Distribution behavior of phosphorus in the coal-based reduction of high-phosphorus-content oolitic i

International Journal of Minerals, Metallurgy and MaterialsVolume 21, Number 4, April 2014, Page 331DO|:10.1007/s12613-0140913-XDistribution behavior of phosphorus in the coal-based reduction ofhigh-phosphorus-content oolitic iron oreYong-sheng Sun, Yue-xin Han, Peng Gao, and Duo-zhen RenCollege of Resources and Civil Engineering, Northeastern University, Shenyang 110819, China(Received: 3 September 2013; revised: 23 October 2013: accepted: 28 October 2013)Abstract: This study focuses on the reduction of phosphorus from high-phosphorus-content oolitic iron ore via coal-based reduction. Thedistribution behavior of phosphorus (i.e, the phosphorus content and the phosphorus distribution ratio in the metal, slag, and gas phases)during reduction was investigated in detail. Experimental results showed that the distribution behavior of phosphorus was strongly influencedby the reduction temperature, the reduction time, and the c/o molar ratio. a higher temperature and a longer reaction time were more favorable for phosphorus reduction and enrichment in the metal phase. An increase in the C/O ratio improved phosphorus reduction but also hindered the mass transfer of the reduced phosphorus when the C/O ratio exceeded 2.0. According to scanning electron microscopy analysis, theiron ore was transformed from an integral structure to metal and slag fractions during the reduction process. Apatite in the ore was reduced toP, and the reduced P was mainly enriched in the metal phase. These results suggest that the proposed method may enable utilization ofhigh-phosphorus-content oolitic iron ore resourcesKeywords: iron ores; phosphorus; ore reduction; pyrometallurgypublished, and the results suggest that direct reduction fol1 Introductionlowed by magnetic separation is the most effective technique [2-8]. However, the considerable amount of phosThe high-phosphorus-content oolitic iron ore in China is phorus in the ore that transfers into metallic iron during ditypical high-phosphorus-content(with an average phos- rect reduction poses a serious problem [9]. Studies have fo-phorus content of 0.8wt%)iron ore resource with a large re- cused on the removal of phosphorus from high-phosphorus-serve of 3. 72 x 10 t. Between the 1960s and 1970s, inves- content iron ore through the addition of a dephosphorizationigations on the processing of oolitic hematite started in agent during the reduction process [10-11]. However, highChina, and various possible mineral separation methods quantes oof dephosphorization agent are needed due to thehave been explored. However, no satisfactory mineral proc- high phosphorus content in the ore, which results in an exessing methods have been developed because of the low pensive reduction process and the inability to recover phosgrade of the iron ore (35wt% to 5Owt% Fe), the poor libera- phorous from the oretion of iron minerals, the high phosphorous content, and soPhosphorus is one of the most detrimental impurities inon [1-3]. The high-phosphorus-content oolitic hematite ore steels, and most of phosphorus in metallic iron phase isis considered to be one of the most refractory iron oreseliminated into the steelmaking slag [12-13]. However, theWith the rapid development of its iron and steel industry, concentration of P2O5 in these slags is typically lwt%toChina has emerged as the largest steel producer in the world 3wt%, which makes it ineffective as a fertilizer or phospho-for ten consecutive years. Thus, China has intensified its rus resource [14]. a previous study has indicated that thefocus on the utilization of refractory iron ores(especially duplex steelmaking process developed in Japan is favorableoolitic iron ore). Numerous studies on the recovery of iron to dephosphorization [12, 15]. In the case of a normal hotigh-phosphorus-content oolitic iron ore have been metal([P]=0.lwt%), the P2Os content in the dephosphori-Corresponding author: Yue-xin Han E中国煤化工2 Sprringero University of Science and Technology Beijing and Springer-Verlag Berlin Heidelberg 2014CNMHG332Int. J. Miner. Metall. Mater., Vol. 21, No. 4, Apr. 2014zation slag is approximately 5wt%. When the phosphorus (4)are exhibited in Fig. I to reflect the relationship betweerconcentratIondjusted to between the standard Gibbs energies(AG )and the temperature0. 