- Email: [email protected]
Production of magnesium during carbothermal reduction of magnesium oxide by differential condensation of magnesium and alkali vapours
Available online at www.sciencedirect.com
ScienceDirect Journal of Magnesium and Alloys 1 (2013) 323e329 www.elsevier.com/journals/journal-of-magnesium-and-alloys/2213-9567
Full length article
Production of magnesium during carbothermal reduction of magnesium oxide by differential condensation of magnesium and alkali vapours* Cheng-bo Yang a,b,c, Yang Tian a,b,c,*, Tao Qu a,b,c, Bin Yang a,b,c, Bao-qiang Xu a,b,c, Yong-nian Dai a,b,c a
National Engineering Laboratory for Vacuum Metallurgy, Kunming University of Science and Technology, Kunming 650093, China b Key Laboratory for Nonferrous Vacuum Metallurgy of Yunnan Province, Kunming 650093, China c State Key Laboratory Breeding Base of Complex Nonferrous Metal Resources Clear Utilization in Yunnan Province, Kunming 650093, China Received 18 December 2013; accepted 4 January 2014
Abstract Most of researchers believed that the developments on the condensation of magnesium produced by carbothermic reduction just concentrated on two process routes: the “quench” route and the “solvent” route. But this paper will briefly analyzes the major challenges in magnesium vapor condensation during the vacuum carbothermic reduction of calcined dolomite, on equipment upgrade, heat transfers alter, to achieve condensation control and production collection. Solutions are then proposed using theoretical calculations and experiment results. Comparative analysis of the experiment results shows that the burning and even explosion of condensation products during the vacuum carbothermic reduction of calcined dolomite are mainly due to the burning of crystallized powder magnesium, which results from the self-ignition of alkali metals. Finally, this paper proposes a multistage condensation solution to improve traditional vacuum condensation equipment. And result show that the condensation equipment can effectively mitigate the burning and loss during condensation, also the morphology of the condensation products clearly improved, the grain size increased, and the oxidation rate decreased. The potassium/sodium vapor and the magnesium vapor were separately condensed. Copyright 2014, National Engineering Research Center for Magnesium Alloys of China, Chongqing University. Production and hosting by Elsevier B.V. Open access under CC BY-NC-ND license. Keywords: Carbothermic reduction; Calcined dolomite; Magnesium vapor; Condensation
* Foundation item: Supported by National Natural Science Foundation of China (No. 51304095); Science and Technology Planning Project of Yunnan Province, China (No. S2013FZ029); the personnel training Funds of Kunming University of Science and Technology, China (No. 14118665). * Corresponding author. National Engineering Laboratory for Vacuum Metallurgy, Kunming University of Science and Technology, Kunming 650093, China. Tel.: þ86 871 65161583. E-mail addresses: [email protected] (C.-b. Yang), emontian@hotmail. com (Y. Tian). Peer review under responsibility of National Engineering Research Center for Magnesium Alloys of China, Chongqing University
Production and hosting by Elsevier
Carbothermic reduction is an alternative to both silicothermic and electrolytic processes for the production of magnesium. It has the advantages of low reductant cost, high equipment utilization rate, and environment-friendliness compared to the first two techniques. The current domination of world production by the Pidgeon process mainly reflects the economic conditions of China rather than any inherent advantages of the Pidgeon process over electrolytic routes. Under environment pressure, this process has always been favored by a great deal of researchers [1e3]. On condensation of magnesium vapor during Carbothermic reduction aspect, Hansgirg [4] tackled the problem of reversion and consolidation of fine magnesium using a wide variety of condensation techniques including quenching the magnesium vapor with water-cooled pipes and mechanically scraping the condensate from the tubes, shock cooling through
2213-9567 Copyright 2014, National Engineering Research Center for Magnesium Alloys of China, Chongqing University. Production and hosting by Elsevier B.V. Open access under CC BY-NC-ND license. http://dx.doi.org/10.1016/j.jma.2014.01.002.
324
C.-b. Yang et al. / Journal of Magnesium and Alloys 1 (2013) 323e329
injection of a cold gas, chilling with hydrocarbon oils and subsequently distilling magnesium from the slurry formed. All of these approaches failed to overcome the key problems of flammable powders and effective consolidation of fine powders, though his papers and numerous patents provide a thorough record of process development in this area and are essential reading for any researcher contemplating carbothermic production of magnesium. Cameron [5,6] and co-workers further investigated this route at the pilot plant level in the1980s, concentrating on the reduction stage of the process and providing useful thermodynamic analysis of impurity distributions between slag and vapor phases. However, the experiment was announced as a failure because the magnesium powders blocked the pipelines. Researchers of the CSIRO recently made a breakthrough in a semi-industrial experiment [7]. They used a Laval nozzle to create an environment where a supersonic gas is cooled rapidly at more than 106 C S1 [8e10], facilitating homogeneous nucleation and condensation of the magnesium vapour [11,12]. Meanwhile, our research group National Engineering Laboratory for Vacuum Metallurgy achieved significant progress in the carbothermic reduction of magnesia, which paved the way for the industrial applications of magnesium metal production using the vacuum carbothermic reduction method. Li [13] and Luo [14] conducted studies on the vacuum carbothermic reduction of magnesite and lateritic nickel ores, respectively. Li [15], Tian [16] and Yu [17] also made significant progress in reaction chamber spraying, catalytic mechanisms, and improvements in the reduction rate. At present, our experimental studies focus on the vacuum carbothermic reduction of calcined dolomite. However, the condensation of magnesium vapor fails to meet the collection requirements because of the small grain size of the obtained magnesium metal. In addition, the high oxidation rate easily leads to burning or even explosion, which significantly increases the damage rate due to burning of magnesium and reduces production safety. Therefore, the effective control of magnesium vapor condensation has become the key factor to the successful application of this method to industrial production. This paper is to probe into the reason of the problem, and try to resolve. 1. Theoretical analysis 1.1. Heat input to the condensing zone Fig. 1 shows that the heat in the reaction zone (high-temperature A) is radiated into the condensing zone (low-temperature B) in three ways: (1) The reaction zone radiates heat to the condensing zone through coupling sleeves (graphite condensing pipe C). Heat radiation is the major heat-transfer process at high temperatures. In addition, heat can be transferred through a vacuum without requiring any material media; (2) Superheated magnesium vapor radiates heat when it enters the condensing zone and is condensed into solid magnesium; and
Fig. 1. Schematic model of a vacuum furnace.
