A Kinetic Study of the Carbothermic Reduction of Zinc Oxide with Various Additives

Materials Transactions, Vol. 47, No. 9 (006) pp. 4 to 46 #006 The Japan Institute of Metals EXPRESS REGULAR ARTICLE A Kinetic Study of the Carbothermic Reduction of Zinc Oxide with Various Additives Byung-Su Kim*, Jae-Min Yoo, Jin-Tae Park and Jae-Chun Lee Minerals & Materials Processing Division, Korea Institute of Geoscience & Mineral Resources, Daejeon, Korea Most processes for recovering zinc from electric arc furnace (EAF) dust employ carbon as a reducing agent for zinc oxide in the dust. In the present work, the reduction reaction of zinc oxide with carbon in the presence of various additives was kinetically studied. The effects of temperature and the additives of Fe O, mill scale, and CaCO on the kinetics of the reduction reaction were measured in the temperature range of 7 7 K under nitrogen atmosphere. The mill scale is one of byproducts generated from the steel rolling process. It was found from the experimental results that all three additives enhance the reaction rate of zinc oxide with carbon, but the effect of CaCO addition is the highest. The increase in the reaction rate is because Fe O, mill scale, and CaCO in the reduction reaction promote the carbon gasification reaction. The spherical shrinking core model for a surface chemical reaction control was also found to be useful in describing the kinetics of the reaction, which had an activation energy of 4 kj/mol (5 kcal/mol) for ZnO system, 75 kj/mol (4 kcal/mol) for ZnO-Fe O system, 84 kj/mol (44 kcal/mol) for ZnO-mill scale system, and 6 kj/mol (9 kcal/mol) for ZnO-CaCO system. [doi:0/matertrans.47.4] (Received February 4, 006; Accepted August, 006; Published September 5, 006) Keywords: recycling, zinc oxide, electric arc furnace dust. Introduction In the viewpoints of environmental protection and resources conservation, the recovery of zinc metal from the electric arc furnace (EAF) dust has been well known to be worth. Therefore, many processes for recovering zinc metal from EAF dust have been developed, and some of them were already commercialized. 7) So far, the major processes applied for EAF dust treatment are known to be pyrometallurgical processes. All the processes applied involve the reduction and volatilization of zinc metal from the dust by the carbothermic reduction method, leaving an iron rich residue. Thus, many investigations had been conducted on the kinetics of the reduction reaction of zinc oxide with carbon. 8 0) The reaction of zinc oxide with carbon can be expressed as: ZnO(s) þ C(s) ¼ Zn(g) þ CO(g) ðþ This reaction is a combination of the following reactions: ZnO(s) þ CO(g) ¼ Zn(g) þ CO (g) ðþ CO (g) þ C(s) ¼ CO(g) ðþ However, very little fundamental information is available in the literature on the effect of additives such as mill scale and CaCO on the above reaction rate. Usually, the additives are used to promote the zinc recovery ratio from the dust by minimizing or avoiding the formation of accretions in the kiln furnace.,) Therefore, in the present work, the reduction reaction of zinc oxide by carbon with additives such as Fe O, mill scale, and CaCO under condition in which the effects of external mass transfer and interstitial diffusion were eliminated was studied using a weight-loss technique in nitrogen atmosphere. The mill scale is one of byproducts generated from the steel rolling process, which mainly contains over 70% for iron metal. This study aims at investigating the effect of additives for zinc recovery from EAF dust by conventional pyrometallurgical processes. *Corresponding author, E-mail: bskim@kigam.re.kr. Experimental Work. Materials Materials used in this study were ZnO, Fe O,CaCO, carbon, and mill scale. The ZnO, Fe O, and CaCO were 99.9 mass% pure, and the carbon 99.99 mass% pure. All of the reagents ranged 74 mm in size, except where the size of the mill scale was 00 mm. The mill scale that is one of byproducts generated from a steel rolling process of special steel company in Korea mainly consists of 59.0 mass% FeO and 9.0 mass% Fe O.. Procedure Experiments were carried out in a horizontal tube furnace with cooling system, described in detail