- Email: [email protected]
Thermodynamic analysis of the carbothermic reduction of electric arc furnace dust in the presence of ferrosilicon
CALPHAD: Computer Coupling of Phase Diagrams and Thermochemistry 52 (2016) 143–151
Contents lists available at ScienceDirect
CALPHAD: Computer Coupling of Phase Diagrams and Thermochemistry journal homepage: www.elsevier.com/locate/calphad
Invited article
Thermodynamic analysis of the carbothermic reduction of electric arc furnace dust in the presence of ferrosilicon S.M. Moosavi Nezhad, Ahad Zabett n Department of Material Science and Engineering, Faculty of Engineering, Ferdowsi University of Mashhad, Khorasan Razavi, Iran
art ic l e i nf o
a b s t r a c t
Article history: Received 1 September 2015 Received in revised form 29 November 2015 Accepted 29 November 2015
Thermodynamic analysis of zinc recovery from electric arc furnace (EAF) dust by carbon in the presence of ferrosilicon was studied. A preheating process was performed to remove volatile compounds from dust to avoid impurities in the final zinc product. The main process includes reduction of zinc oxide and gasification and condensation of zinc vapor. Equilibrium condition was computed using FactSage program for the processes of reduction and gasification. The effect of ferrosilicon addition on the recovery of zinc was investigated for the temperature range of 950–1050 °C at the constant pressure of one atmosphere. It is shown that ferrosilicon provides a better kinetic condition for zinc recovery by facilitating slag formation. The preferred condition for formation of liquid slag was predicted based on thermodynamic calculations and verified by performing some experiments. The experimental results show that good agreement between calculations and experiments can be obtained with higher amounts of silicon contents and at higher temperatures where liquid slag is presented. A mixture of the dust with 10% C and 7% FeSi(75%) can be used to achieve over 97% zinc recovery. & 2015 Elsevier Ltd. All rights reserved.
Keywords: Electric arc furnace dust Ferrosilicon Zinc recovery Slag formation Kinetic condition
1. Introduction Electric arc furnace dust is classified as a hazardous material because of the presence of heavy metal constituents (i.e., leachable lead, cadmium and hexavalent chromium) [1–4]. More than 6 million tons of EAF dust is annually produced worldwide, which has to be treated to comply with the environmental regulations. One common choice of treating this toxic material is stabilization and landfilling. EAF dust contains significant amounts of zinc and iron. Since the use of galvanized steel continues to increase in the automotive industry, the amount of zinc in the dust will rise concurrently. There are many different processes to recover zinc from EAF dust [5–12]. Today, about 30% of the industrial zinc is produced by recovery from secondary zinc resources such as galvanized sheets and steelmaking dusts [13]. Zinc recovery from EAF dust is economically feasible when zinc concentration is more than 15% [14]. Pyro-metallurgical treatments are main industrial technologies for zinc recovery from EAF dust whereas 80% of recycled dust is treated by Waelz kiln process. In pyro-metallurgical methods EAF dust is reduced at a high temperature. Zinc can be recovered from the vapor and the iron can be recovered in the form of liquid metal. n
Corresponding author.
http://dx.doi.org/10.1016/j.calphad.2015.11.003 0364-5916/& 2015 Elsevier Ltd. All rights reserved.
Many researchers have studied different aspects of zinc recovery from EAF dust in pyrometallurgical processes. In most of these works carbonaceous materials are used as the reductant. Pickles [15] studied the reaction of EAF dust with solid iron powder in argon atmosphere and reported zinc recovery of 87% at 1050 °C. In another work [16] he considered the reaction of EAF dust with molten pig iron containing carbon and silicon. He showed that in this process, zinc can be completely reduced and recovered in the form of crude zinc oxide. Silicon can be an alternative reducing agent for carbonaceous materials. Fig.1 shows changes in standard free energy of oxidation for silicon, zinc, iron, wustite and magnetite by considering the following reactions. 4Fe3O4(s) þO2(g) ¼6Fe2O3(s)
(1)
6FeO(s) þ O2(g) ¼2Fe3O4(s)
(2)
2Fe(s) þO2(g) ¼ 2FeO(s)
(3)
2Zn(s) þO2(g) ¼2ZnO(s)
(4)
Si(s) þO2(g) ¼SiO2(s)
(5)
As it can be seen in Fig. 1, silica is more stable than zinc and iron oxides and therefore reduction of these oxides by silicon is
144
S.M. Moosavi Nezhad, A. Zabett / CALPHAD: Computer Coupling of Phase Diagrams and Thermochemistry 52 (2016) 143–151
Fig. 1. Standard free energy changes of oxidation for Zn, Fe, FeO, Fe3O4 and Si according to reactions (1–5).
Fig. 2. Calculated zinc loss during the pretreatment of EAF dust.
thermodynamically possible. Using silicon for reduction of EAF dust increases silica content of the residue and forms low melting temperature slag. This could enhance the rate of reactions by increasing the mass and heat transfer in the condense phases. Due to the higher cost of silicon use of a mixture of silicon and carbon are proposed as the reducing agent for the zinc recovery from EAF dust. The aim of the present research is to investigate the thermodynamic conditions for reduction of EAF dust by a mixture of carbon and ferrosilicon. Alkali halides and lead oxide are usually problematic in the recovery of zinc from EAF dust, especially in the zinc metal recovery processes. Therefore, according to the literature [17,18] a preheating step for removing the more volatile species from EAF dust at a lower temperature was suggested. In this pretreatment, volatile impurities such as PbO, NaCl, KCl and CdO are removed from EAF dust.
2. Thermodynamic calculations and experiments Fig. 3. Calculated amounts of air needed for 95% removal degree of volatile species during the pretreatment.
2.1. Calculations FactSage 6.1 program was used to study the equilibrium conditions of the reactions involved in the heat treatment process. EQUILIB (a module of FactSage) [19] utilizes the chemical composition of reactants in the form of compounds or elements as the input and determines the combination of the most stable products at any desired temperature and pressure. Calculations are based on minimization of Gibbs free energy of the system at isothermal and isobaric conditions. For the pretreatment, heating of 100 g dust in the presence of different amounts of air as a carrier gas was studied. Then, computations of thermodynamic equilibrium of the pretreated dust (solid residue remaining from the pretreatment) with different amounts of carbon and ferrosilicon at different temperatures were performed. The amounts of carbon and ferrosilicon were increased from 1 to 20 g by a step size of 1 g. Temperature was increased from 850 °C to 900 °C in steps of 5 °C for pretreatment and from 950 °C to 1050 °C in steps of 10 °C for the reduction. Input data was
introduced to EQUILIB program in the form of elemental composition. The chemical compositions of the real EAF dust samples before and after pretreatment were used. 2.2. Experiments EAF dust was obtained from an alloy steelmaking company in Iran. Graphite powder was provided by Sinopharm Chemical Reagent Co. Ltd. and ferrosilicon powder containing 75% silicon and 25% iron from Iran Ferrosilice Co. was used. Chemical composition of the EAF dust and the residues of the pretreatment were analyzed by an inductively coupled plasma-mass spectrometer (ICPMS, Spectro MS). The carbon content of the dust was determined by a Leco CS 244 analyzer. For the pretreatment process, 100 g EAF dust was heated in an alumina boat (200 mm in length, 100 mm in width and 30 mm in height) by an electric furnace at 875 °C for 4 h and cooled inside
Table 1 Chemical composition of the EAF dust. Element
Fe
Zn
Ca
Pb
Na
Mg
Al
Cd
Cr
K
Si
C
Cl
wt%
30.0
19.02
4.5
0.99
3.38
4.99
0.53
0.04
0.34
0.16
0.16
2.1
6.45
S.M. Moosavi Nezhad, A. Zabett / CALPHAD: Computer Coupling of Phase Diagrams and Thermochemistry 52 (2016) 143–151
Fig. 4. Calculated recovery of (a) zinc and (b) iron from dust at 950 °C.
Fig. 5. Calculated recovery of (a) zinc and (b) iron from dust at 1050 °C.
