Isothermal reduction of titanomagnetite concentrates containing coal

International Journal of Minerals, Metallurgy and MaterialsVolume 21, Number 2, February 2014, Page 131Do:10.1007/s12613-014-0875-zIsothermal reduction of titanomagnetite concentrates containing coalTu Hu.2, Xue-wei Lil, Chen-guang Bai", and Gui-bao Qiu"1)College of Materials Science and Engineering, Chongqing University, Chongqing 400044, China2)Faculty of Metallurgical and Energy Engineering, Kunming University of Science and Technology, Kunming 650093, China(Received: 9 September 2013: revised: 29 October 2013: accepted: 3 November 2013)Abstract: The isothermal reduction of the Panzhihua titanomagnetite concentrates(Prc) briquette containing coal under argon atmospherewas investigated by thermogravimetry in an electric resistance furnace within the temperature range of 1250-1350.C. The samples reduced iargon at 1350C for different time were examined by X-ray diffraction(XRD) analysis. Model-fitting and model-free methods were used toevaluate the apparent activation energy of the reduction reaction. It is found that the reduction rate is very fast at the early stage, and then, at alater stage, the reduction rate becomes slow and decreases gradually to the end of the reduction. It is also observed that the reduction of PTCby coal depends greatly on the temperature. At high temperatures, the reduction degree reaches high values faster and the final valueachieved is higher than at low temperatures. The final phase composition of the reduced PTC-coal briquette consists in iron arof about 68 kJ-mol'and by three-dimensional diffusion for reduction degree greater than 0.7 5 withe d wustite(FeO)are intermediaterous-pseudobrookite(FeTi2O ), while Fe275Ti02504, Fe2sTiosO4, Fe2.25Ti07504, ilmenite(FeTiO3)anproducts. The reaction rate is controlled by the phase boundary reaction for reduction degree less than 0. 2 with an apparent activation energy134 kJ mol. For the reduction degree in the range of 0. 2-0.75, the reaction rate is under mixed control, and the activation energy increaseswith the increase of the reduction degreeKeywords: titanomagnetite; carbothermal reduction; thermogravimetric analysis; kineticsThe direct reduction of Panzhihua titanomagnetite con1 Introductionentrates(PtC) by solid reductants has been widely studiedfor many decades [1-6]. It was found that the reduction ofVanadium-titanium-bearing magnetite of Panzhihua, titanomagnetite was slower than that of magnetite due to theChina, is a complex iron ore composed by the coexistence different crystal structures and the presence of titarof vanadium and titanium that accounts for more than 90% sulting in a higher thermodynamic stability of titanomagnetof the titanium reserves in China. Most of the titanomagnet- ite [7-13]. Here, the kinetics of the carbothermal reductionte concentrates are now used as the main materials for the of PtC was studied by using thermogravimetry. During theblast furnace process in Panzhihua area. Most of the iron heating process of the composite briquette of PTc and coaland, partly, vanadium can be reduced into the hot metal; the mass loss is caused by the release of volatiles from coalhowever, almost all of the titanium remains in the slag, carbon loss, and removal of oxygen by reduction. Aforming a high-titanium slag with contents of TiO2 varying pseudo-kinetic parameter, fraction of reaction () has beenfrom 22wt% to 25wt %. There is no an appropriate and used to measure the reaction rate. It has been observed thateconomic method to deal with such slag [l] so far. Recently, the estimation of the pseudo-kinetic parameter ( is notmost of the studies focused on developing an alternative identical to that of the degree of reduction and can only pro-route to use the titanomagnetite concentrates highlighting