(3) Heat is radiated to the condensing zone through the crucible surface in the reduction zone. Given that the crucible in the entire reduction zone is sealed and wrapped by an insulation quilt, only coupling sleeve C is connected to the condensing zone. Therefore, compared with the two previously described input processes, this process is not considered here. Assume that high-temperature zone A is the surrounded object and condensing zone B is the peripheral object. Then [18], " 4 4 # T1 T2 Q12 ¼ C12 A1 : 100 100
Therefore; C12 ¼ h
1
31
þ AA12
C0
1
32
1
i
where C1e2 is the emission coefficient, C0 is the blackbody emission coefficient, 3 1 is the blackness of the surrounded object, 3 2 is the blackness of the peripheral object, A1 is the radiation area of the surrounded object, A2 is the radiation area of the peripheral object, T1 is the temperature in the reaction zone, and T2 is the temperature in the condensing zone. When the condenser size is specified and the temperature exceeds 1000 C, then A1, A2, 3 1, and 3 2 can all be considered as fixed values [19], and T1 is the only variable. By satisfying the technological conditions as well as the small variation range of the temperature, the reduction temperature T1 contributes to the relative stability of the radiation heat in the condensing zone. 1.2. Heat released by the condensation of magnesium vapor into solid magnesium The condensation of magnesium vapor into solid magnesium involves five steps [20]. The heat that enters the condensing zone is then calculated based on the five-step process. The five steps are as follows: (1) The heat released by the cooling of the superheated vapor to the boiling temperature is denoted as QI: QI ¼ Cvapor ðts tb Þ$m ¼ 0:204ð1350 1090Þ$m ¼ 53:04m
C.-b. Yang et al. / Journal of Magnesium and Alloys 1 (2013) 323e329
325
where Cvapor is the vapor emission coefficient, ts is the superheated vapor temperature, tb is the vapor boiling temperature, and m is the magnesium vapor in the condensing zone per unit time;
(1) Transfer between the residual gas in the furnace and the furnace wall; (2) Heat release from the cooling water in the condensing area.
(2) The evaporation latent heat released by the magnesium vapor at the normal boiling temperature is denoted as QII:
The water coefficient of the thermal conductivity is tens of times higher than the air coefficient. Thus, the heat lost due to cooling water is higher than those due to other factors. Therefore, the heat output QO is
QII ¼ 1337m; (3) The heat released by the cooling liquid magnesium from the normal boiling temperature (tb) to the melting temperature (tm) is denoted as QIII. Cl is the specific heat of liquid magnesium. QIII ¼ Cl ðtb tm Þ$m ¼ 0:305ð1090 651Þ$m ¼ 133:9m; (4) The melting latent heat released by the condensation of liquid magnesium is denoted as QIV: QIV ¼ 88:8m; (5) The heat released by the cooling solid condensation product from the melting temperature to the condenser temperature (tc) is denoted as QV. Cs is the specific heat of solid magnesium. QV ¼ Cs ðtm tc Þ$m ¼ 0:295ð651 600Þ$m ¼ 15:045m Therefore, the total heat released during the transfer of magnesium vapor from the reduction zone to the condensing zone is QM ¼ QI þ QII þ QIII þ QIV þ QV ¼ 1627:785m In general, when the technological conditions of the reduction remain unchanged while the reduction period m changes according to a certain rule, the heat released by the magnesium vapor during its entry into the condensing zone is constant. Thus, the available QIN heat input should be the sum of Q1e2 and QM; that is, QIN ¼ Q12 þ QM : When T1 ¼ 1350 C, T2 ¼ 600 C, A1 ¼ 0.0005 m2, A2 ¼ 0.001 m2, 3 1 ¼ 0.85 w/m2k4, 3 2 ¼ 0.95 w/m2k4 [21], and m ¼ 0.024 kg/h. Then, C1e2 ¼ 3.98 w/m2k4, Q1e2 ¼ 265.9 kJ/ h, QM ¼ 163 kJ/h, and QIN ¼ 428.9 kJ/h. Thus, the radiation heat in the traditional vacuum equipment accounted for 62% of the input heat of the condensation, whereas the heat radiated into the condensing zone by the magnesium vapor accounted for 38%. 1.3. Heat output in the condensing zone Two output processes occur in the condensing area when this area is known and specific:
QO ¼ C$Dt$M: where C is the specific heat of water, M is the water flow, and Dt is the temperature gradient between the inlet and outlet water. M is related to the diameters of the inlet and outlet water gates as well as the water pressure. Dt is affected by the inlet and outlet water temperatures, the heat-transfer medium of the cold plate, the heat-transfer area, and other factors. If the temperature in the condensing zone is kept within the technological index, the heat input is balanced by the heat output. As shown from the previous calculation, QIN is constant when the filling materials in the reduction zone are specific. Therefore, QO is vital in maintaining the temperature in the condensing zone. 1.4. Analysis of the temperature in the condensing zone The heat composition in the condensing zone was analyzed. By starting from QIN and QO, the control of the temperature in the condensing zone requires the adjustment of the corresponding heat input and output equipment. After the condensation system is designed and set up, the cooling water flow is restricted by the feed pressure as well as by the water inlet and outlet diameters. The range of M is narrow. Therefore, the experiment can only begin by changing the radiation heat input. From the calculations, the input heat from the magnesium vapor accounted for 38% of the total input heat. Therefore, if we want to change the equipment and prevent the unusual magnesium vapor crystallization due to the change in the input heat, the heat balance and temperature gradient in the condensing zone must be guaranteed. Under the experimental conditions, the heat that affected the potassium/sodium crystallization after magnesium vapor crystallization mainly originated from radiation heat transfer. Under the operation conditions, the dew-point temperature of the magnesium vapor exhibited a 200 C temperature difference from the crystallization temperature of the alkali metal vapors [22]. Thus, the separate condensation of the magnesium vapor and the potassium/sodium alkali metal vapors ensured that the condensation area had a sufficient temperature difference. Finally, the experimental schemes were developed based on the premise that the setting up of a condensing tower with a multistage temperature (rather than changing the input radiation heat in the condensing zone) to collect the magnesium vapor and alkali metal vapor separately in different condensation levels, also could reduce the oxidation rate of the magnesium metal and prevent magnesium loss due to burning
326
C.-b. Yang et al. / Journal of Magnesium and Alloys 1 (2013) 323e329
when the material is discharged during the separation of the magnesium crystals from the potassium/sodium particles. 2. Experiment