145
146
S.M. Moosavi Nezhad, A. Zabett / CALPHAD: Computer Coupling of Phase Diagrams and Thermochemistry 52 (2016) 143–151
Fig. 6. Calculated CO:CO2 ratio in vapor phase at (a) 950 °C and (b) 1050 °C.
the furnace to room temperature. For the recovery process, 2 g of the pretreated dust was mixed with different amounts of graphite and ferrosilicon powders and placed in a cylindrical alumina crucible (35 mm in diameter and 130 mm in height). To avoid direct contact of the mixture with air, three grams of Al2O3 powder was placed over the mixture. The furnace was preheated and the crucible was inserted into the furnace at 1000 °C or 1050 °C. Samples were air cooled after heat treatment. For each sample, recovery or removal efficiency (η) of the metals was calculated as follows:
shows zinc loss as a function of temperature. As it can be seen, zinc loss increases with increasing temperature which is not desired. Zabett and Lu [18] also reported that reducing condition and higher temperatures, increase zinc loss due to reduction and evaporation. Based on Fig. 2, heating of the dust at a temperature lower than 900 °C is appropriate to avoid zinc loss. In order to determine the proper temperature for removal of volatile species, number of moles of air required for 95% removal of lead, potassium and cadmium at 850–900 °C is plotted in Fig. 3. As it can be seen, the number of moles of air at 875 °C to achieve removal degree of 95% for Pb, K and Cd, are 73, 93 and 108, respectively.
ηE (wt%) = 100⋅(Wt Ei ⋅MiT − Wt Ef ⋅MfT )/(Wt Ei ⋅MiT )
3.1.2. Recovery process The recovery percentage of zinc and iron as a function of carbon and ferrosilicon at 950 °C is shown as a 3-D plot in Fig. 4. The figure shows that at zero amount of ferrosilicon, complete zinc recovery can be achieved with the addition of 14% carbon to the dust. At low amounts of reductants, there is a region at which zinc recovery is relatively constant. The boundaries of this triangular region are approximately limited to 9% carbon and 15% ferrosilicon. This corresponds to a region in Fig. 4b at which iron recovery increases very rapidly by increasing the amount of reductants. Next to this region, increasing the amounts of carbon or ferrosilicon enhances zinc recovery. For example at 950 °C, in the presence of 5% carbon, zinc recovery increases after 95% iron recovery. It can be concluded that at low amounts of reductants (carbon and silicon), reduction of zinc oxide will thermodynamically enhance after substantial iron recovery. This trend can be explained by considering the following reactions:
WtEi
(6)
WtEf
where and are the weight percent of the element E before and after the heat treatment. MiT and MfT are the total weights of the dust before and after the heat treatment. The percentage of iron recovery is calculated as follows: T Fe T ηE (wt%) = 100⋅Wt Fe f ⋅Mf /Wt i ⋅Mi
(7)
Fe where WtFe i and Wtf are the weight percent of iron in dust before treatment and the weight percent of metallic iron in the final solid residue.
3. Results and discussion 3.1. Thermodynamic calculations 3.1.1. Pretreatment process Chemical composition of the dust is reported in Table 1. Presence of 2.1% carbon in the dust, results in increasing reduction condition and thus higher zinc loss in the preheating step. Fig. 2
FeO þCO(g)¼Fe(s)þCO2(g)
(8)
ZnOþCO(g)¼Zn(g)þ CO2(g)
(9)
S.M. Moosavi Nezhad, A. Zabett / CALPHAD: Computer Coupling of Phase Diagrams and Thermochemistry 52 (2016) 143–151
Fig. 7. Calculated CO:CO2 ratio in vapor phase. Mixture of dust and 5% carbon at (a) 950 °C and (b) 1050 °C.
2FeO þSi(s) ¼2Fe(s)þ SiO2(s)
(10)
2ZnO þSi(s) ¼2Zn(g) þSiO2(s)
(11)
In the absence of ferrosilicon, zinc and iron recoveries are mainly occurred by indirect reduction reactions (8) and (9) [20,21]. Free energy change of reaction (9) is directly related to the partial pressure of zinc. In the early stages of the dust reduction, when partial pressure of zinc is very low, reaction (9) will be thermodynamically more favorable than reaction (8). As the partial pressure of zinc increases, free energy changes of reaction (9) increases and reaction (8) will be thermodynamically more favorable. With the addition of ferrosilicon, zinc and iron oxides of the dust can be reduced according to reactions (10) and (11). While FeO is in the form of pure solid, assuming activity one, and the partial pressure of zinc is not very low, free energy changes of reaction (10) is lower than that of reaction (11). This means that silicon is consumed earlier to reduce iron oxide. At higher amounts of ferrosilicon, solid FeO disappears and enters the slag liquid phase. This decreases FeO activity and increases free energy changes of reaction (10). At this condition, free energy changes of reactions (10) and (11) are comparable and zinc recovery increases. Zinc and iron recoveries at 1050 °C are plotted as a function of carbon and ferrosilicon in Fig. 5a and b, respectively. Results indicate that at zero amount of ferrosilicon, raising the temperature results in decreasing of iron recovery and increasing of zinc recovery. Same results are also reported by Pickles [22]. At this condition, complete zinc recovery can be achieved with the addition of 13% carbon to the dust. At low amount of carbon and
147
higher temperature free energy changes of reaction (9) reduces and CO:CO2 ratio increases. This rises the zinc vapor pressure. Therefore, zinc recovery enhances at higher temperatures. Since carbon is completely used for the reduction of zinc and iron oxides, at a fixed amount of carbon, higher amount of zinc recovery causes lower amount of iron recovery. Ferrosilicon addition to the mixture of EAF dust and carbon improves reducing conditions. This changes oxygen potential and CO:CO2 ratio in the system. Therefore, computing the amount of CO:CO2 ratio can help to better understand the effect of ferrosilicon addition on zinc recovery. CO:CO2 ratio as a function of ferrosilicon weight percent is plotted in Fig. 6. At 950 °C and constant carbon content, by increasing the amount of ferrosilicon, first CO:CO2 ratio remains constant and then rises. At higher ferrosilicon values a rapid enhancement occurs in CO:CO2 ratio and then this ratio becomes higher with an increasing slope. A similar trend with higher values of CO:CO2 ratio is obtained at 1050 °C, as shown in Fig. 6b. Comparing Figs. 4a and 5a with Fig. 6 indicates that there is a close relation between zinc recovery and CO:CO2 ratio. To examine this relation, at fixed carbon content of 5%, zinc and iron recoveries, CO:CO2 ratio, slag weight and the weight of solid FeO is plotted against the amount of ferrosilicon at 950 °C and 1050 °C, see Fig. 7. According to Fig. 7a, four distinct zones are recognized. The first zone with low ferrosilicon where zinc recovery and CO:CO2 ratio do not change. The second zone with medium amount of ferrosilicon where zinc recovery and CO:CO2 ratio increase simultaneously. The third zone with a higher amount of ferrosilicon where zinc recovery rises while iron recovery is completed. The fourth zone with the highest amount of ferrosilicon where both iron and zinc recoveries are completed and CO:CO2 ratios rises slowly. As previously discussed, at a low amount of reductants, zinc recovery increases after a substantial increase in the iron recovery. At the first zone, iron recovery increases according to reaction (10) and therefore CO:CO2 ratio does not change. With increasing ferrosilicon, slag forms and after substantial decrease in the weight of solid FeO, remaining FeO enters the slag phase. At the second zone with elimination of solid FeO and because of lower activity of FeO in the slag, reduction of zinc oxide begins to increase and leads to an increase in the zinc recovery. At this zone, CO:CO2 ratio is under the influence of FeO reduction with CO (reaction (8)). However, these values change very rapidly at the end of the second zone when slag disappears and 100% iron recovery is achieved. After complete reduction of iron oxides (the third zone), CO:CO2 ratio is under the influence of reaction (9). After completion of ZnO reduction (fourth zone) CO:CO2 ratio is affected by Boudouard reaction as follows: C þCO2(g)¼ 2CO(g)
(12)
Fig. 7b shows CO:CO2 ratio as a function of ferrosilicon at 1050 °C. At the higher temperature, the lengths of zones 1 and 2 have increased and zone 3 has disappeared. In the presence of carbon at zero amount of ferrosilicon, iron recovery decreases as a result of temperature increase. Therefore, completion of iron recovery needs higher amount of reductant (e.g. ferrosilicon). As it can be seen in Fig. 7, iron recovery completes with 14% ferrosilicon at 950 °C and with 16% ferrosilicon at 1050 °C. On the other hand, at fixed amounts of reduction agents, CO:CO2 ratio at 1050 °C is higher than 950 °C in the fourth zone. For example at ferrosilicon amount of 20%, CO:CO2 ratios at 950 °C and 1050 °C are 137 and 476, respectively. This is due to the fact that CO is more stable than CO2 at higher temperatures and therefore CO:CO2 ratio is larger at higher temperatures. According to Fig. 7, slag weight and composition affect the zinc and iron oxides reduction reactions. This effect is represented in
148
S.M. Moosavi Nezhad, A. Zabett / CALPHAD: Computer Coupling of Phase Diagrams and Thermochemistry 52 (2016) 143–151
Fig. 8. Calculated weight of liquid slag at (a) 950 °C and (b) 1050 °C.