vide a rough estimate of the extent and characteristics of thethe rotary hearth furnace(RHF)process, which involves the reduction [141reduction of composite briquette of titanomagnetite concenIn the present work, the isothermal reduction of Prc bytrates with coal, and the smelting of the reduced sample in coal was investigated by thermogravimetry, and the reducan electric arc furnacetion degree was used to analyze the kinetic dataorrespondingauthorTuHuE-mail:hutu1219@@126.comO University of Science and Technology Beijing and Springer-Verlag Berlin Heidelberg 2014中国煤化工 springerCNMHGIntJ. Miner. Metall. Mater., Vol. 21, No. 2, Feb. 20142. Experimentalthe Ptc was mixed homogenously with the coal with a mo-2. 1. Specimen preparation and experimental procedurees lar ratio(Cfixed/O obonded with Fe)of 1.2. Themixture was madeinto spheroid briquettes at 15 MPa with a briquette makerThe chemical compositions and size distributions of the The diameter of the briquette was about 30 mm, and theraw materials were examined and are presented in Tables tal mass of the sample was about 25 g. The other type of1-3. Fig. I shows the X-ray diffraction(XRD) pattern of briquette made of coal and alumina particles, which has thePTC. The mineral compositions were mostly titanomagnet- same amount of coal and the same geometries as the briite, magnetite, and ilmenite. Two types of specimens were quette made of PTC and coal, was made under the sameprepared. One type contained PTC and coal, and the other conditions. All the briquettes were dried at 120 C for 6hcontained alumina particles and coal. For the former type, before the reduction experiTable 1. Chemical composition of Panzhihua titanomagnetite concentrates (tC) and coalwt%oTFe Feo T1O2 V2Os SiO Cao MgO AlO3Volatile526232.0012004.201.302.603.9081.956.9510410.69Table 2. Size distribution of Panzhihua titanomagnetite concentrates (ptc<7474-8080-9696-109109-12020-150Content/wt%3.264.201.810.320.15Table 3. Size distribution of coalDiameter /um<100150-180180-250550-1700Total20.8051.3414.89.871.1argon quickly after the sample was taken out from the fnace. Two types of briquettes were heated under the same150FeTi,O.experimental conditions at each temperature. Three identicalexperiments were performed for each time-temperature pair,and the averaged mass loss was usedThe phase transformations of ptc during the reduction at1350C by coal were investigated by XRD, which wasconducted using a Cu K source0AAAElectronic balance203040506070809020/(°)DATA ACQ. SYSig. 1. X-ray diffraction pattern of Panzhihua titanomagnetite concentrates (PrC)The isothermal reductions of the ptc with coal were car.ried out in a vertical electric resistance furnace. whoseschematic is shown in Fig. 2. a briquette was loaded in abasket made of nichrome wire and hanged up to the furnaceThe furnace was purified by blowing argon at a high ratebefore each experiment and, then the flow rate of argon waskept at I L min The basket with the briquette was loadeddown quickly into the hot zone of the furnace when the furnace temperature reached the desired value. The mass of theThermocouplebriquette was measured and recorded continuously until themass became constant. The reduced briquette was cooled inli H中国煤化工 teCNMHGT. Hu et aL., Isothermal reduction of titanomagnetite concentrates containing coal1332. 