Table 2 Chemical compositions of calcined dolomite (mass fraction, wt%). Ingredient
MgO
CaO
Na
K
Content
41.98
56.80
0.084
0.13
2.1. Raw material The chemical compositions of the reducing agents and of the calcined dolomite are shown in Tables 1 and 2, respectively. 2.2. Experiment equipment The vacuum distillation system was set up by using an internally heated vacuum furnace (Fig. 2). The longitudinal section of the multistage condenser is shown in Fig. 3. 2.3. Experiment method Coking coal and calcined dolomite were used as raw materials. These materials were pulverized into 200 mesh particles, uniformly mixed, and then placed in a graphite crucible after compression molding. The crucible was then placed in a vacuum furnace, and the cooling water valve was then opened. The furnace was heated when the pressure decreased to below 30 Pa. The heating rate was kept constant at approximately 15 K/min. The furnace was initially heated to 1073 K (which is the optimal coking temperature of coal) and was then insulated for 20 min. The temperature was then increased to 1623 K [16]. The furnace was insulated to allow the vacuum carbothermic reduction reaction. The materials and reduction conditions were not changed throughout several trials. Therefore, the products provide different benefits depending on terms of the change in the condensation equipment when the condensing zone open to the air. To complete the experiments, scanning electron microscopy (SEM) and energy-dispersive spectroscopy (EDS) were performed on the prepared magnesium metal. Meanwhile, the chemical composition of the slag was analyzed. The collect efficiency of the magnesium metal can be obtained using the formula below, the obtained value was then used as the technological index of magnesium vapor condensation, and was show in Fig. 7: h¼
Fig. 2. Schematic diagram of vertical vacuum distillation furnace. 1. Heat system. 2. Condensation system. 3. Vacuum pump.
where h is the collect efficiency of the magnesium metal, M is the mass of the obtained magnesium metal in the condensing zone, m is the mass of the magnesium metal reduced from the materials, n is the magnesium oxide content of the materials, and 3 % is the reduction rate of magnesium oxide. 3. Results and analysis At an average pressure of 60 Pa, the condensation temperature of the magnesium metal vapor in the condensing zone completely depends on the radiant heat loss in the reduction area as well as on the heat transfer of the cold plates. When the inlet water flow and temperature at the condensing mouth are fixed, the effect of the radiant heat loss in the reduction area on the magnesium vapor condensation is significant. Given that excessive powder crystallization is caused by an excessive temperature gradient in the condensing zone (Fig. 4), the equipment in this experiment were properly adjusted, and the temperature gradient of the magnesium vapor was reduced from a high temperature to a low temperature. At the same
M 100% m
m ¼ n 3%
3 5
Table 1 Chemical compositions of the reducing agents (mass fraction, wt%). Sample
Coking Coal
Fixed Carbon
63.58
Volatile
16.34
Water
0.78
Ash (19.21) SiO2
Al2O3
CaO
60.75
26.3
2.47
Fig. 3. Longitudinal section of the multistage condenser.
C.-b. Yang et al. / Journal of Magnesium and Alloys 1 (2013) 323e329
Fig. 4. Condensation product of a traditional condenser.
Fig. 5. Condensation products at the fourth condensation level.
time, the atomic clusters of the nucleated magnesium vapor were given sufficient time to grow after the magnesium vapor went through a mitigatory temperature interval. These largegrain magnesium metals have higher chemical potentials, and can thus continuously consume fine grains to complete their self-growth. On the other hand, compared with the crystallized magnesium with low-porosity holes, the magnesium crystallization products with bigger grain sizes reduced the contact area between magnesium and CO, which was beneficial in reducing the oxidation rate of the magnesium metal. Figs. 5 and 6 show the physical map of the multistage condensation products that were collected at the fourth and third condenser levels, respectively. The rich alkali metals in the condensation products crystallized at the fourth level because of the constant mitigating temperature gradient. The
327
Fig. 6. Condensation products at the third condensation level.
condensation temperature at the third level was relatively closer to the gaseliquid transition temperature. In addition, the subsequent vapor nucleation became easier when the phase transition decreased the nucleation energy. Thus, the magnesium vapor tend to crystallize at the third level. The following are the condensation results for all condensation equipment at the same reduction temperature. Table 3 shows the average of each result from several experiments. The condensation equipment uses the traditional vacuum carbothermic method and thus does not have a separate potassium/sodium collection device. Therefore, potassium/sodium condenses with the magnesium vapor on the cold plates and accumulates at the region with a relatively lower temperature compared with the magnesium vapor crystallization temperature in the condensing zone. A comparison of the experimental data (Table 3) shows that the burning or even explosion of the condensation products occurred during the vacuum carbothermic reduction, and got the average of data after two groups of contrast experiment for 10 times, respectively. In addition, prior to the condensation equipment upgrade, the calcined dolomite mainly resulted from the burning of magnesium powder, which was due to the self-ignition of the alkali metals. The temperature gradient in the condensing zone was excessively large. At 2 h insulation time, the uneven heat transfer caused a large temperature gradient in the cold plates, which increased the difficulty in separating the magnesium crystals from those of potassium/sodium. Thus, when the condensing products were collected, some unoxidized alkali metals burned, which led to a further loss of crystallized magnesium powder due to burning. Fig. 7 shows that because of the combustion or even explosion of the condensation products, the magnesium metal was oxidized into magnesium oxide. Thus, at 2 and 3 h
Table 3 Comparison of the condensation products of two kinds of condensation equipment after 2 h insulation time under the same reduction conditions. Equipment Traditional condenser Multistage condenser
Third Level Fourth Level
Mg content (MMg %)
Na/K content (MA%)
Average particle size
Burning probability
33% 92% 3%
0.58% 0.033% 0.47%
<50 mm >500 mm <50 mm
100% 0% 10%
328
C.-b. Yang et al. / Journal of Magnesium and Alloys 1 (2013) 323e329
Fig. 7. Collect efficiency of the condensation products of two kinds of condensation equipment under different holding times.
Fig. 8. Traditional condensation products.