Table 2 Calculated composition of the slag obtained from heating of 100 g dust at different conditions. Conditions
Slag composition and weight
Temperature (°C)
C (wt%)
FeSi (wt%)
Slag weight (g)
FeO (wt%)
ZnO (wt%)
SiO2 (wt%)
MgO (wt%)
Al2O3 (wt%)
CaO (wt%)
950 950 950 950 950 950 1050 1050 1050 1050 1050 1050
0 0 5 10 15 20 0 0 5 10 15 20
7 14 7 7 7 7 7 14 7 7 7 7
13.47 28.88 12.57 8.77 7.52 7.52 20.71 75.29 20.85 10.19 8.65 8.65
14.54 22.01 9.78 0.55 0.17 0.17 33.82 21.61 18.20 21.57 1.66 0.11
16.32 20.20 13.77 4.44 1.36 1.36 20.06 28.89 20.06 2.79 0.17 0.17
30.74 30.74 35.20 42.37 43.49 43.49 26.33 30.01 26.37 41.88 43.56 43.56
31.50 18.21 34.73 48.12 50.94 50.94 20.12 12.21 20.14 42.53 47.35 47.35
5.29 3.24 2.55 1.31 1.13 1.13 4.72 1.25 4.81 3.72 2.64 2.64
1.38 5.26 3.81 3.11 2.81 2.81 6.92 9.28 6.94 7.33 6.10 6.10
Fig. 8 where the slag weight is plotted against carbon and silicon at 950 °C and 1050 °C. Fig. 8a shows that at small amounts of ferrosilicon (o5%), there is no slag phase in the system. At higher amounts of ferrosilicon, SiO2 increases and leads to the formation of slag with low melting point. At this condition, SiO2, MgO, ZnO and FeO form the main part of the slag. The weight and composition of the slag phase at different conditions are given in Table 2. The weight of slag becomes higher with increasing silica, reaches a maximum and then decreases as its FeO content decreases. These changes rise the slag melting point and reduce the slag weight. After completing the FeO reduction, the weight of slag will become zero. Any change in the carbon content of the dust also has an effect on slag formation and zinc recovery. At a medium amount of ferrosilicon (5–10%), less slag is available at higher carbon content.
Since zinc and iron recoveries increase at higher carbon contents, lower amounts of zinc and iron oxides are available to form the slag phase. In the absence of carbon, complete removal of the slag occurs at 18% of ferrosilicon while in the presence of 10% of carbon this elimination occurs at 11% of ferrosilicon. On the other hand at relatively large amounts of carbon, where complete zinc and iron recoveries occur, iron and zinc oxides do not exist in the slag and it mainly contains MgO and SiO2 (see Table 2). Changes in the slag weight as a function of ferrosilicon and carbon at 1050 °C is approximately similar to that observed at 950 °C. Temperature elevation facilitates slag formation leading to slag weight increase with temperature. Under such a condition slag remains up to higher amount of ferrosilicon. According to Min et al. [23] since the molten phase (slag) provides extensive interfacial area with different reductants, slag
S.M. Moosavi Nezhad, A. Zabett / CALPHAD: Computer Coupling of Phase Diagrams and Thermochemistry 52 (2016) 143–151
149
Fig. 9. Calculated zinc recovery from dust mixed with 10% carbon.
Fig. 10. Comparison of the removal degrees of volatiles in pretreatment from calculations and experiments.
formation plays an important role on the reduction rate of metal oxides in ironmaking and steelmaking processes. Zabett and Lu [18] have shown that progress in the removal of volatile species from EAF dust is directly related to the change in the weight of the liquid phase (slag phase). Therefore, formation of a slag phase in the pyrometallurgical zinc recovery from EAF dust can affect the kinetics of ZnO reduction. For the formation of the slag phase, at
Fig. 12. Zinc recovery from dust mixed with fixed amount of Cþ Si¼ 10 wt%.
least 5% ferrosilicon is needed. Due to higher price of ferrosilicon in comparison with carbonaceous materials such as coal or coke, unnecessary use of this reductant for non-essential iron recovery should be avoided. Therefore a mixture of carbon bearing materials and ferrosilicon, as reducing agents, is economically proper when carbon is sufficient to complete reduction of iron oxides. According to Figs. 4a and 5a, at zero amount of ferrosilicon,
Fig. 11. Comparison of the zinc recoveries from calculations and experiments at (a) 950 °C and (b) 1050 °C.
150
S.M. Moosavi Nezhad, A. Zabett / CALPHAD: Computer Coupling of Phase Diagrams and Thermochemistry 52 (2016) 143–151
complete recovery of iron is obtained at 10–12% carbon in the temperature range of 950–1050 °C. Fig. 9 shows zinc recovery as a function of temperature and amount of ferrosilicon at carbon content of 10%. At least 7% ferrosilicon is necessary to achieve zinc recovery of 97% at 950 °C, as can be seen in Fig. 9. The weight of slag under this condition is approximately 13%. At 1050 °C, addition of 4% ferrosilicon is sufficient to achieve 97% zinc recovery. However, under this condition slag phase does not exist. In the absence of slag, because of limited contact between reactants, the expected zinc recovery value will not be achieved. With the addition of 7% ferrosilicon, 21% slag will be produced at 1050 °C. The above discussion indicates that a mixture of 10% of C–7%FeSi could be proper as a reductant to recover zinc from the dust in the temperature range of 950–1050 °C. 3.2. Verifying thermodynamic computations To verify computer calculations, the preheating and the main heating processes were experimentally examined. The computation and experimental results for removal of Pb, K, Na and Cd are given in Fig. 10. As it can be seen, the results are in good agreement with calculations. Based on the experimental results, a removal degree of 95% for K, Pb and Cd could be achieved by performing a preheating process at 875 °C for 4 h. Zinc recovery from pretreated dust mixed with different amounts of carbon and ferrosilicon heated for 1 h at 950–1050 °C is shown in Fig. 11. Results of calculations are also represented for comparison. Based on the calculations, complete zinc recovery can be achieved at 950 °C in both samples containing 20% of C and 10% of C þ10% of FeSi. However zinc recovery percentages of 79% and 53% were achieved after 1 h heating at 950 °C for samples with and without ferrosilicon, respectively. Higher zinc recovery of the sample containing ferrosilicon is due to slag formation. Moreover, at the higher temperature better agreement between computations and experiments was attained. Because of higher rate of mass and heat transfer and larger amount of slag phase at higher temperatures, the rate of reactions has increased and experimental conditions have approached the equilibrium condition. To better investigate simultaneous effect of carbon and silicon as reducing agents, a set of experiments was designed. In these experiments, mixtures of pretreated dust with the addition of 10% of C þSi, at different C:Si ratios, were treated at 1000 °C for 1 h. Zinc recovery from these samples is plotted versus weight percentage of silicon in Fig. 12. For the purpose of comparison, the results of thermodynamic calculations for same process are shown in this figure. Calculations indicate that in a fixed amount of the reductant, substitution of carbon by silicon decreases zinc recovery. To achieve same zinc recovery 1% carbon should be replaced with 2% ferrosilicon. Experiments show that zinc recovery increases with increasing the Si:C ratio, reaches a maximum and then decreases. At higher amounts of silicon, the weight of slag has increased and experimental conditions have approached the thermodynamic predictions. Based on the experimental observations, partial melting has happened in the samples with higher amounts of ferrosilicon. This verifies the prediction that ferrosilicon addition increases zinc recovery by increasing the weight of slag which provides better contact between reductants and metal oxides.
