T = 50 ◦ C; P = 10 bar; t = 24 h; L/S = 0.3 T = 100 ◦ C; P = 10 bar; t = 24 h; L/S = 5

Ca = 171.5%a , d CO2uptake = 40.3%

This study

BOF

BOF

a

Calculated from data provided in the paper according to the equation: ca =

b

Ladle slag. Continuous casting slag. See comments in the text about the calculation of the conversion yield.

c d

100% vol. CO2

CO2uptake (%)/44 [CaO](%)/56

[41], would amount to 1350, 1385 and 2200 MJ/t CO2 for the BOF2, BOF1 and EAF slag, respectively. These requirements include the grinding and milling of the residues, the mixing of the residues with water, the heating and pumping of the slurry into the reactor, the pressurization of CO2 in a multiple stage compressor and the solid/liquid separation of the slurry at the outlet of the reactor. As for the thin-film process, the total energy requirements calculated considering the operating conditions discussed in the present study (T = 50 ◦ C; P = 10 bar; L/S = 0.3 L/ kg), a carbonation time of 1 h, and applying the approach described in [41], would amount to 2100, 2230 and 2570 MJ/t CO2 for the EAF, BOF1, and BOF2 slag, respectively. In this case, the total requirements include the grinding and milling of the residues, the pressurization of CO2 in a multiple stage compressor and the heating and operation of the carbonation reactor envisioned for this application as a rotary drum. As may be noted hence, even though for the

× 100.

slurry-phase process a higher operating temperature is employed, for the types of steel slag analyzed in this work, the energy requirements for the slurry-phase route are generally equal or lower than those estimated for the thin-film one. This result can be explained considering the significantly higher reactivity exhibited by these residues (the BOF2 slag in particular) for the conditions applied in the slurry-phase route, which lead to a lower requirement of residues and water to submit to the different unit operations. Furthermore, it should be noted that for the slurry-phase process it was assumed that the heat generated by the exothermic carbonation reaction could be recovered, while for the thin-film one this hypothesis was deemed unrealistic, since, differently from the slurry phase route where the carbonation reaction takes place in the same medium to be heated, the heat transfer efficiency for low L/S ratios was considered to be difficultly predictable and exploitable [41].

308

R. Baciocchi et al. / Journal of Hazardous Materials 283 (2015) 302–313

Fig. 4. Liquid phase composition (slurry-phase carbonation) for major (upper part of the figure) and trace elements (lower part of the figure).

Another important aspect to consider for the slurry-phase carbonation process is that in aqueous carbonation the relevant reactions occur in the liquid phase after dissolution of the original minerals from the solid material (see e.g., [15]), hence additional useful information may be derived from the analysis of the aqueous solution. The constituents of concern followed similar behaviors over time (see Fig. 4), although with some differences in solution concentrations for key elements. In particular, for the BOF2 slag lower concentrations of Ca, Mn and Fe in solution were detected, which may be taken as an indirect indication of the decreased solubility of the controlling solid phases likely due to the common ion (i.e., CO3 2− ) effect. As for trace elements, the results of slurry-phase tests indicated that the concentrations of Mo for the EAF slag and V for the BOF slags (especially the BOF2 samples) may be of potential environmental concern in view of disposal of the final wastewater stream resulting from the process. 3.3. Carbonation kinetics The differences between the thin-film and slurry-phase process were also analyzed in terms of process kinetics. Lumped kinetic parameters were derived by fitting CO2 uptake data as a function of time with a theoretical model. To this aim, the formulation proposed by Lee [42] was adopted, in which the rate of carbonation conversion, x (i.e., CO2 uptake), over time is expressed by:



dx x =r 1− xmax dt

n (3)

where t = reaction time, xmax = maximum theoretical uptake at infinite time, ri = initial rate of CO2 binding by the solid material and n = reaction order. The attenuation term (1 − x/xmax )n in Eq. (3) reflects the reaction rate reduction caused by the formation of a precipitate coating around the shrinking reactive core [11,42]. While either first- or second-order type kinetics are usually assumed to describe the evolution of the carbonation process [42], numerical fitting of our data showed that the latter was more appropriate to match the experimental results, confirming the findings of a number of previous researchers [20,43,44]. In case of n = 2, integration of Eq. (3) yields: x=