elsewhere. ) A covered alumina crucible that has mm hole on the cover surface was also used to hold the powder sample, the reason being to eliminate the effect of concentration gradient of CO(g) and CO (g) in the bed height of powder sample by minimizing the reaction zone as small as possible. The sample tray is a cylindrical alumina crucible having.7 cm diameter and.7 cm depth, and the size of hole on the cover surface of the alumina crucible was fixed at mm diameter. In the experiment, a weighed amount of solid reactants was first mixed thoroughly and placed in the alumina crucible, and then the sample crucible was introduced in the less hot zone of the reactor. After 0 min, the sample crucible was pushed into the central hot zone of the reactor that was maintained at the desired temperature. The temperature of this hot zone was constantly maintained by a temperature controller connected to a R-type (Rh-Rh %/Pt) thermocouple located in the constant temperature zone on the inside of the reactor tube. The temperature in the hot zone was maintained within K under a steady flow of nitrogen (purchased by Air Products Company in Korea). The reactor system was made oxygen-free with nitrogen gas flow inside the reactor. Zero time was counted when the crucible containing the sample was pushed into the reactor that was maintained at the desired temperature. After a stipulated time

4 B.-S. Kim, J.-M. Yoo, J.-T. Park and J.-C. Lee Bed height (cm) 5 0.50.00.50 Temp. = 7 K Temp. (K) 7 7 7 Fig. Effect of bed height on the ZnO system. Fig. Effect of temperature on the ZnO system. period, the reaction was stopped by sliding the crucible from the central hot zone to the outer cool zone. The cooled residual solid was weighed and analyzed by X-ray diffractometer (Rigaku D-max00PC, Rigaku/MSC, Inc., TX, U.S.A) and induction coupled plasma method (JY-8 plus, Horiba Ltd, Kyoto, Japan). In addition, Fe analysis was performed by wet chemistry. The reduction ratio at a particular time was determined by dividing the mass change of zinc in the solid sample at the time by the initial zinc mass in the solid sample.. Results and Discussion It is generally expected that the rate of the reduction reaction of zinc oxide by carbon is dependent on the bed height of solid sample because the zinc oxide is reduced carbothermically either with solid carbon or with carbon monoxide gas according to the reactions () and (). 8,9) Thus, to evaluate the effect of bed height on the rate of the reduction reaction, sample beds of different height (5.50 cm) were tested at 7 K in nitrogen of. L/min. Figure shows the reduction ratio versus time relationships. However, as shown in Fig., the reaction rate was independent of the bed height tested in the experiment. The unexpected results may be due to eliminating the concentration gradient of CO(g) and CO (g) in the sample bed by using a covered sample holder as explained in the previous part. Thus, in all of the subsequent runs, a bed height below.5 cm was chosen to avoid the diffusion effects. The bed height was obtained by mixing.0 g (0:000 g) of ZnO and 0.475 g (0:000 g) of carbon for ZnO system,.0 g (0:000 g) of ZnO, 0.475 g(0:000 g) of carbon and g (0:000 g) of Fe O for ZnO-Fe O system,.0 g (0:000 g) of ZnO, 0.475 g (0:000 g) of carbon and g (0:000 g) of mill scale for ZnO-mill scale system, and.0 g (0:000 g) of ZnO, 0.475 g (0:000 g) of carbon and g (0:000 g) of CaCO for ZnO-CaCO system. Here, the mole ratio of carbon and ZnO was fixed at, which is the stoichiometry amount for ZnO reduction. On the other hand, through preliminary experiments for the effect of nitrogen flow rate in the range of. L/min, it can be seen that there is no effect of the gas flow rate on the Temp. = K Fig. ZnO(g) : Fe O (g) : 0 : 5 : 0.0 : 0.5 : 0 Effect of Fe O addition on the ZnO system. reaction rate. Thus, in all of the subsequent runs, a working gas flow rate of. L/min was chosen to ensure that the rate of the reduction reaction of zinc oxide by carbon was not dependent on the rate of zinc vapor removal from the reaction zone in the covered sample holder. The effect of reaction temperature on the reduction reaction of zinc oxide by carbon without