4. Conclusion Zinc recovery from EAF dust using carbon and ferrosilicon was studied. Thermodyanaic equilibrium conditions were calculated to study the process and to obtain the proper conditions for zinc reduction and vaporization. Results show that complete zinc
recovery in the temperature range of 950–1050 °C may be achieved by addition of about 14% carbon to the dust. However there is no liquid slag in the system when only carbon is used for reduction and therefore kinetics of the process could be slow. Ferrosilicon addition could improve kinetic conditions for ZnO reduction by facilitating low melting temperature slag formation. A liquid slag forms with the addition of at least 5% ferrosilicon to the dust. When a mixture of carbon and ferrosilicon is used for reduction, same zinc recovery can be achieved by substitution of about 2 g ferrosilicon (Si 75%) for every gram carbon. Thermodynamic computation shows that zinc recovery of over 97% could be attained with a mixture of 10% of C þ7% of FeSi(75%). Prediction of better kinetic condition by ferrosilicon addition was verified by experiments. Higher amount of ferrosilicon used for reduction of oxides lead to higher amount of liquid slag and higher amount of zinc recovery. As the amount of silicon in the mixture increases, results of thermodynamic computations and experiments show better convergence.
Acknowledgments The authors are thankful to Amirkabir University of Technology for their support in using FactSage 6.1. Financial support from Isfahan Alloy Steel Company is also greatly appreciated.
Appendix A. Supplementary material Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.calphad.2015.11.003.
References [1] M. Al-harahsheh, S. Kingman, L. Al-Makhadmah, I.E. Hamilton, Microwave treatment of electric arc furnace dust with PVC: dielectric characterization and pyrolysis-leaching, J. Hazard. Mater. 274 (2014) 87–97, http://dx.doi.org/ 10.1016/j.jhazmat.2014.03.019. [2] T. Suetens, B. Klaasen, K. Van Ackerw, B. Blanpain, Comparison of electric arc furnace dust treatment technologies using exergy efficiency, J. Clean. Prod. 65 (2014) 152–167, http://dx.doi.org/10.1016/j.jclepro.2013.09.053. [3] B.S. Yu, Y.R. Wang, T.C. Chang, Hydrothermal treatment of electric arc furnace dust, J. Hazard. Mater. 190 (2011) 397–402, http://dx.doi.org/10.1016/j. jhazmat.2011.03.056. [4] R.A. Shawabkeh, Hydrometallurgical extraction of zinc from Jordanian electric arc furnace dust, Hydrometallurgy 104 (2010) 61–65, http://dx.doi.org/ 10.1016/j.hydromet.2010.04.014. [5] C.F. Pereira, Y. Luna, X. Querol, D. Antenucci, J. Vale, Waste stabilization/solidification of an electric arc furnace dust using fly ash-based geopolymers, Fuel 88 (2009) 1185–1193, http://dx.doi.org/10.1016/j.fuel.2008.01.021. [6] R.N.C. De Siqueira, E. De Albuquerque Brocchi, P.F. De Oliveira, M.S. Motta, Hydrogen reduction of zinc and iron oxides containing mixtures, Metall. Mater. Trans. B 45 (2014) 66–75, http://dx.doi.org/10.1007/s11663-013-9951-4. [7] M.K. Jha, V. Kumar, R.J. Singh, Review of hydrometallurgical recovery of zinc from industrial wastes, Resour. Conserv. Recycl. 33 (2001) 1–22, http://dx.doi. org/10.1016/S0921-3449(00)00095-1. [8] C.A. Pickles, Thermodynamic modelling of the formation of zinc–manganese ferrite spinel in electric arc furnace dust, J. Hazard. Mater. 179 (2010) 309–317, http://dx.doi.org/10.1016/j.jhazmat.2010.03.005. [9] V.P. Korneev, V.P. Sirotinkin, N.V. Petrakova, V.G. Dyubanov, L.I. Leont'ev, Physicochemical properties of the zinc-containing dusts of electric furnace steelmaking, Russ. Metall. 7 (2013) 507–512, http://dx.doi.org/10.1134/ S0036029513070094. [10] C.A. Pickles, Thermodynamic analysis of the selective chlorination of electric arc furnace dust, J. Hazard. Mater. 166 (2009) 1030–1042, http://dx.doi.org/ 10.1016/j.jhazmat.2008.11.110. [11] N. Tsubouchi, H. Hashimoto, N. Ohtaka, Y. Ohtsuka, Chemical characterization of dust particles recovered from bag filters of electric arc furnaces for steelmaking: Some factors influencing the formation of hexachlorobenzene, J. Hazard. Mater. 183 (2010) 116–124, http://dx.doi.org/10.1016/j. jhazmat.2010.06.122. [12] N.Y. Ma, On in-process separation of zinc from EAF dust, in: S.N. Monteiro, D. E. Verhulst, P.N. Anyalebechi, J.A. Pomykala (Eds.), EPD Congress, John Wiley & Sons; Hoboken, NJ, USA, 2011, pp. 947–952.
S.M. Moosavi Nezhad, A. Zabett / CALPHAD: Computer Coupling of Phase Diagrams and Thermochemistry 52 (2016) 143–151
[13] N. Antunano, D. Herrero, P.L. Arias, J.F. Cambra, Electrowinning studies for metallic zinc production from double leached Waelz oxide, Process Saf. Environ. Prot. 91 (2013) 495–502, http://dx.doi.org/10.1016/j.psep.2012.10.014. [14] M.H. Morcali, O. Yucel, A. Aydin, B. Derin, Carbothermic reduction of electric arc furnace dust and calcination of waelz oxide by semi-pilot scale rotary furnace, J. Min. Metall. B 48 (2012) 173–184, http://dx.doi.org/10.2298/ JMMB111219031M. [15] J.R. Donald, C.A. Pickles, Reduction of electric arc furnace dust with solid iron powder, Can. Metall. Q. 35 (1996) 255–267, http://dx.doi.org/10.1016/ 0008-4433(96)00009-2. [16] C.A. Pickles, Reaction of electric arc furnace dust with molten iron containing carbon and silicon, ISS Trans. Iron Steelmak. 29 (2002) 55–67. [17] J.M. Yoo, B.S. Kim, J.C. Lee, M.S. Kim, C.W. Nam, Kinetics of the volatilization removal of lead in electric arc furnace dust, Mater. Trans. 46 (2005) 323–328, http://dx.doi.org/10.2320/matertrans.46.323. [18] A. Zabett, W.K. Lu, Thermodynamical computations for removal of alkali halides and lead compounds from electric arc furnace dust, Calphad 32 (2008)
151
535–542, http://dx.doi.org/10.1016/j.calphad.2008.05.001. [19] C.W. Bale, E. Belisle, P. Chartrand, S.A. Decterov, G. Eriksson, K. Hack, I.-H. Jung, Y.-B. Kang, J. Melançon, A.D. Pelton, C. Robelin, S. Petersen, FactSage thermochemical software and databases-recent developments, Calphad 33 (2009) 295–311, http://dx.doi.org/10.1016/j.calphad.2008.09.009. [20] M. Sh., Bafghi, M. Karimi, M. Adeli, A kinetic study on the carbothermic reduction of zinc oxide from electric arc furnace dust, Iran. J. Mater. Sci. Eng. 10 (2013) 18–30. [21] B.S. Kim, J.M. Yoo, J.T. Park, J.C. Lee, A kinetic study of the carbothermic reduction of zinc oxide with various additives, Mater. Trans. 47 (2006) 2421–2426, http://dx.doi.org/10.2320/matertrans.47.2421. [22] C.A. Pickles, Thermodynamic analysis of the selective carbothermic reduction of electric arc furnace dust, J. Hazard. Mater. 150 (2008) 265–278, http://dx. doi.org/10.1016/j.jhazmat.2007.04.097. [23] D.J. Min, J.W. Han, W.S. Chung, A study of the reduction rate of FeO in slag by solid carbon, Metall. Mater. Trans. B 30 (1999) 215–221, http://dx.doi.org/ 10.1007/s11663-999-0050-5.