Fig. 9. Fourth-level products in the multistage condenser.
insulation time, the recovery efficiency was zero. However, as the insulation time and heat diffusion increased, the temperature in the condensing zone gradually stabilized. As a result, the alkali metal stopped condensing around the cold plates at relatively high temperatures. In addition, the direct-recovery rate slightly increased. Although the morphology of the condensation products improved, the risks of burning and explosion were considerably reduced. But, the cold plate did not capture the alkali metal vapor at increased temperatures, which resulted in the non-accumulation of the alkali metal in the enrichment sites such as in the furnace-wall cracks and the internal vacuum pump. This phenomenon decreases the furnace life and increases the cost of equipment maintenance. Fig. 8 shows the SEM and EDS results at 5 h insulation time. Compared with the traditional equipment, the multistage condensation equipment has more advantages, and its directrecovery rate is stable and reaches 90% within 3 h insulation. This condenser separates magnesium vapor from the alkali metal and achieves a low oxidation rate for the obtained metal. In addition, the use of this condenser results in bigger grain sizes as well as significantly smaller amounts of powder particles because of the controllable condensing gradient. The rich alkali metals in the condensation products at the fourth level can be collected only after all alkali metals in the condensing cover are completely oxidized. The alternative replacement of the condensing cover at this level does not affect the continuity of production. The risk is significantly reduced when magnesium powder burning is excluded in the process. Figs. 9 and 10 show a comparison of the SEM and EDS results for the third and fourth levels in the multistage condensing tower after 5 h insulation. The third level is the major site of magnesium vapor crystallization. It contains both regular and large grain sizes and has a 98% metal content. On the other hand, no potassium/sodium or other alkali metals were found in the visible area. Most of the potassium/sodium accumulated at the fourth level of the condensing zone at lower temperatures, which resulted in low magnesium vapor loss. However, after calculations, the loss was lower than 3% of the total amount.
Fig. 10. Third-level products in the multistage condenser.
C.-b. Yang et al. / Journal of Magnesium and Alloys 1 (2013) 323e329
4. Conclusions (1) Based on theoretical calculations and analysis, the radiation heat in the traditional vacuum equipment accounted for 62% of the input heat of condensation, whereas the heat that radiated into the condensing zone by the magnesium vapor accounted for 38%. The larger temperature gradient in the condensing zone is the primary reason for the difficulty in separately condensing the alkali metal vapor and the magnesium vapor. The radiation heat transfer control plays an important role in the condensation of metal vapors. (2) The vacuum carbothermic reduction of dolomite calcination experiment that use traditional equipment exhibit a high burning loss rate as well as high risks of burning and explosion. The low direct-recovery rate subjects the production process to high risks. Meanwhile, in the comparison experiment, the multistage condensation equipment exhibits a stable direct-recovery rate, and the condensation products meet the collection requirements. This proposed solution not only increases the crystal grain size and lowers the oxidation rate, but also enables easy collection and safe production. These results can be used as a reference for the design of medium- and largescale experiments on magnesium metal production using the vacuum carbothermic reduction of calcined dolomite. References
[4] [5] [6] [7] [8] [9]
[10]
[11]
[12]
[13] [14] [15] [16] [17]
[18] [19] [20]
[1] Hori Fumio, Method for obtaining Mg and Ca through carbon reduction. US, 4147534[P], 1979. [2] Tassios Steven, Production Process, WO, 2020012042A1 [P], 2010. [3] L.I. Rong-ti, P.A.N. Wei, Masamichi Sano, Metal. Mater. Trans. 8 (2003) 433e437.
[21] [22]
329
F. Hansgirg, The Iron Age, 1943, pp. 56e63. G.F. Warren, A.M. Cameron, EU, 842017410[P], 1985. A.M. Cameron, D.L. Canham, V.G. Aurich, JOM 42 (1990) 46e48. G.A. Brooks, S. Trang, P.J. Witt, M.N.H. Khan, M. Nagle, JOM 58 (10) (2006) 51e55. A. Khan, C.H. Heath, U.M. Dieregsweiler, J. Chem. Phys. 119 (6) (2006) 3138e3147. L. Prentice, H.A. Kinetic, Model for the Carbothermal Production of Magnesium [C], CSIRO Process Science and Engineering, Clayton, VIC, 2009 (Internal Report). A. Koo, G.A. Brooks, M. Nagle, Nucleation and growth of Mg condensate during supersonic gas quenching [J], J. Cryst. Growth 310 (2008) 2659e2667. H. Khan, P. Witt, G.A. Brooks, T. Barton, M. Nagle, Design of supersonic nozzles for ultra-rapid quenching of metallic vapors [C], in: Proceedings of Material Processing Fundamentals: Process Modelling (TMS Annual Meeting), 2006, pp. 699e709. H.P. Leon, W.N. Michael, R.D. Barton, T. Steven, T.K. Benny, J.W. Peter, K.C. Keri, Carbothermal production of magnesium: CSIRO’S MAGSONICTM Process [C], in: TMS2012, 2012, pp. 31e34. Li Yi-fu, Experimental Study on Carbothermic Reduction of Magnesite [D], Kunming University of Science and Technology, 2008 (in Chinese). Luo Qi, Qu Tao, Liu Da-chun, Xu Bao-qiang, Yang Bin, Dai Yong-nian, J. Central South Univ. (Sci. Technol.) 10 (2012) (in Chinese). Li Zhi-hua, Xue Huai-sheng, Dai Yong-nian, Energy Metal. Indust. 23 (2004) 20e22 (in Chinese). Tian Yang, Qu Tao, Yang Bin, Dai Yong-nian, Xu Bao-qiang, Geng Seng, Metal. Mater. Trans. B 43 (3) (2012) 657e661. Yu Qing-chun, Yang Bin, Xu Bao-qiang, Liu Da-chun, Li Zhi-hua, Dai Yong-nian, Study on the thermodynamic and experimental carbothermic reduction magnesia [C], in: Proceedings of the 9th Vacuum Metallurgy and Surface Engineering Conference, 2009, pp. 428e433. S.D. Kirkpatrick, Chem. Met. Eng. (1944) 54. A.V. Kostanovskii, M.G. Zeodinov, M.E. Kostanovskaya, High Temp. 43 (2005) 793e795. Xu Ri-yao, Magnesium Production Process [M], Central South University Press, Chang Sha, 2003. Tan Tian-en, Chemical Engineering [M], Chemical Industry Press, Beijing, 1984. Wang Yu-zhi, Cheng Xu, The Technology of Vacuum [M], Beijing University of Aeronautics and Astronautics Press, 2007, p. 34.
