Thermodynamic Study on Process in Copper Converters (The Copper-making Stage)

Journal of University of Science and Technology BijingVol. 7(2000), No.3, p.184MetallurgyThermodynamic Study on Process in Copper Converters(The Copper-making Stage)Chunlin Chen" Jiayun Zhang", Tuping Zhou", Shoukun Wei", Xingxiang La?, Meng Baif, Jinhong Jiang"1) Metalurgy School, University of Science and Technology Beiing, Beijing 100083, China 2) Guixi Smeter, Jiangxi, 335424(Received 1999-09-27)Abstract: Theoretical caleulations were based on thermodynamic equilbriumn in the multi-component and rmulti-phase system with heatand mass balance as well as the oxygen fficiency to take account for the ffects of process kinetics. The variations of temperature, massfractions of dissolved oxygen and sulfur in blister copper partial pressures for O, S. SO2 in gas phase for the copper-making stage werecalculated. The mode! predicted temperature, time of blowing 35 well as mass of the blister copper at end points for 6 heats showed afairly good agreements with corresponding plant data. The calculated O content of 0.065% and S content of 0.87% in blister copper wereboth at reasonable levels. Compared with the So called Goto model, the present model bas very much improved process description ofcopper-making stage as well as the prediction of end points for a copper converter by introducing the oxygen efficiency.Key words: Peirce Smith copper converter; themodynamic model; copper-making stageIn reference [1], the thermodynamic model devel-low.oped by present authors for Peirce-Smith copper con-logf= -(3113. 0%c.cn + 242.6.S%rinC) )-一verter, which will be called as copper converter below,was described and applied to the slag making stage ofthe process of copper converter No. 1 and 3 in Guixilogfs 一-(485.2 . 0%in.) + 281.6 S%inCc )-一(2)Smelter. The model is based on the thermodynamicThe underline "”in equations (1) and (2) indicatesequilibrium in the multi-component, multi-phase sys-he dissolved chemical species in molten copper sol-tem and combining with the mass and heat balance asution.well as the oxygen eficiency introduced by presentauthors to take account for the effects of the reactionThe standard Gibbs energy with reference state of akinetics. For the slag making stage, good agreementssolution containing oxygen of 1% (mass fraction) inhave been achieved between the calculated and meas-liquid copper is expressed as [5]ured temperatures as well as the contents for main ele-△.- - 85350+ 18.54T(3)ments in molten matte and slag phase. As a continu-ation, the present paper describes the model applicationFor dissolved sulfur, the corresponding equation isto the copper-making stage of the same.presented as [5]▲G=一119 660 + 25.32T(4)1 Thermodynamic DataUsing equations (1) (4), the temperature depend-Most of the thermodynamic data adopted in the cal-ence of Rault activity coefficients for oxygen and sul-culations can be found in references [2, 3]. Some newlyfur, 2 and '5 in Cu-O-S system could be obtained asre- assessed thermodynamic data used for species in thefollows:matte, slag and gas phase have been tabulated []. Inthe copper-making stage, molten copper is formed, ad-为=25.18 exp (一10265/T + 2.23)- 107一2240.ETding one more phase in the system, As it was assumed% = 50-26 xe (-14 2021T+ 21021 0202-421m7 (6)中国煤化工that only sulfur and oxygen dissolved in the moltenwhereCNMH Gfraction for oxygenblister copper, so that the thermodynamic data in liquidand sulxurwpeNy HIN wuum activity for copper inMHCu-O-S system could be used for the calculations of theCu-O-S system was assumed to be unity.copper-making stage. The Henry activity coefficientsfor oxygen and sulfur, f and f, have been taken from a2 Plant Datathermodynamic study of Cu-O-S system by Sano andSakao [4] The corresponding expressions are given be-The industrial trials of copper converts caried out inC. Chen, J. Zhang, T. Zhou, et al:185Guixi Smelter for a month were described in a previousthe slag-tapping at the end of slag-making stage. Thepaper [1] and report [6].chemical compositions of 'white metal' and coldAs