Contents lists available at ScienceDirect
CALPHAD: Computer Coupling of Phase Diagrams and Thermochemistry journal homepage: www.elsevier.com/locate/calphad
Invited article
Thermodynamic analysis of the carbothermic reduction of electric arc furnace dust in the presence of ferrosilicon S.M. Moosavi Nezhad, Ahad Zabett n Department of Material Science and Engineering, Faculty of Engineering, Ferdowsi University of Mashhad, Khorasan Razavi, Iran
art ic l e i nf o
a b s t r a c t
Article history: Received 1 September 2015 Received in revised form 29 November 2015 Accepted 29 November 2015
Thermodynamic analysis of zinc recovery from electric arc furnace (EAF) dust by carbon in the presence of ferrosilicon was studied. A preheating process was performed to remove volatile compounds from dust to avoid impurities in the final zinc product. The main process includes reduction of zinc oxide and gasification and condensation of zinc vapor. Equilibrium condition was computed using FactSage program for the processes of reduction and gasification. The effect of ferrosilicon addition on the recovery of zinc was investigated for the temperature range of 950–1050 °C at the constant pressure of one atmosphere. It is shown that ferrosilicon provides a better kinetic condition for zinc recovery by facilitating slag formation. The preferred condition for formation of liquid slag was predicted based on thermodynamic calculations and verified by performing some experiments. The experimental results show that good agreement between calculations and experiments can be obtained with higher amounts of silicon contents and at higher temperatures where liquid slag is presented. A mixture of the dust with 10% C and 7% FeSi(75%) can be used to achieve over 97% zinc recovery. & 2015 Elsevier Ltd. All rights reserved.
Keywords: Electric arc furnace dust Ferrosilicon Zinc recovery Slag formation Kinetic condition
1. Introduction Electric arc furnace dust is classified as a hazardous material because of the presence of heavy metal constituents (i.e., leachable lead, cadmium and hexavalent chromium) [1–4]. More than 6 million tons of EAF dust is annually produced worldwide, which has to be treated to comply with the environmental regulations. One common choice of treating this toxic material is stabilization and landfilling. EAF dust contains significant amounts of zinc and iron. Since the use of galvanized steel continues to increase in the automotive industry, the amount of zinc in the dust will rise concurrently. There are many different processes to recover zinc from EAF dust [5–12]. Today, about 30% of the industrial zinc is produced by recovery from secondary zinc resources such as galvanized sheets and steelmaking dusts [13]. Zinc recovery from EAF dust is economically feasible when zinc concentration is more than 15% [14]. Pyro-metallurgical treatments are main industrial technologies for zinc recovery from EAF dust whereas 80% of recycled dust is treated by Waelz kiln process. In pyro-metallurgical methods EAF dust is reduced at a high temperature. Zinc can be recovered from the vapor and the iron can be recovered in the form of liquid metal. n
Corresponding author.
http://dx.doi.org/10.1016/j.calphad.2015.11.003 0364-5916/& 2015 Elsevier Ltd. All rights reserved.
Many researchers have studied different aspects of zinc recovery from EAF dust in pyrometallurgical processes. In most of these works carbonaceous materials are used as the reductant. Pickles [15] studied the reaction of EAF dust with solid iron powder in argon atmosphere and reported zinc recovery of 87% at 1050 °C. In another work [16] he considered the reaction of EAF dust with molten pig iron containing carbon and silicon. He showed that in this process, zinc can be completely reduced and recovered in the form of crude zinc oxide. Silicon can be an alternative reducing agent for carbonaceous materials. Fig.1 shows changes in standard free energy of oxidation for silicon, zinc, iron, wustite and magnetite by considering the following reactions. 4Fe3O4(s) þO2(g) ¼6Fe2O3(s)
(1)
6FeO(s) þ O2(g) ¼2Fe3O4(s)
(2)
2Fe(s) þO2(g) ¼ 2FeO(s)
(3)
2Zn(s) þO2(g) ¼2ZnO(s)
(4)
Si(s) þO2(g) ¼SiO2(s)
(5)
As it can be seen in Fig. 1, silica is more stable than zinc and iron oxides and therefore reduction of these oxides by silicon is
144
S.M. Moosavi Nezhad, A. Zabett / CALPHAD: Computer Coupling of Phase Diagrams and Thermochemistry 52 (2016) 143–151
Fig. 1. Standard free energy changes of oxidation for Zn, Fe, FeO, Fe3O4 and Si according to reactions (1–5).
Fig. 2. Calculated zinc loss during the pretreatment of EAF dust.
thermodynamically possible. Using silicon for reduction of EAF dust increases silica content of the residue and forms low melting temperature slag. This could enhance the rate of reactions by increasing the mass and heat transfer in the condense phases. Due to the higher cost of silicon use of a mixture of silicon and carbon are proposed as the reducing agent for the zinc recovery from EAF dust. The aim of the present research is to investigate the thermodynamic conditions for reduction of EAF dust by a mixture of carbon and ferrosilicon. Alkali halides and lead oxide are usually problematic in the recovery of zinc from EAF dust, especially in the zinc metal recovery processes. Therefore, according to the literature [17,18] a preheating step for removing the more volatile species from EAF dust at a lower temperature was suggested. In this pretreatment, volatile impurities such as PbO, NaCl, KCl and CdO are removed from EAF dust.
2. Thermodynamic calculations and experiments Fig. 3. Calculated amounts of air needed for 95% removal degree of volatile species during the pretreatment.
2.1. Calculations FactSage 6.1 program was used to study the equilibrium conditions of the reactions involved in the heat treatment process. EQUILIB (a module of FactSage) [19] utilizes the chemical composition of reactants in the form of compounds or elements as the input and determines the combination of the most stable products at any desired temperature and pressure. Calculations are based on minimization of Gibbs free energy of the system at isothermal and isobaric conditions. For the pretreatment, heating of 100 g dust in the presence of different amounts of air as a carrier gas was studied. Then, computations of thermodynamic equilibrium of the pretreated dust (solid residue remaining from the pretreatment) with different amounts of carbon and ferrosilicon at different temperatures were performed. The amounts of carbon and ferrosilicon were increased from 1 to 20 g by a step size of 1 g. Temperature was increased from 850 °C to 900 °C in steps of 5 °C for pretreatment and from 950 °C to 1050 °C in steps of 10 °C for the reduction. Input data was
introduced to EQUILIB program in the form of elemental composition. The chemical compositions of the real EAF dust samples before and after pretreatment were used. 2.2. Experiments EAF dust was obtained from an alloy steelmaking company in Iran. Graphite powder was provided by Sinopharm Chemical Reagent Co. Ltd. and ferrosilicon powder containing 75% silicon and 25% iron from Iran Ferrosilice Co. was used. Chemical composition of the EAF dust and the residues of the pretreatment were analyzed by an inductively coupled plasma-mass spectrometer (ICPMS, Spectro MS). The carbon content of the dust was determined by a Leco CS 244 analyzer. For the pretreatment process, 100 g EAF dust was heated in an alumina boat (200 mm in length, 100 mm in width and 30 mm in height) by an electric furnace at 875 °C for 4 h and cooled inside
Table 1 Chemical composition of the EAF dust. Element
Fe
Zn
Ca
Pb
Na
Mg
Al
Cd
Cr
K
Si
C
Cl
wt%
30.0
19.02
4.5
0.99
3.38
4.99
0.53
0.04
0.34
0.16
0.16
2.1
6.45
S.M. Moosavi Nezhad, A. Zabett / CALPHAD: Computer Coupling of Phase Diagrams and Thermochemistry 52 (2016) 143–151
Fig. 4. Calculated recovery of (a) zinc and (b) iron from dust at 950 °C.
Fig. 5. Calculated recovery of (a) zinc and (b) iron from dust at 1050 °C.