ScienceDirect Journal of Magnesium and Alloys 1 (2013) 323e329 www.elsevier.com/journals/journal-of-magnesium-and-alloys/2213-9567
Full length article
Production of magnesium during carbothermal reduction of magnesium oxide by differential condensation of magnesium and alkali vapours* Cheng-bo Yang a,b,c, Yang Tian a,b,c,*, Tao Qu a,b,c, Bin Yang a,b,c, Bao-qiang Xu a,b,c, Yong-nian Dai a,b,c a
National Engineering Laboratory for Vacuum Metallurgy, Kunming University of Science and Technology, Kunming 650093, China b Key Laboratory for Nonferrous Vacuum Metallurgy of Yunnan Province, Kunming 650093, China c State Key Laboratory Breeding Base of Complex Nonferrous Metal Resources Clear Utilization in Yunnan Province, Kunming 650093, China Received 18 December 2013; accepted 4 January 2014
Abstract Most of researchers believed that the developments on the condensation of magnesium produced by carbothermic reduction just concentrated on two process routes: the “quench” route and the “solvent” route. But this paper will briefly analyzes the major challenges in magnesium vapor condensation during the vacuum carbothermic reduction of calcined dolomite, on equipment upgrade, heat transfers alter, to achieve condensation control and production collection. Solutions are then proposed using theoretical calculations and experiment results. Comparative analysis of the experiment results shows that the burning and even explosion of condensation products during the vacuum carbothermic reduction of calcined dolomite are mainly due to the burning of crystallized powder magnesium, which results from the self-ignition of alkali metals. Finally, this paper proposes a multistage condensation solution to improve traditional vacuum condensation equipment. And result show that the condensation equipment can effectively mitigate the burning and loss during condensation, also the morphology of the condensation products clearly improved, the grain size increased, and the oxidation rate decreased. The potassium/sodium vapor and the magnesium vapor were separately condensed. Copyright 2014, National Engineering Research Center for Magnesium Alloys of China, Chongqing University. Production and hosting by Elsevier B.V. Open access under CC BY-NC-ND license. Keywords: Carbothermic reduction; Calcined dolomite; Magnesium vapor; Condensation
* Foundation item: Supported by National Natural Science Foundation of China (No. 51304095); Science and Technology Planning Project of Yunnan Province, China (No. S2013FZ029); the personnel training Funds of Kunming University of Science and Technology, China (No. 14118665). * Corresponding author. National Engineering Laboratory for Vacuum Metallurgy, Kunming University of Science and Technology, Kunming 650093, China. Tel.: þ86 871 65161583. E-mail addresses: [email protected] (C.-b. Yang), emontian@hotmail. com (Y. Tian). Peer review under responsibility of National Engineering Research Center for Magnesium Alloys of China, Chongqing University
Production and hosting by Elsevier
Carbothermic reduction is an alternative to both silicothermic and electrolytic processes for the production of magnesium. It has the advantages of low reductant cost, high equipment utilization rate, and environment-friendliness compared to the first two techniques. The current domination of world production by the Pidgeon process mainly reflects the economic conditions of China rather than any inherent advantages of the Pidgeon process over electrolytic routes. Under environment pressure, this process has always been favored by a great deal of researchers [1e3]. On condensation of magnesium vapor during Carbothermic reduction aspect, Hansgirg [4] tackled the problem of reversion and consolidation of fine magnesium using a wide variety of condensation techniques including quenching the magnesium vapor with water-cooled pipes and mechanically scraping the condensate from the tubes, shock cooling through
2213-9567 Copyright 2014, National Engineering Research Center for Magnesium Alloys of China, Chongqing University. Production and hosting by Elsevier B.V. Open access under CC BY-NC-ND license. http://dx.doi.org/10.1016/j.jma.2014.01.002.
324
C.-b. Yang et al. / Journal of Magnesium and Alloys 1 (2013) 323e329
injection of a cold gas, chilling with hydrocarbon oils and subsequently distilling magnesium from the slurry formed. All of these approaches failed to overcome the key problems of flammable powders and effective consolidation of fine powders, though his papers and numerous patents provide a thorough record of process development in this area and are essential reading for any researcher contemplating carbothermic production of magnesium. Cameron [5,6] and co-workers further investigated this route at the pilot plant level in the1980s, concentrating on the reduction stage of the process and providing useful thermodynamic analysis of impurity distributions between slag and vapor phases. However, the experiment was announced as a failure because the magnesium powders blocked the pipelines. Researchers of the CSIRO recently made a breakthrough in a semi-industrial experiment [7]. They used a Laval nozzle to create an environment where a supersonic gas is cooled rapidly at more than 106 C S1 [8e10], facilitating homogeneous nucleation and condensation of the magnesium vapour [11,12]. Meanwhile, our research group National Engineering Laboratory for Vacuum Metallurgy achieved significant progress in the carbothermic reduction of magnesia, which paved the way for the industrial applications of magnesium metal production using the vacuum carbothermic reduction method. Li [13] and Luo [14] conducted studies on the vacuum carbothermic reduction of magnesite and lateritic nickel ores, respectively. Li [15], Tian [16] and Yu [17] also made significant progress in reaction chamber spraying, catalytic mechanisms, and improvements in the reduction rate. At present, our experimental studies focus on the vacuum carbothermic reduction of calcined dolomite. However, the condensation of magnesium vapor fails to meet the collection requirements because of the small grain size of the obtained magnesium metal. In addition, the high oxidation rate easily leads to burning or even explosion, which significantly increases the damage rate due to burning of magnesium and reduces production safety. Therefore, the effective control of magnesium vapor condensation has become the key factor to the successful application of this method to industrial production. This paper is to probe into the reason of the problem, and try to resolve. 1. Theoretical analysis 1.1. Heat input to the condensing zone Fig. 1 shows that the heat in the reaction zone (high-temperature A) is radiated into the condensing zone (low-temperature B) in three ways: (1) The reaction zone radiates heat to the condensing zone through coupling sleeves (graphite condensing pipe C). Heat radiation is the major heat-transfer process at high temperatures. In addition, heat can be transferred through a vacuum without requiring any material media; (2) Superheated magnesium vapor radiates heat when it enters the condensing zone and is condensed into solid magnesium; and
Fig. 1. Schematic model of a vacuum furnace.