2. Determination of the degree of reductiondetermined by the kinetic model. The kinetic model maytake various forms. some of which are shown in Table 4The degree of reduction(R) is usually defined as the ratiothe accumulated mass of oxygen removed at time t to the The integral form of Eq (6)is generally used for evaluatingthe activity energy under isothermal conditions, which istotal mass of removable oxygen bonded with iron, which isgiven byg18, (a)=k, (T,)t(1)where g (a)=L L,(a)] 'da is the integral form of thewhere AMo(g) is the mass of oxygen removed at time tduring the isothermal heating process of the briquette, andkinetic model j(Table 4). The rate constant can be deterAMo(g)is the total mass of removable oxygen bonded withmined from the slope of the plot of g(a) versus t For aniron in the briquette. The mass loss of the PTC-coal briappropriate kinetic model, the rate constants are evaluated atquette during heating consists of the volatile emission, carmperatures TiT, and the activation energy is debon, and oxygen loss upon reduction. Therefore, Eq (1)cantermined using the Arrhenius equation in its logarithmicbe re-written asform, which is given as follows△M2-△MC-△Mlnk,(T)=A、k△Mwhere AM:(g)is the mass loss of the PTC-coal briquetteTable 4. Set of kinetic models commonly applied to describesolid state reactionsat time t, AMc(g) is the mass loss of carbon at time t, andAM(g) is the mass loss of the volatile emission at time tNo. Modelf(a)g(a)The value of AMy can be measured according to the massD1(x)loss of the Al2O3-coal briquette in argon at time t, and hereit is represented by AMAc(g)2D2(a)-ln(1-a)(1-a)ln(1-a)+aIt has been established that there is little COz during the 3 D3(a)5L1-(1-a)]d-a)3[1-(1-a)3]reduction of the PTC-coal briquette at temperature higherthan 1000oC, due to the drastic carbon gasification reaction 4 D(a)by CO2[9]. Therefore, in present study, the reduction of iron5F(a)I-aoxides by carbon at high temperature can be generally represented by6 R(a)2(1-a)1-(1-a)Fe o +C=Feo,+co7 R(a)(1-a)n(1-a)2(1-a)Therefore8A2(a)-ln(1-a)}29 A3(a) 3-In(l-a)]a-a) [-In(I-a)]3△MC=,△MTo obtain the most appropriate kinetic model, a can beReplacing△Mand△ M: with,△ Me and△MAplotted as a function of the dimensionless reduction timein Eq (2), respectively, and combining Eq (2)with Eq (1), term t/t, for each model and the experimental data,the degree of reduction can be expressed aswhere to is the time required to reach a specified degree of4△M-△MAcconversion(e.g, a=0. 8). The kinetic model by which the(5) experimental data is well fitted is regarded as the most ap-propriate one. This method, the so-called model-fitting2.3. Methods for the evaluation of the activation energymethod, is widely applied in solid-state kinetics [16-17Only a single average value of the activation energy forThe reaction rate can be generally expressed by Eq.(6) an overall process can be obtained by the model-fittingmethod. It is difficult to obtain the activation energy if the=k(7)f(a)(6) experimental data cannot be fittedtion models, which reveals the complexity of the reactionwhere a is the degree of conversion, t the time, T the tem- mechanism. Thiperature, k(n) the rate constant, and f(a) the pattern function isoconversional中国煤化工CNMHGl34Int. Miner. Metall. Mater., Vol. 21, No 2, Feb. 2014to be determined as a function of the degree of conversion cepted that the mechanism of carbothermic reduction of ironand without any kinetic model assumption [15]oxides is a two-stage mechanism involving the participationFor the model-free isoconversional method, Eqs. (7)and of gaseous intermediates (e.g, CO and CO2)according to(8) can be combined to eliminate k, (T,) to obtain the fol- the following equationslowing equationFe, O, +Co=Fe Oy-+CO2Inta,, =In/_AE(9)g(a)RTThe endothermic carbon gasification reaction plays anwhere R is the gas constant, Aa is the pre-exponential important role in the carbothermal reduction of iron oxidesfactor, and tai is the time required to reach a specified ex- During reduction. the high heat transfer rate induces the intent of conversion at temperature T Ea represents the acti- crease of the rate of gasification reaction and, thus, of thevation energy at a specified a and is evaluated from the reduction rate. a high