15wt% and 0. 25wt%, the P2O5 content in the slag exceeds The initial temperatures at atmospheric pressure at which10wt%[12]. Hence, we proposed a new method that enables reactions(1),(2),(3), and (4) will occur are 1495, 1224,the possible comprehensive exploitation of high-phosphorus- 1269, and 1287C, respectively. In addition, the standardcontent oolitic iron ore. Phosphorus in the iron ore will be Gibbs free energies of these reactions significantly decreaseforced to migrate into the metallic iron phase during the re- with increasing temperature, which demonstrates that calduction process, which allows the collection of reduced iron cium phosphate is more easily reduced at high temperaturesder with high phosphorus content by1000tion. The high-phosphorus reduced iron powder is then re-Reaction(1).- Reaction(2)fined by a duplex steelmaking process. Afterward, molterReaction (3)steel is produced, and high-phosphorus-content steelmakingslag that can be used as a phosphate fertilizer is also obtained. The migration characteristics of phosphorus and thephosphorus content in the metallic iron phase are the keypoints to this methodWith this background, we here report the results of our-200research into the distribution behavior of phosphorus (i.e-400the phosphorus content in the slag and metal phases as well8001000120014001600as the phosphorus distribution ratio in the metal, slag, andgas phases). In addition, the microstructure and compositionFig. 1. Correlation between AG and temperature for reac-characteristics of reduced products were studied. The pre-tions(1)to(4sent study aims to provide a basis for the rational utilizationIn this study, the reduction temperature ranged from 1175of high-phosphorus-content oolitic iron oreto 1275.C, which is close to(or above) the initial temperatures of reactions(2)to(4). Thus, the partial apatite in the2. Thermodynamic basis for phosphorus reduc- ore is reduced to P. Because the reduction process was pertionformed in an open MoSi2 muffle resistance furnace, thesmall amount ofp reduced from the apatite will volatilizeIn the iron ore, phosphorus usually exists as apatite, into the gas phase and the nonvolatilized portion will miwhich is a double salt formed by Ca3 (PO4)2 and CaF2, CaCl2, grate into the metal phase. The unreduced apatite and someor Ca(OH)2, and Ca3(PO4)2 is the main component. Accord- P atoms that have not yet migrated into metallic iron willng to the Ellingham diagram [16], the lines for the reactions remain in the slag phase. These analyses indicate that apatite4/5P+ O2=2/5P2Os and 6Feo+ O2- 2Fe3O4 are close to in the iron ore can be reduced using coal-based reductioneach other. Hence, phosphorus is almost reduced in the blast and that phosphorus will be present in the metal, slag, andfurnace during iron making and transfers into the hot metal. gas phases after the reduction processDuring reduction, calcium phosphate can be reduced to Pby carbon. SiO2 can displace P2Os in calcium phosphate, 3. Experimentaland P2Os is then reduced by carbon. Therefore, the reaction3.1. Materialsof calcium phosphate can be accelerated. The reactionmechanism of calcium phosphate to P can be expressed byThe representative samples of high-phosphorus-contentthe following reactionsoolitic iron ore were collected from the guandian iron mineCa3(PO4)(S)+5C(s)=3CaO(s)+2[P]+ 5Co(g)(1) Hubei province, China. a total of I t ore was collected. The/3Ca3(PO4)2(s)+SiO2(s)+5/3C(s)=ore was then crushed, classified, homogenized, and sampledCaO SiO2(s)+2/3[P]+5/3C0(g)(2) The final sample had a particle size of 100% passing 2 mmCa3(PO4)(s)+ 2Sio2(s)+ 5C(s)and its chemical composition is shown in Table 1. The re-3Cao. 2SiO2(s)+ 