xmax ∗ t t1/2 + t

(4)

where t1/2 is the time at which half of xmax is reached and has the expression t1/2 = xmax /ri . The results of model fitting are summarized in Table 3. The ri and t1/2 values indicate for the two BOF samples faster carbonation kinetics under slurry-phase conditions both in the initial stages and during the whole process. The above mentioned effects of enhanced CO2 solvation and dissolution of alkaline reactive phases and reduced formation of the coating layer on the unreacted core may explain the different carbonation kinetics observed under thin-film and slurry-phase conditions. For the EAF slag, however, the carbonation process proceeded faster under thin-film conditions, which was an unexpected result that still needs to be elucidated.

R. Baciocchi et al. / Journal of Hazardous Materials 283 (2015) 302–313

309

Table 3 Kinetic parameters of the carbonation process as derived from Eq. (3).

Thin-film route xmax (g CO2 /100 g slag) t1/2 (h) ri (g CO2 /100 g slag·h) Overall fitting parameters Slurry-phase route xmax (g CO2 /100 g slag) t1/2 (h) ri (g CO2 /100 g slag·h) Overall fitting parameters

BOF1

BOF2

EAF

Estimated value Standard error Estimated value Standard error Estimated value Standard error R2 Fit standard error

21.4 0.9 0.71 0.12 30.0 4.3 0.968 0.99

21.8 1.3 1.07 0.22 20.4 3.3 0.959 1.24

16.3 0.6 0.16 0.05 102.1 31.7 0.813 0.91

Estimated value Standard error Estimated value Standard error Estimated value Standard error R2 Fit standard error

32.7 0.7 0.41 0.04 79.0 6.8 0.989 0.77

40.7 1.6 0.78 0.14 52.1 7.8 0.965 1.58

27.8 1.2 0.57 0.13 49.1 9.7 0.937 1.23

3.4. Mineralogical and microstructural changes induced by carbonation Due to their better carbonation performance, a more detailed mineralogical investigation was conducted on the BOF samples (see Figs. 5–7). For the EAF slag, the reader is referred to previous work [5]. Mineralogical investigations showed appreciable changes upon carbonation. The content of some original minerals variously decreased depending on their specific reactivity with CO2 , and additional phases were concomitantly observed to form. The FT-IR spectra of the untreated and carbonated BOF materials (Fig. 5(a) and (b)) reveal the strong influence of carbonation on slag mineralogy, although the interpretation of FT-IR results is often made extremely difficult by the matrix complexity [45]. For the untreated slags, peaks of different intensities at wavenumbers associated to silicates were identified, namely 830–860, 890–910, 950–980, 720–740 and 500–530 cm−1 which correspond to Si O Si stretching and bending modes [46,47]; differences in position for some of these bands may be associated to different degrees of silicate polymerization [47]. However, the presence of either Si O Al or Al O bonds, which are also known to display absorption bands in the above mentioned wavelength ranges [46], may not be excluded. The BOF2 sample also displayed a sharp peak at ∼3640 cm−1 , associated to the O H stretching vibrations mode in portlandite. The FT-IR spectra for both slags also showed an increase in IR absorbance at low wavenumbers, likely related to the presence of Fe oxides [45]. The carbonated materials displayed some distinctive intense absorption bands peaking at 1390 (C O asymmetric stretching, ␯3 mode), 870 (C O out-of-plane bending, 2 mode), and 707 cm−1 (C O in-plane bending, 4 mode), clearly related to carbonate minerals [48], with stronger absorption characteristics for the slurry-phase carbonated materials. Additional new absorption bands for the carbonated materials were detected at ∼1020 cm−1 (with a shoulder at ∼ cm−1 ) for the BOF1 slag and 990 cm−1 (with a shoulder at ∼1050 cm−1 ) for the BOF2 slag; a broad band at 3340 (BOF1) and 3360 cm−1 (BOF2) was also identified, more distinctly for the thin-film route. Silica gel is reported to display characteristic absorption features at 800 (weak band), 948 (weak band), 1090 (strong band), 1190 (shoulder), 1640 (very weak band) and 3330 cm−1 (medium-intensity band) [49], therefore such bands may well explain the composition of the carbonated slags, considering that weaker bands may have been masked by other overlapping peaks. Such new bands