(3) Heat is radiated to the condensing zone through the crucible surface in the reduction zone. Given that the crucible in the entire reduction zone is sealed and wrapped by an insulation quilt, only coupling sleeve C is connected to the condensing zone. Therefore, compared with the two previously described input processes, this process is not considered here. Assume that high-temperature zone A is the surrounded object and condensing zone B is the peripheral object. Then [18], " 4 4 # T1 T2 Q12 ¼ C12 A1 : 100 100
Therefore; C12 ¼ h
1
31
þ AA12
C0
1
32
1
i
where C1e2 is the emission coefficient, C0 is the blackbody emission coefficient, 3 1 is the blackness of the surrounded object, 3 2 is the blackness of the peripheral object, A1 is the radiation area of the surrounded object, A2 is the radiation area of the peripheral object, T1 is the temperature in the reaction zone, and T2 is the temperature in the condensing zone. When the condenser size is specified and the temperature exceeds 1000 C, then A1, A2, 3 1, and 3 2 can all be considered as fixed values [19], and T1 is the only variable. By satisfying the technological conditions as well as the small variation range of the temperature, the reduction temperature T1 contributes to the relative stability of the radiation heat in the condensing zone. 1.2. Heat released by the condensation of magnesium vapor into solid magnesium The condensation of magnesium vapor into solid magnesium involves five steps [20]. The heat that enters the condensing zone is then calculated based on the five-step process. The five steps are as follows: (1) The heat released by the cooling of the superheated vapor to the boiling temperature is denoted as QI: QI ¼ Cvapor ðts tb Þ$m ¼ 0:204ð1350 1090Þ$m ¼ 53:04m
C.-b. Yang et al. / Journal of Magnesium and Alloys 1 (2013) 323e329
325
where Cvapor is the vapor emission coefficient, ts is the superheated vapor temperature, tb is the vapor boiling temperature, and m is the magnesium vapor in the condensing zone per unit time;
(1) Transfer between the residual gas in the furnace and the furnace wall; (2) Heat release from the cooling water in the condensing area.
(2) The evaporation latent heat released by the magnesium vapor at the normal boiling temperature is denoted as QII:
The water coefficient of the thermal conductivity is tens of times higher than the air coefficient. Thus, the heat lost due to cooling water is higher than those due to other factors. Therefore, the heat output QO is
QII ¼ 1337m; (3) The heat released by the cooling liquid magnesium from the normal boiling temperature (tb) to the melting temperature (tm) is denoted as QIII. Cl is the specific heat of liquid magnesium. QIII ¼ Cl ðtb tm Þ$m ¼ 0:305ð1090 651Þ$m ¼ 133:9m; (4) The melting latent heat released by the condensation of liquid magnesium is denoted as QIV: QIV ¼ 88:8m; (5) The heat released by the cooling solid condensation product from the melting temperature to the condenser temperature (tc) is denoted as QV. Cs is the specific heat of solid magnesium. QV ¼ Cs ðtm tc Þ$m ¼ 0:295ð651 600Þ$m ¼ 15:045m Therefore, the total heat released during the transfer of magnesium vapor from the reduction zone to the condensing zone is QM ¼ QI þ QII þ QIII þ QIV þ QV ¼ 1627:785m In general, when the technological conditions of the reduction remain unchanged while the reduction period m changes according to a certain rule, the heat released by the magnesium vapor during its entry into the condensing zone is constant. Thus, the available QIN heat input should be the sum of Q1e2 and QM; that is, QIN ¼ Q12 þ QM : When T1 ¼ 1350 C, T2 ¼ 600 C, A1 ¼ 0.0005 m2, A2 ¼ 0.001 m2, 3 1 ¼ 0.85 w/m2k4, 3 2 ¼ 0.95 w/m2k4 [21], and m ¼ 0.024 kg/h. Then, C1e2 ¼ 3.98 w/m2k4, Q1e2 ¼ 265.9 kJ/ h, QM ¼ 163 kJ/h, and QIN ¼ 428.9 kJ/h. Thus, the radiation heat in the traditional vacuum equipment accounted for 62% of the input heat of the condensation, whereas the heat radiated into the condensing zone by the magnesium vapor accounted for 38%. 1.3. Heat output in the condensing zone Two output processes occur in the condensing area when this area is known and specific:
QO ¼ C$Dt$M: where C is the specific heat of water, M is the water flow, and Dt is the temperature gradient between the inlet and outlet water. M is related to the diameters of the inlet and outlet water gates as well as the water pressure. Dt is affected by the inlet and outlet water temperatures, the heat-transfer medium of the cold plate, the heat-transfer area, and other factors. If the temperature in the condensing zone is kept within the technological index, the heat input is balanced by the heat output. As shown from the previous calculation, QIN is constant when the filling materials in the reduction zone are specific. Therefore, QO is vital in maintaining the temperature in the condensing zone. 1.4. Analysis of the temperature in the condensing zone The heat composition in the condensing zone was analyzed. By starting from QIN and QO, the control of the temperature in the condensing zone requires the adjustment of the corresponding heat input and output equipment. After the condensation system is designed and set up, the cooling water flow is restricted by the feed pressure as well as by the water inlet and outlet diameters. The range of M is narrow. Therefore, the experiment can only begin by changing the radiation heat input. From the calculations, the input heat from the magnesium vapor accounted for 38% of the total input heat. Therefore, if we want to change the equipment and prevent the unusual magnesium vapor crystallization due to the change in the input heat, the heat balance and temperature gradient in the condensing zone must be guaranteed. Under the experimental conditions, the heat that affected the potassium/sodium crystallization after magnesium vapor crystallization mainly originated from radiation heat transfer. Under the operation conditions, the dew-point temperature of the magnesium vapor exhibited a 200 C temperature difference from the crystallization temperature of the alkali metal vapors [22]. Thus, the separate condensation of the magnesium vapor and the potassium/sodium alkali metal vapors ensured that the condensation area had a sufficient temperature difference. Finally, the experimental schemes were developed based on the premise that the setting up of a condensing tower with a multistage temperature (rather than changing the input radiation heat in the condensing zone) to collect the magnesium vapor and alkali metal vapor separately in different condensation levels, also could reduce the oxidation rate of the magnesium metal and prevent magnesium loss due to burning
326
C.-b. Yang et al. / Journal of Magnesium and Alloys 1 (2013) 323e329
when the material is discharged during the separation of the magnesium crystals from the potassium/sodium particles. 2. Experiment
Table 2 Chemical compositions of calcined dolomite (mass fraction, wt%). Ingredient
MgO
CaO
Na
K
Content
41.98
56.80
0.084
0.13
2.1. Raw material The chemical compositions of the reducing agents and of the calcined dolomite are shown in Tables 1 and 2, respectively. 2.2. Experiment equipment The vacuum distillation system was set up by using an internally heated vacuum furnace (Fig. 2). The longitudinal section of the multistage condenser is shown in Fig. 3. 2.3. Experiment method Coking coal and calcined dolomite were used as raw materials. These materials were pulverized into 200 mesh particles, uniformly mixed, and then placed in a graphite crucible after compression molding. The crucible was then placed in a vacuum furnace, and the cooling water valve was then opened. The furnace was heated when the pressure decreased to below 30 Pa. The heating rate was kept constant at approximately 15 K/min. The furnace was initially heated to 1073 K (which is the optimal coking temperature of coal) and was then insulated for 20 min. The temperature was then increased to 1623 K [16]. The furnace was insulated to allow the vacuum carbothermic reduction reaction. The materials and reduction conditions were not changed throughout several trials. Therefore, the products provide different benefits depending on terms of the change in the condensation equipment when the condensing zone open to the air. To complete the experiments, scanning electron microscopy (SEM) and energy-dispersive spectroscopy (EDS) were performed on the prepared magnesium metal. Meanwhile, the chemical composition of the slag was analyzed. The collect efficiency of the magnesium metal can be obtained using the formula below, the obtained value was then used as the technological index of magnesium vapor condensation, and was show in Fig. 7: h¼
Fig. 2. Schematic diagram of vertical vacuum distillation furnace. 1. Heat system. 2. Condensation system. 3. Vacuum pump.