additives is shown in Fig.. From the figure, it was observed that the reduction ratio of zinc oxide is very low at 7 K and the reduction rate increases with increasing temperature. The results will be used for comparison with them obtained from the reduction reaction of zinc oxide by carbon with Fe O, mill scale, or CaCO. Shown in Figs. and 4 are the effects of the addition of Fe O and mill scale. Here, the input amount was.0 g (0:000 g) of ZnO, 0.475 g (0:000 g) of carbon and 0 0 g (0:000 g) of Fe O for the ZnO- Fe O system and.0 g (0:000 g) of ZnO, 0.475 g (0:000 g) of carbon and 0 0 g (0:000 g) of mill scale for the ZnO-mill scale system. As shown in the figures, the addition of a proper amount of Fe O and mill scale enhanced the reduction rate of zinc oxide. However, the reduction rate decreased when the weight ratio of ZnO and Fe O is over :5 and that of ZnO

A Kinetic Study of the Carbothermic Reduction of Zinc Oxide with Various Additives 4 ZnO(g) : Mill-scale(g) : 0 : 5 : 0.0 : 0.5 : 0 Temp. = K ZnO(g) : CaCO (g) : 0 : 5 : 0.0 : 0.5 : 0 Temp. = K Fig. 4 Effect of mill scale addition on the ZnO system. Fig. 6 Effect of CaCO addition on the ZnO system. Intensity 000 500 000 500 0 0 0 0 40 50 60 70 and mill scale over :0.0. The increase in the reaction rate may be because Fe O and mill scale easily produce CO(g) and CO (g) by reacting with solid carbon, and thus reaction (), Boudouard reaction, is promoted on the carbon surface, resulting in a fast reduction rate. Otherwise, the decrease in the reaction rate may be because the contact possibility between CO(g) produced by Boudouard reaction and the added Fe O and mill scale increases with increasing the addition amount, the CO(g) being consumed to reduce the ZnO as well as the iron oxides. That is indirectly verified by the fact that after reacting, unreacted ZnO, FeO, and Fe O 4 in the cooled residual solid phase are detected. Figure 5 shows the X-ray pattern of the cooled residual solid obtained from the ZnO-Fe O for 60 min at 7 K. On the other hand, it should be expected in the ZnO-Fe O -C and ZnOmill scale systems that the iron oxides is reduced to iron by CO(g), and the reduced iron is, in turn, oxidized by ZnO according to the solid-solid reaction of which the reaction rate is relatively slow. ) Thus, the reduction behavior of iron oxide through the quantitative analysis of iron oxides was not investigated due to the complication in the present research. Figure 6 shows the effect of CaCO addition. Here, the input amount was.0 g (0:000 g) of ZnO, 0.475 g (0:000 g) of carbon and 0 0 g (0:000 g) of CaCO. The results indicate that the addition of CaCO accelerates θ FeO Fe O 4 ZnO Fig. 5 X-ray diffraction pattern of the reacted solid from ZnO-Fe O -C reaction system for 60 min at 7 K. (The weight ratio of ZnO and Fe O is :5.) Intensity 000 600 00 800 400 0 0 0 0 40 50 60 70 the reduction reaction of ZnO. The increase in the reaction rate may be the reason why CaCO is decomposed to CaO and CO (g), and thus Boudouard reaction is promoted on the carbon surface. This was verified by X-ray analysis. Otherwise, the increase was no significant difference when the weight ratio of ZnO and CaCO is over :0.. That might be the reason why carbon as a reducing agent is limited in the reaction system. Thus, it was considered that over the weight ratio of ZnO and CaCO of :0., the reduction reaction rate of ZnO is not affected by the addition amount of CaCO. Figure 7 shows the X-ray pattern of the final solid obtained from the ZnO-CaCO for 60 min at K. It is seen that the solid contains only unreacted ZnO and CaO. This result indicates that CaCO is easily decomposed to CaO and CO (g), and thus the produced CO (g) promotes Boudouard reaction with solid carbon. It was thus considered that the increase in the reaction rate would be due to forming a relatively high reduction atmosphere by Boudouard reaction in the reaction system. Figure 8 presents, for comparison, a plot of the reduction ratio curves with the various additives at K. The results θ - ZnO - CaO Fig. 7 X-ray diffraction pattern of the reacted solid from ZnO-CaCO -C reaction system for 60 min at 7 K. (The weight ratio of ZnO and CaCO is :0..)