2[P]+ 5CO(g)(3) sults indicate that the contents of total iron, SiO2, Al2O3, and2/3Ca3(PO4)2(s)+SiO2(s)+10/3C(s)=Cao were approximately 42.21 wt%, 21.80wt%, 5.47wt%2 Cao.sio2(s)+4/3[P]+103CO(g)(4)and4.33wt%,H中国煤化工 tent of harmfulThe calculated thermodynamic results for reactions (1)to phosphorus wasCNMHGc ore wasY.S. Sun et al, Distribution behavior of phosphorus in the coal-based reduction of high-phosphorus-content oolitic iron ore 333mainly composed of iron oxides and silicates. The minera- X-ray diffraction(XRD). The main crystalline phases werelogical analysis of the iron ore sample was performed using hematite, quartz, chamosite, and apatite( Fig. 2)Table 1. Chemical composition of high-phosphorus-content oolitic iron oreSiO AL,OMgo PKMn4.330.1310000appeared as oolitic layers; the phosphate was mainly assocated with hematite. The details of apatite in the ooid were8000· Chamositeinvestigated by line-by-line scanning; the results are shownin Fig. 4. The distribution of p was identical to that of Caand alternated with Fe. This result further shows that apatiteis closely associated with hematite in the ore200013.70pm10203040506070809020/(°)Fig. 2. XRD pattern of the oolitic iron oreThe SEM image of the sample is illustrated in Fig. 3, inwhich the ooid structure can be readily observed. Thus, theiron ore was a typical oolitic iron ore. Hematite was mostpredominant in the image, followed by quartz. This resultalso indicates that hematite was the most important ironmineral in the ore and that the primary silicate mineral wasFig 3. SEM image of the ore sample: A-apatite; Q--quartzquartz. Apatite was the only phosphate mineral and usually H-hematite(d)中国煤化工Fig 4. Line-by-line scanning images of the ooid structure:(a)image of sCNMHGCa334Int. J. Miner. Metall. Mater., Vol. 21, No. 4, Apr. 2014An anthracite coal that was crushed to 100% passing 2 distribution ratios in the metal, slag, and gas phases are de-nm was used as a reductant. The proximate analysis results fined by eqs. (5),(6), and (7), respectivelyof the coal are listed in Table 2 and indicate that the coal Pam=Pm X mm/(Po x mo)x%contains a high content of fixed carbon and comparatively Pa=P、Xm、/(P。×m)×100%low content of ash, volatiles, sulfur, and phosphorus. The Pas=1-Pam-Pdscoal is therefore a good reducing agent for coal-based re- where Pdm is the p distribution ratio in the metal phase, Pm isductionthe p content in the metal, mm is the mass of the metal in theTable 2, Proximate analysis of the coal wt% reduced sample, Po is the P content in the ore sample, mo isthe mass of the ore sample in the mixture, Pas is the P dis-tribution ratio in the slag phase, Ps is the P content in the67.83184512021.480.0040.028slag, ms is the mass of the slag in the reduced sample, andNote: Mar -moisture; Arash; Va-volatile matter; FCd fixed Pag is the P distribution ratio in the gas phaseThe morphology and microstructure of the reduced sam-3. 2. Experimental methodple was studied using a SSX-550 scanning electron micro-scope (SHIMADZU). The composition was analyzed brThree main factors that influence the distribution of energy-dispersive spectrometry(EDS)on an Inca X-rayphosphorus were investigated on the basis of the character- spectrometer combined with the scanning electron micro-istics of coal-based reduction. These parameters are the re-scope.duction temperature, the reduction time, and the c/O molarratio (i.e, the molar ratio of fixed carbon in the coal to re- 4. Results and discussionible oxygen in iron oxides). The iron ore samples and theoal were thoroughly mixed in specific proportions by rig4.1. Distribution behavior of phospheorous stirring for approximately 30 min. The prepared mix- 4.1.1. Effect of reduction temperaturetures were then reduced at various temperatures and time inThe ore samples were reduced at temperatures betweenan open special MoSi2 furnace(Fig. 5), in which the tem1175 and 1275C to determine the effect of temperature onperature fluctuation was controlled to within 1.5C. Coal the distribution behavior of