sintering rate of the oxides at highslope of the plot of In ta. against Ttemperatures, enhancing the heat transfer inside the briIn this study, both the model-fitting and model-free quette, could also be responsiblemethods were used to evaluate the apparent activation energy of the isothermal reduction of ptc by coal, and theal350°Cdegree of conversion(a) was substituted for the degree ofs6°120reduction(R)0.63. Results and discussion1350°1250°3.1. Effect of the temperature on the carbothermal re-24681000duction of PrcThe mass loss of the PTC-coal briquette and Al2O3-coal210briquette during reduction at different temperatures areTime/minshown in fig 3. There is little difference in the mass loss ofFig 3. Mass losses of panzhihua titanomagnetite concentrates1O3-coal briquettes heated at different temperatures be- (PTC)-coal briquette and Al2Oj-coal briquette during reduccause the volatilization rate of the volatiles in coal is very tion at different temperatures.fast at the involved temperatures. According to the mass lossof briquettes, the degree of reduction of the PTC-coal briquette was calculated. The curves of reduction degree as a0.8function of time for the ptc- coal briquettes at different temperatures are shown in Fig 4. The reduction rate was very0.6fast at the early stage and, then, it became slow and de-creased gradually to the end of the reduction process. At theearly stage, the reduction of magnetite to iron mainly oc-curred, and the carbon gasification by CO2 was very fast,0.2resulting in a high reaction rate at this stage. While the reduction proceeded, the reduction of titanium bearing iron0.0oxides, whose reaction rate is slower than that of iron oxidesTime/minbecame the main reduction reaction, in which the carbon Fig 4. Degree of reduction vs. reduction time for Panzhihuacontent was decreased due to the high consumption at the titanomagnetite concentrates(PTC)-coal briquettes reduced atearly stage. Therefore, the reduction rate decreased at the different temperatures3. 2. Phase transitions of the isothermal reduction of ptcIt is also observed that the reduction of ptc by coalby coalgreatly depends on the temperature. For higher temperatures,the reduction degree reaches high values faster and the finalThe briquettes were heated at 1350C for different timevalue achieved is higher. This can be attributed to the higher periods, and theples were examined by XRDheat transfer rate at higher temperatures. It is widely ac- the patterns of中国煤化工 -he main phasesCNMHGT. Hu et aL., Isothermal reduction of titanomagnetite concentrates containing coall35of the sample reduced for 3 min are magnetite(Fe3O4),Panzhihua titanomagnetite concentrateswustite (FeO), metallic iron, ilmenite(FeTiO3), and ti-FeTiOFe-Ti, OaFe Oatanomagnetite(Fe2.75Ti0. 25O4). Traces of Fe2. Tio sO4 wereobserved after 7 min. Fe2. 25T107504 appeared in the samplereduced for 10 min. Wustite (FeO) and ilmenite(FeTiO3)disappeared in the samples heated for 13 and 19 min, re-spectively. Traces of ferrous-pseudobrookite(FeTi2O5)wereexsTiasOadetected in the XRd pattern of the sample heated for 19 minIt is noticed that ulvospinel(Fe]TiO4) was not detected inFe,?sT1o7sOahe reduced samples because the ulvospinel formed duringthe heating process is quickly further reduced to metalliciron and ilmeniteFeTi, O.Fe★Fe275Tia25O4△Fe2 sLosh4oFe225F▲FeO4·FeO◆ FeTio, x FeTi,O530 minFig. 6. Phase transitions of panzhihua titanomagnetite con-centrates(Ptc) during carbothermal reduction at 1350.C.1916 min3.3. Evaluation of the activation energy13 minTo ascertain the most appropriate reaction model for usthe model-fittiethod. the10evaluated against the standard reduced time plots of the reaction models listed in table 4. The results are shown in7 in which the dasheding from the average of three isothermal experiments per:人人formed at 1250. 