44 B.-S. Kim, J.-M. Yoo, J.-T. Park and J.-C. Lee Reduction, X Control 5 % Fe O 0 % mill scale 0 % CaCO Temp. = K -(-X) (/) 0. 0. For ZnO 7 K K 7 K 7 K Fig. 8 Effect of various additives on the ZnO system. shown there indicate that the effect of the addition of CaCO on the reaction rate is the highest. Reduction of zinc oxide without any additives was around 60% in 60 min. After the same amount of time, the reduction of zinc oxide with CaCO was about 70%. This is the reason that as explained above, CO(g) produced by Boudouard reaction in the ZnO-CaCO - C reaction system is consumed to only reduce ZnO. On the other hand, the effects of temperature on the reduction reaction of zinc oxide by carbon with Fe O, mill scale, or CaCO were investigated. Here, the input amount was.0 g (0:000 g) of ZnO, 0.475 g (0:000 g) of carbon and 5 g (0:000 g) of Fe O for the ZnO-Fe O system,.0 g (0:000 g) of ZnO, 0.475 g (0:000 g) of carbon and 0.0 g (0:000 g) of mill scale for the ZnO-mill scale system, and.0 g (0:000 g) of ZnO, 0.475 g (0:000 g) of carbon and 0.0 g (0:000 g) of CaCO for the ZnO-CaCO system. The results are shown in Fig. 9. It is generally accepted that the reduction reaction of zinc oxide with carbon is one of the gas-solid reactions in which no solid product is formed. 8 0) The reduction reaction mechanism includes the following sequential steps: ) carbon is vaporized with carbon dioxide to form carbon monoxide as in reaction (), ) the carbon monoxide diffuses towards the surface of the zinc oxide, ) the zinc oxide is reduced by carbon monoxide to zinc as in reaction (), and 4) the zinc vapor diffuses into the bulk gas and the carbon dioxide takes part in the reaction as in reaction () again. The overall reaction rate is controlled by the slowest step. Thus, the particle size of zinc oxide and carbon as reactants is diminished as the reduction reaction proceeds. In general, for such a system, spherical shrinking core model (SCM) is very well known to be useful to analyze the reaction rate data. For reactions following the SCM, two resistances may influence the reaction rate: mass-transfer and surface chemical reaction. However, in the study, the mass-transfer could be neglected by using a covered sample holder as explained in the previous part and using a sufficiently high gas flow rate. Thus, the reaction rate data obtained in the study were measured in the absence of the mass-transfer effects. Based on these observations, the interpretation of the rate data was -(-X) (/) -(-X) (/) -(-X) (/) 0. 0. 0.5 0. 0. 0.5 0. 0. ZnO-Fe O 7 K K 7 K 7 K For ZnO-mill scale 7 K K 7 K 7 K For ZnO-CaCO 7 K K 7 K 7 K t/min Fig. 9 Plot of the reduction ratios obtained from the ZnO-C, ZnO-Fe O - C, ZnO-mill scale-c, and ZnO-CaCO systems according to eq. (4). (The dot lines including symbol are for experimental data, the solid lines being for the best-fit lines according to eq. (4).) carried out using a number of different rate expressions such as Jander equation and nucleation and growth equation, from which the SCM that is useful for a surface chemical reactioncontrolled process was proved to yield the best results. In the SCM rate equation for a surface chemical reaction-controlled process, the reduction ratio of zinc oxide is related to the reaction time by 4) ð XÞ = ¼ k r t ð4þ Here, X is the reduction ratio of zinc oxide, t is the reaction

A Kinetic Study of the Carbothermic Reduction of Zinc Oxide with Various Additives 45 time (min), and k r is a rate constant (min ) that is a function of temperature. It is apparent from eq. (4) that a plot of ð XÞ = versus t should be linear with k r as the slope. The applicability of the SCM rate expression for the effects of temperature on the reduction reaction of zinc oxide by carbon with Fe O, mill scale, or CaCO were verified by first plotting the rate data obtained from each reaction systems according to eq. (4). Figure 9 presents the results. Examination of these figures reveals that the rate data follow relatively well eq. (4), as shown in Fig. 9. The values of k r were thus determined from the slopes of the figures. The slopes were calculated by regression analysis. Figure 0 is Arrhenius plots of the rate constants. The slopes of the straight line placed through the experimental