phosphorus. The results are ilwas placed at both the inlet and outlet sides of the furnace to lustrated in Fig. 6. The data reveal that the reduction temensure that a reducing atmosphere was present within the perature significantly affected the distributiofurnace. After the reduction process, the reduced samples phosphorus. Both the phosphorus content in the metal andwere rapidly removed and immediately quenched in water, the Pdm rapidly increased when the temperature was in-filtered, and dried at 80C in a vacuum oven. In addition, creased from 1175 to 1275.C. Conversely, a sharp decreasepurge gases were not utilized during the reduction and the in the Pas and the phosphorus content of the slag was obreduction experiments were run at atmospheric conditionsserved. The Pdg gradually increased as the reduction temperature was increased and then leveled off after the reduc-Im [n [mtion temperature reached 1250C. The results can be derived卿卿from thermodynamic data in Fig. 1, in which the AG(1)to(4)decrease as the temperacreases. This phenomenon demonstrates that the reductionreactions (1)to(4)can be improved through an increase intemperature. Therefore, more apatite in the ore was reducedto P at higher temperatures. Meanwhile, when the reductionFig. 5. Schematic diagram of the reduction apparatustemperature is increased, the diffusion coefficient of P willfurnace; 2-prepared mixtures; 3- coal; 4 outlet; 5-MoSiz significantly increase, which will accelerate the enrichmentheating element; 6-hearth: 7-thermocouple for temperature of P in the metal and gas phasesmeasurement: 8-inlet: 9--control cabinet: 10-bed.4. 1.2. Effect of reduction timeThe metallic iron content in the reduced samples was deA series of experiments were conducted at various reduc-termined by titrimetry The phosphorus content in the metals tion time to determine their effect on the distribution behavand slag were measured by spectrophotometry and X-ray ior of phosphor中国煤化工 n time on thefluorescence spectrometry, respectively. The phosphorus phosphorus contTHCNMHGpIotted in FigY.S. Sun et al, Distribution behavior of phosphorus in the coal-based reduction of high-phosphorus-content oolitic iron ore 3357(a). The phosphorus content increased in the metal but de- phorus contents in the metal and slag. After the maximumcreased in the slag as the reduction time was increased. The reduction time(70 min), the phosphorus content was as highphosphorus distribution ratios in the metal, slag, and gas as 2. 14wt% in the metal and as low as 0.314wt% in the slagphases as functions of reduction time are shown in Fig. 7(b). The phosphorus distribution reached 70.08% and 16.21% inWhen the reduction time was increased, the phosphorus dis- the reduced metals and the slag, respectively. This phe-tribution ratios in both the metal and gas increased, whereas nomenon occurs because more apatite in the ore is reducedthe phosphorus distribution ratio in the slag decreased. This and more reduced P migrates into the metal phase as the reenomenon was consistent with the changes in the phos- duction time is extended▲ P in metal● P in slagMeta郾Sg1.0895E604020116011801200122012401260128011751200122512501275Fig. 6. Effect of reduction temperature on the distribution behavior of phosphorus(/o molar ratio of 2. 5; reduction time of 50min):(a) phosphorus content;(b)distribution ratio.(b)Metal 8 Slag Gas1.00.50.0Fig. 7. Effect of reduction time on the distribution behavior of phosphorus(C/o molar ratio of 2.5; reduction at 1250 C): (a) phosphorus content; (b)distribution ratio4.1.3. Effect of the c/o molar ratioand the Pas substantially decreased and then slightly deThe effect of the C/O molar ratio (i.e, the molar ratio creased when the C/O ratio was greater than 2.0. The Pdg inbetween the fixed carbon in the coal and the reducible oxy- creased with increasing C/O ratios(Fig. 8(b) because thegen in the ore)on the distribution behavior of phosphorus content of volatile matter increases with the increase of thewas studied at C/O values that varied from 1.0 