1300 and 1350%C. It is found that the data2030405060708090by any of the readFig. 5. X-ray diffraction patterns of Panzhihua titanomagnet- which reveals the complexity of the carbothermal reductionite concentrates(PTC)-coal briquettes reduced at 1350C for of PTC and indicates that the traditional model-fittingdifferent time periods.method for the activation energy evaluation is not feasible inFe.75Ti0. O4 Fe,5TiosO4, and Fe225T1075O4 are three diffethis case. However. R values less than 0.2 fit well into therent forms of titanomagnetite(,- Ti, O4), which can be writ- three-dimensional phase boundary reaction model(r3)andten as 3 Fe3O Fe TiO4, Fe O4 Fe, TiO4, and Fe3 O4 3Fe TiO4, R values greater than 0.75 fit well into the three-dimensionalrespectively. The phase transformation from 3Fe3O4 Fe2T1O4 diffusion model D3). This indicates a change in the mechato Fe3 O4 3Fe2T1O4 indicates that the proportion of magnetite nism during the reduction of the PtC-coal briquette. It is(Fe O4)in the solid solution Fe -Ti, O4 of magnetite(Fe3 O4) suggested that the reaction rate is controlled by the phaseand ulvospinel(Fe2T1O4)decreases during the heating proc- boundary reaction at the first stage and by the three-dimeness, according to the following reactions:sional diffusion at the final stage. At the second stage, the3Fe, O4. Fe,T1O4+2CO(C)reaction rate is under the mixed control of phase boundaryFe,O, Fe, T1O, +6FeO+2C0,(CO)(12) reaction and three-dimensional diffusion3(Fe,O Fe, TIO)+2CO(C)The apparent activation energy of the carbothermal reCO, (CO)13) duction of iron oxides is generally a composite value deter-mined by the activation energies of various reactions and byFe O4. 3Fe,TIO4+CO(C)their influence on the overall reaction rate. Even if the tem3Fe, TIO4 +3Feo+CO,(CO)(14) perature is kept constant, the relative contributions of theBased on the XRD analysis, as shown in Fig. 6, the phase elementary steps into the overall reaction rate vary with thetransitions of PTc during the isothermal carbothermal re- reduction degreeluction under the present conditions can be suggestedeffective activatiH中国煤化工 pendence of theCNMHG136IntJ. Miner. Metall. Mater., Vol. 21, No. 2, Feb. 20140.90.8 Experimental valueconcentrates with coal were investigated by thermogravimR=0.7etry in argon atmosphere at 1250, 1300, and 1350.C.Thefollowing conclusions are obtained(1)The reduction rate is very fast at the early stage. Thenat the later stage the reduction rate becomes slow and de-creases gradually to the end of the reduction. The reductionof Ptc by coal depends greatly on the temperature. Foree rev0.1ues faster, and the final value achieved is high0.0(2)The PtC is reduced to iron at 1350C in argon at-0.00.2040.60.81.01.214mosphere along a stepwise sequence with Fe2. 75T102504ig. 7. Reduced time plots for the reaction models(solid Fe2sTios O4, Fe2 25 Tio 7504, ilmenite(FeTiO3), wustite (Feo)curves 1-9 are obtained using kinetic models as enumerated in and ferrous-pseudobrookite(FeTi2O5)Table 4) and isothermal experimental data for ptc car-()The reduction rate is controlled by a phase boundarybothermal reduction. The dashedcurve corressponds to the reaction for R less than 0.2 with an apparent activation endata averaged from three isothermal experiments performed at ergy of about 68 kJ-mol and by three-dimensional diffu1250,1300,and1350C.sion for R greater than 0.75 with an apparent activation enFig8 shows the dependency of Eg on R obtained using ergy of about 134 kJ. l For R in the range from 0.2 tothe model-free isoconversional method. For R less than 0.2, 0.75, the reaction rate is under a mixed control and the ap-the activation energy is almost constant (-68 kJ-mol). parent activation energy increases with increasing reductionWith the progress of the