shown in the paper earlier [4]，here charge 173 ischarges during the copper-making stage in charge 173again to be Lused as an example to examine the validityare listed in table 1, and the corresponding operationof the model. The main 'input' in copper making stagetime schedule in table 2.is the 'white metal' which remains in a converter afterTablel Compositions of chemical species in materials charged during the copper- making stage of charge 173 (mass fraction in %)ElementWhite metalSlagAirAnode scrapC opper scrapCu78.440.50 I0.00070.00099.440094.86F0.4749.1414.7000.002519.060.090.1000.00262b0.270.940.0800.0280Zn0.152.160.3000.00000.0015.4421.0004.8000.120078.8200.0700,0000C0.001SiO,21.728.000T/K14461 446298Table 2 Operation time schedule during the copper-makingresults using an Po. of 100%，ie. using the so-calledstage of charge 173Goto model are also demonstratedTime/minOperation stepsAs listed in table 2, the cold materials such as anode0Charge: white metal 158 t, slag 2 t,scrap, copper scrap and skull were charged into theoxygen-enriched air 511 m/minconverter frequently aiming at recycling copper scrap2Charge: anode scrap 5tand controlling the temperature. The saw-toothed cur-Charge: anode scrap 4.5tve for the temperature variation ilustrated in figure 112Charge: copper scrap 5twas caused by melting the cold materials. It is apparent20C harge: anode scrap 5.5↑hat the temperature was controlled in the range ofCharge: anode scrap 7t1 400- 1 450 K in order to avoid over-heating the melts,90Charge: anode scrap 8tfor which may reduce the life of the converter lining.105Charge: anode scrap St132Charge: skull 10t1500200(end) Gas- blowing is stopped. 142.5 t blister copper istapped, the temperature of blister copperis 1 418 K1450As described earlier []，the oxygen blown into the1400converter was not completely utilized. A parameter ofPresent modeloxygen eficiency, Mo, was introduced to present the ra-1350Experimentaltio between the consumed and the total amount of O2Goto modelMa was statistically evaluated using the plant data dur-1300ing the one month of industrial trials in Guixi Smelter.200250300350400450For the copper-making stage, the average oxygen effi-ti minciency took the value of 81%[I, 6].Figure中国煤化工perature during copper-makin3 Results and DiscussionHCNMHGNo Ccontinuous temperature raise observed during the3.1 Charge 173copper-making stage was due to several reasons. Firs-The model predicted results for the copper-makingtly, the enthalpy change caused by the reactions wasstage of charge 173 of converter No.3 are ilustrated inmuch less than that during the slag making stage. Sec-the following figures. For comparison, the calculatedondly, the oxygen fficiency for copper-making stage186J. of Univ. of Sci. cand Tech. Beijing, 7(2000), No.3usually was 10% less than that for the slag. makingstage. Thirdly, in order to avoid the overheating, airwith lower oxygen enrichment was used and cold ma-terials were frequently charged in operating practice.Figure 2 shows that the calculated oxygen and sulfur-- Present mode)contents in molten blister copper were 0.065 % and.... .Goto modelpP=0.1 MPa0.87 % respectively. At the end of the charge, the sulfurcontent lowered down and the oxygen content rose sud-0denly, which gave the sign for the end point.2002503003SO4004S0t! minFigure 4 Variation of Pso in gas phase during copper makingstage.0.1. ... Goto model4. .... Goto model0.01Present modelpe -0.1 MPa留-6240280320360 400 440t!minFigure 2 Variation of 0 and S content in blister copper dur--10ing copper-making stage.350450timinIt can be seen in figures 3- 5 that the partial pressuresFigure 5 Variation of Pa i证gas pbase duriag copper m akingof S., SO., O: in gas phase were approximately unchan-stage,ged. The sharp decline for Ps, and Pso, and the suddenrise for Po at the end of the charge were also the sym-160bols for the end point. Unfortunately,no plant data for140the gas phase and, therefore,the comparison of these120calculated partial pressures to the measured data could100not be provided.3 80E 6040Experimental20..... Goto model400450-20 It/min...