145
146
S.M. Moosavi Nezhad, A. Zabett / CALPHAD: Computer Coupling of Phase Diagrams and Thermochemistry 52 (2016) 143–151
Fig. 6. Calculated CO:CO2 ratio in vapor phase at (a) 950 °C and (b) 1050 °C.
the furnace to room temperature. For the recovery process, 2 g of the pretreated dust was mixed with different amounts of graphite and ferrosilicon powders and placed in a cylindrical alumina crucible (35 mm in diameter and 130 mm in height). To avoid direct contact of the mixture with air, three grams of Al2O3 powder was placed over the mixture. The furnace was preheated and the crucible was inserted into the furnace at 1000 °C or 1050 °C. Samples were air cooled after heat treatment. For each sample, recovery or removal efficiency (η) of the metals was calculated as follows:
shows zinc loss as a function of temperature. As it can be seen, zinc loss increases with increasing temperature which is not desired. Zabett and Lu [18] also reported that reducing condition and higher temperatures, increase zinc loss due to reduction and evaporation. Based on Fig. 2, heating of the dust at a temperature lower than 900 °C is appropriate to avoid zinc loss. In order to determine the proper temperature for removal of volatile species, number of moles of air required for 95% removal of lead, potassium and cadmium at 850–900 °C is plotted in Fig. 3. As it can be seen, the number of moles of air at 875 °C to achieve removal degree of 95% for Pb, K and Cd, are 73, 93 and 108, respectively.
ηE (wt%) = 100⋅(Wt Ei ⋅MiT − Wt Ef ⋅MfT )/(Wt Ei ⋅MiT )
3.1.2. Recovery process The recovery percentage of zinc and iron as a function of carbon and ferrosilicon at 950 °C is shown as a 3-D plot in Fig. 4. The figure shows that at zero amount of ferrosilicon, complete zinc recovery can be achieved with the addition of 14% carbon to the dust. At low amounts of reductants, there is a region at which zinc recovery is relatively constant. The boundaries of this triangular region are approximately limited to 9% carbon and 15% ferrosilicon. This corresponds to a region in Fig. 4b at which iron recovery increases very rapidly by increasing the amount of reductants. Next to this region, increasing the amounts of carbon or ferrosilicon enhances zinc recovery. For example at 950 °C, in the presence of 5% carbon, zinc recovery increases after 95% iron recovery. It can be concluded that at low amounts of reductants (carbon and silicon), reduction of zinc oxide will thermodynamically enhance after substantial iron recovery. This trend can be explained by considering the following reactions:
WtEi
(6)
WtEf
where and are the weight percent of the element E before and after the heat treatment. MiT and MfT are the total weights of the dust before and after the heat treatment. The percentage of iron recovery is calculated as follows: T Fe T ηE (wt%) = 100⋅Wt Fe f ⋅Mf /Wt i ⋅Mi
(7)
Fe where WtFe i and Wtf are the weight percent of iron in dust before treatment and the weight percent of metallic iron in the final solid residue.
3. Results and discussion 3.1. Thermodynamic calculations 3.1.1. Pretreatment process Chemical composition of the dust is reported in Table 1. Presence of 2.1% carbon in the dust, results in increasing reduction condition and thus higher zinc loss in the preheating step. Fig. 2
FeO þCO(g)¼Fe(s)þCO2(g)
(8)
ZnOþCO(g)¼Zn(g)þ CO2(g)
(9)
S.M. Moosavi Nezhad, A. Zabett / CALPHAD: Computer Coupling of Phase Diagrams and Thermochemistry 52 (2016) 143–151
Fig. 7. Calculated CO:CO2 ratio in vapor phase. Mixture of dust and 5% carbon at (a) 950 °C and (b) 1050 °C.
2FeO þSi(s) ¼2Fe(s)þ SiO2(s)
(10)
2ZnO þSi(s) ¼2Zn(g) þSiO2(s)
(11)
In the absence of ferrosilicon, zinc and iron recoveries are mainly occurred by indirect reduction reactions (8) and (9) [20,21]. Free energy change of reaction (9) is directly related to the partial pressure of zinc. In the early stages of the dust reduction, when partial pressure of zinc is very low, reaction (9) will be thermodynamically more favorable than reaction (8). As the partial pressure of zinc increases, free energy changes of reaction (9) increases and reaction (8) will be thermodynamically more favorable. With the addition of ferrosilicon, zinc and iron oxides of the dust can be reduced according to reactions (10) and (11). While FeO is in the form of pure solid, assuming activity one, and the partial pressure of zinc is not very low, free energy changes of reaction (10) is lower than that of reaction (11). This means that silicon is consumed earlier to reduce iron oxide. At higher amounts of ferrosilicon, solid FeO disappears and enters the slag liquid phase. This decreases FeO activity and increases free energy changes of reaction (10). At this condition, free energy changes of reactions (10) and (11) are comparable and zinc recovery increases. Zinc and iron recoveries at 1050 °C are plotted as a function of carbon and ferrosilicon in Fig. 5a and b, respectively. Results indicate that at zero amount of ferrosilicon, raising the temperature results in decreasing of iron recovery and increasing of zinc recovery. Same results are also reported by Pickles [22]. At this condition, complete zinc recovery can be achieved with the addition of 13% carbon to the dust. At low amount of carbon and
147
higher temperature free energy changes of reaction (9) reduces and CO:CO2 ratio increases. This rises the zinc vapor pressure. Therefore, zinc recovery enhances at higher temperatures. Since carbon is completely used for the reduction of zinc and iron oxides, at a fixed amount of carbon, higher amount of zinc recovery causes lower amount of iron recovery. Ferrosilicon addition to the mixture of EAF dust and carbon improves reducing conditions. This changes oxygen potential and CO:CO2 ratio in the system. Therefore, computing the amount of CO:CO2 ratio can help to better understand the effect of ferrosilicon addition on zinc recovery. CO:CO2 ratio as a function of ferrosilicon weight percent is plotted in Fig. 6. At 950 °C and constant carbon content, by increasing the amount of ferrosilicon, first CO:CO2 ratio remains constant and then rises. At higher ferrosilicon values a rapid enhancement occurs in CO:CO2 ratio and then this ratio becomes higher with an increasing slope. A similar trend with higher values of CO:CO2 ratio is obtained at 1050 °C, as shown in Fig. 6b. Comparing Figs. 4a and 5a with Fig. 6 indicates that there is a close relation between zinc recovery and CO:CO2 ratio. To examine this relation, at fixed carbon content of 5%, zinc and iron recoveries, CO:CO2 ratio, slag weight and the weight of solid FeO is plotted against the amount of ferrosilicon at 950 °C and 1050 °C, see Fig. 7. According to Fig. 7a, four distinct zones are recognized. The first zone with low ferrosilicon where zinc recovery and CO:CO2 ratio do not change. The second zone with medium amount of ferrosilicon where zinc recovery and CO:CO2 ratio increase simultaneously. The third zone with a higher amount of ferrosilicon where zinc recovery rises while iron recovery is completed. The fourth zone with the highest amount of ferrosilicon where both iron and zinc recoveries are completed and CO:CO2 ratios rises slowly. As previously discussed, at a low amount of reductants, zinc recovery increases after a substantial increase in the iron recovery. At the first zone, iron recovery increases according to reaction (10) and therefore CO:CO2 ratio does not change. With increasing ferrosilicon, slag forms and after substantial decrease in the weight of solid FeO, remaining FeO enters the slag phase. At the second zone with elimination of solid FeO and because of lower activity of FeO in the slag, reduction of zinc oxide begins to increase and leads to an increase in the zinc recovery. At this zone, CO:CO2 ratio is under the influence of FeO reduction with CO (reaction (8)). However, these values change very rapidly at the end of the second zone when slag disappears and 100% iron recovery is achieved. After complete reduction of iron oxides (the third zone), CO:CO2 ratio is under the influence of reaction (9). After completion of ZnO reduction (fourth zone) CO:CO2 ratio is affected by Boudouard reaction as follows: C þCO2(g)¼ 2CO(g)
(12)
Fig. 7b shows CO:CO2 ratio as a function of ferrosilicon at 1050 °C. At the higher temperature, the lengths of zones 1 and 2 have increased and zone 3 has disappeared. In the presence of carbon at zero amount of ferrosilicon, iron recovery decreases as a result of temperature increase. Therefore, completion of iron recovery needs higher amount of reductant (e.g. ferrosilicon). As it can be seen in Fig. 7, iron recovery completes with 14% ferrosilicon at 950 °C and with 16% ferrosilicon at 1050 °C. On the other hand, at fixed amounts of reduction agents, CO:CO2 ratio at 1050 °C is higher than 950 °C in the fourth zone. For example at ferrosilicon amount of 20%, CO:CO2 ratios at 950 °C and 1050 °C are 137 and 476, respectively. This is due to the fact that CO is more stable than CO2 at higher temperatures and therefore CO:CO2 ratio is larger at higher temperatures. According to Fig. 7, slag weight and composition affect the zinc and iron oxides reduction reactions. This effect is represented in
148
S.M. Moosavi Nezhad, A. Zabett / CALPHAD: Computer Coupling of Phase Diagrams and Thermochemistry 52 (2016) 143–151
Fig. 8. Calculated weight of liquid slag at (a) 950 °C and (b) 1050 °C.