reduction, for R greater than 0.2, degreethe activation energy increases gradually because the reduc-tion of titanium bearing oxides, which are more difficult to Acknowledgementsreduce than magnetite and wustite. becomes dominant at in-creasing R. Moreover, the reactants are covered by the re-This work was financially supported by the Major Pro-duced product, resulting in the reduction occurring through gram of National Natural Science Foundation of China(Nosthe diffusion of gaseous species. When the degree of reduc- 51090383 and 51090382) and the Scholarship Award fortion is up to 0.75, the activation energy is constant(134 Excellent Doctoral Student granted by the Ministry of EdukJ mol), indicating that diffusion becomes the rate-con-tion of china(No.0903005109081-8)trolling step. This is consistent with the observation in Fig. 7References[1 D.S. Chen, L N. Wang, T. Qi, and B Song, Study on preoxianadium-bearing titanomagnetite concentrates jHunan Univ. Sci. Technol. Nat. Sci., 26(2011), No 3, p. 951002 Z.J. Liu, G.Q. Yang, Q.G. Xue, J. L. Zhang, and T.J. Yang,Research on direct reduction of coal-containing pellets of va-nadic-titanomagnetite by rotary hearth furnace, Chin. JProcess Eng, 9(2009), No. 1, p 5170[3] L.H. Zhou, D P. Tao, M.X. Fang, F.H. Zeng, and x. Pu,Carbothermic reduction of v-Ti magnetite ore. chin. Rare0.00.10.20.30.40.50.60.70809Met.,332009),No.3,p.40[4] G.H. Zhang, Z. Yan, Y.J. Feng, M. Guo, X.D. Wang, L.F. Li,Fig 8. Dependency of the activation energy onind G.Z. Zhou, Reduction kinetics of vanadic titanomagnetitereduction determined using the model-free isoconversionalin Panzhihua, J. Chin. Rare Earth Soc., 26(2008), p. 21([5] X. Xue, Research on direct reduction of vanadic titanomagnetite, Iron Steel Vanadium Titanium, 28(2007), No 3, p. 3716] Y.Z. Lan and C P. Liu, Kinetics of carbon catalytic reduction4. Conclusionsof titanomagnetite concentrate, Acta Metall. Sin., 32(1996).No.5,p.502中国煤化工The isothermal reductions of Panzhihua titanomagnetite [7] E Park andHa-ferrous ore byCNMHGT. Hu et aL., Isothermal reduction of titanomagnetite concentrates containing coall37hydrogen, // Int, 44(2004), No 6, p. 999[13 J.L. Zhang, X.D. Xing, M. M. Cao, K.X. Jiao, C L. Wang, and[8] E. Park and O. Ostrowski, Reduction of titania-ferrous ore byS. Ren, Reduction kinetics of vanadium titano-magnetitecarbon monoxide, IS/ In., 43(2003), No 9, p. 1316carbon composite pellets adding catalysts under high tem[9 T. Hu, X.W. Lv, C.G. Bai, Z.G. Lun, and g B Qiu, Reduc-perature, J. Iron Steel Res. Inf, 20(2013). No. 2, p. Ition behavior of Panzhihua titanomagnetite concentrates with [14]S K Dey, B. Jana, and A. Basumallick, Kinetics and reductioncoal, Metall. Mater. Trans. B, 44(2013), No 2, p. 252characteristics of hematite-noncoking coal mixed pellets under[10] X W. Lv, Z.G. Lun, J.Q. Yin, and C G. Bai, Carbothermicrogen gas atmosphere, ISI In, 33(1993), No. 7, p. 735reduction of vanadium titanomagnetite by microwave irra- [15]s. Vyazovkin and C.A. Wight, Model-free and model-fittingdiation and smelting behavior, ISTJ Int, 53(2013). No. 7, papproaches to kinetic analysis of isothermal and nonisothermal data, Thermochim. Acta, 340-341( 1999), p 53[11] H.Y. Sun, X.J. Dong, X.F. She, Q.G. Xue, and J.S. Wang, [16]ME. Brown, D. Dollimore, and A K Galwey, Reactions inSolid state reduction of titanomagnetite concentrate by graphthe solid state, [in] Comprehensive Chemical Kinetics, Elte, IST In,53(2013),No.4,p.564ever,22(1980),p.347[12 H.Y. Sun, J.S. Wang, Y.H. Han, X F. She, and Q.G. Xue, [7 J.H. Sharp, G.w. Brindley, and B.N. Narahari, NumericalReduction mechanism of titanomagnetite concentrate by hy-data for some commonly used solid state reaction equations Jdrogen, Int J Miner. Process, 125(2013), No. 12, P. 122Am Ceram Soc., 49(1966), No. 7, P. 379中国煤化工CNMHG