， Goto model30Figure 6 Variation of mass of blister copper duriog copper-户m0.1 MPamaking stage.-40Compared the solid and the dashed line in each of theth minfigures, it can be seen that the oxygen and sulfur con-Figure 3 Variation of Ps in gas phase during copper makingtents in blister copper as well as Ps, Pso and Pa in gasphase given by the two models were quite close duringthe blowin中国煤化Ige before the endFigure 6 shows that the mass of liquid blister copperpoint. WhiYH| CN MH Ged using the pre-increased continuously due to the reduction ofCuS. Atsent model was somewna Lowe and Pa was higherthe end of the stage, the calculated mass of blister CoP-than those using the Goto model. These were caused byper was 140 t, and the measured data was 142.5 t whichusing a lower oxygen eficiency of 81% in this study,indicates a good agreement between the two data.rather than 100% in Goto model. Moreover, the time,C. Chen, J. Zhang, T Zhou, et at.:187temperature as well as the mass of blister copper for themaking stage in six charges during the industrial trials.end point given by the present model were much moreFor the sake of brevity, the calculated results for the sixclose to the relevant plant data than the predictionscharges with the relevant plant data as well as thoseusing the Goto model. This may demonstrate the valid-predicted using Goto model are summarized in table 3.ity of the present model.As shown in table 3, the oxygen and sulfur contents3.2 The six chargesin blister copper during the blowing in the copper-mak-The model has been used to simulate the copper-ing stage were around 0.06 % and 0.87 % respectively.Table 3 Calculated results of copper-making stageCalculated orCharge nummbermeasured0%S%1 (blowing)/ minTKm(Cu)/tCalc.(this model)0.0650.872961 444142.03- 173Calc.(Goto model)274I 464142.5Measured3051423138.0Calc .this model)0.853241 438158.51-503041 480159.03151 498154.00.0663341 460163.01-51Calc Goto model)310l 485163.53451 448156.5Calc (this model)0.862881430164.01-58Calc.(Goto mode)0.06S26014762951 395155.0Calc.(his mode)}2851412167.51- 601 450168.01 347160.0Calc .(this model)1 430172.03-16831414681 416Note: O %, S % (mass fraction).The sulfur content is somewhat lower than the satu-molten blister copper is also reasonable. For the equip-rated sulfur content (1.1% at I 200 C) obtained fromment reason,the samples could not be taken during thethe phase diagram of Cu(Cu:S system [7], showing thatblowing. Samples could only be taken from blister cop-the sulfur content in molten blister copper given by thisper stream during slag tapping. The content data obta-study is reasonable.ined fom sample assays were around 0.5% for oxygen,The solubilities for oxygen and sulfur in liquid cop-and 0.03 % for sulfur. As the dramatic changes of oxy-gen and sulfur contents at the end of a copper-makingper can be described as:stage ocurred, the 0 and S contents in the solid copperSanCcu 20meo SO:(7)samples could not be used to compare with the relevantThe equilibrium constant, K for the above reactioncalculated data during the blow ing.The deviation between the calculated and measuredcan be presented as;temperature which varies from土1 K to about=15K isPSo /PK=(%S)-(%O)(8)considered to be reasonable for the random nature ofthe industrial onerations In charge 60, this deviationwhere P =0. MPa.was中国煤化工id be atributed to theThe value of K in equation (4) is around 40 at 1 200lowerYHC N M H Gire measurement usingC [8]. As shown in figure 4, Pso in copper-makinga radiative pyrometer at the start of the blowing. Goodstage is around 0.02 MPa. With the values of K and Psu,agreement between the calculated and measured massthe oxygen content in blister copper at 1 200 C calcu-of blister copper can be seen from table 3. The excep-lated using equation (8) was 0.067 %. This indicatestion was in charge 168 where the large deviation of thethat the model predicted content, 0.065% for oxygen inblister copper mass could be seen, which might result188J. of Uniu. of Sei. and Tech. Beifing, 7(2000)，No.3from the fluctuated operating factors.trial trials