to 3.0( Fig 8). amount of reductant and more reduced phosphorus willThe distribution behavior of phosphorus was considerably volatilize into gas with the volatile matteraffected by the amount of reductant. The phosphorus con-tent in the metal increased with increasing C/O ratio and de4.2. Characterization of the reduced samplecreased when the C/O ratio exceeded 2.0. The changes in 4.2.1. Phase composition of the reduced samplethe pam as a function of the c/o ratio were consistent withThe XRD pattern of a selected reduced sample is shownthe phosphorus content in the metal. This phenomenon is at- in Fig. 9. The XRD pattern shows that iron appeared mosttributed to more reductant promoting the reduction of phos- often in the form of metallic iron in the reduced sampleof the reduced p metallic iron iscovered bv low-intensity magneticinto the metal phase. Both the phosphorus content in the slag separation after中国煤化工 vas still the priCNMHG336Int. J. Miner. Metall. Mater., voL. 21, No. 4, Apr. 2014mary impurity. However, in comparison with the Xrd pat- intense, and the peaks of chamosite disappeared after thetern of the original iron ore(as shown in Fig. 2), the pattern reduction process. Peaks attributable to Ca2SiO4 andof the reduced sample contains no apatite peaks, which re- Ca(Al2Si2Os) appeared in the Xrd pattern of the reducedduction process. Moreover, the peaks of Sio2 became less had reacted during the reduction proces/osite in the oreveals that apatite did react with the reductant during the re100(a)(b)80■Mtal圈Slag國Ga-P in metal60二o P in sl0.0Fig 8. Effect of C/O molar ratio on the distribution behavior of phosphorus(reduction at 1250C for 50 min): (a) phosphorus content; (b) distribution ratio.▲ Metallic ironphase duline-by-line scanning results of the reduced sample, the★Fe3O4tribution of p clearly coincided with Fe, whereas Ca was≌5000v Ca SiO(Mg, Fe)SiOascattered in the space of P, which further indicated that apa84000·Ca(Al2Si2O3tite was reduced to p and that the reduced p was enriched inthe metal phasc5 Conclusions1000The reduction of phosphorus in high-phosphorus-content1020304050607080oolitic iron ore was examined using coal-based reduction20/(°)The effects of reduction temperature, reduction time, andFig 9. XRD pattern of the reduced sample(C/O molar ratioof 2.0, reduction at 1250C for 50 min)coal content on the distribution behavior of phosphorus wered the results are summarized as follow4. 2. SEM analysis of the reduced sample(1)Reduction temperature, reduction time, and coal conWe also investigated the selected sample by SEM to bet- tent significantly influenced the distribution behavior ofter understand the previously discussed experimental results. phosphorus. Both the phosphorus content and the distribuThe SEM images of the reduced sample and the results of tion ratio in the metal phase increased with increasing reac-EDS measurements at different points and of surface scans tion temperature and time. The phosphorus content and disof Fe, O, P, and Ca are illustrated in Fig. 10. The SEM im- tribution ratio in the slag phase decreased with an increase inages of the original ore and reduced samples showed that the three aforementioned factors, while those in the gas phaseiron minerals in the ore were reduced to metallic iron and increased. With an increase in the C/O ratio, the phosphorusthat the metal set in the slag in the form of spherical parti- content and the distribution ratio in the metal initially incles(Fig. 3 and Fig. 10). The complicated particles of creased, but decreased when the C/O ratio exceeded 2.0(2)During the reduction process, the original ore samplemetal and slag phases after reduction. On the basis of the was reduced to the metal and slag phases. Apatite was reEDS analyses of spots I and 2, P exists in the metal phase duced to P, and most of the reduced P enriched into theand the slag is mainly composed of Si, Al, P, Ca, and O. metal, with a small amount of P volatilized into the gas phaseThese results demonstrate that apatite in the ore was reduced The unreducedHEhat did not mito p and that the reduced p transferred to the metallic iron grate into the me中国煤化工 slag phaseCNMHGY.S. Sun et al, Distribution behavior of phosphorus in the coal-based reduction of high-phosphorus-content oolitic iron ore337F会E/kevFig 10. SEM analysis results of the reduced sample(c/o molar ratio of 2.0, reduction at 1250C for 50 min ): (a)image of selectedarea;(b)EDS spectra;(e)Fe;(d)O;(e)P;(n Ca.