may thus be related to the formation of amorphous silica due to dissolution of the original silicates, followed by precipitation of the excess silicate ions. The values of differential absorption between the carbonated and untreated slags (Fig. 5(c)–(f)) show more clearly the enhanced formation of carbonates and amorphous silica upon carbonation, particularly for the slurry-phase process. The peaks area in the low wavelength ranges also decreased, indicating the disappearance of additional constituents (presumably Al and Fe oxides) upon carbonation. Reaction time affected slag mineralogy more appreciably for the thin-film process due to the slower dissolution kinetics. According to XRD analyses, for the BOF1 sample (Fig. 6(a)) the main phases affected by carbonation included dicalcium and tricalcium silicates as well as some oxide minerals (Ca2 Fe1.4 Al0.6 O5 , Ca2 Cr0.5 Fe1.5 O5 , wüstite), again more appreciably for the slurryphase treatment. For BOF2 (Fig. 5(b)) the portlandite content considerably decreased upon carbonation, completely disappearing after 4 h in the slurry-phase process. The presence of a highly reactive phase such as portlandite is believed to explain the significantly higher CO2 uptake of the BOF2 slag, although the finer mean particle size may also have played a role. Other minerals affected by slurryphase carbonation included, as already noted for the BOF1 slag, the oxide phases Ca2 Fe1.4 Al0.6 O5 , Ca2 Cr0.5 Fe1.5 O5 and wüstite, in addition to Mg silicate (MgSiO3 ), Mg–Mn oxide ((MgO)0.725 (MnO)0.275 ) and magnetite. As for the newly formed carbonates, for both BOF samples their amount and variety in the slurry-phase route was appreciably higher, with other complex carbonate minerals appearing in addition to Ca carbonates (calcite and aragonite). Although univocal association of diffraction peaks to individual carbonate phases involves some margins of uncertainty due to the overlapping XRD patterns of several complex carbonates, a tentative interpretation of XRD results on the basis of the likelihood of mineral presence in the materials was made. For the thinfilm-carbonated BOF1 sample, only calcite and aragonite were detected, while mixed carbonates containing Fe, Mg and Mn were present in the slurry-phase carbonated slag; among these, the most probable phases appeared to be ankerite (CaMg0.27 Fe0.73 (CO3 )2 ) and kutnohorite ((Ca0.86 Mn0.14 )(Ca0.14 Mn0.5 Fe0.36 )(CO3 )2 ). For the BOF2 sample, the newly formed carbonates included calcite, aragonite and Mg-calcite ((Mg0.064 Ca0.936 )(CO3 )) for the thin-filmtreated sample; for the slurry-phase process, in addition to such phases, dolomite (CaMg(CO3 )2 ) and possibly ankerite were formed.

310

R. Baciocchi et al. / Journal of Hazardous Materials 283 (2015) 302–313

Fig. 5. FT-IR spectra of untreated and carbonated (4 h) BOF1 (a) and BOF2 (b) slags and differential spectra of carbonated versus untreated BOF1 (c, d) and BOF2 (e, f) slags.

However, for both slag samples Ca carbonate phases were still observed to be the prevalent carbonate minerals in the final material (see Fig. 6). For the EAF slag (see [5] for details), the XRD analysis indicated, for both the thin-film and the slurry-phase process, the disappearance of phases such as Ca–Al oxide and periclase, and a relevant

reduction in peak intensities of dicalcium silicate, cuspidine and chromium oxide; on the other hand, no evident variations in the contents of akermanite, Cr–Mg oxide, magnetite and calcium fluoride could be detected. The only phase for which a significant increase in peak intensity could be retrieved after carbonation was calcite, indicating that, although MgO showed to decrease upon