Table 2 Calculated composition of the slag obtained from heating of 100 g dust at different conditions. Conditions
Slag composition and weight
Temperature (°C)
C (wt%)
FeSi (wt%)
Slag weight (g)
FeO (wt%)
ZnO (wt%)
SiO2 (wt%)
MgO (wt%)
Al2O3 (wt%)
CaO (wt%)
950 950 950 950 950 950 1050 1050 1050 1050 1050 1050
0 0 5 10 15 20 0 0 5 10 15 20
7 14 7 7 7 7 7 14 7 7 7 7
13.47 28.88 12.57 8.77 7.52 7.52 20.71 75.29 20.85 10.19 8.65 8.65
14.54 22.01 9.78 0.55 0.17 0.17 33.82 21.61 18.20 21.57 1.66 0.11
16.32 20.20 13.77 4.44 1.36 1.36 20.06 28.89 20.06 2.79 0.17 0.17
30.74 30.74 35.20 42.37 43.49 43.49 26.33 30.01 26.37 41.88 43.56 43.56
31.50 18.21 34.73 48.12 50.94 50.94 20.12 12.21 20.14 42.53 47.35 47.35
5.29 3.24 2.55 1.31 1.13 1.13 4.72 1.25 4.81 3.72 2.64 2.64
1.38 5.26 3.81 3.11 2.81 2.81 6.92 9.28 6.94 7.33 6.10 6.10
Fig. 8 where the slag weight is plotted against carbon and silicon at 950 °C and 1050 °C. Fig. 8a shows that at small amounts of ferrosilicon (o5%), there is no slag phase in the system. At higher amounts of ferrosilicon, SiO2 increases and leads to the formation of slag with low melting point. At this condition, SiO2, MgO, ZnO and FeO form the main part of the slag. The weight and composition of the slag phase at different conditions are given in Table 2. The weight of slag becomes higher with increasing silica, reaches a maximum and then decreases as its FeO content decreases. These changes rise the slag melting point and reduce the slag weight. After completing the FeO reduction, the weight of slag will become zero. Any change in the carbon content of the dust also has an effect on slag formation and zinc recovery. At a medium amount of ferrosilicon (5–10%), less slag is available at higher carbon content.
Since zinc and iron recoveries increase at higher carbon contents, lower amounts of zinc and iron oxides are available to form the slag phase. In the absence of carbon, complete removal of the slag occurs at 18% of ferrosilicon while in the presence of 10% of carbon this elimination occurs at 11% of ferrosilicon. On the other hand at relatively large amounts of carbon, where complete zinc and iron recoveries occur, iron and zinc oxides do not exist in the slag and it mainly contains MgO and SiO2 (see Table 2). Changes in the slag weight as a function of ferrosilicon and carbon at 1050 °C is approximately similar to that observed at 950 °C. Temperature elevation facilitates slag formation leading to slag weight increase with temperature. Under such a condition slag remains up to higher amount of ferrosilicon. According to Min et al. [23] since the molten phase (slag) provides extensive interfacial area with different reductants, slag
S.M. Moosavi Nezhad, A. Zabett / CALPHAD: Computer Coupling of Phase Diagrams and Thermochemistry 52 (2016) 143–151
149
Fig. 9. Calculated zinc recovery from dust mixed with 10% carbon.
Fig. 10. Comparison of the removal degrees of volatiles in pretreatment from calculations and experiments.
formation plays an important role on the reduction rate of metal oxides in ironmaking and steelmaking processes. Zabett and Lu [18] have shown that progress in the removal of volatile species from EAF dust is directly related to the change in the weight of the liquid phase (slag phase). Therefore, formation of a slag phase in the pyrometallurgical zinc recovery from EAF dust can affect the kinetics of ZnO reduction. For the formation of the slag phase, at
Fig. 12. Zinc recovery from dust mixed with fixed amount of Cþ Si¼ 10 wt%.
least 5% ferrosilicon is needed. Due to higher price of ferrosilicon in comparison with carbonaceous materials such as coal or coke, unnecessary use of this reductant for non-essential iron recovery should be avoided. Therefore a mixture of carbon bearing materials and ferrosilicon, as reducing agents, is economically proper when carbon is sufficient to complete reduction of iron oxides. According to Figs. 4a and 5a, at zero amount of ferrosilicon,
Fig. 11. Comparison of the zinc recoveries from calculations and experiments at (a) 950 °C and (b) 1050 °C.
150
S.M. Moosavi Nezhad, A. Zabett / CALPHAD: Computer Coupling of Phase Diagrams and Thermochemistry 52 (2016) 143–151
complete recovery of iron is obtained at 10–12% carbon in the temperature range of 950–1050 °C. Fig. 9 shows zinc recovery as a function of temperature and amount of ferrosilicon at carbon content of 10%. At least 7% ferrosilicon is necessary to achieve zinc recovery of 97% at 950 °C, as can be seen in Fig. 9. The weight of slag under this condition is approximately 13%. At 1050 °C, addition of 4% ferrosilicon is sufficient to achieve 97% zinc recovery. However, under this condition slag phase does not exist. In the absence of slag, because of limited contact between reactants, the expected zinc recovery value will not be achieved. With the addition of 7% ferrosilicon, 21% slag will be produced at 1050 °C. The above discussion indicates that a mixture of 10% of C–7%FeSi could be proper as a reductant to recover zinc from the dust in the temperature range of 950–1050 °C. 3.2. Verifying thermodynamic computations To verify computer calculations, the preheating and the main heating processes were experimentally examined. The computation and experimental results for removal of Pb, K, Na and Cd are given in Fig. 10. As it can be seen, the results are in good agreement with calculations. Based on the experimental results, a removal degree of 95% for K, Pb and Cd could be achieved by performing a preheating process at 875 °C for 4 h. Zinc recovery from pretreated dust mixed with different amounts of carbon and ferrosilicon heated for 1 h at 950–1050 °C is shown in Fig. 11. Results of calculations are also represented for comparison. Based on the calculations, complete zinc recovery can be achieved at 950 °C in both samples containing 20% of C and 10% of C þ10% of FeSi. However zinc recovery percentages of 79% and 53% were achieved after 1 h heating at 950 °C for samples with and without ferrosilicon, respectively. Higher zinc recovery of the sample containing ferrosilicon is due to slag formation. Moreover, at the higher temperature better agreement between computations and experiments was attained. Because of higher rate of mass and heat transfer and larger amount of slag phase at higher temperatures, the rate of reactions has increased and experimental conditions have approached the equilibrium condition. To better investigate simultaneous effect of carbon and silicon as reducing agents, a set of experiments was designed. In these experiments, mixtures of pretreated dust with the addition of 10% of C þSi, at different C:Si ratios, were treated at 1000 °C for 1 h. Zinc recovery from these samples is plotted versus weight percentage of silicon in Fig. 12. For the purpose of comparison, the results of thermodynamic calculations for same process are shown in this figure. Calculations indicate that in a fixed amount of the reductant, substitution of carbon by silicon decreases zinc recovery. To achieve same zinc recovery 1% carbon should be replaced with 2% ferrosilicon. Experiments show that zinc recovery increases with increasing the Si:C ratio, reaches a maximum and then decreases. At higher amounts of silicon, the weight of slag has increased and experimental conditions have approached the thermodynamic predictions. Based on the experimental observations, partial melting has happened in the samples with higher amounts of ferrosilicon. This verifies the prediction that ferrosilicon addition increases zinc recovery by increasing the weight of slag which provides better contact between reductants and metal oxides.