(3)For all the experiments, the phosphorus content in the Referencesmetal phase was greater than 1.Owt%. Therefore, when thereduced iron powder(which can be separated by magnetic [] Y.s. Sun, Y.X. Han, P. Gao, Z.H. Wang, and D.Z. Ren, Re-separation)is refined by a duplex steelmaking process,acovery of iron from high phosphorus oolitic iron ore usingslag that contains more than 1 Owt% PoS will be obtainedcoal-based reduction followed by magnetic separation, Int JThis slag can be used as a fertilizer or as a phosphorus reMiner. Metall. Mater., 20(2013), No. 5, p[2] S.F. Li, Y.S. Sun, Y.X. Han, G.Q. Shi, and P. Gao, Fundasource. Our analysis demonstrates that the recovery of phosmental research in utilization of an oolitic hematite by deepphorus from high-phosphorus-content oolitic iron ore is feasireduction, Adv. Mater. Res, 158(2011), p. 106ble. The results of this study indicate that the rational utiliza- [3]YS Sun, P. Gao, Y X Han, and D.Z. Ren, Reaction behavtion of high-phosphorus-content oolitic iron ore is possibleior of iron minerals and metallicgrowthcoal-based reduction of an oolitic iron ore, Ind. Eng. ChemAcknowledgementsRes,52(2013),No.6,p.23234] P. Gao, Y.S. Sun, D Z. Ren, and Y.X. Han, Growth of metalThis work was financially supported by the Nationallic iron particles during coal-based reduction of a rare-earthsNatural Science Foundation of China(No. 51134002)andbearing iron ore, Miner:. Metall. Process, 30(2013), No. 1, pthe Fundamental Research Funds for the Central Universi- [5] KQ Li,wties of China(No. N120601004)tion from oo中国煤化工 l, Iron extrac-process, J. IronCNMHGInt. J. Miner. Metall. Mater., Vol. 21, No. 4, Apr. 2014Steel Res Int, 18(2011), No. 8, p. 9Li, Dephosphorization mechanism in a roasting process forH.Q. Tang, Z.C. Guo, and Z.L. Zhao, Phosphorus removal ofdirect reduction of high-phosphorus oolitic hematite in westhigh phosphorus iron ore by gas-based reduction and meltHubei Province, China, Univ. Sci. Technol. Beijing,separation, J. Iron Steel Res. Int, 17(2010), No 9, p. I32(2010),No.8,p.968[7] Y.F. Yu and C.Y. Qi, Magnetizing roasting mechanism and [12] J Diao, B. Xie, Y.H. Wang, and X Guo, Recovery of phoseffective ore dressing process for oolitic hematite ore, J .Wuphorus from dephosphorization slag produced by duplex highhan Univ. Technol. Mater. Sci. Ed, 26(2011), No. 2, p. 176phosphorus hot metal refining, /S/ Int, 52(2012), No. 6, p[8] J. Wu, Z.J. Wen, and M.J. Cen, Development of technologiesfor high phosphorus oolitic hematite utilization, Steel Res Int, [13 H. Kubo, Matsubae-Yokoyama K, and T Nagasaka, Mag82(2011),No.5,p.494.netic separation of phosphorus enriched phase from multi[9 C.R. Manning and r.J. Fruehan, The behavior of phosphorusphase dephosphorization slag, // Inf, 50(2010), No. 1, p. 59in direct-reduced iron and hot briquetted iron, Iron Steel- [14]K. Morita, M.X. Guo, N Oka, and N. Sano, Resurrection ofaker,30(2003),p.62the iron and phosphorus resource in steel-making slag, J.[10J C.Y. Xu, T.C. Sun, C.Y. Qi, Y L. Li, X L Mo, D w. Yang,Mater:. Cycles Waste Manage, 4(2002), p. 93Z.X. Li, and B L. Xing, Effects of reductants on direct reduc- [15] Z.H. Tian, B H. Li, X M. Zhang, and Z.H. Jiang, Double slagtion and synchronous dephosphorization of high-phosphorousperation dephosphorization in BOF for producing lowoolitic hematite, ChinJ, Nonferrous Met, 21(2011), No. 3, pphosphorus steel, J. Iron Steel Res Int, 16(2009), No. 3, p.6[16] D.R. Gaskell, Introduction to the Thermodynamics of Mate[ll] D W. Yang, T.C. Sun, H.F. Yang, C.Y. Xu, C.Y. Qi, and Z.X.rials, Taylor and Francis, Washington, 1995中国煤化工CNMHG