R. Baciocchi et al. / Journal of Hazardous Materials 283 (2015) 302–313

311

Fig. 6. XRD patterns of carbonated (4 h): (a) BOF1 slag and (b) BOF2 slag. Legend: (a) portlandite [Ca(OH)2 ]; (c) calcite [CaCO3 ]; (d) larnite [Ca2 SiO4 ]; (e) magnetite [FeO·Fe2 O3 ]; (h) Ca–Fe–Al oxide [Ca2 Fe1.4 Al0.6 O5 ]; (k) corundum [Al2 O3 ]; (l) Ca–Cr–Fe oxide [Ca2 Cr0.5 Fe1.5 O5 ]; (m) wüstite [FeO]; (n) dolomite [CaMg(CO3 )2 ]; (o) aragonite [CaCO3 ]; (p) ankerite [CaMg0.27 Fe0.73 (CO3 )2 ]; (q) kutnohorite [(Ca0.86 Mn0.14 )(Ca0.14 Mn0.5 Fe0.36 )(CO3 )2 ]; (r) Mg-calcite [(Mg0.064 Ca0.936 )(CO3 )]; (s) Mg silicate [MgSiO3 ]; (t) Mg–Mn oxide [(MgO)0.725 (MnO)0.275 ]; (v) Cr oxide [CrO0.87 ]. Carbonate phases are reported in bold red. (For interpretation of the references to color in text, the reader is referred to the web version of this article.)

carbonation, for the EAF slag the chemical conditions of the carbonation process were unfavorable to the formation of carbonate species different from calcium carbonate. SEM observations (Fig. 7) confirmed the previously discussed results. For the slurry-phase carbonated BOF1 slag, Ca

carbonate crystals covering the surface of the original slag grains were clearly visible (Fig. 7(a) and (b)), with residual unconverted material remaining after 24-h treatment. In addition to the lower intrinsic reactivity of mineral phases in the BOF1 slag, the presence of larger particles in the carbonated material may also be explained

Fig. 7. SEM images of 24-h slurry-phase carbonated BOF1 (a, b) and BOF2 (c, d) slags at different magnifications.

312

R. Baciocchi et al. / Journal of Hazardous Materials 283 (2015) 302–313

by the different initial grain size distribution compared to the BOF2 slag. For the BOF2 sample, SEM analyses indicated that Ca carbonate crystals dominated the microstructure of the carbonated material, and only minor amounts of unreacted particles could be detected in the SEM images (Fig. 7(c) and (d)). 4. Conclusions The study showed that, under suitable carbonation conditions, steelmaking slag exhibits a high CO2 sequestration potential. Compared to previous literature studies, the carbonation yield attained for the BOF slag was more than 20% higher than the best-documented results for the same type of slag. The chemical properties of the original slag (mineralogical characteristics more than elemental composition), the grain size and the treatment conditions adopted showed all to affect the sequestration yield. Maximum CO2 uptake values of 176 (EAF), 201 (BOF1) and 209 g CO2 / kg slag (BOF2) were attained after 24-h carbonation under thin-film conditions, as opposed to values of 280, 325 and 403 g CO2 / kg slag for the slurry-phase route. The corresponding conversion yield also indicated that, especially for the BOF2 sample carbonated adopting the slurry-phase route, a virtually complete conversion of the potentially reactive minerals into carbonates was attained. Under the best conditions tested and based on slag production data from European BOF-type steelmaking plants, the CO2 uptake values attained would result in a theoretical sequestration capacity of 4.2 Mt CO2 /y. Based on the experimental findings reported in this work for an operating time of 1 h, it was estimated that the energy requirements of CO2 sequestration via steel slag carbonation would range, depending on the specific type of slag employed, between 1350 and 2200 MJ/t CO2 sequestered for the slurry-phase route, and between 2100 and 2570 MJ/t CO2 sequestered for the thin-film one. Hence, making reference to a natural gas-fired power plant, assuming a heat release of 20 GJ per ton of CO2 emitted, the energy penalty associated to the carbonation process (excluding CO2 capture), expressed as % of CO2 emitted per CO2 sequestered, would lie in the ranges 7–11% and 10.5–13% for the slurry-phase and the thin-film processes, respectively. These values indicate the theoretical feasibility of the carbonation processes analyzed in this work, although it should it be noted that 100% CO2 flows were employed, hence a preliminary CO2 capture step, with some additional associated energy requirement, would be necessary. FT-IR and XRD investigations indicated that the carbonation process was able to induce some significant rearrangement of the original mineralogical structure of the residues and, unlike noted in previous studies, for the BOF slags the carbonation reactions appeared to involve not only Ca, but also Mg, Fe and, to some extent, Mn silicates and oxide phases. It was probable that the newly formed carbonates included, in addition to the commonly reported minerals calcite and aragonite, mixed Ca–Mg, Ca–Mg–Fe or Ca–Mn–Fe carbonates (Mg-calcite, (Mg0.064 Ca0.936 )(CO3 ); dolomite, CaMg(CO3 )2 ; ankerite, CaMg0.27 Fe0.73 (CO3 )2 ; kutnohorite (Ca0.86 Mn0.14 )(Ca0.14 Mn0.5 Fe0.36 )(CO3 )2 ) depending on both the slag type and the carbonation conditions. The formation of this set of carbonate phases is regarded as the main innovative finding of the present study, which demonstrates that by properly adjusting the carbonation conditions the mineral constituents of the original slag can be largely involved in the carbonation process. While on one instance the significant mineralogical changes upon carbonation mirror a high dissolution degree of the original slag constituents that make the reactive alkali metals available for carbonate precipitation, it would be very important to assess the extent to which the environmental behavior of the materials is also affected by such mineralogical alterations. As a result, an important issue which deserves further attention relates to the