4. Conclusion Zinc recovery from EAF dust using carbon and ferrosilicon was studied. Thermodyanaic equilibrium conditions were calculated to study the process and to obtain the proper conditions for zinc reduction and vaporization. Results show that complete zinc
recovery in the temperature range of 950–1050 °C may be achieved by addition of about 14% carbon to the dust. However there is no liquid slag in the system when only carbon is used for reduction and therefore kinetics of the process could be slow. Ferrosilicon addition could improve kinetic conditions for ZnO reduction by facilitating low melting temperature slag formation. A liquid slag forms with the addition of at least 5% ferrosilicon to the dust. When a mixture of carbon and ferrosilicon is used for reduction, same zinc recovery can be achieved by substitution of about 2 g ferrosilicon (Si 75%) for every gram carbon. Thermodynamic computation shows that zinc recovery of over 97% could be attained with a mixture of 10% of C þ7% of FeSi(75%). Prediction of better kinetic condition by ferrosilicon addition was verified by experiments. Higher amount of ferrosilicon used for reduction of oxides lead to higher amount of liquid slag and higher amount of zinc recovery. As the amount of silicon in the mixture increases, results of thermodynamic computations and experiments show better convergence.
Acknowledgments The authors are thankful to Amirkabir University of Technology for their support in using FactSage 6.1. Financial support from Isfahan Alloy Steel Company is also greatly appreciated.
Appendix A. Supplementary material Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.calphad.2015.11.003.
References [1] M. Al-harahsheh, S. Kingman, L. Al-Makhadmah, I.E. Hamilton, Microwave treatment of electric arc furnace dust with PVC: dielectric characterization and pyrolysis-leaching, J. Hazard. Mater. 274 (2014) 87–97, http://dx.doi.org/ 10.1016/j.jhazmat.2014.03.019. [2] T. Suetens, B. Klaasen, K. Van Ackerw, B. Blanpain, Comparison of electric arc furnace dust treatment technologies using exergy efficiency, J. Clean. Prod. 65 (2014) 152–167, http://dx.doi.org/10.1016/j.jclepro.2013.09.053. [3] B.S. Yu, Y.R. Wang, T.C. Chang, Hydrothermal treatment of electric arc furnace dust, J. Hazard. Mater. 190 (2011) 397–402, http://dx.doi.org/10.1016/j. jhazmat.2011.03.056. [4] R.A. Shawabkeh, Hydrometallurgical extraction of zinc from Jordanian electric arc furnace dust, Hydrometallurgy 104 (2010) 61–65, http://dx.doi.org/ 10.1016/j.hydromet.2010.04.014. [5] C.F. Pereira, Y. Luna, X. Querol, D. Antenucci, J. Vale, Waste stabilization/solidification of an electric arc furnace dust using fly ash-based geopolymers, Fuel 88 (2009) 1185–1193, http://dx.doi.org/10.1016/j.fuel.2008.01.021. [6] R.N.C. De Siqueira, E. De Albuquerque Brocchi, P.F. De Oliveira, M.S. Motta, Hydrogen reduction of zinc and iron oxides containing mixtures, Metall. Mater. Trans. B 45 (2014) 66–75, http://dx.doi.org/10.1007/s11663-013-9951-4. [7] M.K. Jha, V. Kumar, R.J. Singh, Review of hydrometallurgical recovery of zinc from industrial wastes, Resour. Conserv. Recycl. 33 (2001) 1–22, http://dx.doi. org/10.1016/S0921-3449(00)00095-1. [8] C.A. Pickles, Thermodynamic modelling of the formation of zinc–manganese ferrite spinel in electric arc furnace dust, J. Hazard. Mater. 179 (2010) 309–317, http://dx.doi.org/10.1016/j.jhazmat.2010.03.005. [9] V.P. Korneev, V.P. Sirotinkin, N.V. Petrakova, V.G. Dyubanov, L.I. Leont'ev, Physicochemical properties of the zinc-containing dusts of electric furnace steelmaking, Russ. Metall. 7 (2013) 507–512, http://dx.doi.org/10.1134/ S0036029513070094. [10] C.A. Pickles, Thermodynamic analysis of the selective chlorination of electric arc furnace dust, J. Hazard. Mater. 166 (2009) 1030–1042, http://dx.doi.org/ 10.1016/j.jhazmat.2008.11.110. [11] N. Tsubouchi, H. Hashimoto, N. Ohtaka, Y. Ohtsuka, Chemical characterization of dust particles recovered from bag filters of electric arc furnaces for steelmaking: Some factors influencing the formation of hexachlorobenzene, J. Hazard. Mater. 183 (2010) 116–124, http://dx.doi.org/10.1016/j. jhazmat.2010.06.122. [12] N.Y. Ma, On in-process separation of zinc from EAF dust, in: S.N. Monteiro, D. E. Verhulst, P.N. Anyalebechi, J.A. Pomykala (Eds.), EPD Congress, John Wiley & Sons; Hoboken, NJ, USA, 2011, pp. 947–952.
S.M. Moosavi Nezhad, A. Zabett / CALPHAD: Computer Coupling of Phase Diagrams and Thermochemistry 52 (2016) 143–151
[13] N. Antunano, D. Herrero, P.L. Arias, J.F. Cambra, Electrowinning studies for metallic zinc production from double leached Waelz oxide, Process Saf. Environ. Prot. 91 (2013) 495–502, http://dx.doi.org/10.1016/j.psep.2012.10.014. [14] M.H. Morcali, O. Yucel, A. Aydin, B. Derin, Carbothermic reduction of electric arc furnace dust and calcination of waelz oxide by semi-pilot scale rotary furnace, J. Min. Metall. B 48 (2012) 173–184, http://dx.doi.org/10.2298/ JMMB111219031M. [15] J.R. Donald, C.A. Pickles, Reduction of electric arc furnace dust with solid iron powder, Can. Metall. Q. 35 (1996) 255–267, http://dx.doi.org/10.1016/ 0008-4433(96)00009-2. [16] C.A. Pickles, Reaction of electric arc furnace dust with molten iron containing carbon and silicon, ISS Trans. Iron Steelmak. 29 (2002) 55–67. [17] J.M. Yoo, B.S. Kim, J.C. Lee, M.S. Kim, C.W. Nam, Kinetics of the volatilization removal of lead in electric arc furnace dust, Mater. Trans. 46 (2005) 323–328, http://dx.doi.org/10.2320/matertrans.46.323. [18] A. Zabett, W.K. Lu, Thermodynamical computations for removal of alkali halides and lead compounds from electric arc furnace dust, Calphad 32 (2008)
151
535–542, http://dx.doi.org/10.1016/j.calphad.2008.05.001. [19] C.W. Bale, E. Belisle, P. Chartrand, S.A. Decterov, G. Eriksson, K. Hack, I.-H. Jung, Y.-B. Kang, J. Melançon, A.D. Pelton, C. Robelin, S. Petersen, FactSage thermochemical software and databases-recent developments, Calphad 33 (2009) 295–311, http://dx.doi.org/10.1016/j.calphad.2008.09.009. [20] M. Sh., Bafghi, M. Karimi, M. Adeli, A kinetic study on the carbothermic reduction of zinc oxide from electric arc furnace dust, Iran. J. Mater. Sci. Eng. 10 (2013) 18–30. [21] B.S. Kim, J.M. Yoo, J.T. Park, J.C. Lee, A kinetic study of the carbothermic reduction of zinc oxide with various additives, Mater. Trans. 47 (2006) 2421–2426, http://dx.doi.org/10.2320/matertrans.47.2421. [22] C.A. Pickles, Thermodynamic analysis of the selective carbothermic reduction of electric arc furnace dust, J. Hazard. Mater. 150 (2008) 265–278, http://dx. doi.org/10.1016/j.jhazmat.2007.04.097. [23] D.J. Min, J.W. Han, W.S. Chung, A study of the reduction rate of FeO in slag by solid carbon, Metall. Mater. Trans. B 30 (1999) 215–221, http://dx.doi.org/ 10.1007/s11663-999-0050-5.