points yield an activation energy of 4 kj/mol (5 kcal/mol) for ZnO-C reaction system, 75 kj/mol (4 kcal/mol) for ZnO-Fe O -C reaction system, 84 kj/mol (44 kcal/mol) for ZnO-mill scale system, and 6 kj/mol (9 kcal/mol) for ZnO-CaCO system. These lines are represented by the following equations: For ZnO system k r ¼ :6 0 6 exp½ 6;900=TŠ (min ) ð5þ For ZnO-Fe O system k r ¼ :8 0 4 exp½ ;00=TŠ (min ) ð6þ For ZnO-mill scale system k r ¼ 7:74 0 4 exp½ ;00=TŠ (min ) For ZnO-CaCO system k r ¼ :5 0 4 exp½ 9;400=TŠ (min ) The activation energies obtained are relatively high, which are less than those obtained for the pure zinc oxide with CO(g) (5 kj/mol) 0) and for the pure zinc oxide with iron metal (0 kj/mol). ) The results indirectly indicate that the reduction reaction of zinc oxide with solid carbon without additives or with additives such as Fe O, mill scale, and CaCO is controlled by a surface chemical reaction. 8 0) However, the effects of reaction () and () on the reduction rate of zinc oxide with carbon were not distinguished in the present research. 4. Conclusions The reduction reaction of zinc oxide with carbon in the presence of various additives such as Fe O, mill scale, and CaCO was studied in the temperature range of 7 K 7 K under nitrogen atmosphere using a weight-loss technique. The addition of a proper amount of Fe O and mill scale enhanced the reduction rate of zinc oxide, but the reduction rate decreased when the weight ratio of ZnO and Fe O is over :5 and that of ZnO and mill scale over :0.0. The effect of the addition of CaCO on the reaction rate was the most effective, but the increase in the reaction rate was no significant difference when the weight ratio of ZnO and CaCO is over :0.. The increase in the reaction rate is because Fe O, mill scale, and CaCO in the reduction reaction of zinc oxide with carbon promote the carbon gasification reaction. The spherical shrinking core model ð7þ ð8þ /min - ) /min - ) /min - ) /min - ) -8-8 -8 For ZnO E = 4 kj/mol (5 kcal/mol) E = 75 kj/mol (4 kcal/mol) For ZnO-Fe O E = 84 kj/mol (44 kcal/mol) For ZnO-mill scale E = 6 kj/mol (9 kcal/mol) For ZnO-CaCO Fig. 0 T - /0-4 K - Arrhenius plot of the rate constants. for a surface chemical reaction control was found to be useful in describing the kinetics of the reaction over the entire temperature range. The reaction has an activation energy of 4 kj/mol (5 kcal/mol) for ZnO system, 75 kj/mol (4 kcal/mol) for ZnO-Fe O system, 84 kj/mol (44 kcal/mol) for ZnO-mill scale system, and 6 kj/mol (9 kcal/mol) for ZnO-CaCO system.

46 B.-S. Kim, J.-M. Yoo, J.-T. Park and J.-C. Lee REFERENCES ) S. Itoh and T. Azakami: J. Min. Metall. Inst. Jpn. 04 (988) 97 0. ) J. R. Donald and C. A. Pickles: Metal. and Mater. Trans. B 7B (996) 6 74. ) D. Mishra, N. Lamaro, I. Gaballah and B. Dupré: Fundamentals of Advanced Materials for Energy Conversion, ed. by D. Charidra and R. G. Bautista, (TMS, Warrendale, PA, 00) pp. 49 57. 4) D. K. Xia and C. A. Pickles: Minerals Engineering (999) 69 700. 5) S. S. Thomas and E. D. Clifton: Economic evaluation of the HTR process for treating EAF Dust, ed., by R. Kaltenhauser, Disposal, recycling and recovery of electric arc furnace dust, (ISS-AIME, USA, 987) pp. 4. 6) J. F. Pusateri, R. Chew and A. E. Stanze: On-site treatment of EAF dust via the St. Joe flame reactor, ed., by R. Kaltenhauser, Disposal, recycling and recovery of electric arc furnace dust, (ISS-AIME, USA, 987) pp. 97. 7) J. A. Aune, K. Berwitz-Larsen, I. J. Eikeland and T. D. Pederson: Elkem multi-purpose furnace technology for reclamation of metal values in the treatment of hazardous materials, Recycling of metalliferrous materials, (IMM, London, 990) pp.. 8) H. Chen: Scandi. J. Metal. 0 (00) 9 96. 9) C. E. Guger and F. S. Manning: Metal. and Mater. Trans. (97) 08 090. 0) G. González, Z. Szczygiel Jordens and S. Escobedo: Thermochimica Acta 78 (996) 9 4. ) N. Tsuneyama, M. Takewaki and M. Yasukawa: Lead-Zinc 90, ed., by T. S. Mackey and R. D. Prengaman, (TMS, Warrendale, PA, 990) pp. 4 4. ) H. Maczek and R. Kola: Journal of Metals (980) 5 58. ) J. M. Yoo, B. S. Kim, J. C. Lee, M. S. Kim and C. W. Nam: Mater. Trans. 46 (005) 8. 4) J. Szekely, J. W. Evans and H. Y. Sohn: Gas-solid reactions (New York, academic press Inc., 976) pp. 65 75.