leaching behavior of potential inorganic contaminants from the carbonated slag. Additional experiments are currently underway to clarify the effects of accelerated carbonation on the release of inorganic contaminants from the material such as Cr, V, Ni, Pb and Mo in order to identify other potential environmental implications of the tested accelerated carbonation process besides CO2 storage.

References [1] K.S. Lackner, C.H. Wendt, D.P. Butt, E.L. Joyce Jr., D.H. Sharp, Carbon dioxide disposal in carbonate minerals, Energy 20 (1995) 1153–1170. [2] J. Sipilä, S. Teir, R. Zevenhoven, Carbon dioxide Sequestration by Mineral Carbonation, Literature Review Update 2005–2007, Åbo Akademi University Faculty of Technology Heat Engineering Laboratory, Report VT 2008-1, 2008. [3] R. Baciocchi, G. Costa, E. Di Bartolomeo, A. Polettini, R. Pomi, The effects of accelerated carbonation on CO2 uptake and metal release from incineration APC residues, Waste Manage. 29 (2009) 2994–3003. [4] R. Baciocchi, G. Costa, E. Di Bartolomeo, A. Polettini, R. Pomi, Carbonation of stainless steel slag as a process for CO2 storage and slag valorization, Waste Biomass Valor. 1 (2010) 467–477. [5] R. Baciocchi, G. Costa, E. Di Bartolomeo, A. Polettini, R. Pomi, Wet versus slurry carbonation of EAF steel slag, Greenh. Gases 1 (2011) 312–319. [6] M. Back, M. Kuehn, H. Stanjeck, S. Peiffer, Reactivity of alkaline lignite fly ashes towards CO2 in water, Environ. Sci. Technol. 42 (2008) 4520–4526. [7] P.J. Gunning, C.D. Hills, P.J. Carey, Accelerated carbonation treatment of industrial wastes, Waste Manage. 30 (2010) 1081–1090. [8] M. Uibu, M. Uus, R. Kuusik, CO2 mineral sequestration in oil-shale wastes from Estonian power production, J. Environ. Manage. 90 (2009) 1253–1260. [9] A. Iizuka, M. Fujii, A. Yamasaki, Y. Yanagisawa, Development of a new CO2 sequestration process utilizing the carbonation of waste cement, Ind. Eng. Chem. Res. 43 (2004) 7880–7887. [10] O.L. Shtepenko, C.D. Hills, N.J. Coleman, A. Brough, Characterization preliminary assessment of a sorbent produced by accelerated mineral carbonation, Environ. Sci. Technol. 39 (2005) 345–354. [11] D.N. Huntzinger, J.S. Gierke, S.K. Kawatra, T.C. Eisele, L.L. Sutter, Carbon dioxide sequestration in cement kiln dust through mineral carbonation, Environ. Sci. Technol. 43 (2009) 1986–1992. [12] R. Baciocchi, G. Costa, A. Polettini, R. Pomi, Influence of particle size on the carbonation of stainless steel slag for CO2 storage, Energy Procedia 1 (2009) 4859–4866. [13] D. Bonenfant, L. Kharoune, S. Sauvé, R. Hausler, P. Niquette, M. Mimeault, M. Kharoune, CO2 sequestration potential of steel slags at ambient pressure and temperature, Ind. Eng. Chem. Res. 47 (2008) 7610–7616. [14] E.