of copper converters in Guixi SmelterUsing the data of columns 5, and 7 in table 3, the av-The oxygen and sulfur contents predicted using pre-erage deviations for temperature and time at the endsent model were respectively 0.065% and 0.87% dur-points respectively took the values of 10K and -ing the blowing, which were found both at reasonablemin in this work, while the corresponding values calcu-levels. Compared with the so called Goto model, thelated using the so called Goto model with 100% oxygencalculated results using present model demonstratedeficiency were +46 K and一29 min. Apparently, com-that the model could very much improve the accuraciespared with the Goto model, the present work has veryof temperature at end points and blowing time requiredmuch improved the process description of the copper-for copper- making stage of a copper converter.making stage and could predict the end point of a con-In addition to the slag-making stage [1]，the presentverting process much more close to that in the conver-work has provided a better understanding for the cop-ter operating practice. The prediction of the terminal ofper-making stage. The present work could be applied asa copper converting process is an important componenta means to optimize the operation, and could be prefer-in the process dynamic control. The present model mayable for the process dynamic control of copper conver-provide a useful tool for the process optimization andters in industy.dynamic control of copper converters.he major feature of the present thermodynamicAcknowledgmentsmodel is the introduction of the oxygen efficiency. TheThe authors wish to thank the National Natural Sci-success model predictions for the slag- making [1] andcopper-making stages of copper converter may confirmence Foundation of China for the financial support tothat the concept of the oxygen efficiency is valid and inthe project, No.59874005. The financial support fromagreement w ith the converting practice in a copper con-Division of Theoretical Metallurgy, Dept. of Metal-lurgy, KTH, Sweden to one of the authors Zhangverter.Jiayun) to initiate the work in this area is gratefully ac-The lower down of the oxygen efficiency during theknowledged. The authors also wish to thank Prof Scopper- making stage may be caused by a kinetic rea-Seetharaman as well as Prof. Du Sichen both in theson. The volume of the melts continuously declines inabove mentioned institution for the helpful discussionscopper making stage due to the sulfur removal fromin this work.white metal'. This would Jead the lower down of thesurface level of the molten bath during copper-makingReferencesstage. The residence time of the bubbles in the moltenU] C. L. Chen, T. P Zhou, J. Y. Zhang, et al.: J. University ofbath may therefore be reduced, which could cause aScience and Technology Beijing. 6 (1999), D 187.continuous decrease of the oxygen efficiency This in-[2] s. Goto: [n:] Copper Metallurgy Practice and Theory, M. J.dicates that oxygen fficiency could be evaluated usingJones [eds). Inst. Min. Met, London, 1974, p.23.a kinetic approach, so that the copper converting pro-[3] s. Goto: [in;] Copper and Nickel Converters, R. E. Johnson[eds). TMS- AIME, Warrendale, PA, 1979, p. 33.cess could be described using a modified thermody-(4] K. Sano, H. Sakao: J. Japan Frast. Metals, 19 (1995), p. 431.namic route. Such an effort has been made later [6] and[S] G. K. Sigworth, J, F. Eliott: Canadian Metallurgical Quar-will be published elsewhere.terly, 13 (1974), p. 455.[6] C. L. Chen: Thermodynemic and Kinetitc Studies on Copper5 Summary and ConclusionsCou中国煤化工yerter: [Doctoral Degreeheence and Technology Be-In this work, oxygen efficiency was introduced in thejingiY片CNMHGthermodynamic calculations of multi-component and[闪H. H Kelog: Canadian Melurgtcal Qutrtery;, 8(1969), p.multiphase systems to predict the yariations of tem-[8] A. K. Biswas, w. D. Davenport: Extractive Metalturgv ofperature and chemical compositions in all the phasesCopper, 3rd Ed, Pergamon Press, Oxford, 1994. p.151.during tihe copper-making stage for six heats of indus-