where h is the collect efficiency of the magnesium metal, M is the mass of the obtained magnesium metal in the condensing zone, m is the mass of the magnesium metal reduced from the materials, n is the magnesium oxide content of the materials, and 3 % is the reduction rate of magnesium oxide. 3. Results and analysis At an average pressure of 60 Pa, the condensation temperature of the magnesium metal vapor in the condensing zone completely depends on the radiant heat loss in the reduction area as well as on the heat transfer of the cold plates. When the inlet water flow and temperature at the condensing mouth are fixed, the effect of the radiant heat loss in the reduction area on the magnesium vapor condensation is significant. Given that excessive powder crystallization is caused by an excessive temperature gradient in the condensing zone (Fig. 4), the equipment in this experiment were properly adjusted, and the temperature gradient of the magnesium vapor was reduced from a high temperature to a low temperature. At the same
M 100% m
m ¼ n 3%
3 5
Table 1 Chemical compositions of the reducing agents (mass fraction, wt%). Sample
Coking Coal
Fixed Carbon
63.58
Volatile
16.34
Water
0.78
Ash (19.21) SiO2
Al2O3
CaO
60.75
26.3
2.47
Fig. 3. Longitudinal section of the multistage condenser.
C.-b. Yang et al. / Journal of Magnesium and Alloys 1 (2013) 323e329
Fig. 4. Condensation product of a traditional condenser.
Fig. 5. Condensation products at the fourth condensation level.
time, the atomic clusters of the nucleated magnesium vapor were given sufficient time to grow after the magnesium vapor went through a mitigatory temperature interval. These largegrain magnesium metals have higher chemical potentials, and can thus continuously consume fine grains to complete their self-growth. On the other hand, compared with the crystallized magnesium with low-porosity holes, the magnesium crystallization products with bigger grain sizes reduced the contact area between magnesium and CO, which was beneficial in reducing the oxidation rate of the magnesium metal. Figs. 5 and 6 show the physical map of the multistage condensation products that were collected at the fourth and third condenser levels, respectively. The rich alkali metals in the condensation products crystallized at the fourth level because of the constant mitigating temperature gradient. The
327
Fig. 6. Condensation products at the third condensation level.
condensation temperature at the third level was relatively closer to the gaseliquid transition temperature. In addition, the subsequent vapor nucleation became easier when the phase transition decreased the nucleation energy. Thus, the magnesium vapor tend to crystallize at the third level. The following are the condensation results for all condensation equipment at the same reduction temperature. Table 3 shows the average of each result from several experiments. The condensation equipment uses the traditional vacuum carbothermic method and thus does not have a separate potassium/sodium collection device. Therefore, potassium/sodium condenses with the magnesium vapor on the cold plates and accumulates at the region with a relatively lower temperature compared with the magnesium vapor crystallization temperature in the condensing zone. A comparison of the experimental data (Table 3) shows that the burning or even explosion of the condensation products occurred during the vacuum carbothermic reduction, and got the average of data after two groups of contrast experiment for 10 times, respectively. In addition, prior to the condensation equipment upgrade, the calcined dolomite mainly resulted from the burning of magnesium powder, which was due to the self-ignition of the alkali metals. The temperature gradient in the condensing zone was excessively large. At 2 h insulation time, the uneven heat transfer caused a large temperature gradient in the cold plates, which increased the difficulty in separating the magnesium crystals from those of potassium/sodium. Thus, when the condensing products were collected, some unoxidized alkali metals burned, which led to a further loss of crystallized magnesium powder due to burning. Fig. 7 shows that because of the combustion or even explosion of the condensation products, the magnesium metal was oxidized into magnesium oxide. Thus, at 2 and 3 h
Table 3 Comparison of the condensation products of two kinds of condensation equipment after 2 h insulation time under the same reduction conditions. Equipment Traditional condenser Multistage condenser
Third Level Fourth Level
Mg content (MMg %)
Na/K content (MA%)
Average particle size
Burning probability
33% 92% 3%
0.58% 0.033% 0.47%
<50 mm >500 mm <50 mm
100% 0% 10%
328
C.-b. Yang et al. / Journal of Magnesium and Alloys 1 (2013) 323e329
Fig. 7. Collect efficiency of the condensation products of two kinds of condensation equipment under different holding times.
Fig. 8. Traditional condensation products.