-E. Chang, S.-Y. Pan, Y.-H. Chen, C.-S. Tan, P.-C. Chiang, Accelerated carbonation of steelmaking slags in a high-gravity rotating packed bed, J. Hazard. Mater. 227/228 (2012) 97–106. [15] W.J.J. Huijgen, G.J. Witkamp, R.N.J. Comans, Mineral CO2 sequestration by steel slag carbonation, Environ. Sci. Technol. 39 (2005) 9676–9682. [16] S. Kodama, T. Nishimoto, N. Yamamoto, K. Yogo, K. Yamada, Development of a new pH-swing CO2 mineralization process with a recyclable reaction solution, Energy 33 (2008) 776–784. [17] S.N. Lekakh, C.H. Rawlins, D.G.C. Robertson, V.L. Richards, K.D. Peaslee, Kinetics of aqueous leaching and carbonization of steelmaking slag, Metall. Mater. Trans. B 39B (2008) 125–134. [18] R.M. Santos, D. Ling, M. Guo, B. Blanpain, T. Van Gerven, Valorisation of thermal residues by intensified mineral carbonation, in: Proceedings of the 50th Conference of Metallurgists and 6th International Symposium on Waste Recycling in Mineral and Metallurgical Industries, COM2011, Montreal, 2011, art. no. 47189. [19] R.M. Santos, D. Ling, A. Sarvaramini, M. Guo, J. Elsen, F. Larachi, G. Beaudoin, B. Blanpain, T. Van Gerven, Stabilization of basic oxygen furnace slag by hot-stage carbonation treatment, Chem. Eng. J. 203 (2012) 239–250. [20] R.M. Santos, D. Franc¸ois, G. Mertens, J. Elsen, T. Van Gerven, Ultrasoundintensified mineral carbonation, Appl. Therm. Eng. 57 (2013) 154–163. [21] S. Teir, S. Eloneva, C.J. Fogelholm, R. Zevenhoven, Dissolution of steelmaking slags in acetic acid for precipitated calcium carbonate production, Energy 32 (2007) 528–539. [22] M. Uibu, R. Kuusik, L. Andreas, K. Kirsimäe, The CO2 -binding by Ca–Mg–silicates in direct aqueous carbonation of oil shale ash and steel slag, Energy Procedia 4 (2011) 925–932. [23] M. Fernández Bertos, X. Li, S.J.R. Simons, C.D. Hills, P.J. Carey, Investigation of accelerated carbonation for the stabilization of MSW incinerator ashes and the sequestration of CO2 , Green Chem. 6 (2004) 428–436. [24] European Slag Association (Euroslag), Position Paper on the Status of Ferrous Slag, April 2012. [25] D.M. Proctor, K.A. Fehling, E.C. Shay, J.L. Wittenborn, J.J. Green, C. Avent, R.D. Bigham, M. Connolly, B. Lee, T.O. Shepker, M.A. Zak, Physical and chemical characteristics of blast furnace, basic oxygen furnace, and electric arc furnace steel industry slags, Environ. Sci. Technol. 34 (2000) 1576–1582. [26] C. Shi, Steel slag – its production, processing, characteristics, and cementitious properties, J. Mater. Civ. Eng. 16 (2004) 230–236.