Fig. 9. Fourth-level products in the multistage condenser.
insulation time, the recovery efficiency was zero. However, as the insulation time and heat diffusion increased, the temperature in the condensing zone gradually stabilized. As a result, the alkali metal stopped condensing around the cold plates at relatively high temperatures. In addition, the direct-recovery rate slightly increased. Although the morphology of the condensation products improved, the risks of burning and explosion were considerably reduced. But, the cold plate did not capture the alkali metal vapor at increased temperatures, which resulted in the non-accumulation of the alkali metal in the enrichment sites such as in the furnace-wall cracks and the internal vacuum pump. This phenomenon decreases the furnace life and increases the cost of equipment maintenance. Fig. 8 shows the SEM and EDS results at 5 h insulation time. Compared with the traditional equipment, the multistage condensation equipment has more advantages, and its directrecovery rate is stable and reaches 90% within 3 h insulation. This condenser separates magnesium vapor from the alkali metal and achieves a low oxidation rate for the obtained metal. In addition, the use of this condenser results in bigger grain sizes as well as significantly smaller amounts of powder particles because of the controllable condensing gradient. The rich alkali metals in the condensation products at the fourth level can be collected only after all alkali metals in the condensing cover are completely oxidized. The alternative replacement of the condensing cover at this level does not affect the continuity of production. The risk is significantly reduced when magnesium powder burning is excluded in the process. Figs. 9 and 10 show a comparison of the SEM and EDS results for the third and fourth levels in the multistage condensing tower after 5 h insulation. The third level is the major site of magnesium vapor crystallization. It contains both regular and large grain sizes and has a 98% metal content. On the other hand, no potassium/sodium or other alkali metals were found in the visible area. Most of the potassium/sodium accumulated at the fourth level of the condensing zone at lower temperatures, which resulted in low magnesium vapor loss. However, after calculations, the loss was lower than 3% of the total amount.
Fig. 10. Third-level products in the multistage condenser.
C.-b. Yang et al. / Journal of Magnesium and Alloys 1 (2013) 323e329
4. Conclusions (1) Based on theoretical calculations and analysis, the radiation heat in the traditional vacuum equipment accounted for 62% of the input heat of condensation, whereas the heat that radiated into the condensing zone by the magnesium vapor accounted for 38%. The larger temperature gradient in the condensing zone is the primary reason for the difficulty in separately condensing the alkali metal vapor and the magnesium vapor. The radiation heat transfer control plays an important role in the condensation of metal vapors. (2) The vacuum carbothermic reduction of dolomite calcination experiment that use traditional equipment exhibit a high burning loss rate as well as high risks of burning and explosion. The low direct-recovery rate subjects the production process to high risks. Meanwhile, in the comparison experiment, the multistage condensation equipment exhibits a stable direct-recovery rate, and the condensation products meet the collection requirements. This proposed solution not only increases the crystal grain size and lowers the oxidation rate, but also enables easy collection and safe production. These results can be used as a reference for the design of medium- and largescale experiments on magnesium metal production using the vacuum carbothermic reduction of calcined dolomite. References
[4] [5] [6] [7] [8] [9]
[10]
[11]
[12]
[13] [14] [15] [16] [17]
[18] [19] [20]
[1] Hori Fumio, Method for obtaining Mg and Ca through carbon reduction. US, 4147534[P], 1979. [2] Tassios Steven, Production Process, WO, 2020012042A1 [P], 2010. [3] L.I. Rong-ti, P.A.N. Wei, Masamichi Sano, Metal. Mater. Trans. 8 (2003) 433e437.
[21] [22]
329
F. Hansgirg, The Iron Age, 1943, pp. 56e63. G.F. Warren, A.M. Cameron, EU, 842017410[P], 1985. A.M. Cameron, D.L. Canham, V.G. Aurich, JOM 42 (1990) 46e48. G.A. Brooks, S. Trang, P.J. Witt, M.N.H. Khan, M. Nagle, JOM 58 (10) (2006) 51e55. A. Khan, C.H. Heath, U.M. Dieregsweiler, J. Chem. Phys. 119 (6) (2006) 3138e3147. L. Prentice, H.A. Kinetic, Model for the Carbothermal Production of Magnesium [C], CSIRO Process Science and Engineering, Clayton, VIC, 2009 (Internal Report). A. Koo, G.A. Brooks, M. Nagle, Nucleation and growth of Mg condensate during supersonic gas quenching [J], J. Cryst. Growth 310 (2008) 2659e2667. H. Khan, P. Witt, G.A. Brooks, T. Barton, M. Nagle, Design of supersonic nozzles for ultra-rapid quenching of metallic vapors [C], in: Proceedings of Material Processing Fundamentals: Process Modelling (TMS Annual Meeting), 2006, pp. 699e709. H.P. Leon, W.N. Michael, R.D. Barton, T. Steven, T.K. Benny, J.W. Peter, K.C. Keri, Carbothermal production of magnesium: CSIRO’S MAGSONICTM Process [C], in: TMS2012, 2012, pp. 31e34. Li Yi-fu, Experimental Study on Carbothermic Reduction of Magnesite [D], Kunming University of Science and Technology, 2008 (in Chinese). Luo Qi, Qu Tao, Liu Da-chun, Xu Bao-qiang, Yang Bin, Dai Yong-nian, J. Central South Univ. (Sci. Technol.) 10 (2012) (in Chinese). Li Zhi-hua, Xue Huai-sheng, Dai Yong-nian, Energy Metal. Indust. 23 (2004) 20e22 (in Chinese). Tian Yang, Qu Tao, Yang Bin, Dai Yong-nian, Xu Bao-qiang, Geng Seng, Metal. Mater. Trans. B 43 (3) (2012) 657e661. Yu Qing-chun, Yang Bin, Xu Bao-qiang, Liu Da-chun, Li Zhi-hua, Dai Yong-nian, Study on the thermodynamic and experimental carbothermic reduction magnesia [C], in: Proceedings of the 9th Vacuum Metallurgy and Surface Engineering Conference, 2009, pp. 428e433. S.D. Kirkpatrick, Chem. Met. Eng. (1944) 54. A.V. Kostanovskii, M.G. Zeodinov, M.E. Kostanovskaya, High Temp. 43 (2005) 793e795. Xu Ri-yao, Magnesium Production Process [M], Central South University Press, Chang Sha, 2003. Tan Tian-en, Chemical Engineering [M], Chemical Industry Press, Beijing, 1984. Wang Yu-zhi, Cheng Xu, The Technology of Vacuum [M], Beijing University of Aeronautics and Astronautics Press, 2007, p. 34.