R. Baciocchi et al. / Journal of Hazardous Materials 283 (2015) 302–313 [27] World Steel Association, Statistics Archive, 2013, http://www.worldsteel.org/ statistics/statistics-archive.html (accessed April 2013). [28] European Commission, Directorate-General for Energy and Transport, EU Energy in Figures, 2010. [29] Intergovernmental Panel on Climate Change (IPCC), Climate Change 2007: Mitigation of Climate Change. Fourth Assessment Report, 2007. [30] I.Z. Yildirim, M. Prezzi, Chemical, Mineralogical and Morphological Properties of Steel Slag, Adv. Civ. Eng. 2011 (2011) 13 pp, article ID 463638. [31] G. Wang, Y. Wang, Z. Gao, Use of steel slag as a granular material: volume expansion prediction and usability criteria, J. Hazard. Mater. 184 (2010) 555–560. [32] H. Shen, E. Forssberg, An overview of recovery of metals from slags, Waste Manage. 23 (2003) 933–949. [33] E.-E. Chang, C.-H. Chen, Y.-H. Chen, S.-Y. Pan, P.-C. Chiang, Performance evaluation for carbonation of steel-making slags in a slurry reactor, J. Hazard. Mater. 186 (2011) 558–564. [34] E.-E. Chang, A.-C. Chiu, S.-Y. Pan, Y.-H. Chen, C.-S. Tan, P.-C. Chiang, Carbonation of basic oxygen furnace slag with metalworking wastewater in a slurry reactor, Int. J. Greenh. Gas Control 12 (2013) 382–389. [35] C. Navarro, M.D. Díaz, M. Villa-García, Physico-chemical characterization of steel slag: study of its behavior under simulated environmental conditions, Environ. Sci. Technol. 44 (2010) 5383–5388. [36] W.J.J. Huijgen, R.N.J. Comans, Carbonation of steel slag for CO2 sequestration: leaching of products and reaction mechanisms, Environ. Sci. Technol. 40 (2006) 2790–2796. [37] O. Shtepenko, C.D. Hills, A. Brough, M. Thomas, The effect of carbon dioxide on ␤-dicalcium silicate and Portland cement, Chem. Eng. J. 118 (2006) 107–118. [38] S.-Y. Pan, P.-C. Chiang, Y.-H. Chen, C.-S. Tan, E.-E. Chang, Ex situ CO2 capture by carbonation of steelmaking slag coupled with metalworking wastewater in a rotating packed bed, Environ. Sci. Technol. 47 (2013) 3308–3315. [39] E.-E. Chang, T.-L. Chen, S.-Y. Pan, Y.-H. Chen, P.-C. Chiang, Kinetic modeling on CO2 capture using basic oxygen furnace slag coupled with cold-rolling

[40]

[41]

[42] [43]

[44]

[45]

[46]

[47]

[48] [49]

313

wastewater in a rotating packed bed, J. Hazard. Mater. 260 (2013) 937–946. R.M. Santos, J. Van Bouwel, E. Vandevelde, G. Mertens, J. Elsen, T. Van Gerven, Accelerated mineral carbonation of stainless steel slags for CO2 storage and waste valorization: Effect of process parameters on geochemical properties, Int. J. Greenh. Gas Control 17 (2013) 32–45. D. Zingaretti, G. Costa, R. Baciocchi, Assessment of accelerated carbonation processes for CO2 storage using alkaline industrial residues, Ind. Eng. Chem. Res. 53 (2014) 9311–9324. D.K. Lee, An apparent kinetic model for the carbonation of calcium oxide by carbon dioxide, Chem. Eng. J. 100 (2004) 71–77. G. Montes-Hernandez, D. Daval, N. Findling, R. Chiriac, F. Renard, Linear growth rate of nanosized calcite synthesized via gas–solid carbonation of Ca(OH)2 particles in a static bed reactor, Chem. Eng. J. 180 (2012) 237–244. A. Garenne, G. Montes-Hernandez, P. Beck, B. Schmitt, O. Brissaud, A. Pommerol, Gas–solid carbonation as a possible source of carbonates in cold planetary environments, Planet. Space Sci. 76 (2013) 28–41. B. Stumpe, T. Engel, B. Steinweg, B. Marschner, Application of PCA and SIMCA Statistical analysis of FT-IR spectra for the classification and identification of different slag types with environmental origin, Environ. Sci. Technol. 46 (2012) 3964–3972. A. Beran, D. Voll, H. Schneider, Dehydration and structural development of mullite precursors: an FTIR spectroscopic study, J. Eur. Ceram. Soc. 21 (2001) 2479–2485. I. Halasz, M. Agarwal, R. Li, N. Miller, What can vibrational spectroscopy tell about the structure of dissolved sodium silicates? Microporous Mesoporous Mater. 135 (2010) 74–81. K.W. Ryu, M.G. Lee, Y.N. Jang, Mechanism of tremolite carbonation, Appl. Geochem. 26 (2011) 1215–1221. F.A. Miller, C.H. Wilkins, Infrared spectra and characteristic frequencies of inorganic ions: their use in qualitative analysis, Anal. Chem. 24 (1952) 1253–1254.